Plutonism and Orogeny in North—Central Washington— Timing and Regional Context By KENNETH F. FOX, JR., C. DEAN RINEHART, and JOAN C. ENGELS GEOLOGICAL SURVEY PROFESSIONAL PAPER 989 UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1977 UNITED STATES DEPARTMENT OF THE INTERIOR THOMAS S. KLEPPE, Serrelmy GEOLOGICAL SURVEY V. E. McKelvey, Director Library of Congress Cataloging in Publication Data Fox, Kenneth F. Plutonism and orogeny in north-centralWashington. (Geological Survey Professional Paper 989) Bibliography: p. 24—27. Supt. ofDocs.no.: I 19.162989 1. Intrusions (Geology)—Washington (State). 2. Geology¥Washi11gton (State). I. Rinehart, Charles Dean, 1922— joint author. 11. Engels, Joan C.,joint author. III. Title. IV. Series: United States Geological Survey Professional Paper 989. QE611.5.U6F67 551.8'09797 76—608343 For sale by the Superintendent of Documents, Us. Government Printing Office Washington, DC. 20402 Stock Number 024001-02925-1 CONTENTS Page Page Abstract ___________________________________________________ 1 Orogenic implications of geologic history ___________________ 16 Introduction _______________________________________________ 1 PreviOus views on orogeny in the Cordillera _____________ 16 Geologic provinces _________________________________________ 2 Late Triassic orogeny and plutonism ___________________ 16 Geologic history ___________________________________________ 3 Association of deformation of the Monashee Group and Permian and Triassic history ___________________________ 3 generation of gneiss domes with Cordilleran thrust Jurassic to Paleocene history ___________________________ 8 faulting _____________________________________________ 17 Discordant ages ___________________________________ 12 Laramide orogeny, formation of gneiss domes, and thrust Age of Omineca crystalline belt _____________________ 13 faulting _______________________________ ~ ______________ 18 Eocene to Miocene history _____________________________ 15 Discussion _________________________________________________ 20 References cited ___________________________________________ 24 ILLUSTRATIONS Page FIGURE 1. Index map showing orogenic provinces in Washington and southern British Columbia ________________________________ . 1 2. Geologic map of Okanogan Highlands and vicinity showing radiometric ages ________________________________________ 4 3. Geologic map of part of northeastern Washington and southern British Columbia ____________________________________ 6 4. Map showing distribution of late Paleozoic and Mesozoic stratified rocks of the eugeosynclinal province ________________ 8 5. Graph showing age determinations within the study area projected to an east-west transect __________________________ 12 6. Graph showing Mesozoic and Cenozoic age determinations within the region plotted with respect to the approximate boundaries of the Omineca crystalline belt and projected to the 49th parallel ____________________________________ 13 7. Map showing distribution of late Mesozoic and early Cenozoic thrust belts and areas of medium-to high-grade meta- morphic rock, including gneiss domes in western United States and southwestern Canada ________________________ 18 8. Diagrams showing stages in the development of the Omineca crystalline belt ________________________________________ 22 TABLES Page TABLE 1. Sources of radiometric age data in figures 2, 3, 5, and 6 ____________________________________________________________ 8 2. Interpretation of radiometric age determinations in part of Okanogan Highlands ____________________________________ 9 HI 1“ 399’? PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON— TIMING AND REGIONAL CONTEXT By KENNETH F. Fox, JR., C. DEAN RINEHART, and JOAN C. ENGELS ABSTRACT Bedrock in north-central Washington comprises (1) weakly to moderately metamorphosed eugeosynclinal rocks of Permian, Trias- sic, and possibly Jurassic ages, (2) high-grade polymetamorphic rocks~gneiss, schist, and amphibolite—that are, at least in part, more highly metamorphosed derivatives of the rocks of the previous category, (3) Mesozoic and Cenozoic plutonic rocks, and (4) Cenozoic lavas and continental sedimentary deposits. A review of the radiometric ages from north-central Washington suggests a complex history of plutonism and metamorphism begin- ning in Late Triassic and extending to Eocene time. However, ages of coexisting minerals from single samples typically are moderately to highly discordant. The discordance reaches a maximum along a zone flanking the Okanogan gneiss dome on the west. This dome forms the southwestern extremity of the Omineca crystalline belt, a north-trending orogenic subprovince about 250 km (150 mi) long and 55 km (35 mi) wide in British Columbia and Washington charac- terized by the presence of Shuswap (Monashee Group) terrane, gneiss domes, and allied metamorphic rocks. The Okanogan gneiss dome and the Shuswap are believed to be products of metamorphism and deformation, in part at least, of Late Cretaceous age. The development of the gneiss domes indicates that metamorphism within the Omineca reached sufficient intensity at depth to cause incipient anatexis and mobilization of the infrastruc- ture. The discordance west of the gneiss dome and the Omineca belt is attributed to weak thermal metamorphism that developed above and west of the zone of more intense metamorphism. Ages from within the Okanogan gneiss dome range from Late Cretaceous (lead-uranium, zircon) to Eocene (K—Ar, biotite and mus- covite; fission track, apatite). Their discordance is attributed to slow cooling after the climax of metamorphism in the Late Cretaceous. Orogeny and plutonism in the north-central area of Washington began almost simultaneously during the Late Triassic with folding of the Permian and Triassic bedded rocks and their intrusion by the 195-million-year-old Loomis pluton. Thermal events subsequent to Late Triassic cannot be linked with specific orogenic deformation, except for deformation associated with the hypothesized Late Cre- taceous metamorphism and mobilization of the Shuswap and the gneiss domes within it. This event is temporally associated with westward-directed overthrusting along the Shuksan thrust to the west, attributed to mid-Cretaceous orogeny, and with eastward- directed overthrusting to the east along the Cordilleran thrust belt, attributed to the Laramide orogeny, which ended in the Late Cre- taceous or early Tertiary. The Shuswap terrane and associated gneiss domes appear to occupy the axial zone between convergent thrusts that show an aggregate crustal contraction of possibly 250 km (150 mi). These relations suggest a genetic model as follows: The Permian volcanic and pyroclastic rocks of the region were probably deposited in island-arc and back-arc basins located east of an east-dipping subduction zone. We speculate that in Late Triassic, the continental plate overrode a rise system embedded in the oceanic plate, after which the eugeosynclinal prisms were invaded by calc-alkalic mag- mas derived through partial melting of upper mantle and lower crust within a zone of high heat flow above the overridden rise. The res- idue remaining in the zone of partial melting probably became progressively more dense as the hyperfusible part was removed and in Late Cretaceous catastrophically sank into the asthenosphere, forming a short-lived convection cell whose axis lay beneath the Okanogan region. The overlying crust was dragged toward this axis which caused thrust faulting to the east (Cordilleran thrusts) and west (Shuksan thrust) and thickening of the crust over the cell by stacking of thrust sheets and plastic flow. Concurrently, elements of the infrastructure were mobilized and penetrated higher levels in the crust, with their culminations forming the Okanogan gneiss dome and other gneiss domes within the Shuswap. After the demise of the convection cell in latest Cretaceous, the thickened crust isos- tatically rebounded. Upper crustal levels over the Omineca were rapidly eroded away, and elements of the Late Cretaceous infra- structure (the gneiss domes and the Shuswap) were exposed in the Eocene. INTRODUCTION Bedrock in north-central Washington (fig. 1) com- prises (1) weakly to moderately metamorphosed eugeoeynclinal rocks of Permian, Triassic, and possibly Jurassic age, (2) high-grade polymetamorphic rocks—— gneiss, schist, and amphibolite—that are, at least in part, more highly metamorphosed derivatives of the rocks of the previous category, (3) Mesozoic and CenoZoic plutonic rocks, and (4) Cenozoic lavas and continental sedimentary deposits that patchily overlie the older rocks. The lavas are products of regional episodes of volcanism, the older of which is believed to be Eocene and the younger late Miocene to early Pliocene. The age of plutonism and metamorphism in the Okanogan area is bracketed by the Permian to Triassic age of the eugeosynclinal host rocks and the early Cenozoic age of the overlying continental deposits. Fossils are so rare in the eugeosynclinal deposits, however, that the older boundary is quite im- precise. Previously reported radiometric ages of rocks from within the study area (fig. 2; table 1) range from Late Triassic to Eocene and taken collectively indicate that the history of magmatism and metamorphism has been 2 PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON 124° 120° 116° I I i, I I , Frenchman’s Cap Rocky Mountain gneiss dome i thrust belt Columbian inter- +0 Purcell Coast plutonic ontane belt 0) 50° ‘ complex 6:1, Omineca .7 Area of crystalline . belt igure 2 Insular _BRITlSH COLUMBIA fANAJ’fi— -— . fO'd WASHINGTON UNITED STATES belt v I l o 48 -_ V____ \ — Area of l 5 figures 3 and 4 ‘ K \ Columbia Plateau l \\,_ province | 46° Cascade K _ I 33.13:; ,,,___—\. a/ I If ‘ I I l I O 100 200 MILES 0 100 200 KILOMETERS FIGURE 1.—Index map showing orogenic provinces in Washington and southern British Columbia. Modified and extended mainly from Wheeler (1970, p. 1). complex. However, the age of many of the plutons in the study area, including some that have been studied intensively, has not been unambiguously determined. These rocks typically show substantial discordance be- tween age determinations on mineral pairs from indi- vidual samples and discordance from sample to sample within individual plutons. In general, we assume that this discordance reflects varied retention of fission tracks, argon, rubidium, strontium, lead, or uranium, either during protracted cooling after initial emplace- ment of the pluton or varied retention during and after later thermal metamorphism. The temperature of a cooling mineral, at which the rate of loss of the radiogenic daughter element by diffusion becomes neg- ligible compared to the rate of its accumulation, is defined as the critical blocking temperature (Macin- tyre and others, 1967, p. 822). This temperature may be a function of various factors, notably grain size, the pressure of argon within the host but external to the dated mineral, and the length of time during which that particular temperature is maintained. The block- ing temperature for radiogenic argon is higher in hornblende than in coexisting biotite, and as a result, in thermally metamorphosed terranes the hornblende commonly yields an older age than the coexisting bio- tite (Hart, 1964; Hanson and Gast, 1967). The concept of a critical blocking temperature can also be extended to the rate of erasure of fission tracks relative to their rate of formation. Thus, the minimum age of crystallization of plutonic rocks yielding discordant K—Ar or fission- track ages is here assumed to be the older of a mineral pair. This paper refines the plutonic and metamorphic history of part of the Okanogan Highlands, mainly through review of the radiometric ages and considera- tion of the implications of their discordance, and relates this history to the depositional and orogenic history of the region. The eastern part of the region includes stratified rocks of Precambrian and early Paleozoic age that are not found within the study area. The distribution of these rocks, along with that of the more widespread late Paleozoic and Mesozoic stratified rocks, and their various dynamically metamorphosed derivatives provides the basis for dividing the region into depositional and orogenic provinces. We define these provinces and then discuss the Mesozoic and Cenozoic geologic history of the study area, where pos— sible in the broader context of the geologic history of the region. We are grateful to R. G. Yates, who introduced us to the geologic problems of the region and who provided advice and encouragement throughout the duration of the study. We are also indebted to M. D. Crittenden for numerous stimulating discussions on tectonic prob- lems of the region. GEOLOGIC PROVINCES The region encompassed by figure 3 comprises parts of two depositional provinces of differing ages. The pre-Tertiary stratified rocks of that part of northeastern Washington lying roughly east of long 118° W. are chiefly Precambrian and lower Paleozoic miogeoclinal (Dietz and Holden, 1966) rocks, in con- trast to those to the west, which are upper Paleozoic and Mesozoic eugeosynclinal rocks. In Washington, the contacts between the eugeosyn- clinal rocks and the miogeoclinal strata to the east are obscured by faulting and folding along much of their extent (Yates and others, 1966, p. 54). According to Campbell (1964), the eugeosynclinal deposits in the Hunters quadrangle and vicinity (south of Kettle Falls, Wash.) have been thrust eastward over the miogeoclinal rocks. In British Columbia however, rocks of the eugeosyncline (Milford Group) concordantly but unconformably overlie rocks probably be— longing to the miogeocline (Lardeau Group, Cairnes, 1934, p. 36-38). In addition to the depositional provinces discussed above, the southern British Columbia and north- central Washington region can also be assigned to three orogenic provinces, the Purcell foldbelt, the Omineca crystalline belt, and the Columbian inter- montane belt (fig. 1). GEOLOGIC HISTORY 3 The Omineca crystalline belt is distinguished chiefly by the presence of medium- to high-grade gneiss and schist of the Monashee Group of the Shuswap terrane of Jones (1959). Several gneiss domes have been iden- tified within the rocks of the Monashee Group and its probable correlatives. The gneiss. domes include among others the Valhalla and Passmore domes in British Columbia (Reesor, 1965) and the Okanogan dome in Washington (Fox and Rinehart, 1971) (fig. 1). The eastern limit of the Monashee Group and Monashee-like rocks within the region serves as the boundary with the Purcell foldbelt, and the western limit serves as the boundary with the Columbian in- termontane belt. The contact between the Purcell foldbelt and the Omineca crystalline belt (figs. 1, 3) nearly coincides with the contact between rocks of the eugeosynclinal facies and the rocks of the miogeoclinal facies. GEOLOGIC HISTORY PERMIAN AND TRIASSIC HISTORY The eugeosynclinal province within the area of figure 3 is chiefly floored by sparsely fossiliferous rocks, prob- ably of late Paleozoic age. Within the study area (fig. 2), these rocks have been assigned to the Anarchist Group (Rinehart and Fox, 1972, p. 8—11), which is composed of variably metamorphosed complexly interfingering and intergrading deposits of argillite, siltstone, sharpstone conglomerate, graywacke, sandstone, and limestone. Deposits exclusively of volcanic origin—metamorphosed lava or pyroclastic material—are also present but are not abundant. However, clasts of volcanic rock are a minor but apparently widespread constituent of the coarser grained clastic rocks, such as the sharpstone conglomerate and graywacke, suggesting that material of volcanic origin may be a significant constituent of the finer grained clastic rocks as well. The Anarchist is be- lieved to have an overall thickness in excess of 4,500 m (15,000 ft). Fossils are rare. Those found, which are probably restricted to a narrow stratigraphic interval in the upper middle of the group, have been dated as Permian, perhaps Late Permian (Rinehart and Fox, 1972, p. 9—10; Waters and Krauskopf, 1941, p. 1364). The Anarchist is therefore, at least in part a temporal equivalent of the lithologically similar Cache Creek Group of south-central British Columbia (fig. 4), from which fossils of Mississippian(?), Pennsylvanian, and Permian ages have been reported (Cockfield, 1961, p. 9—11). Within the northern part of the study area (fig. 2), the Kobau Formation and Palmer Mountain Greenstone overlie the Anarchist Group disconformably or along a slight unconformity (Rinehart and Fox, 1972, p. 11—12, 22). The Palmer Mountain Greenstone consists of metavolcanic rocks estimated to be locally at least 2,100 m (7,000 ft) thick, consisting chiefly of greenstone and metadiabase. The formation probably originated as a pile of mafic extrusive rocks—a volcano—into which mafic magma was abundantly intruded. The Kobau Formation is composed of an unfossili- ferous sequence—locally at least 3,600 In (12,000 ft) thick—of phyllite, greenstone, and massive metachert. The lower parts of the Kobau apparently interfinger with the upper part of the Palmer Mountain, leading to the hypothesis that the Kobau originated "as a belt of marine deposits that formed satellitic to, and partly as a result of, the volcanic activity represented by the Palmer Mountain Greenstone” (Rinehart and Fox, 1972, p. 22). The maximum possible age of the Kobau and Palmer Mountain, as established by the fossils in the subjacent Anarchist, is Permian, probably Late Permian. Okulitch (1973, p. 1514) rejected this conclu- sion and suggested instead that the Kobau Formation as mapped by us is in part older than the Anarchist because the Kobau contains rootless isoclinal folds that he believed were related to a deformational event that predates the Anarchist. In his opinion (p. 1516), the Kobau is mid-Paleozoic. Our mapping (Rinehart and Fox, 1972) revealed no evidence supporting his specula- tions, and in our view the disconformable or unconform- able contact of Kobau over Anarchist is well established. In the northwest corner of the region, near Kamloops (fig. 4), the Cache Creek Group is overlain along an erosional unconformity by the Nicola Group, a sequence of volcanic rocks composed mainly of greenstone and minor sedimentary rocks containing Triassic—locally Late Triassic—fossils (Cockfield, 1961, p. 8—15). The Nicola crops out over an extensive area within south- central British Columbia. The stratigraphic positions of the Nicola over the Cache Creek and Kobau over the Anarchist thus seem comparable and suggest that these two chiefly volcanic and volcaniclastic assemblages— Nicola and Kobau—are correlative. However, Dawson (1879, p. 87B) considered the rocks now referred to the Kobau (along the route of his reconnaissance between the Similkameen and Okanogan Rivers and a few kilometers north of the international boundary) to be Cache Creek, although he recognized that the sequence contained members resembling the Nicola. In View of the relations noted above, the Kobau could be either the temporal equivalent of the Nicola—differing from the Nicola of the type area in that volcaniclastic rocks and bedded chert predominate over lava—or alternatively, older than the Nicola but not present or not recognized in the type areas of the Nicola and the Cache Creek. In the southern part of the study area, the Anarchist is unconformably overlain by the Cave Mountain For- mation (Rinehart and Fox, 1976 ), a sequence of PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON 120°00' 30' , 119°oo' (115.8). L_618 L—704 70.9 (177 49°00' . I I I t ‘ ‘ \ . \ ~0011conull /¢ 4 30‘ I I :66»; ,‘, : ,‘, 7o)-‘ I , (1 \ l \ \ I a l ’ 48 00 Base from U.S. Dept. Commerce, Sectional Geology compiled by K. F. Fox, 11., Aeronautical Chart, Kootenai, Edition of and C. D. Rinehart, 1973 Nov., 1951, revised Dec. 1968 0 5 ‘0 M'LES 0 5 1O KILOMETERS FIGURE 2.—Geologic map of Okanogan Highlands and vicinity showing radiometric ages. weakly to highly metamorphosed limestone, dolomite, siltstone, slate, and capping basaltic volcanic rock. One member contains fossils dated as Late Triassic (Misch, 1966, p. 118). These rocks are similar to a sequence of interlayered quartzite and limestone at Hedley, British GEOLOGIC HISTORY EXPLANATION SURFICIAL DEPOSITS (Quaternary) — Alluvium and glacial drift ' COLUMBIA RIVER BASALT GROUP (upper Miocene) VOLCANIC ROCKS(Eocene) — Dacitic , andesitic, and minor trachytic lava flows and hypabyssal intrusive rocks EPICLASTIC AND VOLCANICLASTIC ROCKS (Eocene) — Sandstone, graywacke, tuff, conglomerate, and shale PARAGNEISS, ORTHOGNEISS, AND ASSOCIATED GRANITIC ROCKS OF THE OKANOGAN GNEISS DOME (age of metamorphism and mobilization probably late Mesozoic) — Includes similar rocks of the Monashee Group of the Shuswap terrane of Jones (1959) in British Columbia.A, paragneiss; B, orthogneiss and associated granitic rocks GRANITIC ROCKS (Triassic to Tertiary) — Includes granodiorite, quartz monzonite, quartz diorite, diorite, monzonite, and minor nepheline syenite, shonkinite, malig- nite, alkali pyroxenite, peridotite, hornblendite, and serpentinite. Interplutonic contacts shown where known. Kruger Alkalic Complex identified by “K” EUGEOSYNCLINAL DEPOSITS (Permian to Cretaceous) —~ Greenstone, greenschist, slate, phyllite, schist, metawacke, metasharpstone conglomerate, marble, and meta- chert GRANITOID GNEISS 0F OKANOGAN RANGE (age unknown) — Trondhjemitic orthogneiss and paragneiss 7__ — ............ Contact Fault Queried where projected through unmapped areas Dotted where concealed KEY TO RADIOMETRIC AGES 120° 119° (Ages i?;%“ff§§$ff“‘°“ «’12 1 9 s 1. Bostock (1940) 49° 2. Fox (1970) Locality “2‘8 /17 3. Fox (unpublished mapping) K-Ar age \ 6 15 2 3 11 4. Fox and Rinehart (unpublished mapping) Biotite 179 5 5. Hawkins (1968) . 6. Hibbard (1971) Musw’m [5”] 7. Huntting and others (1961) Hornblende (194) 1 1 8. Little (1957) Plagioclase 51.19 3 4 9. Little (1961) Fission track age 4 10. Menzer (1964, 1970) A‘patite (53),, 16 ll. Pearson (1967) Sphene (78); 10 12. Rice (1947) . 7 13. Rinehart (unpublished mapping) Allamte 59’5”“ 14. Rinehart and Fox (1970) Zircon (90>: 15. Rinehart and Fox (1972) Epidote <63>epi 7 16. Roberts, R. J ., and Hobbs, S. W. (unpublished mapping) Rb'Sf age 489 | I 17. Staatz and others (1971) Pt” age 07°) SOURCES OF AREAL 6 E0 LOGIC DATA FIGURE 2.—Continued. 5 Columbia, originally described by Camsell (1910) that also contains Late Triassic fossils. The nomenclature and lithology of these rocks have been reviewed by Rice (1947, p. 12—14), who considers them to be part of a dominantly sedimentary facies of the Nicola. Within PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON 50° : 48 - .;. Base from US. Dept. Commerce, Navigational Chart E 16 Edition 5, 1968 Geology compiled by K. F. Fox, IL, and C. D. Rinehart, 1973 0 15 30 MILES O 15 30 KILOMETERS FIGURE 3.—Geologic map of part of northeastern Washington and southern British Columbia showing distribution of localities at which age determinations have been made. GEOLOGIC HISTORY 7 EXPLANATION SEDIMENTARY, VOLCANIC, AND LOW-GRADE METAMORPHIC ROCKS BASALT (upper Miocene and lower Pliocene) EPICLASTIC SEDIMENTARY ROCKS AND VOLCANIC ROCKS (Eocene) — Basaltic, andesitic, dacitic, rhyolitic, and trachytic lava flows, hypabyssal in- trusive rocks, pyroclastic rocks, sandstone, graywacke, shale, and conglomerate EUGEOSYNCLINAL DEPOSITS (Cambrian? to Cretaceous) — Sandstone, gray- wacke, argillite, conglomerate, lava flows, pyroclastic rock, greenstone, greenschist, slate, phyllite, schist, metawacke, metasharpstone conglomerate, marble, and meta- chert MIOGEOCLINAL DEPOSITS (Precambrian to Mississippian?) — Limestone, dolo- mite, argillite, slate, phyllite, schist, marble, quartzite, siltite, and minor greenstone and conglomerate MEDIUM- AND HIGH-GRADE METAMORPHIC ROCKS PARAGNEISS, SCHIST, AND AMPHIBOLITE (Age not specified) MONASHEE‘GROUP 0F SHUSWAP TERRANE 0F JONES (1959), AND SIMILAR ROCKS (age of metamorphism probably Mesozoic) — Paragneiss, ortho- gneiss, and associated granitic rocks, probably in part originating as diapiric intrusions. Includes rock within Thor-Odin, Pinnacles, Valhalla, Passmore, and Okanogan gneiss domes PLUTONIC ROCKS GRANITIC ROCKS (Triassic to Tertiary) ~ Includes granodiorite, quartz monzonite, syenite, monzonite, quartz diorite, diorite, granite, and minor nepheline syenite, shonkinite, malignite, alkali pyroxenite, gabbro, peridotite, dunite, hornblendite, and serpentinite O _._._. 7— —_ ............ Radiometric age locality Boundary of orogenic province Contact Fault Compiled from sources listed in tables Queried where projected through Dotted where concealed 1 and 2 unmapped areas 121° 120° ° ° 117° 51° 1 19 118 1 #1 1. Becraft (1966) 20. Monger (1968) 2. Bostock (1940) 21. Muessig (1967) 7 14 3° 3. Bostock (1941a) 22. Parker and Calkins (1964) 4. Bostock (1941b) 23. Pearson (1967) 5. Campbell and Raup (1964) 24. Reesor (1965) o 13 13‘ 6. Church (1971) 25. Rice (1947) 5° 7. Cockfield (1961) 26. Rinehart (unpublished 8. Fox (1970) mapping) 9. Fox and Rinehart 27. Rinehart and Fox (1970) 25 17 15 (unpublished mapping) 28. Rinehart and Fox (1972) 4 ,3 10. Hawkins (1968) 29. Roberts, R. J. and a“ 16 11. Hibbard (1971) Hobbs, s. w., (unpub- 33 2 20 12. Huntting and others (1961) lished mapping) 49° \ 13. Hyndman (1968b) 30. Ross (1970) w 1128 a] 23122 El 14. Jones (1959) 31. Savage (1967) 10 15. Little (1957) 32. Staatz (1964) 26 27 9 21 31 16. Little (1960) 33. Staatz and others (1971) 12 19 29 12 17. Little (1961) 34. Tabor and others (1968) 32 18. Little (1962) 35. Yates (1964) m 19. Menzer (1964, 1970) 48° 1 I l S SOURCES OF AREAL GEOLOGIC DATA FIGURE 3.—Continued. 8 PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON TABLE 1,—Sources of radiometric age data compiled in figures 2, 3, 5, and 6 Figure 2, 5 References Cannon, 1966 Engels, 1971 Engels, Tabor, Miller, and Obradovich, 1976 Fox, Rinehart, and Engels, 1975 Fox, Rinehart, Engels, and Stern, 1976 Hawkins, 1968 Hibbard, 1970, 1971 Mathews, 1964 Menzer, 1970 Naeser, Engels, and Dodge, 1970 Rinehart and Fox, 1972 Rinehart and Fox, 1976 All above, plus the following: Baadsgaard, 1961 Engels, 1975 Hills and Baadsgaard, 1967 Leech, Lowden, Stockwell, and Wanless, 1963 Lowdon, 1960, 1961 Lowdon, Stockwell, Tipper, and Wanless, 1963 Nguyen, Sinclair, and Libby, 1968 Preto, White, and Harakal, 1971 Roddick and Farrar, 1971, 1972 Roddick, Farrar, and Procyshyn, 1972 Sinclair and White, 1968 Tabor, Engels, and Staatz, 1968 Wanless, Stevens, Lachance, and Rimsaite, 1965 Wanless, Stevens, Lachance, and Edmonds, 1968 White and others, 1967 Yates and Engels, 1968 3,6 the study area (fig. 2), both the Anarchist Group and Kobau Formation were strongly folded along north- northwest axes and afterwards intruded and metamorphosed by the Loomis pluton, a quartz diorite—granodiorite batholith. The Loomis pluton is the oldest pluton so far recognized within the study area (fig. 2). K—Ar ages of 194 my and 179 my were meas— ured on coexisting hornblende and biotite (table 2), re- spectively. The older of these is regarded as a minimum age—and probably close to the true age of the batholith, in view of the modest amount of the dis- cordance. The relations within the study area (fig. 2) indicate that deposition of eugeosynclinal volcaniclastic sedi- ments, carbonate sediments, and minor intercalated lavas was terminated during the Permian. After an interval of minor tilting, and possibly of minor erosion, voluminous lavas were erupted and deposited along with associated pyroclastic sediments. This phase of volcanism was terminated in Late Triassic by uplift and strong folding, followed by batholithic intrusion about 195 my ago. JURASSIC TO PALEOCENE HISTORY The folded Kobau Formation and Anarchist Group were beveled and buried beneath the Ellemeham For- 121° 51° , 120° 119° 118° 117 so I" Anarchist yGroup Hall and Rossland Formations, Sinemuriané Beds Anarchisfl?) Group/WT” Anarchist Group ‘ Bradshaw, Barslow. ngwalo Shoemaker, Independence, 5'0”. and Old Torn Formations Nicole Group Kobsu_(?) ormanon .1 , . r . _ 49 . ‘ 35mm; _ a“ M NITED STATE Anarchist Group Mount Roberts Formation Mountain FormationV nnd Punyten 48 | Groups 1 0 50 MILES 0 50 KILOMETEHS FIGURE 4.—Late Paleozoic and Mesozoic stratified rocks of the eugeosynclinal province in northeastern Washington and south- ern British Columbia. Contrasting patterns used to aid in distinc- tion of map units. Data modified slightly from same sources as figure 3. mation, a unit composed of a basal member of greenstone and tuffaceous metasiltstone, a middle member of monolithologic greenstone conglomerate, and an upper conglomeratic member composed of heterogeneous fragments of metamorphic rock. The El- lemeham Formation is locally at least 850 m (3,000 ft) thick (Rinehart and Fox, 1972, p. 22—25). The El- lemeham is younger than the subjacent Kobau, proba- bly younger than the Late Triassic Loomis pluton, and is clearly older than 157 m.y., the age of biotite (sample L62OR, table 2) in a hornfels formed at the contact of a crosscutting alkalic intrusive body; hence it appears to be of Late Triassic or Jurassic age. With the possible exception of the Ellemeham Forma- tion, there are no known depositional units of Jurassic to Paleocene age within the study area. Rocks older than Ellemeham were eroded to levels deep enough to expose the plutonic rocks, then buried by the lavas, pyroclastic deposits, breccia, and conglomerate of the Ellemeham. Deposition was interrupted in the central and west- ern parts of the eugeosynclinal province during the Late Triassic tectonism, but on the east flank of the province, GEOLOGIC HISTORY TABLE 2.—Interpretation of radiometric age determinations in part of northern Okanogan Highlands Method Mineral Inferred age of age age Source of of event Unit Sample Mineral determinationI (m.y.) mineral age (m_y,) Remarks Swimptkin Creek 0—419 Biotite ,,,,,,,, K—Ar ______ 48.0: 1.5 Fox, Rinehart, Engels, Intrusion, rapid cooling at Swimptkin Creek pluton cuts Okano- pluton, and Stern, 1976. ~48 gan gneiss dome; contains xenoliths Hornblendenn _ do-,, ______ 48.2: 1.5 do. of lineated gneiss of dome and oflow- grade metamorphic rocks believed to have been part of roof of dome when cut by Swimptkin Creek. See NeW< comb (1937). Contains autoliths and fine-grained marginal phases. Andesite of White- L—657 49.1: 1.8 Rinehart and Fox, 1972. Volcanism, ~50. Volganjc rocks cut or overlie associated stone Mountain. sedimentary rocks that contain prob- Andesite-dacite plug L—590 51.4:2.6 do, ably Eocene flora (Rinehart and Fox, Do ,,,,,,,,,,,,,, L—147 52.1:23 do. 1972, p. 61). Although volcanic rocks Dacite of Carter C—550 45,1:20 Rinehart and Fox, 1976. frin e gneiss dome, they do not di- Mountain. rect y overlie it. Intercalated con- Marron Formation AK—112 49 Mathews, 1964, p. 465. glomerate contains abundantdetrital (volcanics). ranitic rock, which northeast of Dacite of C—559 .1: do ____________ do ,,,,,, 42.9:13 Engels, Tabor, Miller, ome resembles rocks ofgneiss dome. Twin Peaks, and Obradovich, 1976i Sanpoil Volcanics_.__ OBP—65—01 Plagioclase _________ do ,,,,,, 51.1 do Okanogan gneiss 0—36A Biotite ________ K—Ar ______ 49.3: 1.6 Fox, Rinehart, Engels, Intrusion and metamor- Samples 0—131, 0—296 of alkalic border dome. and Stern, 1976. phism in Late Cretaceous; phase (monmnite and syenodiorite) Muscovite ___e K—Ar ,,,,,, 49.83.16 do. uplift and slow cooling to at north edge of dome, inter reted Allanite ,,,,,, F.T ,,,,,,,, 59:6 Naeser, Engels, and retentivity threshold ofAr as metasomatite formed uring Dodge, 1970. in biotite at ~50; unro- emplacement of dome. Sample Apatite ...... FiT ________ 51:5 0. ofing of dome after 50. 0.176D and 0—176E are of layered 0—37A Muscovite ____ K—Ar ______ 49.0:2.2 Fox, Rinehart, Engels, paragneiss, and remaining samples and Stern, 1976. are of granodioritic gneiss. Volcanic 53.8: 1.6 do. rocks unconformably overlie country 51.2: 1.6 do, rocks dynamically metamorphosed 50.8: 1.6 do. during forceful intrusion of dome; 0.38A Biotite 54.8: 1.7 do. therefore, volcanics are younger than Allanite ,,,,,, F.T ,_ 59:2 Naeser, Engels, and age of emplacement of dome. The Dodge, 1970. 20‘5Pb/2='3U age of87.3 m.y. is the most precise of the Pb-U ages. The cata- Apatite ______ F.T ________ 53:5 do clastic textures of the rocks of the 7 gneiss dome indicate that pervasive 0—131 Hornblendeu" K—Ar ______ 58.1:1 7 Fox, Rinehart, Engels, and penetrative deformation con- and Stern, 1976. tinned well after most mineral con- Epidote ______ F.T 77777777 63:3 Naeser, Engels, and stituents including zircon had crys- Dodge, 1970 tallized. We conclude that deforma- Sphene ________ FiT ________ 66:7 do. tion and mobilization terminated in 0—176D Hornblendenu K—Ar ______ 49.3:17 Fox, Rinehart, Engels, Late Cretaceous, afizer about 87 m.y. and Stern, 1976. ago, Younger ages of other minerals 0—176E Zircon ________ 206Pb/“L‘U _ . 87.3 do. reflect slow cooling. It is less likely, in 2""Pb/“5U . _ 100.0 do, our opinion, that discordance could be 208Pb/“Wh _. 94.0 do, due wholly or in part to thermal . 0—296D Biotite ________ K—Ar _____ 50.0: 1.5 do. metamorphism of gneiss dome at ~50 Hornblende _ I___ do , 53.9: 1.6 do, m.y. 0—425 Biotite ____________ do ,,,,,, 46.0: 1.4 do. Coyote Creek pluton _,0—424B Biotite ,,,,,,,, K—Ar ,,,,,,,, 49.1: 1.5 Fox, Rinehart, Engels, ~49(?) Granodiorite p0 hyry of Coyote Creek and Stern, 1976 pluton boun s Okanogan gneiss dome on southeast, not delineated separately in figure 2. Probably younger than gneiss dome, since only slight cataclasis apparent, but it is possible that granodiorite porphyry is marginal phase grading inward to gneissic quartz diorite of gneiss dome. Conconully pluton __ C—555 Biotite ________ K—Ar ______ 62.5:2.2 Rinehart and Fox, 1976. Intrusion prior to 80, proba- Menzer (1970, p. 576) concludes that 72~8¢45 d0. bly at about 90- L063] this event is 81 m.y. old, relying on T—156 78.8:2.4 Engels, Tabor, Miller, thermal metamorphism at the Rb-Sr isochron. However, the F.T, and Ohradovich, 1976 or subsequent to 63. and Pb-a ages average 89, suggesting 812:2.4 do. that the pluton is older than 81. The OK—l 89:9 Menzer (1970). discordance between the hornblende 84:8 do. and biotite K—Ar ages at C—555 and 90:20 do. T—156, and between hornblende at OK—4 Apatite 94:12 do. C—555 and T—156, and between OK—l Apatite, hornblende at C—555 and the mineral biotite. ages of OK—l and OK—4, probably re- 0K—4 Whole rock, Rb—Sr 81.1:0.8 do. flect metamorphism by a later ther‘ muscovite, mal event. A steep east—west gradient biotite. in discordancy is required between 0555 and OK—l, and between C—555 and T—156 (fig. 2), suggesting meta- morphism by a nearby source to the east concealed in the subsurface. Trondhiemitic gneiss T—153 Biotite ________ K—Ar ,,,,,, 88.5:2.7 Engels, Tabor, Miller, Intrusion prior to intrusion Trondhjemitic gneiss is enclosed in,and of Tifl‘any Mountain and Obradovich, 1976. ofOld Baldy p1uton(129 or is thus older than Old Baldy pluton Homblende ,,,,,,,, do ,,,,,, 93,5:2.8 do. before) ("granodioritic gneiss,” Menzer, T—154 Biotite _._ do, 91.7:2.8 do. 1970, p. 576). Discordancy ssibly Muscovite _ do" 88.5:2.7 do. due to cumulative metamorphism re- T—155 Biotite ._ _ do" 108:3 do. lated to intrusion of Conconully plu- Muscovite __________ do ______ . . do. ton about 90 m.y. ago and to Late Cretaceous and early Tertiary ther- ("Trondhjemitic 0K—3 Apatite Menzer (1970). mal event. Menzer ascribes its gneiss"; Menzer, Zircon fl d0. anomalously younger isotopic age, 1970) 0K—6 Apatite d0. relative to Old Baldy pluton, as pos~ Zircon ,,,,,,,,,,,,, do ______ do. sibly due to "penetrative rock defor- OK—3 Whole rock, mation that pervades the unit" (p. biotite. } Rb—Sr 1042:05 do. 577). Thus far, however, our observa- OK-6 A." do 7777777 isochron tions reveal no consistent differences between the two units in age of de- formation, its intensity, or its style. 10 PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON TABLE 2.—Interpretation of radiometric age determinations in part of northern Okanogan Highlands—Continued Method Mineral Inferred age of age age Source of of event Remarks Unit Sample Mineral determination‘ (m.y.) mineral age (m.y.) Old Baldy pluton __-1 T—152 Biotite ,,,,,,,, KAAr ______ 90.1:2.7 Engels, Tabor, Miller, Intrusion at 129 or before. The Old Baldy pluton is cut by the Con- and Obradovich, 1976. conully pluton. Discordancy due to Hornblende ,,,,,,,, do ...... 98.5:30 do, cumulative metamorphism including ("Granodioritic OK—5 Sphene ,,,,,,,, F.T ________ 78:8 Menzer (1970). that associated with intrusion of Con- gneiss" of Menzer, conully pluton about 90 m.y. ago and 1970), Whole rock, Rb—Sr 129.0: 1.8 do. that due to Late Cretaceous and early biotite. isochron. Tertiary regional thermal event. Cathedral batholith .1 H—l Biotite ________ KAAr ,,,,,, 97.7:29 Engels, Tabor, Miller, Intrusion at ~95(?). K-Ar ages of biotite provisionally ac- and Obradovich, 1976. cepted, since no younger rocks known (2) .7" do ,,,,,,,,,,,, do” _ 94312.8 Hawkins, 1968. in vicinity, and sample locations ap- (2) -7 98 Misch (cited by Hawkins, pear to be west of area of influence of 1968). Late Cretaceous and early Tertiary regional thermal event. Anderson Creek L—451C Biotite ________ K—Ar ______ 100.1130 Engels, Tabor, Miller, Intrusion at ~115. Intrudes Loomis pluton (Rinehart and pluton. and Obradovich, 1976. Fox, 1972). No younger plutons Hornblende ........ do ______ ll5.6¢3.0 do, known in vicinity; therefore, dis- AC39 Biotite ____________ do ______ 106.5268 Hibbard, 1971 cordancy between biotite and hornblende ages attributed to Late Cretaceous and early Tertiary re- gional thermal event. Aeneas Creek 0—139 Biotite ,,,,,,,, K—Ar ,,,,,, 92.7:66 Rinehart and Fox, 1976. Intrusion prior to 93. Well within area of influence of Late pluton. Cretaceous and early Tertiary re- gional thermal event; therefore, bio- tite age probably substantial] less than actual age ofintrusion ofp uton. Evans Lake pluton e. C—554 Biotite ,,,,,,,, K—Ar ,,,,,, 88.8:2,8 Rinehart and Fox, 1976. Intrusion before 89. Well within area of influence of Late . Cretaceous and early Tertiary re- gional thermal event; therefore, bio- tite age probably substantially less than actual age ofintrusion ofpluton. Osoyoos pluton ...... Osoyoos #2 Biotite ________ K—Ar ,,,,,, 49.0: 1.5 Fox, Rinehart, Engels, Intrusion before 50 m.y.ago. Well within area of influence of Late and Stern, 1976- Cretaceous and early Tertiary re— gional thermal event; therefore, biot- ite age probably substantially less than actual age of intrusion of pluton. Dynamic metamorphism along south border attributed to intrusion of Okanogan gneiss dome, therefore older than dome. Blue Goat pluton .,,, 0246B Biotite ,,,,,,,, KvAr ______ 990:3.0 Rinehart and Fox, 1976. Intrusion at or before 142. Discordance attributed to metamorph- Hornblende ,,,,,,,, do ...... 141.8:82 o. ism during Late Cretaceous and early Tertiary regional thermal event. Quartz diorite gneiss OK72 Apatite ______ F.T Menzer, 1970. Intrusion before 170. Discordance attributed to cumulative (Menzer, 1970). Sphene 7" """" ' "‘ do. effect of metamorphism during Late Zircon """"" Pb—a """"" do, Cretaceous and early Tertiary re- Whole "wk ‘10- gional event and metamorphism as- biotite. isochron. sociated with intrusion of Conconully pluton about 90 my. ago, Menzer im- plies that this rock is a marginal facies of a granodioritic gneiss unit (1970, p. 573); latter is Old Baldy plu- ton of this paper. Toats Coulee L591 Biotite ________ K—Ar ,,,,,, 151:5 Rinehart and Fox, 1972. Intrusion between 170 and Intrudes Loomis pluton (Rinehart and pluton. 195. Fox, 1972). Discordance attributed to Hornblende eeeeeeee do ,,,,,, 170:5 do. cumulative effect of metamorphism during Late Cretaceous and early Tertiary regional thermal event and metamorphism associated with in- trusion of Cathedral batholith about 95 m.y. ago. if the age of the Slocan Group overlaps the Late Triassic Frebold and Little, 1962, p. 3-9), whereas those on the and Early Jurassic as indicated by Little (1960, p. 56— west side (Ladner and Dewdney Creek Groups) were 57), deposition there must have been roughly concur- primarily marine volcaniclastic sediments derived from rent with deformation and plutonism in the central erosion of the Nicola terrane in the central part of the and western parts—possibly contemporaneous with dep- region (Coates, 1970, p. 150—151) and deposited in a osition of the Ellemeham Formation in the study area. subsiding fault-bounded trough (White, 1959, p. 77), the In Middle and probably Late Jurassic time, however, Methow graben. deposits were accumulating on both flanks of the prov- Sedimentation in the western part of the eugeosyn- ince. Those on the east side were primarily lavas and clinal province within the region (fig. 3) continued related rocks (Rossland Group; Little, 1960, p. 62—71; through the Early Cretaceous marked by the appear- GEOLOGIC HISTORY 11 TABLE 2,—Interpretation of radiometric age determinations in part of northern Okanogan Highlands—Continued Method Mineral Inferred age of age age Source of of event Remarks Unit Sample Mineral determination' (m.y.) mineral age (m.y.) Shankers Bend L—620R Biotite ,,,,,,,, K—Ar ,,,,,, 157.4247 Engels, Tabor. Miller, Intrusion before 157. Sample of hornfelsed rock of El- alkalic complex. and Obradovich, 1976. lemeham Formation at contact; mid-Jurassic age surprising, since locality is well within area influ- enced by Late Cretaceous and early Tertiary regional thermal event. El- lemeham Formation probably Upper Triassic, Lower Jurassic, or Middle Jurassic. Similkameen com» L—277Y Biotite ________ K-Ar ______ 74.8:22 Fox, Rinehart, and En- Intrusion between 191 (iso- Thebatholith is zoned, gradingoutward posite pluton, (hornfelsed gels, 1975. chron age of hornblendes; from quartz monzonite and Koban at Hornblende ,,,,,,,, do ,,,,,, 166.7:50 do. see Fox, Rinehart, and granodiorite to monzonite. The mon- contact) Engels, 1975) and 177; zonite in turn grades outward to (Comprises Similka- L—277W Muscovite ,,,,,,,,,, do ,,,,,, 135,7:4.1 do. thermal metamorphism malignite and shonkinite of the bor- meen batholith (peg'matite- between 70 and 50. dering Kruger Alkalic Complex. This (Daly, 1906, 1912) alaskite do, gradation suggests the batholith and and Kruger Al- dike cuts alkalic complex are roughly coeval. kalic Complex pyroxenite The complex sharply crosscuts wall- (Daly, 1906, 1912; phase) rock of the Kobau Formation, of Campbell, 1939; L—277Z Biotite 777777777777 do 777777 1403242 do Permian or Triassic age. Discordance FOX; Rinehart, and (pyroxenite 7 , attributed to Late-Cretaceous or Engels, 1975). phase) early Tertiary regional thermal L—589C ..__ do ____________ do ,,,,,, 83,4:25 do_ .metamorphlsm‘ (granodiorite interior Hornblende ,,,,,,,, do ,,,,,, 155.5:4.7 do. phase) b301 Biotit/e ____________ do ,,,,,, 69.9:2.1 Engels, 1971. (shonkinite alkalic Hornblende ________ do ,,,,,, 170.9251 do. border (hastingsite). phase) 14-618 Biotite ____________ do ,,,,,, a7091-21 do. (granodiorite, interior Hornblende ,,,,,,,, do ______ 3 177.2253 do. phase) L_704 ,,,, do ,,,,,,,,,,,, do ______ 115.8136 Fox, Rinehart, and En— (granodiorite, gels, 1975, interior do. phase) W—65—1 Hornblende- e," do ______ 152:9 Cannon, 1966, (Kruger augite syenite; mixture. shonkinite from al- kalic bor— der phase) Loomis pluton ,,,,,, L—498A Biotite ,,,,,,,, K—Ar 179:5 Rinehart and Fox, 1972 Intrusion about 195 m.y. Discordance attributed to cumulative Hornblende ________ do ,,,,,, 194:6 do, effect of metamorphism during Late Cretaceous and early Tertiary re- gional thermal event, and to metamorphism associated with in- trusion of the Toats Coulee between 195 and 170 m.y. ago. and the meta- morphism associated with intrusion of Cathedral batholith about 95 m.y. ago, Chopaka Intrusive CG10 Actinolitic K—Ar ,,,,,, 19052156 Hibbard, 1971. Metamorphism about 195 Probably the age of metamorphism as- Complex of Hibbard hornblende. m.y. ago. sociated with intrusion of the nearby (1971). Loomis pluton. ‘F.T., fission track. 2Sample number unknown. “Corrected for cross-contaminationl of biotite and hornblende in mineral separates (see Engels, 1971). ance of material eroded from sources west of the basin as well as from the east and by the appearance of granitic detritus (Coates, 1970, p. 151). The central part of the province, which includes the study area (fig. 2), was apparently positive and was the site of continued plutonism through the Jurassic and Early Cretaceous. By mid-Cretaceous most of the province had become positive, and the area of plutonism had spread eastward and westward through the entire extent of the province within the region (fig. 3). Within the study area (fig. 2), the Late Triassic Loomis pluton is cut by the Toats Coulee pluton on the west and the Anderson Creek pluton on the east. Coexisting hornblende and biotite from the Toats Coulee were dated at 170 m.y. and 151 m.y., respec- tively, and coexisting hornblende and biotite from the Anderson Creek were dated at 115.6 m.y. and 100.1 m.y., respectively (sample L—45lC, fig. 2 and table 2). The discordance between ages of coexisting minerals in other plutons increases dramatically to the north and 12 east of the Loomis. The discordance between hornblende and biotite ages from some samples of the Similkameen batholith and the coeval Kruger Alkalic Complex is as much as 106 m.y., hornblende being in- variably the older of the mineral pair. In some samples of the Similkameen and Kruger, hornblende shows maximum ages of 155 to 177 my (table 2). In fact, all the other plutons within the study area (fig. 2) from which coexisting minerals have been dated show significant discordance, except for the Swimptkin Creek pluton of Eocene age, which yielded hornblende and biotite ages of48.2 my and 48.0 m.y., respectively. The data summarized in table 2 suggest that the plutonism in the study area, which had begun with intrusion of the Loomis pluton, continued sporadically through the Jurassic, Cretaceous, and early Tertiary. DISCORDANT AGES The discordancy between apparent ages of coexisting minerals reaches a maximum in a zone that borders the Okanogan gneiss dome on the west (fig.5). The Okanogan gneiss dome occupies 2,500 km2 (950 mi?) within the study area (fig. 2). Rocks of the gneiss dome previously have been considered—at least in part—as part of the Colville batholith (Pardee, 1918; Waters and Krauskopf, 1941; Staatz, 1964). The gneiss dome consists chiefly of layered paragneiss, augen gneiss, and granodiorite, marked over most of the area by a penetrative cataclastic fabric whose chief elements are a very regular low-dipping foliation and a persistent wes‘t-northwest-trending lineation. Country rocks ad- jacent to the dome at the western, northwestern, and southwestern contacts have been crushed and broken, and rocks of the dome at and near its upper surface along the western contact are mylonitized. The mylonitization and gneissic fabric of the gneiss dome and the brecciation of the wallrock were attrib- uted by Waters and Krauskopf (1941) to emplacement of the mass as a protoclastic batholith. However, Snook (1965, p. 775) concluded that these features resulted from development of a distributed flat thrust at depth within a regionally metamorphosed sedimentary or volcanic terrane that was later upfaulted against the lower grade metamorphic rock on the west. We have suggested a third alternative—that the gneiss dome formed through metamorphism of volcanic and sedimentary strata at depth, a metamorphism that culminated in the mobilization and diapiric emplace- ment of the dome (Fox and Rinehart, 1971). Ages from within the gneiss dome are themselves discordant, with the 21 age determinations previously reported (Fox and others, 1976) ranging from 100 my (2°7Pb/235U, zircon) down to 46.0 m.y. (K—Ar, biotite). PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON WEST EAST ° 1‘ _ c _ m 50. .‘ . .. I . ,_- m ,. < — 1:1 ~ LU . _ 1 - IL _ _ ° I m — I _ z . 9 ,oo_. 1' I _ .1 _ ' I _ :' . 2 _ _ g _ _ ui _ _ 0 I < 150 - [ . _ 200 O 10 20 MILES l—I—l_l—1 o 10 20 KILOMETERS FIGURE 5.—Plot of age determinations (exclusive of uranium- thorium-lead ages of sample 0—176E, table 2) within study area (fig. 2) projected to east-west transect. Ages of coexisting minerals connected by vertical lines. Rb—Sr ages disregarded where less than fission-track ages of coexisting apatite, sphene, or zircon. Westward extremity of Okanogan gneiss dome at point C. The terminal stages of metamorphism and emplace- ment of the dome are believed to have been Late Cretaceous, after which the dome cooled slowly through the successive temperature thresholds (block- ing temperatures) of the dated minerals, which in- cluded zircon, sphene, epidote, allanite, hornblende, apatite, muscovite, and biotite. The possibility that these minerals were variably degraded as a result of later mild reheating of the dome during a thermal event associated with regionwide Eocene volcanism can be evaluated through comparison of age data from elsewhere within the region. A compilation of age data shows that Tertiary ages are abundant within the three orogenic provinces rep— resented within the region (figs. 1, 3)—from east to west, the Purcell foldbelt, the Omineca crystalline belt, and the Columbian intermontane belt. However, pre- Tertiary ages, which are abundant within both the Purcell foldbelt and the Columbian intermontane belt, are strikingly rare in the Omineca crystalline belt. The contrast is apparent in figure 6, a plot of ages projected to an idealized east-west transect across the region shown in figure 3. GEOLOGIC HISTORY 13 o FOMINECA CRYSTALLINE BELTé: 121 1170 - | | - .., . | - . l g o p I: ' .' u 'I o I . : .. l . 3 < 50— ‘- , ‘f‘ "P .5 .0. ' ‘5'. ’ t' 3: . n.‘ I on»? ' ’ ’ — 50 u: ' ' ‘1' z >- ’ O l . 2 l s . ’= . | " | "- I o». o té’wo—x. ° -° .1 I l ' '3' ' ‘.”‘°° 9 ' ' . | | .l :' ' | | i 150 _ . a . 0 : I l l— 150 g,- f’" ' I I <9 ' ° . : -- ‘ ,. ' | < - ~:’ ‘ - | l 200 ff“. '1'? . l I — zoo ' I | o 15 30 MILES I-fi—J—I—‘ o 15 so KILOMETERS FIGURE 6.—Mesozoic and Cenozoic age determinations within region shown in figure 3 plotted with respect to the approximate boundaries of the Omineca crystalline belt and projected to the 49th parallel. Apparently the zone of discordant ages parallel to the west side of the Okanogan gneiss dome (fig. 5) is part of a regional pattern of discordance associated with the Omineca crystalline belt. AGE OF OMINECA CRYSTALLINE BELT The age of the Omineca crystalline belt refers to the ages of the lithologic elements by which it is defined— the Monashee Group of the Shuswap terrane of Jones (1959) (and correlative rocks) and the gneiss domes included within that terrane. Three ages must be con- sidered: (1) the age of the parent material; (2) the age of metamorphism, deformation, and mobilization; and (3) the age of cooling through the blocking tempera- tures of the dated minerals. The Okanogan gneiss dome is one of a family of gneiss domes located within the Shuswap—or Shuswap-like—terrane of the Omineca. These include, from north to south, Frenchman’s Cap (Wheeler, 1965), Thor-Odin (Froese, 1970; Reesor and Moore, 1971), Pinnacles (Reesor and Froese, 1968), Valhalla (Reesor, 1965), and Passmore (Reesor, 1965) gneiss domes (figs. 1, 3). As redefined by Jones (1959), the Shuswap, in addi- tion to the Monashee Group, also includes on the west flank two outliers of low-grade metamorphic rocks named the Mount Ida Group and the Chapperon Group. Their lower metamorphic grade and lack of evidence of profound tectonic involvement set them in marked contrast with rocks of typical Monashee ter— ranc, and for that reason they are here considered part of the Columbian intermontane belt, rather than part of the Omineca. The Mount Ida Group is also litholog- ically dissimilar to the Monashee Group (Jones, 1959, p.30). Thus, the Mount Ida and Chapperon Groups may not share a common age and origin with the Monashee Group. Hence, age constraints imposed on the Mount Ida and Chapperon Group do not necessar- ily apply to the Monashee. We refer to the K—Ar biotite ages of 135, 140, and 140 my (samples GSC 61—1, 2, 3; Lowdon and others, 1963) from rocks of the Mount Ida Group and to the fact that the Mount Ida and Chappe- ron both are unconformably overlain by the Permian Cache Creek Group—the Mount Ida at the Glenemma(?) unconformity (Jones, 1959, p. 48) and the Chapperon at the Salmon River and Dome Rock unconformities (Jones, 1959, p. 28, 29)—thus placing firm younger age limits on both groups. Again, these restrictions do not place a younger limit on either the age of the parent material or the age of metamorphism and mobilization of the gneiss domes or the Monashee Group. The material composing the Frenchman’s Cap dome lithologically resembles part of the late Precambrian, Cambrian, and post-Cambrian succession (Wheeler, 1965, p. 9), as does that of the Thor-Odin (Froese, 1970, p. 173; Reesor and Moore, 1971, p. 113). Reesor (1970, p. 85) suggests that Pinnacles dome may be composed in part of rocks equivalent to the Milford Group (Mis- sissippian to Triassic?), and Giovanella (1968) suggests 14 that the Malton gneiss may include small bodies of late Precambrian Kaza rocks. Reesor (1970, p. 85) notes that changes in sedimentation from Windemere pelite, grit, and pebble conglomerate to Cambrian quartzite, pelite, and limestone can be tentatively identified by parallel changes in composition of rocks within the gneiss domes. Thus, the age of the material composing some of the domes appears to be late Precambrian to late Paleozoic. The age of mobilization and emplacement of the domes has not been narrowly bracketed. Rocks refer- red to the Shuswap Series east of the Frenchman’s Cap dome in the “Clachnacudainn Salient” are cut by a stock giving a K—Ar biotite age of 110 my (Wheeler, 1965, p. 15), and north-trending joints in the dome it- self are occupied by lamprophyre dikes dated at 41 my (Wheeler, 1965, p. 16). Unconformable contacts of the Cache Creek Group over the Monashee Group at Lavington and at B. X. Creek were reported by Jones (1959, p. 47—48), but the contacts have been reinterpreted as probable faults by Preto (1965). The Monashee Group, however, is patch- ily overlain by early Tertiary epiclastic deposits and volcanic rocks (Jones, 1959, p. 52). The Monashee Group has been contact metamor- phosed by a 69-m.y.-old pluton in the Nakusp area (Hyndman, 1968b, p. 68—69). Hyndman (1968a) con— cluded that the age of deformation of the Monashee Group in the Nakusp area was the same as that of the adjacent rocks of the Triassic Slocan Group because the orientation of the most prominent axis of folding of schistosity and cleavage is similar in both groups. Hyndman noted that the foliation in the Slocan pre- dates contact metamorphism by Cretaceous plutons and therefore concluded that the age of this foliation, and hence the age of metamorphism and deformation of both the Slocan and Monashee Groups, is Jurassic. A contrary view—namely that the main deformation of the Monashee Group did not involve the Slocan Group——has been put forward by Ross (1970). He suggests that in the Kootenay arc three phases of deformation can be discerned, each producing morpho- logically distinct folds and associated foliation and lineation. The earliest folds are isoclinal and formed easterly verging allocthonous nappes cored with Shuswap gneiss. Since in his opinion these folds affect only pre-Slocan rocks, he believes they are of pre- Triassic age. They were refolded during a second phase of deformation, also of pre-Triassic age, and again dur- ing a third and final phase of deformation. The folds produced in the third phase of deformation are open structures that involve rocks in the Slocan Group, as well as older rocks. Since the deformed Slocan is in- truded by the undeformed Nelson batholith, isotopi- PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON cally dated at about 164 my (Nguyen and others, 1968), this third phase of deformation is bracketed be- tween Late Triassic and Early Jurassic. Ross attri- butes the refolded isoclinal folds of rocks of the Slocan Group, reported by Fyles (1967, p. 34—35) and Hyndman (1968a), to local deformation associated with emplacement of the Kuskanax and Nelson batholiths. The K—Ar ages of 17 biotite samples from the Val- halla gneiss dome range from 11 to 66 my (Reesor, 1965, p. 51), and K—Ar ages of both biotite and hornblende from the core zone of the Thor-Odin gneiss dome range from 60 to 70 my. (Reesor, 1970, p. 86). K—Ar ages of 15 samples of the Shuswap Series (Monashee Group) reported by the Geological Survey of Canada are as follows: Age Dated Sam le In my. mineral num er Reference 62 Biotite GSC 60—1 Lowdon, 1961 57 Do. 61—4 Lowdon, Stockwell, Tipper, and Wanless, 1963 52 Do. 61—5 Do. 71 Do. 61—6 Do. 102 Do. 61—7 Do. 81 Muscovite 61-8 Do. 89 Biotite 62—36 Leech, Lowden, Stockwell, and Wanless, 1963 64 Do. 62—35 Do. 70 Do. 62—44 Do. 65 Muscovite 62—45 Do. 73 Do. 62—46 Do. 76 Biotite 62—47 Do. 81 Do. 62—48 Do. 61 Hornblende 66—43 Wanless, Stevens, Lachance, and Edmonds, 1968 79 Do. 66—44 Do. The apparent cooling age of the Monashee Group thus spans an interval from middle Cretaceous to Eocene, overlapping the cooling age of the gneiss domes. However, the average of the ten biotite ages listed above of about 72 my significantly exceeds that of the six biotite ages from the Okanogan gneiss dome (table 2), which range from 46 my to 56.8 my. and average about 50 my One striking feature of the Omineca crystalline belt is the contrast in metamorphic grade between the Monashee and Monashee-like rocks and the neighbor— ing Permian and Triassic metavolcanic and metasedimentary rocks. Jones (1959, p. 130) notes that “unmetamorphosed Cache Creek rocks in many places lie in direct contact with sillimanite gneisses of the Monashee Group***.” The transition from fine-grained greenchist facies country rock to coarsely recrystal- lized amphibolite facies rocks of the Okanogan gneiss dome is also abrupt. Rocks typical of the Okanogan dome near this con- GEOLOGIC HISTORY tact are mylonitized quartz-oligoclase-orthoclase- biotite gneiss or diopside-hornblende-plagioclase— biotite gneiss, or granodiorite. Alumino-silicate min— erals are rare, but locally, toward the interior of the dome, the gneiss is composed of sillimanite, cordierite, orthoclase, biotite, garnet, quartz, and plagioclase and in places contains accessory muscovite. This as- semblage is referable to the sillimanite-cordierite- muscovite-almandine subfacies of the amphibolite facies, Abukuma-type facies series (Winkler, 1967, p. 121), indicating metamorphism under moderate pressures, estimated by Winkler (1967, p. 187) as 3—5 kilobars. Country rocks at the west contact show little evi- dence of contact metamorphism, a point emphasized by Waters and Krauskopf (1941, p. 1376). However, there is a metamorphic halo from 100 m (110 yd) to at least 2 km (1.2 mi) wide bordering the dome on the north and northeast, within which garnet, biotite, and staurolite are present. Nevertheless it is clear from the contrast in metamorphic grade at the contact that the rocks of the dome could not have been metamorphically formed in situ. The rocks 10 km (6 mi) west of Curlew (fig. 3) are an exception to the above generalization that there is a sharp contrast in metamorphic grade at contacts of the Shuswap (Monashee Group) and its country rock. Parker and Calkins (1964, p. 5—24) described these rocks as the metamorphic rocks of Tenas Mary Creek and suggested their correlation with the Shuswap (Monashee Group). The Tenas Mary Creek rocks form a homoclinal sequence estimated by them to be about 5,200 In (17,000 ft) thick that dips northward about 25°. The sequence shows a gradual stratigraphic upward de- crease in degree of metamorphism from orthoclase- quartz-oligoclase gneiss at the base, through sillimanite-, cordierite-, and muscovite-bearing schist and gneiss, to sericite-, chlorite-, and biotite-bearing phyllite at the top, where the sequence is conformably overlain by Permian to Triassic greenstone. Presuming that the processes of metamorphism and penetrative deformation were at least in part simul- taneous with emplacement of the Okanogan gneiss dome, the contact relations with the Triassic Cave Mountain Formation and the Eocene sedimentary and volcanic rocks place older and younger limits on the age of these processes. The cataclastic texture preva- lent throughout much of the dome suggests that these processes continued well after most mineral con- stituents, including zircon, had crystallized. Thus, the 87-m.y. (2°6Pb/238U) age of zircon from one sample (table 2) is regarded as an older limit on the terminal age of penetrative deformation and mobilization of the gneiss (Fox and others, 1976). 15 The distribution pattern of the Eocene volcanic rocks, which are Widely scattered over the entire region shown in figure 3, does not coincide with that of the Omineca crystalline belt or its bordering zone of highly discordant ages. Therefore we see no correlation between the present outcrop areas of the Eocene rocks and the belts of discordant ages and conclude that the thermal event presumably associated with the forma- tion of these volcanic rocks is not a factor, except perhaps locally, in the origin of the discordancies. Thus, the Shuswap-like rocks of the Omineca crystal- line belt probably cooled through the successive block- ing temperatures for the minerals dated in Late Cretaceous and early Tertiary as a result of uplift and progressive unroofing by erosion. The many gneiss domes and the Shuswap (Monashee) are products of an intense regional defor- mational and metamorphic event that was localized along the Omineca crystalline belt. It follows that the discordance in the ages of plutonic rocks bordering the Omineca is at least in part a manifestation of thermal degradation caused by reheating during this deforma- tional and metamorphic event. Locally, however, the effect of contact metamorphism by younger plutonic rocks seems to be a factor in the origin of the discord- ance. For example, east of the Omineca crystalline belt in northern Washington and Idaho, the biotite and muscovite ages of pre-Tertiary plutonic rocks show increasing degradation towards areas of extensive Eocene plutonic rocks (Miller and Engels, 1975). The Omineca metamorphic event could have begun in British Columbia prior to 110 my. ago, if the K—Ar age determination of biotite in the stock cutting the Shus- wap in the Clachnacudainn Salient reported by Wheeler (1965, p. 15) is accurate. In the Okanogan area, that event persisted into the Late Cretaceous, when both the mobilized elements of the metamorphosed rocks and also a thermal front of sufficient intensity to affect the K—Ar age of some minerals in the rocks of the terrane west of the belt reached high levels in the crust. The metamorphic event had been concluded and the high-grade metamorphic rocks had cooled through the blocking temperature of biotite by the end of the Cre- taceous in British Columbia and by Eocene in Washington. EOCENE TO MIOCENE HISTORY The Tertiary volcanic rocks of the region (fig. 3) have been shown by Mathews and Rouse (1963), Rouse and Mathews (1961), Mathews (1964), and Hills and Baadsgaard (1967) to include two groups, one Eocene and the other late Miocene and early Pliocene. Most of the Eocene volcanic rocks within the region were 16 formed during a thermal episode climaxing about 50 m.y. ago. The Eocene episode began with roughly contem- poraneous deposition in local basins of arkose, wacke, and conglomerate. At the base, these deposits consist of quartzo-feldspathic material eroded from nearby sources, but higher in the section these deposits contain pyroclastic material in increasing proportions. The sedimentary beds are typically overlain by pyroclastic rocks and lava flows and intruded by the hypabyssal intrusive equivalents of the volcanic rocks. The Eocene rocks were deposited on a profound angular unconfor- mity beveled on older rocks. Eocene volcanic rocks in the southeastern and north- eastern parts of the study area within Wasington in- clude both the flows of quartz latite, rhyodacite, dacite, and interbedded tuff of the Sanpoil Volcanics and their rhyodacitic and quartz latitic hypabyssal intrusive equivalents, the Scatter Creek Formation (Muessig, 1962; Staatz, 1964; Parker and Calkins, 1964; Muessig, 1967). Coextensive rocks in British Columbia include andesite, trachyandesite, and sodic trachyte of the Marron Formation (Monger, 1968). Correlative vol— canic rocks to the west include dacitic and andesitic flows and their hypabyssal intrusive equivalents, which are patchily distributed along the axis of the Okanogan Valley and also occupy a large tract farther west underlying the much dissected summit plateau of the Okanogan Range. A fossil flora in sedimentary deposits cut by intrusive equivalents of the volcanic rocks northwest of Oroville is regarded as indicating a probable early Eocene age by J. A. Wolfe (in Rinehart and Fox, 1972, p. 61). The fossil flora of the O’Brien Creek Formation, a volcaniclastic unit that underlies the Sanpoil Volcanics near Republic (fig. 3), a few kilometers east of the study area, resembles the Eocene flora of Alaska and of the lower part of the Puget Group of Washington, whereas the fossil flora of the Klondike Mountain Formation, which overlies the Sanpoil, is Oligocene and Miocene(?) (R. W. Brown, in Muessig, 1967). K—Ar ages of volcanic rocks Within the study area, previously reported by Rinehart and Fox (1972), Mathews (1964), and Engels, Tabor, Miller, and Obradovich (1976) range from 42.9 to 52.1 m.y. In the Republic area, Eocene deposition was accom- panied by penecontemporaneous faulting and subsid- ence of a north-northeast-trending graben (Muessig, 1967, p. 95—96). There and elsewhere in the region the Eocene rocks were folded and faulted, partially eroded, and later buried in places by basalt during the late Miocene. The late Miocene and early Pliocene rocks include the plateau basalts of the Columbia River Basalt Group, which overlaps the south edge of the PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON region, and its erosional outliers, and other smaller patches of basalt distributed over the remainder of the region. OROGENIC IMPLICATIONS GEOLOGIC HISTORY PREVIOUS VIEWS ON OROGENY IN THE CORDILLERA Phanerozoic orogeny in the Cordillera was visualized by Gilluly (1965) as a virtually continuous, albeit loc- ally episodic, process. King, while holding that in gen- eral orogeny “was concentrated in a succession of episodes during each orogenic phase***” (1969b, p. 45), concluded that in the northern Cordillera (north of the 49th parallel), Mesozoic time "was one of nearly con- tinuous orogeny from place to place in the eugeosyn- clinal area***” (1969b, p. 67—68). King concludes that climactic middle Mesozoic orogeny in the central Cor— dillera (for example, the Sierra Nevada) was in Late Jurassic but was either earlier or later in other parts of the eugeosynclinal belt. Also, it was "progressively younger eastward across the fold belt, and in the miogeosynclinal area there is no clear separation be- tween a middle Mesozoic orogeny and a terminal Mesozoic orogeny; most of the events were at inter- mediate times***” (1969b, p. 70). White (1959) attributes Mesozoic and early Cenozoic deformation in British Columbia to three orogenies: the Cassiar orogeny, which occurred between Permian and Late Triassic (p. 72); the Coast Range orogeny, whose effects first appear in the earliest Jurassic, culminate in the Early Cretaceous, and persist well into the Late Cretaceous (p. 78); and the Rocky Mountain orogeny, which culminated in Paleocene (p. 98). White later modified these views, stating (White, 1966, p. 187) that ”much of Mesozoic was a time of what might be termed, ‘continuous-intermittent orogeny,’ when orogenic events of one sort or another were happening in one part or another of the Western Cordillera throughout a span of perhaps 150 million years.” He further states (p. 189) that “it is doubtful that Tertiary orogeny was separated from Mesozoic orogeny by any significant period of crus- tal stability ***.” Within the study area, the Cassiar orogeny may be responsible for the unconformable con- tacts ofKobau over Anarchist (Rinehart and Fox, 1972), but the degree of deformation is apparently slight, probably of epeirogenic rather than orogenic propor- tions. No pre-Late Triassic plutonic rocks have been discovered. LATE TRIASSIC OROGENY AND PLUTONISM The Upper Triassic Nicola Group is cut by the Guichon batholith (White, 1959, p. 87), now known to be about 200 m.y. old (White and others, 1967). Within the study area (fig. 2), strongly folded rocks of the Kobau OROGENIC IMPLICATIONS OF GEOLOGIC HISTORY Formation, a probable correlative of the Nicola Group, are similarly cut by the Loomis pluton, believed to be about 195 m.y. old (Rinehart and Fox, 1972; this report, table 2). Thus, orogeny (as defined in Gary and others, 1972, p. 500) probably occurred simultaneously in southern British Columbia and northern Washington during Late Triassic. This event evidently has no chronological counterpart in the Sierra Nevada or in the Klamath Mountains of California. The earliest comparable Mesozoic event in the Sierra Nevada was in Early or Middle Triassic, and the next was in Early or Middle Jurassic (Evernden and Kistler, 1970). Mesozoic plutonism in the Klamath Mountains began in early Middle Jurassic, where it ushered in the Nevadan orogeny (Lanphere and others, 1968, p. 1047, 1050). The ages of plutonic rocks in the Okanogan region suggest that the eugeosynclinal province underwent sporadic plutonism without a clearcut break between the Late Triassic and early Tertiary. However, the plutonism subsequent to Late Triassic and prior to emplacement of the Okanogan gneiss dome cannot be linked with orogenic deformation in the Okanogan region by means of the evidence at hand. ASSOCIATION OF DEFORMATION OF THE MONASHEE GROUP AND GENERATION OF GNEISS DOMES WITH CORDILLERAN THRUST FAULTING The most important thermal and deformational— hence, orogenic—event within the region was the metamorphism climaxed by mobilization of large seg- ments of the lithosphere within and adjacent to the Omineca crystalline belt. Understanding of this event seems pivotal to an understanding of the geologic his— tory of the region. The deformation and mobilization of the Shuswap (Monashee Group) and correlative rocks may be linked to thrusting along the Rocky Mountain thrust belt to the east and along the Shuksan thrust belt of the north- ern Cascade Range to the west (fig. 7). According to Bally, Gordy, and Stewart (1966, p. 366—372), deformation proceeded from west to east, beginning with mobilization of gneiss domes and thrusting in the interior in Late Jurassic to Early Cretaceous and, ex- cept for later uplift, ending with the final stages of thrusting in the Rocky Mountains Foothills province in the Eocene. Price and Mountjoy (1970) postulated con- temporaneous metamorphism and mobilization of the gneiss domes and thrusting in the Rocky Mountain thrust belt to the east. In their view, the origin of the Shuswap and the thrust belts can be explained by “progressive buoyant upwelling and lateral spreading of a hot mobile infrastructure beneath a relatively pas- sive suprastructure*** and equivalent progressive northeasterly growth of the foreland thrust belt that 17 forms the Rocky Mountains” (Price and Mountjoy, 1970, p. 7). The crystalline core of the northern Cascade Range is flanked to the west and east by major fault zones, the Shuksan thrust and associated faults on the west and the Ross Lake fault zone on the east (Misch, 1966). The Shuksan thrust is part of Misch’s (p. 128) “mid- Cretaceous Northwest Cascades System.” The age of this westward thrusting is bracketed between “that of the Nooksack Group (early Early Cretaceous) and that of the Chuckanut Group (late Late Cretaceous and Paleocene)” (McTaggart, 1970, p. 138, 144). The terrane between these two major faults consists mainly of the Cascade River Schist and the Skagit Gneiss, believed by Misch (1966, p. 103, 113) to have been metamorphosed before the Shuksan thrusting of the mid-Cretaceous orogeny in pre-Jurassic or probably pre-Mesozoic time. However, McTaggart (1970, p. 143) suggested that the Skagit migmatites and their correla- tive in Canada, the Custer Gneiss, might be Early to middle Cretaceous in age and closely related to the mid-Cretaceous orogeny. Mattinson (1972, p. 3778— 3779) obtained ages ranging from 90 to 60 m.y. on sphene and zircon from the Skagit Gneiss and correla- tive rocks by Pb-U and Pb-Pb methods, indicating that the Skagit metamorphism was probably middle to Late Cretaceous. In view of the relative ages of Skagit metamorphism and Shuksan thrusting, implied by Misch (1966), it is unlikely that the Shuksan thrusting is older than Late Cretaceous. According to Misch (1966, p. 128) yielding on the Shuksan and associated thrusts is southwestward on the south, westward near the Border, and northwest- ward north of it. Movement on the Ross Lake fault zone was primarily strike slip, although reverse in part with eastward-directed over-thrusting on the subsidiary Jack Mountain fault of up to about 10 km (6 mi) displacement (Misch, 1966, p. 134). The northern exten- sion of the Ross Lake fault zone, the Hozameen fault, seems to be very steep (McTaggart, 1970; Coates, 1970). Slivers of tectonically emplaced ultramafic rock are as- sociated with both the Shuksan and Ross Lake fault systems, causing Misch (p. 133, 134) to postulate exten- sion of these faults to the base of the crust. In Misch’s View (p. 136), a 40-km-wide (25-mi-wide) wedge of the crystalline core of the Cascades has been compressively uplifted between the two faults. Maximum offset of the Shuksan and associated thrusts was estimated as about 50—65 km (30—40 mi) (Misch, 1966, p. 128), whereas the cumulative offset on the Rocky Mountain thrusts was variously estimated as about 190 km (120 mi) by Bally, Gordy, and Stewart (1966, p. 359) and at least 200 km (125 mi) by Price and Mountjoy (1970, p. 16). The Shuswap complex and as- 18 PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON 613° 8: .9 \ ¢4o / / / 7< N a C a a 3' n “l. 3 III \ \ \ San 1200 /” / / «0593’ ’ g ‘06? “h , ’0 , //e° / l l | /PC «4 / a 7L l ,’e‘( \q\ ,/ Wlldhorse \ \ \ /’/S <\,’ ’ // oooesh/b’ Raft River area \ I \’ “\P// b "l /, RU Y Rangee I areaw i 1 Rangav N; area 0 0 PL , complex A) { t’Snakei >0 //’(~,9 0“ ’z/ Canyon 0" N\ 0 area 0 Rincon Mountains 2' 0‘s " Tucson \ \x / Shin-L 3'0 vNVlNOW 1.138 03 r? 0 100 200 300 MILES l’—’fi—'—I——‘r‘L—“—J O 100 200 300 KILOMETERS sociated gneiss domes thus appear to occupy the axial zone between convergent thrusts that show an apparent aggregate crustal contraction of possibly 250 km (150 mi). The Shuswap (Monashee Group) is but one of the several areas of gneiss domes, or of medium- to high- grade metamorphic rock, some exhibiting Late Cretaceous to early Tertiary cooling ages, that flank the interior (western) side of the Cordilleran thrust belt in British Columbia, Washington, Idaho, Nevada, and Utah (fig. 7). These also include the Wolverine complex; Wildhorse Canyon Complex (Dover, 1969); the metamorphic complexes in Idaho, Utah and Nevada grouped by Armstrong and Hansen (1966) into three areas, namely Raft River, Ruby Range, and Snake Range; and the metamorphic complex in the Rincon Mountains area of Arizona (Waag, 1969). We speculate that some of these metamorphic terranes are either (1) uplifted elements of an abscherungzone (Armstrong and Hansen, 1966), (2) diapiric tongues of gneiss, possibly originating at an abscherungzone and intruding supracrustal strata of the overthrust plate, or (3) gneiss domes forming through mobilization of infrastructure farther to the west in the hinterland. The metamorphic terranes appear to be both tempor- ally and geographically associated with the Late Cretaceous to early Tertiary thrust belts. These relations compel the conclusion that the for- mation of the Shuswap (Monashee Group) and the gneiss domes is the regional expression of the wide- spread deformational and metamorphic events that culminated in the formation to the west of the westward-directed Shuksan thrust, and to the east, as advocated by Balley, Gordy, and Stewart (1966) and Price and Mountjoy (1970), of the belt of easterly di- rected Rocky Mountain thrusts. LARAMIDE OROGENY, FORMATION OF GNEISS DOMES, AND THRUST FAULTING In concept, the geographic locus of a deformational event could gradually shift through time so that the beginning and ending of the event would not be everywhere the same age. However, few Mesozoic and FIGURE 7.—Distribution of (1) late Mesozoic and early Cenozoic thrust belts and (2) areas of medium- to high-grade metamorphic rock, including gneiss domes, that show late Mesozoic to early Cenozoic cooling ages and (or) structural fabrics suggesting ex- tensive mobilization. Belts of Jurassic and Cretaceous magmatic activity modified after Kistler, Evernden, and Shaw (1971, p. 858). Thrust faults after King (19693). Section A—A’ shown in figure 8. OROGENIC IMPLICATIONS OF GEOLOGIC HISTORY Cenozoic deformational events have left a structural signature by which they can be correlated from region to region. The Laramide orogeny may be an exception. Eardley (1951, p. 285) noted that Mesozoic movements in the Rocky Mountains systems “although intense in some places, were generally precursory to the climactic ones at the close of the Cretaceous and in the early Tertiary.” He indicated that the Laramide orogeny was "the great compressional disturbance that occurred in very late Cretaceous, Paleocene, and Eocene.” It fol- lows that thrusting of that approximate age along the nearly continuous Rocky Mountain thrust belt in Canada, the Montana disturbed belt, Idaho-Wyoming thrust belt, and the Sevier belt in Nevada and Utah, is referable to the Laramide orogeny, regardless of minor differences in age of thrusting from region to region. This deduction contravenes Armstrong’s (1968, p. 451) conclusion that thrusting in the Sevier belt predates and therefore is not part of the Laramide orogeny. Noting that thrusting in the Sevier belt ended in the Campanian Age of the Late Cretaceous, he introduced the term Sevier orogeny for that defor- mational event and relegated the Laramide orogeny to the succeeding episode of morphogenic uplift ranging in age from the Maestrichtian or Campanian Ages of Late Cretaceous through the middle Eocene. We agree that the term Sevier orogeny should be retained but suggest that it is a regional correlative of the Laramide orogeny as defined by Eardley (1951). The deformational events that resulted in formation of the gneiss domes, Shuswap terrane, and thrust belts therefore include tectogenesis attributed to mid- Cretaceous orogeny (Misch, 1966) on the west and to Laramide orogeny on the east. Price and Mountjoy (1970, p. 23) suggested that the thrusting and folding in the Canadian Rocky Moun- tains spanned the time interval from Late Jurassic to Paleocene or Eocene. Their older limit is apparently based on the supposition that orogenic development in the interior of the Rocky Mountains and deposition of the sequence of elastic wedges to the northwest had begun by Late Jurassic and on the premise that forma- tion of the elastic wedges implies beginning of thrust- ing as well as orogeny. In their view, a minimum age for thrusting and doming of the gneiss of 111 my. is established by the age of porphyroblastic biotite (sam— ple GSC 66—47, Wanless and others, 1968) from meta- morphic rocks believed to be part of the metamorphic halo that envelopes the diapiric Malton Gneiss (see fig. 7), more than 150 km (95 mi) north of the area repre- sented in figure 3. However, the following K—Ar ages have been reported from the gneiss itself: 19 Mineral if 5163/ 557,255 Biotite 72 GSC 65—24 Wanless, Stevens, Lachance, and Edmonds, 1967 Biotite 53 67—43 Wanless, Stevens, Lachance, and Delabio, 1970 Biotite 59 67—43 Do. Hornblende 1 14 67-44 Do. Muscovite 60 70—16 Wanless, Stevens, Lachance, and Delabio, 1972 Biotite 66 70—17 D0. Biotite 57 70—18 Do. The cooling age of mica in the Malton Gneiss is roughly Paleocene, virtually equivalent to that of the gneiss domes to the west. Armstrong and Oriel suggested that thrusting in the Idaho-Wyoming thrust belt spanned the interval from Late Jurassic to Eocene (1965, fig. 19, p. 1861). Evi- dence of thrusting prior to latest Cretaceous appar- ently consists of the presence of a latest Jurassic to earliest Cretaceous conglomerate 24 km (15 mi) east of the trace of the Paris thrust fault. Armstrong and Oriel reasoned that if the conglomerate is a synorogenic de- posit derived from source areas to the west, as is likely, then movement on the Paris thrust is dated as latest Jurassic and earliest Cretaceous (1965, p. 1859). Although the time at which the Laramide orogeny ended is relatively easily established by the cooling ages of the crystalline rocks and the termination of overthrusting, the time at which this orogeny began is more difficult to establish and is partly a matter of definition. We‘ suppose that the Shuswap (Monashee Group) and correlative rocks include metamorphosed equivalents of Precambrian, Paleozoic, and even Mesozoic supracrustal rocks that were converted to infrastructure—that is, were deeply buried, heated, metamorphosed, migmatized, and ultimately formed rheomorphic bodies that later were forced upward into the suprastructure, much as visualized by Campbell (1973, p. 1614). This process may have begun in places in the region as early as the Late Triassic, which may mark a significant milestone in the orogenic cycle. However, the gneiss domes are set apart not by mig- matization or metamorphic grade, but by internal structure and contact relations suggesting rheomor- phic flow and diapiric intrusion. The time at which wholesale upward and lateral flow of parts of the infrastructure began therefore marks the beginning of this deformational event in the Okanogan region. The evidence within this region for the time at which this phase began is nebulous, principally because of diffi- culty in discriminating effects of earlier phases of metamorphism and deformation. 20 PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON The earlier phases of plutonism and metamorphism may have been a necessary precursor to the later inva- sion of supracrustal rocks by gneiss domes and con- comitant overthrusting. However, the diapirism and thrusting seemingly mark a radical, even catastrophic, change in the tenor of compressional tectonics and thereby stand in contrast to the previous history. The rough simultaneity in the cooling age of the gneiss domes and the age of terminal thrusting indicates a similar equivalence in the age of their birth, assuming that the thrusts and gneiss domes are genetically re- lated and that both originate at least in part in re- sponse to a particular episode of regional compression. If the gneiss domes and thrusts are indeed roughly coeval, the probable older limit on the age of the Shuksan thrust of Late Cretaceous suggests a compar- able older limit on the age of the gneiss domes. DISCUSSION A stable—and perhaps at times extensional— margin bordered the continental plate during building of the late Precambrian and early Paleozoic deposi- tional wedge now exposed in the miogeoclinal province (Stewart, 1972; Gabrielse, 1972, p. 533). The tectonic stability that marked this period was succeeded in late Paleozoic or early Mesozoic time by the strongly con- vergent regime that prevailed throughout much of the Mesozoic and Cenozoic, while large areas of oceanic plate were subducted beneath the western side of the North American plate (Hamilton, 1969b). The late Paleozoic Cache Creek strata may represent sediments deposited upon the oceanic plate seaward of an eastwardly dipping subduction zone (Monger and others, 1972, fig. 6). However, the temporal counter- part of these strata in the study area, the Anarchist Group, contains limestone and elastic detritus more compatible with sedimentation in a back-arc basin close to a nearby landmass than with sedimentation far from land at the abyssal depths of the ocean floor. In any case, the predominantly volcanic strata of the overlying Middle to Late Triassic formations and their probable correlative in the study area, the Kobau For- mation, mark the abrupt onset of a period of (1) extru- sion of voluminous lavas, (2) intrusion of batholiths and lesser plutons, (3) strong folding, (4) uplift and subaerial erosion, and deposition of detritus in succes- sor basins, and (5) regional metamorphism. Most of this activity can be ascribed to processes occurring above an eastwardly dipping subduction zone or zones, located along the continental margin to the west (Monger and others, 1972, p. 592—593). Perhaps the most enigmatic of the problems dis- cussed in the preceding pages are those of the origin of the gneiss domes and the thermal metamorphism of the Omineca crystalline belt and the reason for the apparent association of these features with the Cordil- leran thrust faults. In the Canadian Cordillera, the spatial and temporal association of the gneiss domes and thrust faults with the Mesozoic plutonic belts and with the presumed convergent plate boundary suggest that the gneiss domes and thrust faults are in some way related to Mesozoic subduction to the west (Campbell, 1973, p. 1618). But it is puzzling that south of the Canadian Cordillera the belt of gneiss domes and the track of the eastern thrust belts diverge from parallelism with the continental boundary (as marked roughly by the pres- ent coastline, fig. 7) and swing far inland of the J uras— sic and Cretaceous belts of plutonic activity delimited by Kistler, Evernden, and Shaw (1971, p. 858). Coney (1972, p. 620) stresses the probable tectonic impact of the worldwide change in spreading directions that occurred about 80 my ago. He proposed (1972, p. 619—620) that the Late Cretaceous orogeny in the Sevier belt of Armstrong (1968) resulted from failure of the leading edge of the North American plate as it was being driven northwestwardly over a Benioff zone en- trained in the oceanic plate below. Later, after the change in spreading directions, the early Tertiary orogeny resulted from failure of the leading edge of the North American plate as it was driven westwardly to southwestwardly, also over a subduction zone or zones entrained in the subjacent oceanic plate. The latter is a vital stipulation because it frees the process from the possible requirement that subduction-related phenomena occur reasonably near the margin of the overriding plate. The cause of the change in spreading directions is unclear; in our opinion, it is as likely that the orogeny caused the change in spreading directions as vice versa. We do not dispute the geographic association of the tectonic and plutonic features of Laramide orogeny with the generally convergent plate boundary to the west, but this observation falls short of being a solution to the gneiss dome and orogeny problem. Before a genetic mechanism can be suggested that could account for the association of a convergent plate boundary with the Laramide orogeny, the question must be answered as to whether or not the cumulative displacement on the Shuksan and Rocky Mountain thrust systems represents actual crustal shortening. Price and Mountjoy (1970, p. 18) hypothesized that the sole of the Rocky Mountain thrusts climbed the westward-dipping upper surface of the Precambrian crystalline basement in response to gravity loading in the vicinity of what is now the Omineca crystalline belt. DISCUSSION The gravity loading resulted from the upwelling of mobile material of the infrastructure, and in their view no crustal shortening is required. However, the Shus- wap (Monashee Group), which probably represents both the upwelled material and its wrapping of supracrustal rocks and other supracrustal rocks metamorphosed to like degree in situ, spans only 55—105 km (35—65 mi). Furthermore, the high-grade metamorphic rocks of pos- sible diapiric origin are not continuous elements of the hinterland along their north-south trend west of the Rocky Mountain thrust zone. If crustal shortening were not a factor, the dilatational features in the hinterland that would be required to account for the cumulative offset on the thrusts—in places approaching perhaps 250 km (150 mi)—do not appear to be present. There- fore, we agree with Bally, Gordy, and Stewart (1966, p. 360) that crustal shortening must be invoked to exp- lain those offsets. The alternative—that the thrust faults represent gravitational gliding from hypothetical highlands located west of the thrust belts within the hinterland—has been advocated by Mudge (1970). The arguments against this hypothesis have been recently summarized by Price (1971), with special reference to the disturbed belt of northwestern Montana, and by Armstrong (1972), with special reference to the Sevier orogenic belt, and are, in our view, compelling. The required compressional mechanism apparently causes regional crustal contraction and concomitant mobilization and upward injection of infrastructure. Possibly the compressive stresses and high heat flow that caused the thrust faulting and diapirism of gneiss could result from opposing drag of the continental lithosphere by converging convection cells in the underlying asthenosphere. This hypothesis has been offered as an explanation of the bilateral symmetry exhibited by certain paired belts of post-Paleozoic thrust faults in part of the Cordillera by Burchfiel and Davis (1968). According to their interpretation, the magnitude of required crustal shortening in certain segments, such as that across the 40-km-wide (25-mi- wide) span between the east-dipping Shuksan thrust and the vertical or west-dipping Ross Lake fault zone (fig. 7), cannot be accommodated except by swallowing of crustal material at depth. They concluded that crus- tal shortening of large magnitude is required across the thrust complexes and suggest that convergence of sub- crustal convection cells was the responsible orogenic mechanism. . Although we believe that the Ross Lake fault zone is a relatively minor feature and that the contractional axis must lie well to the east, below the Omineca crystalline belt, the convection cell hypothesis does provide a be- 21 lievable thrusting mechanism. In this model, the base- ment rocks at either side of the bilateral thrust belts— Shuksan and Cordilleran—are dragged toward and under the centrally located allochthon by a frictional couple with the converging convection cells in the man- tle below. The crust would be expected to thicken along the axis and to either side of the cells through folding, stacking of thrust sheets, and plastic flow. This thicken- ing is presumably counterbalanced by crustal shorten- ing. The buoyancy of the crust would probably prevent its being swallowed by the mantle at the vortex of the cell. The model is unsatisfactory as it stands, however, because no triggering force, reason for localization, or physical basis for the convection is provided. These de- fects can perhaps be overcome by considering the pos- sible ramifications of the ideas articulated by Vine (1966, p. 1411), Palmer (1968, p. 341), and Kistler, Evernden, and Shaw (1971, p. 864—866), whereby, in the Mesozoic, the westward-drifting North American plate overrode an ancestral East Pacific rise locked be- tween underthrusting oceanic plates. Palmer suggested (1968, p. 344) that the Mesozoic plutonism and volcanism of the Cordillera was at- tributable to heat emanating from the underthrust rise and that the Cordilleran thrusts formed during “the terminal throes of convection before motion from the overridden crust was stifled***.” Kistler, Evernden, and Shaw (1971) suggested that the loci of magmatic activity in the western United States defined linear belts and that these belts were pro- gressively displaced to the southeast from Jurassic to Cretaceous time (fig. 7). They hypothesized (1971., p. 864—866) that these magmatic belts originated as North America moved northwestward across a linear heat source analogous to an oceanic rise. But Rutland (1973, p. 829) rejected this hypothesis, noting that “the similarity of timing and spatial relationships of plutonism in North America, South America, and Japan surely demands a common explanation***.” He con- cluded that the plate relationships postulated by Kis- tler, Evernden, and Shaw (1971) could not have ocurred in all these areas. In our View, what all these areas have in common is an intermittently convergent continental plate—oceanic plate relation during the Mesozoic and Cenozoic. This relation could result in either subduction-zone related plutonism and volcanism of the Andean type (Hamil- ton, 1969a) or in plutonism and volcanism related to overriding of an oceanic rise. Both types would be re- lated in space in that they occur inland from the conver- gent plate boundary; they might also appear regionally to be related in time, as periods of strong convergence would be most likely to end in overriding of a rise system. 22 PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON The oceanic rise system has been considered by some (for example, Dietz and Holden, 1970, p. 4941) as a passive element in the plate tectonic model—simply a fracture along which magma upwells to fill the opening that would otherwise remain while the oceanic crust to either side is rafted or pulled away. For the rise system to exert the profound tectonic and magmatic influence here postulated, it must instead represent an active element—perhaps the crustal expression of a linear zone of high heat flow within the mantle that is not readily extinguished even though underthrust beneath a continental plate. We speculate that in the study area overriding of the rise system marked the beginning of a transition from arc-type eugeosynclinal deposition and volcanism to an orogenic phase, with the area of magmatism—formerly restricted to the area of subduction-related volcanism immediately inland from the convergent plate boundary—now diffused along and to either side of the axis of the former rise system (fig. 8). Magmas gener- ated during this phase were presumably the products of differential melting of mantle material, modified by assimilation of material from the sialic crust of the overlying continental plate. Thus as the rise system is overriden production of basaltic magmas gives way to production of inter- mediate and felsic magmas; the latter are the products of varying degrees of fractional melting of mantle and assimilation of sial. Repeated extraction of this salic differentiate from the mantle and possibly the lower crust above the overridden rise system could leave an increasing accumulation of mafic residue. Possibly the build-up of this dense residue of infusibles could pro- duce a mass that is gravitationally unstable relative to nearby areas within the mantle that lie outside of the zone of fractional melting. Ultimately, at some place along the axis of the overridden rise system, the weight of this more dense residue would overcome the plastic limit or strength of adjacent mantle, would sink, and the mantle to either side would converge catastrophi- cally toward it. The overlying sial would be dragged toward the axis of the sinking mass but not swallowed because of its inherent buoyancy. Once started the cell could extend itself lengthwise along much of the over- ridden rise system, until it terminated against parts of the rise system so recently overridden that no signifi- cant instability existed. In summary, we speculate that the convergence of oceanic and continental plates at a subduction zone would ultimately result in the overriding of an oceanic rise system by the continental plate, continuing a cycle whose succeeding phases would be dominated in turn by (1) eugeosynclinal sedimentation, (2) orogenesis with plutonism, (3) extreme compression with failure by overthrusting, possibly attended by mobilization and diapiric intrusion of the infrastructure, forming gneiss domes, and (4) the relaxation of compression, with uplift through isostatic rebound. This is basically an orogenic magmatic cycle similar to that advocated by DeSitter (1964, p. 397) and Rutland (1973, p. 828). The reality of the concept of a tectonic cycle has been questioned by Coney (1970). We hypothesize that this cycle is not only rationalized by but is indeed a necessary consequence of the plate tectonic model. The Omineca crystalline belt may represent part of the relatively rigid center of the Cordilleran-Cascade allochthon, bridging the axis of sinking upper mantle. We speculate that the special features of the Omineca province—gneiss domes and the metamorphic rocks of the Monashee Group—originated through high heat flow and the forcing of hot plastic crustal material of the infrastructure upward along this axis during mid- Cretaceous to Eocene compression. Presumably, the postulated contraction of the continental plate along FIGURE 8.—Stages in the development of the Omineca crystalline belt (sections drawn along line A—A’ in fig. 7). A, In Permian and Early Triassic time, volcanic and pyroclastic deposits accumulated in island-arc and back-arc basin environments. To the west, an east-dipping subduction zone formed the boundary between oceanic and continen— tal plates. A rise embedded within the oceanic plate was gradually nearing the continental plate. B, In Late Triassic and Jurassic time, the thickened eugeosynclinal prisms were invaded by calc-alkalic magmas derived through partial melting of the upper mantle and lower crust within a zone of high heat flow above the overridden rise. The residue (dotted area on figure) remaining in the zone of partial melting became progressively more dense as the hyperfusible part was removed. In Late Cretaceous time (C), after continued calc-alkalic plutonism through the Early Cretaceous, the dense residue occupying the source area of the magmas catastrophically sank into the asthenosphere, forming a short- lived convection that dragged overlying crust toward the axis of the cell. The crust thickened over the cell, through stacking of thrust sheets and plastic flow. Mobilized elements of the infrastructure penetrated higher levels in the crust, with their culminations forming gneiss domes. With the demise of the convection cell in latest Cretaceous, the thickened crust isostatically rebounded. Upper crustal levels over the Omineca were rapidly eroded away, exposing elements of the Late Cretaceous infrastruc- ture (the gneiss domes and the Monashee Group) in the Eocene. DISCUSSION 23 A. PERMIAN AND EARLY TRIASSIC VOLCANISM EU GEOSYNCLINAL PROVINCE IsIandarc Back arc M'05EOCLINAL PROVINCE A, MESOSPHERE “ ‘—‘\ \‘ ‘\ / Amount of \\\\\ Idafmount of contracE'BL |/// / / /\ \\ \\ contractionon I‘ ~~~~~ "II’OT‘ Shuk‘a" thrust / / / \ \ \f COFdillaran ~~~~~ ”I“? ._———I_1// / / \ \\——I—_Lh’“_‘t’__’_‘ ~~~~ ——‘ ######## —— -\ —— ___‘ ‘‘‘‘‘ flflflflflflfl / / 7° ‘>\° \ I ““‘§ C. LATE CRETACEOUS THRUST FAULTING, METAMORPHISM, AND MOBILIZATION OF INFRASTRUCTURE ROSS LAKE FAULT ZONE OMINECA CRYSTALLINE H NTH UST s UKSA R k— BELT —>I CORD'LLEHAN A’ THRUSTs w OCEANIC NE onus-r ”A ’—__ ~§‘ “\ \‘ 0 100 200 MILES 0 100 200 KILOMETERS 24 the Cordillera must have coincided with a reciprocal addition of new oceanic crust elsewhere, either through accelerated spreading rates at existing oceanic rises or through the formation of new rises. The convulsive orogenic climax, represented in the study area by the emplacement of the Okanogan gneiss dome, was evidently followed in Late Cretaceous and early Tertiary by uplift and rapid erosion—and con- sequent cooling of the high-grade metamorphic rocks through the successive blocking temperatures of the dated minerals. This was followed in the Eocene by the postorogenic intrusion of hypabyssal plutons and the extrusion of related volcanic rocks. The Swimptkin Creek and Coyote Creek plutons are probably repre- sentatives of the postorogenic plutons. The Eocene de- posits fringing the gneiss domes have been warped and locally faulted within the study area. To the east, they form a thick sequence of sedimentary deposits and in- terstratified volcanic rocks occupying the Republic gra- ben. The north-northeast trend of the graben suggests that the former east-southeast to east-northeast direc- tion of compression that had prevailed in the study area from the Late Triassic (Rinehart and Fox, 1972) became an axis of tension in the Eocene. The final magmatic event in the region, the extrusion of nonorogenic late Miocene and early Pliocene plateau basalts, introduced a profound change in chemical com- position of magma, from the more silicic and more highly differentiated Upper Triassic and post-Triassic igneous rocks. REFERENCES CITED Armstrong, F. C., and Oriel, S. S., 1965, Tectonic development of Idaho-Wyoming thrust belt: Am. Assoc. Petroleum Geologists Bull., v. 49, no. 11, p. 1847—1866. Armstrong, R. L., 1968, Sevier orogenic belt in Utah and Nevada: Geol. Soc. America Bull., v. 79, p. 429—458. 1972, Low-angle (denudation) faults, hinterland of the Sevier orogenic belt, eastern Nevada and western Utah: Geol. Soc. America Bull., v. 83, p. 1729—1754. Armstrong, R. L., and Hansen, Edward, 1966, Cordilleran infrastruc- ture in the eastern Great Basin: Am. Jour. Sci., v. 264, p. 112— 127. Baadsgaard, H., 1961, Potassium-argon dates of biotites from Cordil- leran granites: Geol. Soc. America Bull., v. 72, p. 689—701. Bally, A. W., Gordy, P. L., and Stewart, G. A., 1966, Structure, seismic data, and orogenic evolution of southern Canadian Rocky Moun- tains: Bull. Canadian Petroleum Geology, v. 14,p. 337—381. Becraft, G. E., 1966, Geologic map of the Wilmont Creek quadrangle, Ferry and Stevens Counties, Washington: US Geol. Survey Geol. Quad. Map GQ—538, scale 1262,500. Bostock, H. S., 1940, Keremeos, Similkameen district, British Col- umbia: Canada Geol. Survey Map 341A, scale 1:63,360. 1941a, Okanagan Falls, Similkameen and Osoyoos districts, British Columbia: Canada Geol. Survey Map 627A, scale 1:63,360. 1941b, Olalla, Similkameen, Osoyoos, and Kamloops districts, British Columbia: Canada Geol. Survey Map 628A, scale 1:63,360. PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON Burchfiel, B. C., and Davis, G. A., 1968, Two-sided nature of the Cordilleran orogen and its tectonic implications: Internat. Geol. Cong, 23d. Prague, Proc. sec. 3, p. 175—184. Cairnes, C. W., 1934, Slocan mining camp, British Columbia: Canada Geol. Survey Mem. 173, 137 p. Campbell, A. B., 1964, General geology of the Hunters quadrangle and vicinity, northeastern Washington(abs): Geol. Soc. America Spec. Paper 82, p. 322. Campbell, A. B., and Raup, O. B., 1964, Preliminary geologic map of the Hunters quadrangle, Stevens and Ferry Counties, Washington: US. Geol. Survey Mineral Inv. Field Studies Map MF—276, scale 1248,000. Campbell, C. D., 1939, The Kruger alkaline syenites of southern British Columbia: Am. Jour. Sci., v. 237, p. 527—549. Campbell, R. B., 1973, Structural cross-section and tectonic model of the southeastern Canadian Cordillera: Canadian Jour. Earth Sci., v. 10, p. 1607—1620. Camsell, Charles, 1910, Geology and ore deposits of the Hedley min- ing district, British Columbia: Canada Geol. Survey Mem. 2, 218 p. Cannon, R. W., 1966, Geochronology and petrographic studies of the intrusive rocks of the Oliver area, Oliver, B. 0.: British Columbia Univ., (Vancouver), B.S. thesis, 23 p. Church, B. N., 1971, The geology of the White Lake Basin, in Geology, exploration and mining in British Columbia, 1970: British Col- umbia Dept. Mines, p. 396—401. Coates, J. A., 1970, Stratigraphy and structure of Manning Park area, Cascade Mountains, British Columbia, in Wheeler, J. 0., ed., Structure of the southern Canadian Cordillera: Geol. Assoc. Canada Spec. Paper 6, p. 149—154. Cockfield, W. E., 1961, Geology and mineral deposits of the Nicola map area, British Columbia: Canada Geol. Survey Mem. 249, 164 p. Coney, P. J ,, 1970, Geotectonic cycle and the new global tectonics: Geol. Soc. America Bull, v. 81, p. 739—748. 1972, Cordilleran tectonics and North American plate motion: Am. Jour. Sci., v. 272, p. 603—628. Daly, R. A., 1906, The Okanagan composite batholith of the Cascade Mountain system: Geol. Soc. America Bull., v. 17, p. 329—376. 1912, Geology of the North American Cordillera at the forty- ninth parallel: Canada Geol. Survey Mem. 38, parts 1—3, 857 p. Dawson, G. M., 1879, Report on exploration in the southern portion of British Columbia: Canada Geol. Survey Report of Progress for 1877.78, p. 131733. DeSitter, L. U., 1964, Structural geology: New York, McGraw-Hill, 551 p. Dietz, R. S., and Holden, J. C., 1966, Miogeoclines (miogeosynclines) in space and time: J our. Geology, v. 74, p. 566—583. 197 0, Reconstruction of Pangaea—breakup and dispersion of continents, Permian to present: Jour. Geophys. Research, v. 75, p. 4939—4956. Dover, J. H., 1969, Bedrock geology of the Pioneer Mountains, Blaine and Custer Counties, central Idaho: Idaho Bur. Mines and Geol- ogy Pamph. 142, 66 p. Eardley, A. J ., 1951, Structural geology of North America: New York, Harper and Bros., 624 p. Engels, J. C., 1975, Potassium—argon ages of the plutonic rocks, in Miller, F. W., and Clark, L. D., Geology of the Chewelah-Loon Lake area, Stevens and Spokane Counties, Washington: US. Geol. Survey Prof. Paper 806, p. 52—58. 1971, Effects of sample purity on discordant mineral ages found in K-Ar dating: Jour. Geology, v. 79, p. 609—616. Engels, J. C., Tabor, R. W., Miller, F. K., and Obradovich,J. D., 1976, Summary of K—Ar, Rb—Sr, U—Pb, and Pb—a and fission—track ages of rocks from Washington State prior to 1975 (exclusive of Col- REFERENCES CITED 25 umbia Plateau basalts): U.S. Geol. Survey Misc. Field Studies Map MF—710, scale 1:1,000,000. Evernden, J. F., and Kistler, R. W., 1970, Chronology of emplacement of Mesozoic batholithic complexes in California and western Nevada: US. Geol. Survey Prof. Paper 623, 42 p. Fox, K. F., Jr., 1970, Geologic map of the Oroville quadrangle, Washington: US. Geol. Survey open—file map, scale 1:48,000. Fox, K. F., Jr., and Rinehart, C. D., 1971, Okanogan gneiss dome (abs), in Metamorphism in the Canadian Cordillera; Program and abstracts, Canada Geol. Assoc. Cordilleran Section Mtg, Vancouver, British Columbia, p. 10. Fox, K. F., Jr., Rinehart, C. D., and Engels, J. C., 1975, K—Ar age of the Similkameen batholith and Kruger Alkalic Complex, Washington and British Columbia: U.S. Geol. Survey Jour. Re- search, v. 3 no. 1, p. 39—43. Fox, K. F., Jr., Rinehart, C. D., Engels, J. C., and Stern, T. W., 1976, Age of emplacement of the Okanogan gneiss dome, north-central Washington: Geol. Soc. America Bull. (in press). Frebold, Hans, and Little, H. W., 1962, Paleontology, stratigraphy, and structure of the Jurassic rocks in Salmo map area, British Columbia: Canada Geol. Survey Bull. 81, 31 p. Froese, E.,1970, Chemical petrology of some pelitic gneisses and migmatites from the Thor-Odin area, British Columbia: Cana- dian Jour. Earth Sci., v. 7, p. 164—175. Fyles, J. T., 1967, Geology of the Ainsworth-Kaslo area, British Col- umbia: British Columbia Dept. Mines and Petroleum Resources Bull. 53, 125 p. Fyles, J. T., and Eastwood, G. E. P., 1962, Geology of the Ferguson area, Lardeau district, British Columbia: British Columbia Dept. Mines and Petroleum Resources Bull. 45, 92 p. Gabrielse, Hubert, 1972, Younger Precambrian of the Canadian Cor- dillera: Am. Jour. Sci., v. 272, p. 521—536. Gary, Margaret, McAfee, Robert, Jr., and Wolf, C. L., eds., 1972, Glos- sary of geology: Washington, DC, Am. Geol. Inst., 805 p. Gilluly, James, 1965, Volcanism, tectonism, and plutonism in the western United States: Geol. Soc. America Spec. Paper 80, 69 p. Giovanella, C. A., 1968, Structural studies of the metamorphic rocks along the Rocky Mountain trench at Canoe River, British Co- lumbia: Canada Geol. Survey Paper 68-1, pt. A, p. 27—30. Hamilton, Warren, 1969a, The volcanic central Andes—a modern model for the Cretaceous batholiths and tectonics of western North America: Oregon Dept. Geology Mineral Industries Bull., v. 65, p. 175—184. 1969b, Mesozoic California and the underflow of Pacific mantle: Geol. Soc. America Bull., v. 80, p. 2409—2429. Hanson, G. N., and Gast, P. W., 1967, Kinetic studies in contact metamorphic zones: Geochim. et Cosmochim. Acta, v. 31, p. 1119—1153. Hart, S. R., 1964, The petrology and isotopic—mineral age relations of a contact zone in the Front Range, Colorado: J our. Geology, v. 72, p. 493—525. Hawkins, J. W., J r., 1968, Regional metamorphism, metasomatism, and partial fusion in the northwestern part of the Okanogan Range, Washington: Geol. Soc. America Bull., v. 79, p.1785— 1820. Hibbard, M. J ., 1970, Magmatic origin of lower Cretaceous granite gneisses and migmatities, Okanogan Range, Washington (abs): Geol. Soc. America, Abs. with Programs, v. 2, no. 2, p. 101. 1971, Evolution of a plutonic complex, Okanogan Range, WashingtonzGeol. Soc. America Bull., v. 82, p. 3013—3047. Hills, L. V., and Baadsgaard, H., 1967, Potassium-argon dating of some Lower Tertiary strata in British Columbia: Bull. Canadian Petroleum Geology, v. 15, p. 138—149. Huntting, M. T., Bennett, W. A. G., Livingston, V. E., Jr., and Moen, W. S., 1961, Geologic map of Washington: Washington Div. Mines and Geology, scale 1:500,000. Hyndman, D. W., 1968a, Mid-Mesozoic multiphase folding along the border of the Shuswap metamorphic complex: Geol. Soc. America Bull., V. 79, p. 575—588. 1968b, Petrology and structure of the N akusp map area, British Columbia: Canada Geol. Survey Bull. 161, 95 p. Jones, A. G., 1959, Vernon map area, British Columbia: Canada Geol. Survey Mem. 296, 186 p. King, P. B., 1969a, compiler, Tectonic map of North America: US. Geol. Survey, scale 125,000,000. 1969b, The tectonics of North America—A discussion to ac- company the tectonic map of North America, scale 1:5,000,000: U.S. Geol. Survey Prof. Paper 628, 94 p. Kistler, R. W., Evernden, J. F., and Shaw, H. R., 1971, Sierra Nevada plutonic cycle; Part I, origin of composite granitic batholiths: Geol. Soc. America Bull., v. 82, p. 853—868. Lanphere, M. A., Irwin, W. P., and Hotz, P. E., 1968, Isotopic age of the Nevadan orogeny and older plutonic and metamorphic events in the Klamath Mountains, California: Geol. Soc. America Bull., v. 79, p. 1027—1052. Leech, G. B., Lowden, J. A., Stockwell, C. H., and Wanless, R. K., 1963, Age determination and geological studies (including isotopic ages—Report 4): Canada Geol. Survey Paper 63—17, 140 p. Little, H.W., 1957, Kettle River (east half), Similkameen, Kootenay and Osoyoos districts, British Columbia: Canada Geol. Survey Map 6—1957, scale 1:253,440. 1960, Nelson map area, west half, British Columbia (82 F W1/2): Canada Geol. Survey Mem. 308, 205 p. 1961, Geology, Kettle River (west half), British Columbia: Canada Geol. Survey Map 15—1961, scale 1:253,440. 1962, Geologic map of British Columbia [2d ed] Canada Geol. Survey Map 932A, scale 1:1,267,200. Lowdon, J. A., 1960, Isotopic ages, Report 1 of Age determinations by the Geological Survey of Canada: Canada Geol. Survey Paper 60—17, 51 p. 1961, Age determinations by the Geological Survey of Canada; Report 2, Isotopic ages: Canada Geol. Survey Paper 61—17, 127 p. Lowdon, J. A., Stockwell, C. H., Tipper, H. W. and Wanless, R. K., 1963, Age determination and geological studies (including isotopic ages, Report 3): Canada Geol. Survey Paper 62—17, 140 p. Macintyre, R. M., York, Derek, and Moorhouse, W. W., 1967, Potassium-argon age determinations in the Madoc-Bancroft area in the Grenville province of the Canadian shield: Canadian J our. Earth Sci., V. 4, p. 815—828. McTaggart, K. C., 1970, Tectonic history of the northern Cascade Mountains, in Wheeler, J. 0., ed., Structure of the southern Canadian Cordillera: Geol. Assoc. Canada Spec. Paper No. 6, p. 137— 148. Mathews, W. H., 1964, Potassium-argon age determinations of Cenozoic volcanic rocks from British Columbia: Geol. Soc. America Bull., v. 75, p. 465—468. Mathews, W. H., and Rouse, G. E., 1963, Late Tertiary volcanic rocks and plant-bearing deposits in British Columbia: Geol. Soc. America Bull., v. 74, p. 55—60. Mattinson, J. M., 1972, Ages of zircons from the northern Cascade Mountains, Washington: Geol. Soc. America Bull., v. 83, p. 3769—3784. Menzer, F. J., Jr., 1964, Geology of the crystalline rocks west of Okanogan, Washington: Washington Univ., Seattle, Ph.D thesis, 83 p. 1970, Geochonologic study of granitic rocks from the Okanogan Range, north-central Washington: Geol. Soc. America Bull., v. 81, p. 573-578. Miller, F. K., and Engels, J. C., 1975, Distribution and trends of 26 PLUTONISM AND OROGENY IN NORTH-CENTRAL WASHINGTON discordant ages of the plutonic rocks of northeastern Washington and northern Idaho: Geol. Soc. America Bull., v. 86, p. 517—528. Misch, Peter, 1966, Tectonic evolution of the northern Cascades of Washington State—-a west cordilleran case history, in Tectonic history and mineral deposits of the western Cordillera: Canadian Inst. Mining and Metallurgy Spec. Vol. 8, p. 101—148. Monger, J. W. H., 1968, Tertiary stratified rocks, Greenwood map area (82 E/2), British Columbia: Canada Geol. Survey Paper 67—42, 39 p. Monger, J. W. H., Souther, J. G., and Gabrielse, H., 1972, Evolution of the Canadian Cordillera, a plate-tectonic model: Am. Jour. Sci., v. 272, p. 577—602. Mudge, M. R., 1970, Origin of the disturbed belt in northwestern Montana: Geol. Soc. America Bull., v. 81, p. 377—392. Muessig, Siegfried, 1962, Tertiary volcanic and related rocks of the Republic area, Ferry County, Washington: U.S. Geol. Survey Prof. Paper 450—D, p. D56—D58. 1967, Geology of the Republic quadrangle and a part of the Aeneas quadrangle, Ferry County, Washington: U.S. Geol. Sur- vey Bull. 1216, 135 p. Naeser, C. W., Engels, J. C., and Dodge, F. C. W., 1970, Fission track annealing and age determinations of epidote minerals: Jour. Geophys. Research, v. 75, p. 1579—1584. Newcomb, R. C., 1937, Geology of a portion of the Okanogan High- lands, Washington: Cincinnati Univ., M.S. thesis. Nguyan, K. K., Sinclair, A. J., and Libby, W. G., 1968, Age of the northern part of the Nelson batholith: Canadian Jour. Earth Sci., v. 5, p. 955—957. Okulitch, A. V., 1973, Age and correlation ofthe Kobau Group, Mount Kobau, British Columbia: Canadian Jour. Earth Sci., v. 10, p. 1508—1518. Palmer, Harold, 1968, East Pacific rise and westward drift of North America: Nature, v. 220, p. 341—345. Pardee, J. T., 1918, Geology and mineral deposits of the Colville Indian Reservation, Washington: U.S. Geol. Survey Bull. 677, 186 p. Parker, R. L., and Calkins, J. A., 1964, Geology of the Curlew quad- rangle, Ferry County, Washington: U.S. Geol. Survey Bull. 1 169, 95 p. Pearson, R. C., 1967, Geologic map of the Bodie Mountain quad- rangle, Ferry and Okanogan Counties, Washington: U.S. Geol. Survey Geol. Quad. Map GQ—636, scale 1262,500. Preto, V. A., 1965, Structural relations between the Shuswap terrain and the Cache Creek Group in southern British Columbia (abs): Canadian Mining Jour., v. 86, no. 1, p. 90. Preto, V. A. G., White, W. H., and Harakal, J. E., 1971, Further potassium-argon age dating at Copper Mountain, British Co- lumbia: Canadian Mining Metall. Bull., v. 64, no. 708, p. 58—61. Price, R. A., 1971, Gravitational sliding and the foreland thrust and fold belt of the North American Cordillera, (discussion): Geol. Soc. America Bull., v. 82, p. 1133—1138. Price, R. A., and Mountjoy, E. W., 1970, Geologic structure of the Canadian Rocky Mountains between Bow and Athabasca Rivers—a progress report, in Wheeler, J. 0., ed., Structure of the southern Canadian Cordillera: Geol. Assoc. Canada Spec. Paper 6, p. 7—25. Reesor, J. E., 1965, Structural evolution and plutonism in Valhalla gneiss complex, British Columbia: Canada Geol. Survey Bull. 129, 128 p. 1970, Some aspects of structural evolution and regional setting in part of the Shuswap metamorphic complex, in Wheeler, J. 0., ed., Structure of the southern Canadian Cordillera: Geol. Assoc. Canada Spec. Paper 6, p. 73—86. Reesor, J. E., and Froese, E., 1968, Petrology and structure, Pinnacle Peaks map area, British Columbia: Canada Geol. Survey Paper 68—1, pt. A., p. 111—112. Reesor, J. E., and Moore, J. M., Jr., 1971, Petrology and structure of Thor-Odin gneiss dome, Shuswap metamorphic complex, British Columbia: Canada Geol. Survey Bull. 195, 149 p. Rice, H. M. A., 1947, Geology and mineral deposits of the Princeton map area, British Columbia: Canada Geol. Survey Mem. 243, 136 p. Rinehart, C. D., and Fox, K. F., Jr., 1970, Geologic map of the Con- conully quadrangle, Washington: U.S. Geol. Survey open-file map, scale 1:48,000. 1972, Geology of the Loomis quadrangle, Okanogan County, Washington: Washington Div. Mines and Geology Bull. 64, 124 p. 1976, Bedrock geology of the Conconully quadrangle, Okano- gan County, Washington: U.S. Geol. Survey Bull. 1402, 58 p. Roddick, J. C., and Farrar, E., 1971, High initial argon ratios in hornblendes: Earth and Planetary Sci. Letters, v. 12, p. 208—214. 1972, Potassium-argon ages of the Eagle Granodiorite, south- ern British Columbia: Canadian Jour. Earth Sci., v. 9, p. 596— 599. Roddick, J. C., Farrar, E., and Procyshyn, E. L., 1972, Potassium- argon ages of igneous rocks from the area near Hedley, southern British Columbia: Canadian Jour. Earth Sci., v. 9, p. 1632-1639. Ross, J. V., 1970, Structural evolution of the Kootenay arc, in Wheeler, J. 0., ed., Structure of the southern Canadian Cordil- lera: Geol. Assoc. Canada Spec. Paper 6, p. 53—65. Rouse, G. E., and Mathews, W. H., 1961, Radioactive dating of Ter- tiary plant-bearing deposits: Science, v. 133, no. 3458, p. 1079— 1080. Rutland, R. W. R., 1973, On the interpretation of Cordilleran orogenic belts: Am. Jour. Sci., v. 273, p. 811—849. Savage, C. N., 1967, Geology and mineral resources of Bonner County, Idaho: Idaho Bur. Mines and Geology County Rept. 6, 131 p. Sinclair, A. J., and White, W. H., 1968, Age of mineralization and postore hydrothermal alteration at Copper Mountain, British Columbia: Canadian Inst. Mining and Metallurgy Bull., v. 61, p. 633—636. Snook, J. R., 1965, Metamorphic and structural history of “Colville Batholith” gneisses, north-central Washington: Geol. Soc. America Bull., v. 76, p. 759—776. Staatz, M. H., 1964, Geology of the Bald Knob quadrangle, Ferry and Okanogan Counties, Washington: U.S. Geol. Survey Bull. 1161—F, 79 p. Staatz, M. H., Weis, P. L., Tabor, R. W., Robertson, J. F., Van Noy, R. M., Pattee, E. C., and Holt, D. C., 1971, Mineral resources of the Pasayten Wilderness Area, Washington: U.S. Geol. Survey Bull. 1325, 255 p. Stewart, J. H., 1972, Initial deposits in the Cordilleran geosyncline; evidence of a Late Precambrian (<850 m.y.) continental separa- tion: Geol. Soc. America Bull., v. 83, p. 1345—1360. Tabor, R. W., Engels, J. C., and Staatz, M. H., 1968, Quartz diorite- quartz monzonite and granite plutons of the Pasayten River area, Washington—petrology, age, and emplacement: U.S. Geol. Sur- vey Prof. Paper 600—C, p. C45—C52. Vine, F. J ., 1966, Spreading of the ocean floor; new evidence: Science, v. 154, p. 1405—1415. Waag, C. J., 1969, Structural geology of the Mt. Bigelow—Bear Wallow—Mt. Lemmon area, Santa Catalina Mountains, Arizona (abs): Dissert. Abs., Sec. B, Sci. and Eng, v. 29, no. 10, p. 3795B—3796B. Wanless, R. K., Stevens, R. D., Lachance, G. R., and Edmonds, C. M., 1967, Age determinations and geological studies; K—Ar isotopic ages, Report 7: Canada Geol. Survey Paper 66—17, 120 p. REFERENCES CITED 27 1968, Age determinations and geological studies; K-Ar isotopic ages, Report 8: Canada Geol. Survey Paper 67—2, pt. A, 141 p. Wanless, R. K., Stevens, R. D., Lachance, G. R., and Delabio, R. N., 1970, Age determinations and geological studies; K—Ar isotopic ages, Report 9: Canada Geol. Survey Paper 69—2A, 78 p. 1972, Age determinations and geological studies; K—Ar isotopic ages, Report 10: Canada Geol. Survey Paper 71-2, 96 p. Wanless, R. K., Stevens, R. D., Lachance, G. R., and Rimsaite, R. Y. H., 1965, Age determinations and geological studies—Part 1, Isotopic ages, Report 5: Canada Geol. Survey Paper 64—17, pt. 1, 126 p. Waters, A. C., and Krauskopf, Konrad, 1941, Protoclastic border of the Colville Batholith (Washington): Geol. Soc. America Bull., v. 52, p. 1355—1417. Wheeler, J. 0., 1965, Big Bend map area, British Columbia: Canada Geol. Survey Paper 64—32, 37 p. 1970, Introduction, in Wheeler, J. 0., ed., Structure of the southern Canadian Cordillera: Geol. Assoc. Canada Spec. Paper 6, p. 1—5. White, W. H., 1959, Cordilleran tectonics in British Columbia: Am. bus. GOVERNMENT PRINTING OFFICE: 1976 - 789-025/15 Assoc. Petroleum Geologists Bull., v. 43, p. 60—100. 1966, Summary of tectonic history, in Tectonic history and mineral deposits of the western Cordillera: Canadian Inst. Min- ing and Metallurgy Spec. Vol. 8, p. 185—189. White, W. H., Erickson, G. P., Northcote, K. E., Dirom, G. E., and Harakal, J. E., 1967, Isotopic dating of the Guichon Batholith, British Columbia: Canadian Jour. Earth Sci., v. 4, p. 677—690. Winkler, H. G. F., 1967, Petrogenesis of metamorphic rocks [revised 2d ed.]: New York Springer-Verlag, 237 p. Yates, R. G., 1964, Geologic map and sections of the Deep Creek area, Stevens and Pend Oreille Counties, Washington: US. Geol. Sur- vey Misc. Geol. Inv. Map I—412, scale 1:31,680. Yates, R. G., Becraf’t, G. E., Campbell, A. B., and Pearson, R. C., 1966, Tectonic framework of northeastern Washington, northern Idaho, and northwestern Montana, in Tectonic history and min- eral deposits of the western Cordillera: Canadian Inst. Mining and Metallurgy Spec. Vol. 8, p. 47—59. Yates, R. G., and Engels, J. C., 1968, Potassium-argon ages of some igneous rocks in northern Stevens County, Washington: US. Geol. Survey Prof. Paper 600—D, p. D242—D247. 0576' pl, 7 DAYS u. 9% Numerical Model of the Salt-Wedge Reach of the Duwamish River Estuary, King County, Washington GEOLOGICAL SURVEY PROFESSIONAL PAPER 990 Prepared in cooperation with the Municipality of Metropolitan Seattle / (0 DEC 1 71975 o \\/' \4’ . .; ‘ ’7’ 8015ch LEW” Numerical Model of the Salt-Wedge Reach of the Duwamish River Estuary, King County, Washington By Edmund A. Prych, W. L. Haushild, and ]. D. Stoner GEOLOGICAL SURVEY PROFESSIONAL PAPER 990 Prepared in cooperation with the Municipality of Metropolitan Seattle UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1976 UNITED STATES DEPARTMENT OF THE INTERIOR THOMAS S. KLEPPE, .S'w'n'lun' GEOLOGICAL SURVEY V. E. McKelvey, Dirwtor Library of Congress Cataloging in Publication Data Prych, Edmund. A. Numerical model of the salt-wedge reach of the Duwamish River Estuary, King County, Washington. (Geological Survey Professional Paper 990) Bibliography: p. 25—26. Supt. of Docs. no.: I 19.162990 1. Oceanography—Washington (State)—Duwamish River estuary—Mathematical models. 2. Estuarine pollution—Washington (State)—Duwamish River estuary—Mathematical models. 3. Phytoplankton—Washington (State)—Duwamish River estuary. I. Haushild, W. L., joint author. II. Stoner, J. D., joint author. 111. Seattle. IV. Title. V. Series: United States Geological Survey Professional Paper 990. GC856.P77 551.4'66'32 76—608136 For sale by the Superintendent of Documents, Us. Government Printing Oflice Washington, DC. 20402 Stock Number 024—001-02886—7 CONTENTS Page Page Abstract __________________________________________________ 1 Application of the model to the Duwamish Introduction ______________________________________________ 1 River estuary—Continued Acknowledgments ________________________________________ 3 Input data—Continued Flow model ______________________________________________ 3 Constituent models—Continued General description ____________________________________ 3 Salinity ______________________________________ 14 Flow in the wedge ____________________________________ 4 Temperature __________________________________ 14 Flow in the upper layer ________________________________ 5 Phytoplankton ________________________________ 15 Transport model __________________________________________ 6 Biochemical oxygen demand __________________ 16 General description ____________________________________ 6 Dissolved oxygen ______________________________ 17 Transport in the wedge ________________________________ 6 Model verification ____________________________________ 17 Transport in the upper layer __________________________ 6 General remarks __________________________________ 17 Advection ________________________________________ 6 Temperature ______________________________________ 20 Diffusion ________________________________________ 8 Chlorophyll a ____________________________________ 21 Constituent models ________________________________________ 8 Dissolved oxygen __________________________________ 23 General description ____________________________________ 8 Prediction of future dissolved-oxygen concentrations _____ 23 Salinity ______________________________________________ 8 Summary and conclusions __________________________________ 24 Temperature __________________________________________ 8 References ________________________________________________ 25 Phytoplankton ________________________________________ 10 Supplemental information Biochemical oxygen demand __________________________ 11 Phytoplankton ________________________________________ 28 Dissolved oxygen ______________________________________ 12 General __________________________________________ 28 Application of the model to the Duwamish River ‘ Sample collection, preparation, and analysis ___________ 28 estuary ____________________________________________ 12 Abundant phytoplankton taxa ______________________ 28 General ______________________________________ . ________ 12 Production and consumption of oxygen by phyto- Input data ____________________________________________ 12 plankton ________________________________________ 29 Flow model ______________________________________ 12 Estimated influence of nutrients on phytoplankton Constituent models ________________________________ 14 growth ________________________________________ 31 Boundary conditions __________________________ 14 Herbivores ____________________________________________ 33 ILLUSTRATIONS Page FIGURE 1. Map of Green-Duwamish River study area ____________________________________________________________________ 2 2—5. Diagrams showing: 2. Longitudinal profiles of salinity, temperature, and dissolved-oxygen concentration in Duwamish River es- tuary __________________________________________________________________________________________ 3 3. Probable circulation pattern in Duwamish River estuary ______________________________________________ 4 4. Definition of elements in Duwamish River estuary model ______________________________________________ 4 5. Definition of elements in the volume-transfer matrix, U ________________________________________________ 7 6. Graph showing solar-radiation and temperature functions for computing phytoplankton growth rates ____________ 11 7—23. Graphs showing computed and observed data for Duwamish River estuary for: 7. Water temperatures and tide stages during July 1967 ___________ __ ______________________________________ 17 8. Water temperatures and tide stages during August 1967 ______________________________________________ 17 9. Chlorophyll a concentrations during July 1967, 16th Avenue South Bridge ______________________________ 18 10. Chlorophyll a concentrations during July 1967, First Avenue South and Spokane Street Bridges __________ 18 11. Chlorophyll a concentrations during August 1967 ______________________________________________________ 18 12. Dissolved-oxygen concentrations during July 1967, 16th Avenue South Bridge __________________________ 19 13. Dissolved-oxygen concentrations during July 1967, Spokane Street Bridge ______________________________ 19 14. Dissolved-oxygen concentrations during August 1967, 16th Avenue South Bridge ________________________ 19 15. Dissolved-oxygen concentrations during August 1967, Spokane Street Bridge ____________________________ 19 16. Water temperatures and tide stages during July 1968 __________________________________________________ 20 17. Water temperatures and tide stages during August 1968 ______________________________________________ 20 18. Chlorophyll a concentrations during July 1968 ________________________________________________________ 20 19. Chlorophyll a concentrations during August 1968 ______________________________________________________ 20 20. Dissolved-oxygen concentrations during July 1968 _____ _ _______________________________________________ 2 1 IV CONTENTS Page FIGURES 7—23. —-Graphs showing computed and observed data for Duwamish River estuary for—Continued 21. Dissolved-oxygen concentrations during August 1968 __________________________________________________ 21 22. Water temperatures and tide stages during September 1971 ____________________________________________ 21 23. Dissolved-oxygen concentrations during September 1971 ________________________________________________ 22 24—29. Graphs showing: 24. Computed dissolved-oxygen concentrations in top sublayer of Duwamish River estuary during September 1971 for 1971 and future RTP effluent discharges __________________________________________________ 24 25. Longitudinal distributions of Cyclotella sp ____________________________________________________________ 31 26. Longitudinal distribution of oval flagellates ____________________________________________________________ 32 27. Longitudinal distribution of “coccoids” plus “coccoid” clusters __________________________________________ 32 28. Relation between concentrations of excess dissolved oxygen and chlorophyll a ____________________________ 33 29. Relations between Michaelis-Menton factors and concentrations of nutrients ____________________________ 33 TABLES . Page TABLE 1. Data required for computing flow in the Duwamish River estuary ______________________________________________ 13 2. Relation of width to elevation at cross sections for computing geometry of the Duwamish River estuary __________ 13 3. Equations for computing location of wedge toe in Duwamish River estuary ______________________________________ 13 4. Data required for the phytoplankton model __________________________________________________________________ 15 5. Data for the BOD Model ____________________________________________________________________________________ 16 6. Miscellaneous BOD inflows to the Duwamish River estuary from upstream BODT, and downstream, BODM, from First Avenue South Bridge __________________________________________________________________________________ 16 7. Data required for the DO model ______________________________________________________________________________ 17 8. Estimated decreases in the monthly averages of computed DO concentrations in the Duwamish River estuary during June-September 1971 for an increase in the RTP effluent discharge to the probable future maximum ________ 24 9. Abundant taxa in samples collected at four stations in Elliott Bay near Seattle during 1967—69 __________________ 29 10. Abundant taxa in freshwater samples collected in 1967 at two stations on the Green-Duwamish River ____________ 29 11. Five most abundant taxa in freshwater samples obtained at two stations on the Green-Duwamish River during March-August 1967 ____________________________________________________________________________________ 29 12. Abundant taxa during two blooms in the Green-Duwamish River ______________________________________________ 30 13. Concentration of herbivores in water samples from the Green-Duwamish River for 3 days during phytoplankton blooms in 1967 and 1968 ________________________________________________________________________________ 33 DEFINITION OF SYMBOLS A ,- average cross-sectional area of wedge element i, in C,- — constituent concentration in wedge element i. square feet. C ’, constituent concentration in wedge element i before B,- average width of interface between a wedge element i computation. and the upper layer, in feet. cl; 1‘ — constituent concentration in sublayer element 1', j. BODi, ,- 5-day biochemical oxygen demand of sublayer element 9’; j — constituent concentration in sublayer element i, j be- i, j, in milligrams per litre. fore a computation. BOD'L j 5-day biochemical oxygen demand of sublayer element Cf constituent concentration in freshwater inflow at wedge i, j before a computation, in milligrams per litre. toe. BODf 5-day biochemical oxygen demand of freshwater inflow c, a coefficient used to compute incident long-wave radia- to the estuary, in milligrams per litre. tion, dimensionless. BODM miscellaneous 5-day biochemical oxygen demand added CS - constituent concentration in seawater. to the estuary downstream from First Avenue South ct}. _ constituent concentration in water flowmg into up- Bridge, in pounds per day. stream end of sublayer j . BODR 5-day biochemical oxygen demand of Green River water CPL j chlorophyll a concentration in sublayer element i, j, in at Tukwila gage, in milligrams per litre. micrograms per litre. BODRTp 5-day biochemical oxygen demand of Renton Treatment CP’i, j chlorophyll a concentration in sublayer element i, j be- Plant effluent, in milligrams per litre. fore a computation, in micrograms per litre. BODS 5-day biochemical oxygen demand of seawater, in mil- CPf chlorophyll a concentration of freshwater inflow to the ligrams per litre. estuary, in micrograms per litre. BODT miscellaneous 5-day biochemical oxygen demand added CPS chlorophyll a concentration of seawater, in micrograms to the estuary upstream from First Avenue South per litre. Bridge, in pounds per day. CPw, chlorophyll a concentration in wedge element i, in mi- C = constituent concentration. crograms per litre. CIA constituent concentration when the Michaelis-Menton CP’wl. chlorophyll a concentration in wedge element i before a factor for that constituent equals 0.5. computation, in micrograms per litre. DOBOD DOC i,j Do', ,- D01, D05 Dow, DO’ wi E (T) FL FN FT Go HL H3 H*,-,,~ CONTENTS upper layer thickness at the mouth of the estuary, in feet. average thickness of a sublayer element in segment i, in feet. oxygen consumed per unit decay of biological oxygen demand, dimensionless. oxygen produced or consumed during a unit change in chlorophyll a concentration by growth or respiration, in milligrams oxygen per microgram chlorophyll a. dissolved—oxygen concentration of sublayer element i, j, in milligrams per litre. dissolved—oxygen concentration of sublayer element i, j before a computation, in milligrams per litre. dissolved-oxygen-saturation concentration of sublayer element i, j, in milligrams per litre. dissolved-oxygen concentration of seawater, in milli- grams per litre. dissolved-oxygen concentration in wedge element i, in milligrams per litre. dissolved-oxygen concentration in wedge element i be— fore a computation, in milligrams per litre. a constant of proportionality used in computing lon- gitudinal mixing in wedge, dimensionless. saturation vapor pressure in air at temperature T, in millibars. vapor pressure in air, in millibars. solar-radiation function in phytoplankton growth equa- tion, dimensionless. Michaelis-Menton factor for reducing the phytoplank- ' ton growth rate because of the deficiency of some con- stituent, dimensionless. temperature function in phytoplankton growth equa- tion, dimensionless. growth rate of phytoplankton, in reciprocal hours. maximum growth rate of phytoplankton, in reciprocal hours. rate of heat transfer to water from atmosphere, in gram-calories per square centimetre per hour. rate of heat transfer to water by evaporation, conduc— tion, and outgoing long-wave radiation, when the water surface temperature is T1,, in gram-calories per square centimetre per hour. rate of incoming long-wave radiation, in gram-calories per square centimetre per hour. rate of solar energy penetrating water surface, in gram-calories per square centimetre per hour. solar energy passing through a unit area of the bottom of sublayer element i, j during a time step, in gram- calories per square centimetre. a subscript denoting segment number. number of segments in the estuary. number of the time step. a subscript denoting sublayer number. decay rate for biochemical oxygen demand, in recip- rocal days. decay rate for biochemical oxygen demand at 20°C, in reciprocal days. column subscript in volume transfer matrix. change in solar-radiation attenuation coefficient per unit change in chlorophyll a concentration, in recip- rocal feet per micrograms per litre. solar-radiation attenuation coefficient for sublayer element i, j, in reciprocal feet. solar-radiation attenuation coefficient for upper layer in absence of phytoplankton, in reciprocal feet. kow Q RTP V solar-radiation attenuation coefficient for wedge in ab- sence of phytoplankton, in reciprocal feet. solar-radiation attenuation coefficient for wedge ele- ment i, in reciprocal feet. depth-averaged solar-radiation intensity in an element, in gram-calories per square centimetre per hour. solar—radiation intensity for maximum phytoplankton growth, in gram-calories per square centimetre per hour. number of time steps in a day. atmospheric pressure, in millibars. tidal-prism thickness, in feet. = upward flow through base of sublayer element i, j, in cubic feet per second. mean daily freshwater discharge into head of estuary, in cubic feet per second. fraction of Q,» that flows into sublayer j at toe of wedge, dimensionless. flow rate into the upstream face of sublayer element i, j, in cubic feet per second. mean daily discharge of Green River at Tukwilla gage, in cubic feet per second. mean daily discharge of effluent from Renton Treat- ment Plant, in cubic feet per second. rate at which saltwater leaves the wedge and is trans- ported upstream from the toe, in cubic feet per second. fraction of Q s that returns to sublayer j at the toe of the wedge, dimensionless. phytoplankton respiration rate, in reciprocal hours. phytoplankton respiration rate at 20°C, in reciprocal hours. the mean daily relative humidity, in percent. the lesser of the volume ratio V,+1/V,~ and unity, di- mensionless. the lesser of the volume ratio V,/ VH1 and unity, di- mensionless. atmospheric-reaeration coefficient, in reciprocal days. salinity of sublayer element i, j, in parts per thousand. = slope of the interface between wedge and upper layer in feet per foot. total daily incident solar radiation. in gram-calories per square centimetre. fractional amount of total solar radiation penetrating water surface, dimensionless. mean daily air temperature, in degrees Celsius. a reference temperature, in degrees Celsius. temperature of sublayer element i, j, in degrees Celsius. temperature of sublayer element i, j before a computa- tion, in degrees Celsius. water-surface temperature, in degrees Celsius. temperature of wedge element i, in degrees Celsius. temperature of wedge element i before a computation, in degrees Celsius. volume-transfer matrix for a sublayer, in cubic feet. term in matrix U which equals volume in sublayer ele- ment i, j at the end of a time step that was in element (i—10+k), j at beginning of the time step, in cubic feet. entrainment volocity across the interface between wedge and upper layer, in feet per second. vertical velocity at base of sublayer j divided by en- trainment velocity, Ue, dimensionless. volume of a wedge element, in cubic feet. volume of a wedge element for a preceding time step, in cubic feet. volume of a sublayer element, in cubic feet. xb = — longitudinal coordinate of wedge toe, in feet. — longitudinal coordinate of downstream face of wedge — change in biochemical oxygen demand of sublayer ele- CONTENTS volume of a sublayer element in preceding time step, in cubic feet. velocity of water in top sublayer relative to water in wedge, in feet per second. mean daily wind speed, in feet per second. settling velocity of phytoplankton, in feet per second. longitudi ial coordinate measured upstream from es- tuary mouth, in feet. longitudinal coordinate of a cross section bounding a volume that shall move into a sublayer element (see fig. 5), in feet. longitudinal coordinate of a cross section bounding a volume that shall move into a sublayer element (see fig. 5), in feet. longitudinal coordinate of center of wedge element i, in feet. element 1', in feet. longitudinal coordinate of the downstream face of wedge element i, for preceding time step, in feet. latent heat of vaporization at 0°C, in gram-calories per gram. change in latent heat of vaporization per unit change in temperature in gram-calories per gram per degree Cel- sius. constant in the Bowen ratio relating heat transfer by conduction to that by evaporation, dimensionless. ment i, j due to decay during a time step, in milligrams per litre. ACii + 1 ACPL ,- ACPwl. ADOw Ax,- At 6 63/ change during a time step in concentration of a con- stituent in wedge element i due to mixing with water from element i+1. change in chlorophyll a concentration in sublayer ele- ment 1', j due to growth and respiration during a time step, in micrograms per litre. change in chlorophyll a concentration in wedge element i due to growth and respiration during a time step, in micrograms per litre. oxygen-consumption rate in wedge, in milligrams per litre per hour. - length of a wedge element i, in feet. = length of time step in finite difference model, in sec- onds difference between kelvin and Celsius temperature scales, in degrees. emissivity of water, dimensionless. - vertical turbulent diffusion coefficient between sub- layers, in square feet per second. constant in equation for saturation vapor pressure, di- mensionless. constant in equation for saturation vapor pressure, in degrees kelvin. - time angle, in radians. = mass transfer coefficient for evaporation in grams per square centimetre per hour per millibar per foot per second. constant in equation for saturation vapor pressure, in millibars. Stefan-Boltsman constant, in gram-calories per square centimetre per hour per degree kelvin. ENGLISH-METRIC EQUIVALENTS Multiply By To obtain feet (ft) 0.3048 metres (m) square feet (ft?) .0929 square metres (m2) cubic feet (113“) .02832 cubic metres (m3) feet per second (ft/s) .3048 metres per second (m/s) square feet per second .0929 square metres per second (ftz/s) cubic feet per second .02832 cubic metres per second (ft3/s) miles (mi) 1.609 kilometres (km) pounds per day (lb/day) .4536 kilograms per day (kg/day) NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY, KING COUNTY, WASHINGTON By EDMUND A. PRYCH, W. L. HAUSHILD, and ]. D. STONER ABSTRACT A numerical model of a salt-wedge estuary developed by Fischer (1974) has been expanded and used to calculate the distributions of salinity, temperature, chlorophyll a concentration, biochemical oxy- gen demand, and dissolved-oxygen concentration in the Duwamish River estuary, King County, Wash. With this model, which was calibrated and verified with observed data, computed temperatures usually agreed within 2° Celsius of observed temperatures. During a phytoplankton bloom in the summer of 1968, the computed chloro- phyll a concentrations increased and decreased with the observed concentrations; however, during two blooms in 1967 the computed high concentrations persisted farther downstream and lasted a few days longer than the observed concentrations. The computed and observed dissolved-oxygen concentrations usually agreed within 2 milligrams per litre, except during phytoplankton blooms. During the blooms, the differences were often larger, especially when the computed chlorophyll a concentrations were larger than the ob- served concentrations. The model was used to predict the dissolved-oxygen concentrations in the Duwamish River estuary when the Renton Treatment Plant sewage-effluent discharge is increased to its proposed maximum of 223 cubic feet per second (6.31 cubic metres per second). The com- puted monthly average dissolved—oxygen concentrations in the es- tuary decreased by a maximum of 2 milligrams per litre when com- pared with computations for the summer of 1971, when the effluent discharge averaged 37 cubic feet per second (1.05 cubic metres per second). The increase in effluent discharge is not expected to cause large changes in phytoplankton concentrations in the estuary. INTRODUCTION This report describes a numerical model designed for predicting concentrations of DO (dissolved oxygen) and other constituents in the Duwamish River estuary, a salt-wedge estuary (fig. 1). Modeling the upper layer (the layer overlying the salt wedge) is emphasized, but because concentrations of constituents in the wedge affect those in the upper layer, modeling of the salt wedge also is described. ' Earlier, Fischer (1974) presented a generalized numerical model for predicting constituent concentra- tions in salt-wedge estuaries. Stoner, Haushild, and McConnell (1974) then used Fischer’s model to predict DO concentrations in the salt wedge of the Duwamish River estuary. They verified the model with observed DO concentrations in the wedge and with observed sa- linity distributions in the upper layer. In the present study, the model for the Duwamish River estuary was extended to predict DO concentrations in the upper layer of the Duwamish River estuary. To accomplish this goal, it is necessary for the model to also predict the temperature, the 5-day BOD (biochemical oxygen demand), and the phytoplankton concentrations in the estuary. This report is one of several resulting from a study of the effects on estuary water quality of changes in the treatment and the quantity of sewage discharged to the Duwamish River estuary. The effects of effluent from RTP (Renton Treatment Plant), which provides secondary treatment, are of particular interest. The study is being conducted by the US. Geological Survey in cooperation with Metro (Municipality of Metropoli- tan Seattle). As part of its general plan for the treat- ment and disposal of sewage from the Metropolitan Seattle area, Metro operates the RTP and relies on the Duwamish River estuary to transport an increasing quantity of RTP effluent to Elliott Bay and Puget Sound. If the RTP effluent is detrimental to the estuary water quality—especially for its use by anadromous fish—regulatory Federal and State agencies may pro- hibit or limit the discharge of the effluent and (or) re- quire additional treatment of the sewage at the plant. A primary reason for developing the mathematical model of this report was to predict the effects of pro- posed changes in the quantity and quality of the RTP effluent on the estuary water quality. The estuary reach included in the model study ex- tends from the estuary mouth to the wedge toe. The wedge toe is defined as the upstream limit of a specified salinity. The location of the toe depends on tide stage and river discharge; therefore, the location of the mod- el’s upstream boundary is a function of time. During the summer low-flow periods, the toe ranges from a little downstream of 16th Avenue South Bridge to a little downstream of East Marginal Way Bridge. (See fig. 1.) The distance from the mouth to the wedge toe during the low-flow periods averages about 5 mi (8 km). All flows are computed by using equations for the conservation of fluid volume and empirical formulas 1 NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY Ell/"of! Bay (PUGET SOUND) I I22“ 20' East Waterway Wes! Waterway S E ATTLE Spokane Street Bridge, surface and bottom EXPLANATION First Avenue South Bridge ‘ Stream gaging station Quality -of-water site, chemical, temperature and bioIogical measure- ments —|6th Avenue South I , Bridge, Surface flwnh monitor W Without monitor dbtt \an— 0 am _____ __J SEATTLE CITY LIMITS : Boeing Bridge I I _‘__‘-"I I l I ,— — J I | 47° 302-4 East Marginal J Woy Bridge I Renton Treatment 0 I ZMILES Plant |—_l_[__.L__.'—_I____Y_J waif. o I 2 3KILOMETRES _ — ' “‘"_“—I Tukwila Gaging Station Study Renton area Junction monitor WASHINGTON I _____ J ”3919 19A!” FIGURE 1.—-Green-Duwamish River study area. FLOW MODEL 3 developed from data for the Duwamish River estuary; the dynamic equations of motion are not used. The wedge is assumed to be vertically and laterally homogeneous. The upper layer is divided into three sublayers that are also assumed to be vertically and laterally homogeneous. Resolution of constituent con- centrations in the longitudinal direction is limited by the length of the segments into which the estuary is divided; segment lengths average about 600 ft (200 m). The model of the Duwamish River estuary was ver- ified by data from the J une-September periods of 1967—69 and 1971. The summer months are periods of low river discharges when the probability of low DO concentrations are highest. Finally, the model was used to predict constituent concentrations in the es- tuary in the future when the discharge from RTP is expected to be much larger than at present. Information on river discharges, salinities, and other constituent concentrations, and physical characteris- tics of the Duwamish River estuary are given in sev- eral previous reports, including Dawson and Tilley (1972), Santos and Stoner (1972), Stoner (1972), Stoner, Haushild, and McConnell (1974), and Welch (1969). The interested reader is referred to these re- ports for background information and data other than that included in this report. ACKNOWLEDGMENTS The authors express their appreceiation to Glen D. Farris, superintendent, Water Quality and Industrial Waste Division, Municipality of Metropolitan Seattle, and his staff for contributing information and many data needed for the model study and for their continu- ing support and encouragement in the study of the es- tuary’s water quality. The authors acknowledge the contribution made by Hugo B. Fischer, Professor of Civil Engineering, University of California at Berke- ley and part-time employee of the US. Geological Sur- vey, in conceiving and developing the salt-wedge es- tuary model subsequently useful in modeling the Duwamish River estuary and for advice in develop- ment of the present model. An interim version of a computer program for constituent transport in the upper layer of the Duwamish River estuary, by Richard H. French, graduate student at the University of California at Berkeley, expedited the development of the present Duwamish River estuary model. The au- thors depended considerably on K. V. Slack of the US. Geological Survey for guidance and advice in modeling the phytoplankton and their effects on dissolved oxy- gen, for suggestions about presenting the results of the study of phytoplankton and herbivores, and for fur- nishing literature pertinent to the biological modeling. The authors also thank H. E. Jobson, and M. J. Sebetich of the Geological Survey for reviewing this manuscript and for their constructive criticisms. FLOW MODEL GENERAL DESCRIPTION Fischer (1974) developed the general salt-wedge es- tuary model which was used in the present study. The typical salt-wedge estuary is characterized by a lower layer, or wedge, of nearly undiluted seawater of uni- form salinity. For the Duwamish River estuary, the term “sea” is interpreted as Elliott Bay (fig. 1), the body of water outside the mouth of the estuary. Figure 2 shows observed salinity, temperature, and DO-concentration distributions in the Duwamish River estuary. The wedge in the Duwamish River es- tuary is defined as that volume of water with a salinity greater than 25 ppt (parts per thousand). By compari- son, the salinity of the water flowing into the estuary from Elliott Bay is about 28 ppt. The 25 ppt salinity was chosen because the locus of points with this salin- ity was fairly easy to define and is a good approxima- tion of the interface between the nearly homogeneous wedge and the stratified upper layer. The interface is RIVER KILOMETRES DEPTH, IN FEET DEPTH, IN METRES 5 RIVER MILES FIGURE 2.—Observed longitudinal profiles of salinity, temperature, and dissolved—oxygen concentration in Duwamish River estuary during the low high tide of September 13, 1968. 4 NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY relatively stable except near the toe. The location of the toe, defined as the upstream extent of the 25 ppt salinity, was fairly easy to define; however, the up- stream extent of salinities slightly less than 25 ppt’ moved erratically, suggesting more intense mixing up- stream from the toe. Stoner (1972) found that during a tidal cycle, salt— water in the wedge near the toe flows upstream at a net rate of about 200 ft3/s (5.7 m3/s). This saltwater is transported upstream from the wedge toe, is mixed vertically, and is returned through the upper layer to the sea. The increase in salinity of the upper layer toward the sea (fig. 2) suggests entrainment from the wedge to the upper layer over the entire interface. A nearly uniform salinity in the wedge, approximately equal to that in Elliott Bay, indicates that freshwater is not transferred into the wedge from the upper layer. The net circulation pattern in the estuary is probably as shown in figure 3. The waters in both the upper layer and the wedge oscillate upstream and downstream in the estuary dur- ing a rising and falling tide. In addition, water in the upper layer has a net downstream motion because the upper layer carries to the sea both the water dis- charged to the estuary by the river and the water en- trained from the wedge. Water in the wedge has a net upstream motion because the wedge loses water to the upper layer by entrainment. Thus, water entering the upper layer is transported to the sea, and on its way it is mixed with water entrained from the wedge. Water entering the wedge on a rising tide either returns to the sea on a falling tide or remains in the wedge and moves upstream until it is entrained into the upper layer, where it then is returned to the sea. In addition to these advective transport processes there also are dispersive transport processes in the estuary; they are discussed in the section “Transport model.” V I e— \ J layer Upper ‘\ J ——> ’—> Mouth FIGURE 3.—Probable circulation pattern in Duwamish River es- tuary. Arrows indicate direction of mean flow. For purposes of computation, the estuary is divided into control volumes called elements (fig. 4). The ele- ments are constructed by first dividing the estuary ver- tically into the wedge and the upper layer, and the upper layer is in turn divided into three equal-depth sublayers. Next, the estuary is divided longitudinally by vertical planes that are perpendicular to the lon- gitudinal axis of the estuary. Each volume between adjacent vertical planes is called a segment and con- tains one wedge element and three sublayer elements. Typically, there are about 35 segments in the estuary. The segments are numbered longitudinally from the toe to the mouth (i=1 to imax), and the sublayers are numbered vertically from bottom to top (j =1,2,3). All segments move with the velocity of the wedge water so that there is no net flow between wedge ele- ments. However, because of the relative motion be- tween the water in the wedge and the water in the sublayers, there is a net flow across the vertical bound- aries between elements in the sublayers. All horizontal flows are computed from conservation of volume equations for the water. Input information required for computing water flow in the model in- cludes the geometry of the estuary, location of the wedge toe, flow rates across the vertical boundary at the wedge toe, vertical entrainment velocities, eleva- tion of the water surface, thickness of the upper layer at the estuary mouth, and the slope of the interface. In this, as in most numerical models, computations are made for finite steps in time, At. This model uses time steps of 15 minutes. The model computes average flow, mixing and other process rates, and the resulting changes in constituent concentrations during each time step. FLOW IN THE WEDGE In the model, the water surface is assumed to remain ’A v A‘ _ _____ _3__ _:_ _ - Upper layer t__5£91°1°.'-'2. _______ 4/ 3d,- Subluyer element 1' ,2 r. Wedge 1’ Section I-I Wedge element / 3—1"! V = | I I I v 3 ' L. _____ J-— - __-_ .. ___.l-__ ' 1 Y - ‘I' _ _ _____________ — — _ n l'““‘l“ .4___-3 ________________ Sublayer 2 Upper layer I ____________ I 39 ' I Wedge _____._L ___J_\r__.- ’V _______,_ Segment In.“ iol I O Estuary mouth ‘r Wedge roe LONGITUDINAL COORDINATE, 1 FIGURE 4,—Schematic diagram defining elements in Duwamish River estuary model. See text for definition of symbols. FLOW MODEL 5 horizontal and the interface is assumed to have a con- stant slope, So, at all times. During a given day, the upper-layer thickness at the mouth, D, is assumed to be constant even though the wedge becomes deeper and shallower with a rising and falling tide. The wedge toe moves upstream and downstream with the tide. Data from the estuary show that these are reasonable as- sumptions for the periods modeled in this study. At the start of computations for a time step, the new space occupied by the wedge is filled with the wedge elements from the end of the preceding time step. These elements, or water Volumes, are deformed to fit the wedge geometry for the time step. Starting with segment number one at the toe, the model calculates for each wedge element a new length Ari, new lon- gitudinal coordinates of the center xmi, and downstream face xui, a new average cross-sectional area A,, and a new average interface width Bi. The model computes the widths, areas, and other geometri- cal properties by interpolation between data at refer- ence cross sections where the channel geometry is de- fined by input data. If the total volume of all the previ- ously designated wedge elements is insufficient to fill the wedge, a new element of seawater is added at the estuary mouth. On the other hand, if the wedge is filled before all the designated elements are used, the fluid in the extra elements is assumed to have flowed into the sea, and the model deletes these downstream elements from the computations. ' Whenever the number of wedge elements exceeds 50, or whenever the shortest wedge element is less than 200 ft (60 m) long, the number of wedge elements is decreased by combining the shortest element with its shortest neighbor. Also, the element adjacent to the wedge toe is combined with its neighbor whenever its volume is less than the volume that would flow out of it during the time step. The corresponding sublayer ele- ments of two segments are combined at the same time the wedge elements are combined. Near the end of all computations associated with a time step, the volume of each wedge element is de- creased to account for entrainment. The model makes the calculation, Vi=V’i—UeB,~ A x,- At, (1) where V’l- and V,- are the wedge-element volumes be- fore and after this subtraction is made, and Ue is the entrainment velocity across the interface. In addition, the volume of the wedge element adjacent to the toe is decreased by the amount QSA t, where Q, is the net upstream flow rate of saltwater in the wedge in the vicinity of the toe. Becausethe elements in the wedge move with the water, there is no net flow across the boundaries be- tween wedge elements. However, longitudinal con- stituent transport between wedge elements may occur by longitudinal dispersion; the modeling of this process is discussed in the section “Transport in the Wedge.” FLOW IN THE UPPER LAYER The location and length of the elements in the sub- layers are determined by the geometry of the elements in the wedge. The thickness, di, of each sublayer ele- ment in a segment is the same: d, = (D +Soxm 9/3. (2) For simplicity, the widths of all sublayer elements in a segment are assumed to equal the interface Width B i. The volume, Vs,» of a sublayer element is therefore given as VSi=A xiBl-di. (3) Because of the relative motion between the water in the wedge and the water in the upper layer, there is usually a net flow between adjacent elements in a sub- layer. An expression for the flow between elements in a sublayer is obtained by writing an equation for the conservation of water volume for a sublayer element. The resulting expression is (Va—Va), )‘L At Q =Q,},+(Qe, —Qe (4) ”U 1,} i,j+1 where Q; j is the flow rate into the upstream face of sublayer element i, j; Q9, 1- is the upward flow through the base of element i, j; and V’s , and Vs,- are the vol— umes of the sublayer element for the preceding and present times steps, respectively. The upward flows Q; 131‘ are computed with the relation Qei,j = U/ej Ue AxiBi: (5) where U’ej is the vertical velocity at the base of sub- layer j divided by the entrainment velocity across the interface U8. Vertical flow through the water surface is prevented by defining U ’8 4:0. The use of equation 4 requires inflow to each sub- layer at the wedge toe Q1, j. The model uses Q1:j _—_ Qlfj Qf + QSJ' Q39 (6) where Qf is the mean daily discharge of freshwater into the head of the estuary, and Q’fj and Q’s]. are the frac- tions of Q: and Q3 that flow into sublayer j . The parameters Ue, U ’e j, Q’fj, and Q’sj are input in- formation to the model. The selection of numerical values for these and other parameters for the Duwamish River estuary is discussed in the section "Input Data.” 6 NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY TRANSPORT MODEL GENERAL DESCRIPTION This section describes the procedures used in the model to calculate changes in constituent concentra- tions due to advection and diffusion. The effects of other processes on constituent concentrations, such as the effects of photosynthesis on DO concentration and of solar radiation on temperature, are discussed in the section “Constituent Models.” The water-transport processes modeled are longitud- inal advection and dispersion in the wedge, entrain- ment from the wedge to the upper layer and from each sublayer to the sublayer above (vertical advection), vertical turbulent diffusion in the upper layer and lon- gitudinal advection in the upper layer. The assumption of lateral and vertical homogeneity in the wedge pro- hibits modeling vertical or lateral transport processes within the wedge, and the assumption of lateral homogeneity in the upper layer prohibits modeling lateral transport processes there. That part of the longitudinal dispersion in the upper layer that is caused by the combined action of vertical differences in velocity and vertical mixing is repre- sented in the model because both these phenomena are modeled; whereas, that part of the longitudinal disper- sion that is caused by the action of lateral differences in velocity and lateral mixing is not modeled explicitly. However, the finite-difference representations of the advective processes in the model introduce additional longitudinal dispersion in the upper layer; this addi- tional dispersion is often called "numerical disper- sion.” (For example, see Bella and Grenney, 1970.) During each time step the model computes the changes in concentrations due to transport and other processes one at a time. The calculations are performed in the following order: 1. Position elements in the wedge. 2. Change concentrations in the wedge to account for longitudinal dispersion in the wedge. 3. Change concentrations in the upper layer to account for entrainment from the wedge and advection in the upper layer. 4. Change concentrations in the upper layer to account for vertical diffusion. 5. Change concentrations in the wedge and the upper layer to account for processes other than water transport. TRANSPORT IN THE WEDGE After the positions of the elements in the wedge are determined by the method described previously in the section “Flow in the Wedge,” the model need only cal- culate changes in constituent concentrations due to longitudinal dispersion. No advection calculations are required for the wedge because there is no net flow between wedge elements. The model computes the changes due to dispersion with a procedure developed by Fischer (1972, 1974). The change in concentration caused by the mixing be- tween adjacent wedge elements during a time step is assumed proportional to the width of the elements and to the distance traveled by the elements during the time step and inversely proportional to the square of the element lengths. The change in constituent con- centration, ACii+1, of wedge element i due to mixing with water from element i+1 is calculated as I Act-H1 : RV,- (CH1 ‘63,) [2E(Bi+Bi+1) ' xu. _xu. i l /(A x,+A x, )2], (7) l i+1 where E is a dimensionless constant of proportionality with a value of about 0.2; xui and x’ui are longitudingal coordinates of the boundary between wedge elements i and i+1 for the present and previous time steps, re- spectively; 0’, is the constituent concentration in wedge element i before the dispersion computation; and RV,- is the lesser of Vi+1/V,- and unity. To avoid occa- sional instabilities in the computations, an upper limit of 0.5 is imposed on the term in square brackets in equation 7. The corresponding change in concentration of a constituent in wedge element i+1 due to mixing with water from wedge element i is AC" —Ac."+1R' /R i Vi ’ i+1 = Vi (8) where R ,Vi equals the lesser of the values Vi/Vi+1 and unity. After calculating the quantities AOL-”1 and AC,“1 for all i, the model computes a new constituent concentra- tion, 0,, for each wedge element, using the expression C,- = C’,+ A Cii+1 + A Cii"1, (9) Where C ’i and C ,- are constituent concentrations before and after this computation. TRANSPORT IN THE UPPER LAYER ADVECTION The present model treats longitudinal advection a little differently than does the original model of Fischer (1974). The present model uses an extension of the upstream difference scheme or the method of characteristics that is computationally stable when a water particle moves more than one element length during a time step. The model computes a volume transfer matrix, U, of TRANSPORT MODEL 7 dimension imax by 21, for each sublayer. Each term Ui, k in the matrix equals the volume in element 1', j at the end of a time step that was advected horizontally from element (i—10+k), j during the time step (fig. 5). Presently, the movement of water is limited to nine elements in the downstream (k=1—9) or upstream (k=11—19) directions. The term with the subscript k=10 is the volume that remains in element i during the time step. Terms with k=20 and 21 represent vol- umes that originate seaward of the estuary mouth and upstream of the wedge toe, respectively. The model evaluates terms in a row of the matrix U as follows. If the flow is into the upstream face of sub- layer element i, j (Qi, j>0), one sums the sublayer vol- umes from the preceding time step, V’s, either wholly or in parts, upstream from the element until the ac- cumulated volume equals the inflow volume Qi, jAt. If the upstream boundary of the model is reached before the desired volume is obtained, a temporary element with sufficient volume to make up the deficit is at- tached to the upstream end of the sublayer. Starting at xa—the location of the cross section at the end of the accumulated volume in figure 5—the volumes V’s are summed again, this time proceeding in the down- stream direction until the accumulated sum equals Vs ,- —(Qe U —Qe ”+1 )At, or until the upstream face of element i is reached. The location of the cross section at the end of this summation is designated as xb. The parts of the volumes V’s contained in the interval x1, to xa are assigned to the appropriate terms in U i, k. If flow is into the downstream face of element i, j (Qi+1,j<0), similar computations are performed on the elements downstream from element i, j. The volume remaining in an element during a time step, Ui,1o, is computed by subtracting the outflow vol- umes from the element volume and assuring that the difference is not negative: Ui,10 = Vsl—0-5 (Qi+1,j + 10 ) lQi+1,j |_Qi,j+ lQi,ji)At ( a or U,, 10 = 0 if the above is negative (10b) With the above information the constituent concen- trations can be changed to account for advection dur- ing a time step. The program uses the equation, Ci,j = (QeiJ-At ci,j—1_Qe I At cw. +Ui’20 Cs +Ui’21 c,j + 19 kgl Ui,k Ci—10+k,j) /Vsl.9 Lj+1 (11) These volumu move to alemenn’ . any 01,] ._ __ Segment I'm, ~i¢2 NI /.n «:2 i-3 r— 4 k=20 IS'” IZ II IO 9 8 6 "'l 2| ”1,8 ”I,7l:/I,6 V51 (”d/ll - Laud M ”Li" CUMULATIVE VOLUME Estuary mouth '0 ’a LONGITUDINAL CO0RD|NATE,K Wedge foe FIGURE 5,—Elements in the volume-transfer matrix, U. See text for definition of symbols. where c ’,-, J- and cid- are the constituent concentrations in sublayer elements before and after this computation (when j =1, the wedge concentration C,- replaces the term c ’i, j_1) and CS and ctj are the boundary conditions for the constituent concentration at the estuary mouth and at the wedge toe, respectively. (They are discussed in the section “Boundary Conditions”) The matrix U appearing in equation 11 is different for each of the three sublayers. In those cases where U,:,1o and Ugg or U131 are the only two nonzero terms in a row of the matrix U, this numerical scheme for modeling longitudinal advection is identical to the common explicit upstream finite- difference scheme. When there are other nonzero terms, the schemes are different, and the present scheme is stable while the common upstream differ- ence scheme is not. When water volumes from two or more elements combine in one element, as would occur in the example shown in figure 5, the inherent mixing of the volumes is numerical mixing in the longitudinal direction. The numerical scheme used for computing vertical advec- tion similarly introduces numerical mixing in the ver- tical direction. The boundary conditions for the constituent concen- trations (the concentrations in the inflowing water) at the estuary mouth, 03, are the same for the entire upper layer and the wedge. These boundary conditions are input data for the model. Boundary conditions for constituent concentrations at the wedge toe are required only for the upper layer. The model computes these boundary conditions with the equation 8 NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY C ’ +c ' C — lej Q3 f ij Qf t- —‘ ————_ i ’ 62;}. (93+ng Q, (12) which determines the concentration of a mixture of (1) water from the wedge at the toe with a concentration of C 1 and (2) freshwater entering the upper layer at the toe with a concentration of c f. The value of c ffor each constituent is input data for the model. DIFFUSION The final water-transport computation in a time step is for vertical turbulent diffusion between sublayers; there is no diffusion between the wedge and the upper layer in the model. To compute vertical transport by turbulent diffusion, the model uses a constant vertical diffusion coefficient, 6y, in the equation, ' €y~At d2. 1 I I [(Ci,j—1 _ci,j)j¢1 I / +(ci,j+1 _ci,j ) #3 ]’ (13) where c’LJ- and cid- now represent constituent concen- trations before and after this computation. CONSTITUENT MODELS GENERAL DESCRIPTION This section describes processes, other than water transport, that affect the constituent concentrations. The constituents discussed are salinity, temperature (heat), chlorophyll a (phytoplankton), 5-day BOD, and DO. Although information on the DO concentrations is the goal of the computations, the concentrations of the other constituents must also be computed because they affect the DO concentrations. SALINITY Salt enters the estuary from the sea and is trans- ported through the estuary by the processes already described. No other sources of salt are considered. The model computes salinities of the sublayer elements in the upper layer, but the salinity in the entire wedge is assumed to equal the salinity of the inflowing seawa- ter. The model treats sea salt as a conservative sub- stance; computations for chemical reactions or other processes are not necessary. TEMPERATURE The model computes the temperature of each of the elements in the upper layer, Tia" and also in the wedge, Tw,» In addition to heat transfer by advection and diffu- sion within the estuary, heat transfer occurs between the top sublayer and the atmosphere, and all sublayers and the wedge are heated (though not equally) by solar radiation. The program computes the heat transfer at the water surface by using the procedure given by Yotsu- kura, Jackman, and Faust (1973). The rate of heat transfer between the atmosphere and the water per unit area of water surface, H, is approximated by the linear expression, de H = HS+HL+Hb+(Ts—Tb) , 14 dTb ( ) H S = rate of solar energy penetrating a unit area of water surface; rate of incoming long-wave radiation per unit area water surface; heat transfer rate per unit area of water surface due to outgoing long-wave radiation, evaporation, and conduc- tion when the water surface tempera- ture is Tb; Tb a reference temperature; and Ts = water surface temperature. Positive values of these quantities denote heat trans- fer from the atmosphere to the water, and negative values denote heat transfer from the water to the at- mosphere. The program computes the quantities H b and de/ dTb once at the beginning of every day using daily av- erage meteorological data and T1, =T1, 3. Meteorological data required by the model are S H E II Ra = mean daily relative humidity, in per- cent; SR = total daily incident solar radiation, in gram-calories per square centimetre (g-cal/cm2); Ta = mean daily air temperature, in degrees Celsius; and Wu = mean daily wind speed, in feet per sec- ond. Also required is coefficient 0,, which is used to compute the incoming long-wave radiation. The heat transfer rate H1, and its derivative are com- puted using the following equations, which were given by Yotsukura, J ackman, and Faust (1973): H, = —[ea-(Tb+5)4] —[AWa(a—,8Tb)(E(Tb)-Ea)] _ [nyWa (a— 3Tb) (Tb—m] (15) CONSTITUENT MODELS 9 and de 3 dE(Tb) __ = _. T +5 _)\ —-aA'yPW . (16) dTb 460'( b ) Wa dTb a The three bracketed terms on the right side of equation 15 represent heat-transfer rates due to outgoing long-wave radiation, evaporation, and conduction, re- spectively. The variables appearing in equations 15 and 16 are defined as follows: E, = vapor pressure in air, in millibars; E (Tb) = saturation vapor pressure in air at temperature T1,, in millibars; P = atmospheric pressure, assumed to be 1,013 millibars; a = 595.9 g-cal/g, latent heat of vaporization at 0°C; ,8 = 0.545 g-cal-g-1-°C-1, change in latent heat of vaporization per unit change in temperature; y = 0.61X 10-3/°C, constant in the Bowen ratio relating heat transfer by conduc- tion to evaporation; 8 = 273 K (kelvin), constant to convert de- grees Celsius to kelvin units; 6 = 0.97, emissivity of water; A = 3.6><10-4 g-cm-2 hr-l-mb-l (ft/s)-1, mass-transfer coefficient for evapora- tion; and o- : 4.88x10-9 g-cal'cm—Z- hr-l-K-l, Stefan-Boltsman constant. With the exception of A the above coefficients were taken from Yotsukura, Jackman, and Faust (1973). The present value of A is based on more recent work by H. E. Jobson (oral commun., 1974) and is about twice the value given by Yotsukura, Jackman, and Faust (1973). The saturation vapor pressure at temperature T1, and its derivative are approximated with the expres- Sions _ _ 1) E(Tb) — EeXPQ Tb +6) (17) and dE(Tb) '1) = E T — 18 dTb ( 1’) (T,,+5)2 ( ) where g = 23.38 mbar; 7;: 5303.3 K; and §=18.1. The model approximates heat transfer to the water by incoming long-wave radiation, H L, by the following equation, proposed by Koberg (1964): HL 2 (C, +0.0263V173a)o-(Ta+ (3)4. (19) The coefficient C, is a function of the air temperature and the ratio of the actual to the clear-sky incident solar radiation. Koberg (1964) presents C, in graphical form; it has a numerical value of the order 0.7. The model computes the quantities H L, H b, and de /dT;J for each day using mean-daily meteorological data and the values of T13 at the start of the day for Tb. During each time step the temperature of each element in the top sublayer is changed according to the relation 73,3 = Tia + [HL+H,, +(Ti:3—Tb)(de/dTb)] At /(30.48 di), (20) where T’i,3 and T,,3 are the temperatures of an element before and after this heat-transfer computation. The numerical constant in equation 20 converts the dimen- sions of the element thickness, d i, from feet to cen- timetres. The amount of solar radiation penetrating a unit area of water surface during a time step, H3, is calcu- lated with equation 21a or 21b, HS = 81', SR Ni sin 0, Osesw (21a) t HS = 0, 0<0 or 0>77, (21b) where 0 is the angle 6 = 277(1,_ E _l)/N,, (22) 4 2 where I, is the number of the time step; N t is the total number of time steps in a day; and S ’R is the fraction of the incident solar radiation that penetrates the water surface. The model computes the effect of solar heating on each sublayer element in the upper layer by taking the difference between the solar energy entering a unit area of the top, H*,~, j+1, and leaving the bottom, H*,-,J~, of each element. These quantities are related by the equation H*i,j = H*i,j+1 eXP(—ki,jdi ), (23) where kid- is the solar-radiation attenuation coefficient for a sublayer element. Thus, during every time step the model changes the temperature of each sublayer element by the equation Ti,j = T'i,j + (H*i,j+1_H*i,j)/(30-48 di) , (24) where T’LJ- and TM are now the temperatures before and after a solar-heating calculation. The model changes the temperatures of elements in 10 the wedge, Twi, according to equation 25, which was derived with the assumption that all solar energy pass- ing through the wedge—upper layer interface heats the wedge elements: Twi = lei +H*i,1Bi /(30.48 At), (25) where T’Wi and Twl- are the temperatures before and after the solar-heating calculation. PHYTOPLANKTON In addition to calculating the transport of phyto- plankton by advection and diffusion, the model con- siders the growth, respiration, and settling of the phytoplankton. In this report growth is defined as the increase in phytoplankton biomass by both reproduc- tion and by the increase in size of the individual plankters. Respiration includes all those processes that cause a decrease in phytoplankton biomass and all those processes by which living and dead (decompos- ing) phytoplankton use oxygen. All phytoplankton are represented in the model as chlorophyll a. Chlorophyll a concentrations are denoted in the upper layer by CPL}, and in the wedge by CPwi. Because of insufficient data, no attempt is made to distinguish between differ- ent taxa, even though it is known that species of both freshwater and saltwater origin bloom in the estuary (see section “Supplemental Information” under "Phy- toplankton”), and that the different species probably respond differently to changes in salinity, tempera- ture, and other parameters. Growth and respiration are modeled by adding the amount ACPi’j = CP'i’J‘ (G *R )At (26) 01‘ ACPwi = CP’wi (G—R)At (27) to the chlorophyll a concentrations in the sublayer or the wedge elements, respectively. The quantities CP’W- and CP’wi represent chlorophyll a concentrations be- fore this addition. The variables R and G are respira- tion and growth rates. The respiration rate used by the model doubles with every 10°C rise in temperature (see for example McKee and Wolf, 1963, p. 284) and is given by R = Rzoexp[0.0693(T—20)], (28) where R20 is R at 20°C and T is the local water temper- ature. The growth rate is a function of the local tem- perature and solar-radiation intensity, G = FTFL Go, (29) where Go is the maximum growth rate and FT and FL NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY are functions of temperature and solar-radiation in- tensity, respectively. The temperature function used by the model is FT = 0, (T<10°) FT = 0.1 T—1.0, (10°ST<20°) (30) FT = 1.0, (T>20°). The solar-radiation function is FL: —exp(1———) (31) Where L is the depth-averaged solar-radiation inten- sity in the wedge or sublayer element where the growth rate is desired and Lo isL for maximum growth. Both FT and FL are shown in figure 6. Equation 30 is the authors’ approximation of the ef- fect of temperature on phytoplankton growth; it is based on data by J itts, McAllister, Stephens, and Strickland (1964) for some marine phytoplankton and on knowledge that phytoplankton blooms are normally observed in the Duwamish estuary when the freshwa- ter temperatures are about 20°C or higher. Equation 31, which describes the effect of solar radi- ation on phytoplankton growth, was previously used by Steele (1962) for computing the effect of light on photosynthesis. The model computes an average value of L for a sub— layer element with the equation H:j+1 1_exp(—ki’j di) L = (32) At kid dl. and for a wedge element with Hi*1 l—exp(—kw.Ai /Bi) L = _’ ——__—.‘ , (33) At kw, Ai/Bi where kid- and kw,- are solar-radiation attenuation coef- ficients for the sublayer and wedge elements. The attenuation coefficients for the sublayer and wedge elements are approximated by kw = kou +k CCPi’j (34) and kw,- = k0w+k cCPwi, (35) where kc” and kow are the attenuation coefficients of the sublayers and the wedge in the absence of phytoplank- ton (chlorophyll a) and kc is the increase in the attenu- CONSTITUENT MODELS TEMPERATURE, IN DEGREES CELSIUS 20 6 SOLAR-RADIATION FUNCTION FL, AND TEMPERATURE FUNCTION,FT O on O l 2 3 4 L/La,DlMENSlONLE$S SOLAR-RADIATION INTENSITY FIGURE 6,—Solar-radiation and temperature functions for comput- ing phytoplankton growth rates. ation coefficient due to a unit increase in chlorophyll a concentration. The effects of nutrient concentrations on phyto- plankton growth were not modeled in this study be- cause present nutrient concentrations in the Duwamish River estuary are believed high enough, even during phytoplankton blooms, not to have an im- portant effect on the growth rate. Data on phytoplank- ton in the estuary and a discussion of the probable effects of nutrients on phytoplankton growth rates in the estuary appear in the section "Estimated Influence of Nutrients on Phytoplankton Growth.” The net amount of oxygen produced in the upper lay- er by the phytoplankton through photosynthesis and respiration is assumed to be proportional to ACE-J. (See equation 45.) Because of settling of phytoplankton, the wedge re- ceives chlorophyll a from the upper layer and loses chlorophyll a by deposition of phytoplankton on the channel bottom; resuspension from the bottom is not modeled. For every time step, the model performs the following settling computation for each wedge ele- ment, equation 36, and each sublayer element, equa- tion 37: CPwi = CP’wl. + (CP’i,1—CP’wi )w A t B, /A,- (36) and CPi’J' = CP'LJ' + (CP'i.j+1 —CP,L"J' ) WA t/di, (37) where CP’wi and CPwi, and CP’i, j+1 and CPU are the chlorophyll a concentrations before and after the set- tling computation and w is a constant settling velocity. BIOCHEMICAL OXYGEN DEMAND The present model computes the BOD (5-day biochemical oxygen demand) in the upper layer of the estuary. BOD was not modeled in the wedge because the processes that transport BOD into the wedge are 11 largely unknown. A constant BOD, equal to the BOD of the seawater, was assumed for the entire wedge. Data from the Duwamish River estuary (Welch, 1969) show that the BOD is typically 1—2 mg/l (milligrams per litre) and exceeds 5 mg/l only during phytoplank- ton blooms. The present low BOD has only a small effect on the dissolved oxygen in the upper layer of the estuary. However, because increased discharges from RTP probably will increase the BOD in the estuary, BOD is modeled for the upper layer. Only the car- bonaceous oxygen demand was modeled because the time of travel from the RTP outfall to the mouth of the estuary is about 5 days. The nitrogenous oxygen de- mand, which was not modeled, becomes important only for longer periods. The sources of BOD in the model are those coming into the estuary from the river, the sea, and miscel- laneous sources such as small industrial or domestic sewage discharges within the estuary. The miscellane- ous sources downstream from First Avenue South Bridge are combined and distributed uniformly along the lengths of the three most seaward elements of the top sublayer. For every time step, the model performs the compu- tation 13013.3 = BOD ,( 3+BODM/(N, 62.4>< 10‘6 imax 2 Vs ), _‘ _ n n—Lmx 2 (38) for i = imax —2 to imax, where BODM is the miscellane- ous BOD, in pounds per day, added to the estuary downstream from the First Avenue South Bridge; the numerical constant in equation 38 converts the units of the term in which it appears to milligrams per litre; and V8,, is the volume of a sublayer element. The BOD of the freshwater flow at the wedge toe, BOD f, is computed within the model by the equation QR BODR +QRTP BODRTP+ 0.185 BOD QR +Q RTP T. (39) The quantity BOD f has units of milligrams per litre and is the BOD of a mixture of (1) the mean daily flow of the Green River at the Tukwila gage, QR, with a concentration of BOD R; (2) the mean daily flow rate from RTP, QRTP, with a concentration of BODRTP; and (3) miscellaneous BOD inputs to the estuary upstream from First Avenue South Bridge, BOD T, expressed in pounds of BOD per day. The model treats the decay of BOD as a first-order reaction and computes the change in‘BOD due to decay 12 with the equation ABODi’J- = —KBODBOD’,,J-A t, (40) where ABODM is the change in BOD of an element due to decay during a time step and KBOD is a temperature-dependent decay rate. The oxygen con- sumed by the BOD is assumed to be proportional to ABODi,J-. (See equation 45.) The decay rate at 20°C, K2OBOD, is data input to the model, and the decay rate at other temperatures is computed by the equation KBOD = KZOBOD exp [0.046(T—20°)], (41) which is derived from information given by Fair, Geyer, and Okun (1971, p. 645). DISSOLVED OXYGEN The processes in the model that affect the DO (dissolved-oxygen) concentrations are the exchange of oxygen between the water and the atmosphere, com- monly called atmospheric reaeration; the production and consumption of oxygen by phytoplankton; and the consumption of oxygen by BOD. These processes are modeled explicitly only in the upper layer. D0 in the wedge is modeled by using an oxygen-consumption rate as was done by Stoner, Haushild, and McConnell (1974). The effect of reaeration is computed—only for the top sublayer—by the linear equation D0,,3 = D013 +7‘ (DOsi,3—DO’,~I3) At, (42) where the quantities D013 and D0,,3 are DO concen- trations of elements in the top sublayer before and after the reaeration computation; r is a temperature- dependent reaeration coefficient; and DOS,,3 is the sat- uration concentration of the top sublayer element in segment i. The DO saturation concentration is com— puted by the equation 130s,3 = (487—2.65S,-,3)/(33.5+Ti’3), (43) where Si,j is the salinity in parts per thousand and TM is the temperature, in degrees Celsius, of sublayer element i, j.The equation was taken from a report by the Thames Survey Committee and the Water Pollu- tion Research Laboratory (1964, p. 349), but one coeffi- cient was changed to give a better fit to the data of G. C. Whipple and M. C. Whipple (American Public Health Association and others, 1971, p. 480). The model computes the reaeration coefficient, r, by the equation r = (5|v,|/d,5/3) exp [0.024(T,,3—20)], (44) where v,- is the velocity of the water in the top sublayer NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY relative to the velocity of that in the wedge. The di- mension of r in the equation is days —1 when U,- is in feet per second and d,- is in feet. The first term in paren- theses is an adaptation of an equation for the reaera- tion coefficient in homogeneous streams by M. A. Churchill, H. L. Elmore, and R. A. Buckingham, and the exponential function is a temperature correction made by H. G. Becker; both are reported by Fair, Geyer, and Okun (1971, p. 651). To account for the effects of BOD and phytoplankton, the model changes the DO concentrations of sublayer elements according to the equation DOW: DO’i,J-+DOC ACPi,J-+DOBOD ABODM, where the differences ACPL-J and ABODiJ are given by equations 26 and 40, respectively, and the coefficients DOC and DOBOD are the changes in D0 concentration associated with unit changes in chlorophyll a concen- tration and BOD. The DO concentration of each wedge element, Bowl, is changed for every time step by the equation now, = Do'w, —ADOwAt, (45) , (46) where ADOw is a rate of oxygen consumption in the wedge, which includes the effects of all dissolved-, suspended-, and benthic-oxygen demands. APPLICATION OF THE MODEL TO THE DUWAMISH RIVER ESTUARY GENERAL Computations for the model of the Duwamish River estuary were made using historical input data for parts of the periods June—September 1967—69 and 1971. A second set of computations was made using most of the input data for 1971 and increasing the effluent dis- charge from RTP to the maximum magnitude expected in the future. Parameters in the model that could be varied were adjusted to give the best agreement between computed and observed data for the 1968 period. These param- eters were not changed for the other years. INPUT DATA FLOW MODEL The items required by the model for computing flow in the estuary are listed in table 1. Values or equations for some items are given either in this table or in tables 2 and 3. The estuary geometry, wedge-toe location, tide stages, tidal-prism thickness, upper-layer thickness at mouth, slope of upper layer-wedge interface, and freshwater inflow are identical to those used by Stoner, Haushild, and McConnell (1974). Data for the cross APPLICATION OF THE MODEL TO THE DUWAMISH RIVER ESTUARY TABLE 1.—Data required for computing flow in the Duwamish River 13 TABLE 3.—Equations for computing location of wedge toe in estuary Duwamish River estuary [Equations were fit to data collected mostly when Qf was between 200 and 600 (13/5, with some Item Value data between 600 and 1,100 ftn/s] Geometry of the estuary____ See table 2_ x, = longitudinal coordinate of wedge toe, in feet upstream from Location of wedge toe ______ See table 3. 'estuar y mouth. Tide stage ________________ See section “Input Data,” Y = tlde stage, In feet above mean lower low water. paragraph 2- Qf = mean daily freshwater discharge, in cubic feet per second. Tidal-prism thickness (ft) __ Pfip :3: 3:31;); “Input Data,” X1 = 25,151+1,626 y_79_95 Y2 - 2 Thickness of upper layer at X2 = 2218814210 Y_36-102Y mouth (ft) ______________ D=2.5+0.3P,+0.003Qf. X3 = 21,276+768 Y+0.60 Y 5191»: 3f nigge/R)layer-wedge s 0 00006 X,1 = 20,400+638 Y in e ace __________ = . . = Freshwater inflow at head of 0 X5 20’200+526 Y estuary (ft3/s) ____________ Qf, see section “Input Data,” paragraph 3. Distribution of freshwater in- I . flow among sublayers ___- ij =0.1, 0.4, 0.5(J =1, 2, 3). x =19400, Y<—1.6, Average rate of transport of t-X Y> 6 (200 saltiwater ifitpstream from Q 200 39— 1’ X X Y/—i.6’ 200 g \400 wegetoe(3/s) __________ = . = _ 1—2 _ >—., < < Distribution of saltwater in- s x, X‘ 200 (Qt 200)’ f flow among sublayers _ ___ Entrainment velocity across upper layer-wedge inter- face (ft/s) ________________ Ue=—6.5><10‘7 +3.25x10“5P, +4.5 X 10 _BQf. Qs’j =O.65, 0.20, 0.15(j =1, 2, 3), Dimensions vertical veloc- ities at base of sublayers Ue’. =1.0, 0.8, 0.5 (j=1, 2, 3). Vertical diffusion coefficient I (fth) __________________ e,=0.0001. TABLE 2.—Relation of width to elevation at cross sections for computing geometry of the Duwamish River estuary [Elevations are in feet above mean lower low water, and widths are in feet. The mouth is defined as 3,500 ft downstream from the Spokane Street Bridge] Cross-section 1, at mouth:1 Elevation ____________ —31 —20 —9 2 15 Width ______________ 175 302 378 476 640 Cross-section 2, 2,000 feet upstream from mouth: Elevation ____________ —31 —20 —9 2 15 Width ______________ 175 302 378 476 640 Cross-section 3, 4,000 feet upstream from estuary mouth: Elevation ____________ ‘34 —22 — 10 2 15 Width ______________ 50 319 485 590 706 Cross-section 4, 5,300 feet upstream from estuary mouth: Elevation ____________ —49 —33 — 17 —3 15 Width ______________ 11 640 984 1,068 1,148 Cross-section 5, 9,900 feet upstream from estuary mouth: Elevation ____________ —31 —20 -9 2 15 Width ______________ 304 425 528 642 700 Cross-section 6, 14,700 feet upstream from estuary mouth: Elevation ____________ —30 —19 —8 3 15 Width ______________ 153 229 274 358 500 Cross-section 7, 19,000 feet upstream from estuary mouth: Elevation ____________ — 19 — 8 — 1 6 15 Width ______________ 128 368 435 506 587 Cross—section 8, 24,700 feet upstream from estuary mouth: Elevation ____________ — 19 — 8 — 1 6 15 Width ______________ 223 281 346 427 500 Cross-section 9, 28,700 feet upstream from estuary mouth: Elevation ____________ ~12 —2 4 10 15 ' 478 572 622 668 690 Cross-section 10, 31,100 feet upstream from estuary mouth: Elevation ____________ — 1 1 —2 3 8 15 Width ______________ 229 242 252 267 280 Cross-section 11, 38,200 feet upstream from estuary mouth: Elevation ____________ — 10 — 1 4 9 15 Width ______________ 120 141 162 183 200 lFor modeling, this cross section was assumed the same as cross-section 2. xt=X2— Xzzooxs @400), Y2—1.6, 400sof<600 xt=X3—X§ —){4 (Qf—600)y YB—1.6, 600 SQf<1100 X4—X 900 Y2—1.6, 1100sQf<2000 x,=X,— 5 (Qf—1100), x,=X,.,, szzooo sections were obtained from 1971 maps by the US. Army Corps of Engineers of the dredged part of the estuary and from measurements by the US. Geological Survey of cross sections upstream from the dredged channel. The slope of the interface and the equations used to compute the location of the wedge toe and the thickness of the upper layer at the mouth were deter- mined from observed data for the Duwamish River es- tuary. The errors in the upper layer thickness and wedge—toe location computed with the equations and the error in the interface slope probably average less than 25 percent. Tide stages at hourly intervals are input data for the model; they are computed using the data for Seattle, Wash., and the procedures published by the [US] En- vironmental Science Services Administration (1967— 71). The model computes the tide stage at each time step by linear interpolation between the hourly stages. The tidal-prism thickness, Pt, is used by the model to compute the upper layer thickness at the mouth and the entrainment velocity across the upper layer-wedge interface. P, is defined here as the difference between the sum of the two daily high and the sum of the two daily low tide stages. The mean daily freshwater inflow to the estuary, Qf, is the sum of the mean daily flow of the Green River at the Tukwila gaging station and the mean daily outflow from RTP, which was obtained from Metro records. Other freshwater inflows, such as that from the relic Black River, are negligibly small. 14 The values of the other required parameters, Q 'fJ-, Q3, Q’sj, Ue, U ’21-, and 6y given in table 1 were determined by trial and error. They were varied, within some logi- cal limits, until computed salinities agreed with ob- served salinities for the three sublayers at 16th Av- enue South, First Avenue South, and Spokane Street Bridges. In order to improve the agreement between computed and observed salinities at Spokane Street Bridge, the values of some of these parameters differ a little from those used by Stoner, Haushild, and McConnell (1974). These investigators did not verify model salinities with the data from Spokane Street Bridge because calculations of advective transport in the upper layer near the estuary mouth sometimes were unstable in the model they used. In general, the agreement between the computed and observed salinities at 16th Avenue South and First Avenue South Bridges is equal for the present and the earlier reported values of these parameters. The values of UK computed by the equation in table 1 are 25 percent higher than comparable values used by Stoner, Haushild, and McConnell (1974); the dimen- sionless quantities, Q’fj, Q’sj, and U’ej, also differ by about 25 percent. Vertical diffusion was not in the ear- lier model but was used in the present model to in- crease the computed salinities in the top sublayer near the mouth without further changing the flow rate there. A value for the vertical turbulent diffusion coef- ficient, 6y, was estimated by computing the average depth value for a homogeneous density flow (Jobson and Sayre, 1970) with a depth and velocity typical of the upper layer in the Duwamish River estuary and multiplying this value by about 0.01 to account for the stabilizing effect of the density stratification in the es- tuary. The reducing factor could only be estimated within one order of magnitude by extrapolating data summarized by Nelson (1972). CONSTITUENT MODELS BOUN DARY CONDITIONS The modeling of each constituent requires that the concentrations in the water flowing into the estuary at its mouth and at the toe of the wedge be known. These concentrations, called boundary conditions, are neces- sary input to the model and were determined from data for estuary water sampled by project personnel or au- tomatic monitors at the locations shown in figure 1. The monitors, presently (1975) operated by Metro, re- cord the conductivity, temperature, DO concentration, and other data at 6- to 12-minute intervals. In the model, the boundary condition is the same for the wedge and the entire upper layer at the estuary mouth. Constituent concentrations there were usually NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY determined from the data for water sampled near the bottom at Spokane Street Bridge. Boundary conditions at the wedge toe are required for only the upper layer. They are computed in the model by equation 12, which requires freshwater con- centrations, Cf, as input data to the model. Because transport and other processes act on the brackish water upstream from the wedge toe, it is seldom correct to equate of to the concentration observed in the fresh- water part of the river. In this study, values of Cf for temperature, DO, and chlorophyll a were calculated by using data collected in the vicinity of the wedge toe in the equation 0: (Q's Qs+Q7Qf> —CiQ'st- Q'er This equation was derived by solving equation 12 for of and deleting the j subscripts. The concentration Ct in equation 47 is the constituent concentration at the 3-ft (l-m) depth at the computed location of the wedge toe and was determined by interpolation between observed concentrations for the same depth at 16th Avenue South Bridge and East Marginal Way Bridge, and oc- casionally at Boeing Bridge. The observed concentra- tion near the bottom at 16th Avenue South Bridge was substituted for C1. Values of Q’s and Q’f for the 3-ft (1-m) depth were estimated by interpolation in the ver- tical direction between the centers of the sublayers. Thus, except for salinity or for some other conserva- tive constituents with slowly varying or steady con- centrations for which there are no sources in the reach between the freshwater part of the river and the wedge toe, the concentrations Cf are not necessarily the con- centrations in the true freshWater inflow. The concen- trations Cf are artificial concentrations computed from available observed concentrations so that when they are used in equation 12, the computed sublayer con- centrations at the wedge toe, ctj, approximate the observed concentrations there. Changes in concen- trations in the brackish waters between the end of freshwater flow and the toe are incorporated in the definition of Cf. (47) Cir- SALIN ITY The only additional input data needed to model sa- linity are the salinities of the seawater and the river water. In the Duwamish River estuary they are 25 and 0 ppt, respectively. TEMPERATURE Input data required by the temperature model are the temperatures of the seawater and of the freshwater inflow to the estuary, and the meteorological data that is used to compute heat transfer at the water surface. APPLICATION OF THE MODEL TO THE DUWAMISH RIVER ESTUARY The temperature of the sea is taken as the daily minimum temperature measured by the bottom monitor at Spokane Street Bridge. The temperature of the freshwater is hourly input data. These are computed—by the method described earlier—from data from the surface monitors at the 16th Avenue South and East Marginal Way Bridges. The meteorological data required by the model are the total incident solar radiation for the day and the mean daily values of air temperature, wind speed, rel- ative humidity, and a coefficient for computing inci- dent long-wave radiation. These data were determined from information collected at the Seattle-Tacoma In- ternational Airport by the [U.S.] Environmental Sci- ence Services Administration (1967—69) and the [U.S.]. National Oceanic and Atmospheric Administration (1971). The coefficient for computing incident long- wave radiation was calculated using the information given by Koberg (1964) and the other meteorological data. PHYTOPLANKTON The data required by the phytoplankton model, which computes chlorophyll a concentrations, are listed in table 4. The chlorophyll a concentration in the seawater is typical of what was found in samples from Elliott Bay. The chlorophyll a concentrations of the freshwater are daily input data for the model. These concentrations were calculated using equation 47 and data for water sampled from the 3-ft (1-m) depth at 16th Avenue South, Boeing, and East Marginal Way Bridges, and from near the bottom at 16th Avenue South Bridge. Data are available fOr samples collected twice daily about two times a week during the years 1967—69. Although hourly data are preferred, the data were sufficient only to estimate daily values of CPf. Usually, data were not available to evaluate directly many of the remaining parameters required in the phytoplankton model. Values in table 4 often were selected because they gave the best agreements be- tween the computed and observed chlorophyll a and DO concentrations for the summer of 1968. No syste- matic attempt was made to determine the combination of parameters that give a statistically best agreement between computed and observed data; instead, deci- sions were made by visually comparing the plotted data. Attempts were always made to select values of the parameters within the range of those published in reports on research of the individual phenomena or in other reports on ecosystem models (for example, DiToro and others, 1971; or Chen and Orlob, 1972). The phytoplankton settling velocity was chosen so that computed chlorophyll (1 concentrations agreed with observed concentrations in the wedge. The tabu- lated value is at the lower end of the range of the val- 15 TABLE 4.—Data required for the phytoplankton model Item Symbol Value Chlorophyll a concentration of seawater ____________ CPS 1 ng/l Chlorophyll a concentration of freshwater inflow at toe ____________________ CPf Variable Phytoplankton settling ve- locity __________________ w 0.00002 ft/s Solar-radiation-attenuation coefficient in upper layer in the absence of phyto- plankton (chlorophyll a) k: 0.25 ft—1 Solar-radiation-attenuation coefficient in wedge in the absence of phyto- plankton (chlorophyll a) kg, 0.125 ft’1 Change in solar radiation- attenuation coefficient due to a unit change in chlorophyll a concentra- tion Maximum growth rate of phytoplankton Solar-radiation intensity for maximum phyto- plankton growth Respiration rate at 20°C __ Oxygen produced during growth or consumed dur- 0.005 ft‘l/(Mg chlorophyll a/l) 0.2 hr—1 10 g-cal-cm—Z-hr—1 0.004 hr—1 ing respiration for a unit change in chlorophyll a concentration __________ c 0.125 mg 02mg chlorophyll a ues reported by Hutchinson (1967, p. 277) for marine diatoms. The solar-radiation attenuation coefficient in the ab- sence of phytoplankton in the upper layer, k0,“ was estimated, by using a few Secchi-disc depths in equa- tion 48, which was suggested by Poole and Atkins (see Sverdrup and others, 1961, p. 92). The Secchi-disc depths averaged about 6 ft (2 m) and were obtained during periods of low chlorophyll a concentrations: _ 1.7 Secchi-disc depth . The constant of 0.73 in place of 1.7 in this equation as given by Welch (1969, p. 15) was not used because the computed solar-radiation intensity at a depth of 10 ft (3 m) would be one-third the intensity at the surface. According to D. L. Todd of Metro, who scuba dives in the Duwamish River estuary, the light intensity seems much less at that depth (oral commun., Nov. 1973). Furthermore, if one used 0.73 in equation 48, the com- puted bottom of the photic zone (estimated by Ryther, 1963, as the depth where the solar-radiation intensity is 1 percent of the intensity at the surface) would be about 40 ft (12 m). However, according to Welch (1969, p. 7), the photic zone is about 12 ft (4 m) deep. Because Todd also noted that the water in the wedge was usually clearer than the water in the upper layer, the attenuation coefficient of the wedge in the absence kou (48) 16 of chlorophyll a was selected arbitrarily to be one-half the value used in the upper layer. Data from the Duwamish River estuary were not available to estimate the increase in the attenuation coefficient due to the presence of phytoplankton. The chosen value of kc is an estimate based on data given by Megard (1973), Platt (1969), Ryther (1963), and Ryther and Yentsch (1957). A maximum growth rate of 0.2 hr-1 used in the model gave the best agreement between computed and observed chlorophyll a concentrations in the upper layer. This value is about twice the maximum value computed with data from J itts, McAlister, Stephans, and Strickland (1964) after adjusting their data for the length of their light and dark cycles. A growth rate of 0.2 hr-1 is equivalent to about 3.5 doublings per 12- hour day, which is a little greater than the usual range of 1 to 3 doublings per day (Talling, 1962, p. 744). The solar-radiation (light) intensity for maximum growth, L0 = 10 g-cal -cm-2-hr-1, was selected so that maximum growth rates would occur during the clear summer days of the bloom periods. The computed daily average growth rate is a maximum at 3-ft (l-m) depth when the total daily solar radiation is 500 g-cal/cm2 and the chlorophyll a concentration is about 45 [Lg/l (micrograms per litre). J itts, McAlister, Stephans, and Strickland (1964) found that the optimum light inten- sity for cell division varied for different species and ranged from 4.5 to 18 g-cal-cm‘z-hr‘l. Strickland (1960, p. 11) reports that the optimum light intensity for photosynthesis is probably between 6 and 9 g-cal-cm‘2~hr—1. However, data by Ryther (1956) shows that optimum light intensities ranged from about 1 to 4 g-cal-cm-Z-hr-l. The respiration rate, R, was calculated by using an oxygen-use rate during respiration of 3.5 m1 (mil- lilitres) oxygen per hour per gram dry weight of phytoplankton (estimated from data in Gibbs, 1962, p. 64), a mass density of oxygen at 1.43 g/l (grams per litre), a phytoplankton dry weight-to-chlorophyll a ratio of 100 (estimated from data in Strickland, 1960), and the tabulated value of DOC=0.125 mg (milli- grams) oxygen per microgram chlorophyll a (obtained by the procedure described in the section “Production and Consumption of Oxygen by Phytoplankton.” The data yield ml 02 1.43 g R =3.5 ——————— g dry weight/hour l (100 g dry weight Mg chlorophyll a g chlorophyll a 0,125 mg 02 > =0.004/hr. At maximum growth, the computed oxygen- NUMERICAL MODEL OF THE SALT—WEDGE REACH OF THE DUWAMISH RIVER ESTUARY production rate by photosynthesis is Go DOC = 0.025 mg oxygen per microgram chlorophyll a per hour. Data by Verduin (1956) yields 0.0064 mg oxygen per micro- gram chlorophyll a per hour. For a photosynthetic quo- tient (moles of CO2/moles of 02) equal to 1.2, the com- puted maximum production rate of carbon is 8 grams carbon per gram chlorophyll a per hour. Barlow, Lorenzen, and Myren (1963) report values between 8.7 and 16.3 grams carbon per gram chlorophyll a per hour for a eutrophic estuary. However, both Welch (1969, p. 17) and Shimada (1958) report values of the order 4 grams carbon per gram chlorophyll a per hour. BIOCHEMICAL OXYGEN DEMAND Data required by the BOD model are listed in tables 5 and 6. The values for BODS and BODR were based on observed data. Mean daily values for the BOD of the RTP effluent were determined from plant records. The BOD added to the estuary by miscellaneous discharges (table 6) was estimated from data furnished by Metro (Cecil Whitmore, written and oral communs., 1973). The decay rate, K2030D=0.25 dayrl, was calculated with data obtained from laboratory analyses of RTP effluent by Metro (R.I. Matsuda, written commun, 1973). The amount of oxygen consumed during the decay of a unit amount of BOD, DOBOD=1.4, is related to the decay rate through the definitions of a first-order reaction and the 5-day BOD. The relationship is TABLE 5.—Data for the BOD model Item Symbol Value BOD of seawater ______________ BODS 1 mg/l BOD of Green River at Tukwila gage ______________________ BODR 1 mg/l Mean daily BOD of RTP efflu- ent ________________________ BODRTP Variable Miscellaneous BOD added to es- tuary upstream of First Ave- nue South Bridge __________ BODT See table 6 Miscellaneous BOD added to es- tuary downstream of First Avenue South Bridge ______ BODM See table 6 BOD decay rate ______________ K 123%1) 0.25 day—1 Oxygen consumed per unit de— cay of BOD ________________ DOBOD 1.4 TABLE 6.—Miscellaneous BOD inflows to the Duwamish River es- tuary from upstream, BODT, and downstream, BODM, from First Avenue South Bridge [BOD is in pounds per day] Year BODT BODM 1967 ________________________ 2,757 6,685 1968 ________________________ 824 6,685 1969 ________________________ 762 6,400 1971 ________________________ 762 0 APPLICATION OF THE MODEL TO THE DUWAMISH RIVER ESTUARY DOBOD =[1—eXP (—5 [(20300)]"1. (49) DISSOLVED OXYGEN A list of the data required by the DO model appears in table 7. Both the DO concentrations of the saltwater, DOS, and of the freshwater, DOf, are hourly inputs- to the model. D05 is set equal to the concentration mea- sured by the bottom monitor at Spokane Street Bridge, and DOf is computed using equation 47 and data from the surface monitors at 16th Avenue South and East Marginal Way Bridges and the bottom monitor at 16th Avenue South Bridge. The rate of oxygen consumption in the wedge, ADOw, is 25 percent greater than that reported by Stoner, Haushild, and McConnell (1974). The larger value was used because the entrainment velocity is higher and the residence time of water in the wedge is shorter in the present model than in the original model. The method of selecting DOBoD was described in the preceding section, and the procedure for estimating DOC is described in the section “Production and Con- sumption of Oxygen by Phytoplankton.” MODEL VERIFICATION GENERAL REMARKS Constituent concentrations for the Duwamish River estuary were computed by the model for most of the June-September periods of 1967, 1969, and 1971 and for most of the June-August period of 1968. In figures 7—23, the computed temperatures, DO concentrations, and chlorophyll (1 concentrations in the upper layer for parts of these periods are compared with the appro- priate observed data. The figures are grouped chron- ologically. Tide stages are plotted on the same figures as the temperature data. No chlorophyll a data were available for 1971; there- fore, the chlorophyll a concentration in the inflowing freshwater was assumed to be lug/l. As a consequence, TABLE 7.—Data required for the D0 model Item Symbol Value Year DO concentration of seawater_- DOS Variable ____ DO concentration of freshwater inflow at toe ________________ DOf Variable -___ Oxygen consumed per unit de- cay of BOD Oxygen produced or consumed during the growth or respira- tion of phytoplankton ______ DOC 0.125 mg Oz/ug chlorophyll (1 Oxygen consumption in wedge ADOw 0.035 (mg/D/hr 1967 0.029 (mg/D/hr 1968 0.033 (mg/D/hr 1969 0.020 (mg/l)/hr 1971 17 IL STAGE, IN METRES STAGE, IN FEET Illl'lllllllIllllllllll'lllll I— 'VIVll "IIIV' UIOUOU O 26 - - Freshwater N b 22 I6Ih Avenue South Bridge IN DEGREES CELSIUS a TEMPE RATURE. observed, 3-foat(l-me1re)depth -—— - CompuIed, top sublayer Bridge l2 I I I I l I l I I l l I 1 I l I I I I l I I I I l l I I I I l 5 IO I5 20 25 3| JULY I967 FIGURE 7.—-Water temperatures and tide stages in Duwamish River estuary at times of high and low tides during July 1967. '5 I I I I l I I I I I I I I | I I I I I I I I I I I I I lo I I I 3 5 a 0 O wry: 'lVl' STAGE, m FEET IN METRES STAGE, Freshwater (n 2 U) .I m o In Id In E9 [6 _ lg IGIh Avenue South Bridge 2 III a: D E E Observed, 3-fooI(I-mefre)dep1h —-—— 2 ————— E 20 - Spokane SIree1 Winn“. Ion wblayer .1 '4 Bridge l2IIIlllllJJAillllI|llllllllllll 5 IO IS 20 25 3| AUGUST l967 FIGURE 8.—Water temperatures and tide stages in Duwamish River estuary at times of high and low tides during August 1967. H 00 lllllllllllIlT llllllllllflTTeo Freshwater Computed top sublayer —— 120 - Computed wedge -------- - Observed: L A 3-foot(|-me1re) depth 3—feet(l-metre) above bed 0 II II Ioo - * ll 1 I IIO‘ CHLOROPHYLL fl CONCENTRATION, 4N MICROGRAMS PER LITRE I\ 30 - lI / l l l l 6° ' ll‘l I A l i l ' i' ‘H 40 _ |6ih Avenue III \l\/\ _ Sout.h Bridge ’Ijl‘ V ' \‘IA \A _ \Il _ 2° ”Aux/If M o #2???” ~4- ~-i>+~k +- ~--r~t 1’9‘4I I' ‘I ‘4' VIM-“104‘ l“r\:1 i‘zl’l' IO I5 20 25 JULY l967 FIGURE 9,—Chlorophyll (1 concentrations in Duwamish River es- tuary at 16th Avenue South Bridge computed at times of high and low tides and observed about twice a day two times per week during July 1967. g 1111IIIIIIIIIII|ITITIIIIIIri—rII'20 : Computed top sublayer—— 'l .1 ‘Compu’red wedge -------- ll h '4 _ IOO g Observed: ll 4 III\ E. — 3-fooI(I-meIre)dopIh A II I\ \ l _ so u) 3-feei(i-mefre) above bed 0 III ”1 li IlI é ‘ I‘ [fl/H ' Vi E) First Avenue ‘ \' \I\ _ 40 South Bridge “I \l \,’\ I\ Z ‘ V f\ — 20 _ A J/\V‘/A\’\/f AW’ 5 .rslb-(TLI.~_.- ..,.,,._L.-.-..--~‘-~"‘~3~"-4 V4,,- e ‘W: ._.._ “Elf/Y; 0 I. < 5 I00 - ii I _ 2 ii I g 60 — |\|\ ll I 8 ! ”\IIIIL e60- Miuillill - 40 — V ll I [I l - E l’ I II I II 8 Spokane Sireei JV ' I| I ll Illr1 I: 20- Bridge l A | I H Ile'\,/\_ 7 § w ” l ‘IIIII \«M 0 . JULY I967 FIGURE 10.—Chlorophyll a concentrations in Duwamish River es- tuary at First Avenue South and Spokane Street Bridges com- puted at times of high and low tides and observed about twice a day two times per week during July 1967. the computed phytoplankton (chlorophyll a) concentra- tions were sufficiently low so that neither photosyn- thesis nor respiration affected the computed DO con- centrations. Although comparisons are not shown, the generally good agreements between computed and ob- served salinities were about the same as those of NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY T 1 I I | I I I I I l I I I I r I T I l I I I I I 40 — Freshwater q 20 ._ " O E 5 so I ‘ a: :ompuIejtopdsublayer—-— I I II I u ompu e we gs -------- _ Ze°‘°”:°':°i;. I. . II‘I‘II‘Il‘iI I I‘III‘II‘I - so -me re a E 40i- 3-feet(l-mefre)ab:ve bed 0 ”V/«xfllll d ‘U lAI' ll v \V/ V U U} l, \KL ézoL |61h Avenue [\{x’ V l i _ 9 South Bridge fi/V V’VN ‘ I ”J : ora‘fl’c’V-‘(f “‘53” ---------------- $'“"“‘” ““5"?" ‘30 -. AA 5 ' Ai‘ i‘ so _ P 5 - \IPN‘J‘I W \“W’I/illiuw '2 i ' / v VI” In First Avenue I\ ‘l J V Hill, g ' South Bridge 1” AJ\)‘ 2° :30 ‘30 £1503: -:£:\~"“"*"‘“"‘“ " f “,4 x ‘ HIT-4nd o -J I'll >-_' 60 ' l‘l‘ I '1 I i ll l ll éqo— lTIMIIfIl,IilIlII~in\/\UI]\J// \JJV‘V ”T, g _ Spokane Street f‘lM ”'1' |I H I: in], II OZONNEI'IEO 'VrlIIIlIlIIII ll _ \I \I ‘1an A\ :1"— FJI I ‘~ - _ . . ‘ I 0 | 5 IO I5 20 25 3| AUGUST |967 FIGURE 11.—Chlorophyll a concentrations in Duwamish River es- tuary computed at times of high and low tides and observed about twice a day two times per week during August 1967. Stoner, Haushild, and McConnell (1974). Data for BOD also are not shown; the computed and observed BOD values were always low (about 2 mg/l or less). The low BOD had only a small effect on the computed DO con- centrations. For each period, graphs show the constituent con- centrations in the freshwater inflow, which are input data, and the observed and computed concentrations at 16th Avenue South and Spokane Street Bridges. Chlorophyll a concentrations at First Avenue South Bridge also are given. Although the model can supply output concentrations for every 15 minutes, all graphs were drawn through data plotted only for each high and low tide. At some low tides, the upstream end of the model (wedge toe) was downstream from 16th Av- enue South Bridge. For these times, the computed con- centrations at the wedge toe were plotted in place of the computed data at 16th Avenue South Bridge. The observed temperature and DO-concentration data are from the automatic monitors that sample water from about the 3-ft (1-m) depth; the chlorophyll a data are from samples taken from a similar depth and from near the streambed. All computed data are for top- sublayer elements which almost always contain the observed data-collection points near the water surface. Because 16th Avenue South Bridge is in the vicinity APPLICATION OF THE MODEL TO THE DUWAMISH RIVER ESTUARY _ IGIII Avenue Q IIIIIIIIIII-IIIIIIIIIIIIIIIIrITI8 IIIIIIII‘IIIIIII1IIIIIIII [III 20 _ - I6 - ‘ I3 ' -I4 3 - - I6 m '5 E - ‘ I2 " " '4 j Freshwater 3:" a. ~ - - I2 5 '0 0. °' 2 (I! _ < , _ I 3 3 g 0 E :I o :I I- - 8 3 6 5 i E Freshwater s 20 - 4 _ g z o ile - I; IS - - 4 9 n: " I— 7 l6 - 2 I6 - ff LIJ Observed, 3—toot(I-metre)depth — z ‘2’ Computed, top subloyer ----- 3 l4 - 8'4 .— z 8 5 I2 - z '2 ISNI Avenue 8 I' II I III . x I So th Bud e § Io - u 9 9 IO _ ,I ‘? § I 8 5‘ 3 ' I > (I) -‘ m 8 a U, 5 Observed, 3-foot(I-metre)depth — ’ Computed, top suboner O I l I I L I I I I I I I I I I I I I I I I I I I I I I I I I I 5 I0 I!) 20 25 3| JULY I967 FIGURE 12.~Dissolved-oxygen concentrations in Duwamish River estuary at 16th Avenue South Bridge at times of high and low tides during July 1967. 20 rIIIlIIIIIIIIIIIIIIIIIIIIIIIIT I8 ' Observed, 3-foot(I-metre )depth -— _Computed, top sublayer 6 3 I I3 I Spokane Street Brque 6 DISSOLVED-OXYGEN CONCENTRATION, IN MILLIGRAMS PER LITRE a 1 6 1 my 4 _ 2 - J O L I I I I I I I I I I J I I I I I I I J I I I I I I I I I I I 5 I0 I5 20 25 3I JULY '967 FIGURE 13.—Dissolved oxygen concentrations in Duwamish River estuary at Spokane Street Bridge at times of high and low tides during July 1967. of the wedge toe, the differences between the observed and computed data at that station are measures of the South Bridge A llLIllllILlllIllll IIIIIIIIIIII 2 25 so l5 AUGUST I967 FIGURE 14,—Dissolved oxygen concentrations in Duwamish River estuary at 16th Avenue South Bridge at times of high and low tides during August 1967. l6IIIIIIIIIIIIIIII'IITIIII‘III11 Observed, 3-foot(I-metre)depth——— I '4 ' Computed, top eublayer ---- I" i ‘ r l| I: '2’ » “IIIIIIIII [II ,‘I'JIIIH‘II IO- II-IIII :‘IJIIII‘IHI’II Hum 'I “I” w" v’ I: I h Spokane Street Bridge IN MILLIGRAMS PER LITRE 0| a) I DISSOLVED-OXYGEN CONCENTRATION, N A E: I 0 IIIIIIIIIIIIIIIIIIIIIIIIIIIII | 5 IO 25 3| l5 AUGUST |967 FIGURE 15.—Dissolved oxygen concentrations in Duwamish River estuary at Spokane Street Bridge at times of high and low tides during August 1967. errors in the boundary conditions at the wedge toe. The differences between the observed and computed data at Spokane Street Bridge at low tide are measures of the errors of the upper layer models. The concentrations at Spokane Street Bridge at high tide are strongly depen- dent on the concentrations in Elliott Bay. In each graph, the concentrations at 16th Avenue N) O U) “fut—J'sllllllllllllllllllIllllIlllll “IE 0 m IO - *3 u w a r 5 s 2 mg m 0 V v g I l l I v r v 0 E .5— - 24 “ Freshwater 22 20 w 2 I8 0) _J 3 I6 I6Ih Avenue Observed, 3-foot(I-metre)depth —- 8 South Bridge Computed, top sublayer ----- In 22 n: o u o E u? n: E < n: E $20 ' Spokane Street ‘ '4 Iu . I— Bridge Ie ' 15 JULY I968 FIGURE 16.—Water temperatures and tide stages in Duwamish River estuary at times of high and low tides during July 1968. (I) |‘J-Ellf’IIIIIIIIIIIIIII'III'IIIIIIIIIIIY “IE g In I0 -33 u I—"' 5 - I—5 "’ E m z o . , , l . . o _ .5 > .. 24 _ Freshwater _ 22 20 (I) 2 (I: I8 .J u: 0 I6 U) DJ '6: (9 [4 Lu 0 I6th Avenue E ' South Bridge Observed, 3-foot(l~metre)depth ‘ 22 ~ Computed, top sublayer ----- m _ [I 3 I— < I: LL! 0. 2 Lu I- 20 " ii Spokane Street Ie —:‘. 9. Bridge 4 I I.“ 3| | 5 IO 15 20 2 AUGUST ISGB FIGURE 17.—Water temperatures and tide stages in Duwamish River estuary at times of high and low tides during August 1968. NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY I I I l I I l I I I I I I I I I I I I I T I I I I I I I I I 40 ' - 20 Freshwater - 0 4o - - 20 ' l6th Avenue AIL South Bridge J 0 r+__++b~*—— firxfim‘fibfiz‘ ...... : Com uted to s bla e —— _ P P U Y Y _ 60 Computed wedge Observed: 3-foot(l—metre) depth A ' 4O 3-teet(I-metre)above bed 0 ' First Avenue / 20 $0th Bridge k/ +~~—+~r~a~n~~+am c -—-. o __ ’- .— CHLOROPHYLL a CONCENTRATION, IN MICROGRAMS PER LITRE 4O - — 20 - Spokane Street _ Bridge All 0 . ._ . I 5 IO 5 20 25 3| JULY l968 FIGURE 18,—Chlorophyll a concentrations in Duwamish River es- tuary computed at times of high and low tides and observed about twice a day two times per week during July 1968, E IIIIIIIIIIIIIIITIIIIIIIIIII40 5 _ Freshwater E” U’ S 3 60 I. Computed top Subloyer —— _ g f‘ A Computed wedge -------- g 40 AJ‘I H i Observed: 9 fl] l\,’\ I6th Avenue 3-toot(i-rnetre) depth A i 5 i w South BrIdge 2 J I ‘\.’\ 3-feet(I-metre) above bed 0 ‘ 2° — o “A .. V E OI? , -W,o~e.4~r_‘,~,, f‘fiflfiraf:n*cxr€.fi$_qnn_ . _ 60 '5 I “5 I- 4 \ _ E #V \J/ \l/\ First Avenue 40 g _ Ar South Bridge ‘ 20 8 A 'l\f\/\ w-v"~”~—~’~ m ........ l” -:E*‘x—'. merfincxee‘rfiflua_ u 60 O i t M i >- 40 r I — E A] \I \|\ Spokane Street 0 J i I i‘ B 'd 0:20 - i \‘I “ 9" ~ g I I III/m ,\ ~,\ 0 V \v If ‘ O | 5 IO 20 25 3| l5 AUGUST |968 FIGURE 19.—Chlorophyll (1 concentrations in Duwamish River es- tuary computed at times of high and low tides and observed about twice a day two times per week during August 1968. South and Spokane Street Bridges follow the trends in the concentrations in the freshwater inflow. Variations within a day, due to the effects of tides or solar- radiation intensity, are apparent in both the computed and observed data. TEMPERATURE The computed temperatures as shown in figures 7, 8, APPLICATION OF THE MODEL TO THE DUWAMISH RIVER ESTUARY 21 'Gn—IfIIIIIIIIIIIIIIIIIIIIIIrIIII l4— Freshwater Observed, 3-foot(I—metre)depfh —- Computed, Iop sublayer 16th Avenue South Bridge (—3 I l .h Spokane Street Bridge I DISSOLVED-OXYGEN CONCENTRATION, IN MILLIGRAMS PER LITRE m I l Oi llIII 4| 5 l I I I1 I II I I I IO I5 20 25 3| JULY I968 FIGURE 20.—Dissolved oxygen concentrations in Duwamish River estuary at times of high and low tides during July 1968. 16, 17, and 22 usually agree with the observed temper- atures within 2°C, and the daily averages usually agree within about 1°C. High temperatures typically are associated with the less saline water at low tides, and low temperatures with the more saline water at high tides. The association is expected because the seawater in summer is usually cooler than the river water. However, for some unknownreason, high tem- peratures occasionally are associated with high tide as observed at Spokane Street Bridge in August 1967 (fig. 8); the water there is usually a few degrees cooler than it is at 16th Avenue South Bridge. The temperatures in the wedge, which are not shown in these figures, usu- ally did not vary spacially by more than 1°C. During the summer months, the wedge temperatures were usually between 10° and 14°C. CHLOROPHYLL a Computed and observed chlorophyll a concentrations are compared in figures 9, 10, 11, 18, and 19. Three phytoplankton blooms were observed during these periods, one each in July and August 1967 and one in late J uly-early August 1968. High DO concentrations, often in excess of the saturation concentrations, are associated with each of the blooms (figs. 12—15, 20—21). During blooms, the chlorophyll a concentrations were '4 IlI‘IIIIIIIIIIIIIIIIIIIIIIIII Freshwater — l4 Observed, 3—foot(l-metre)depth —— _ |2 Computed, top sublayer ----- _ IO IGIh Avenue South Bridge 5 K) 3 I DISSOLVED-OXYGEN CONCENTRATION, IN MILLIGRAMS PER LITRE w 2 I I I I | l I I I l I l I I I I I I 1 I I I I | 5 IO I5 20 25 3| AUGUST I968 FIGURE 21.—Dissolved-oxygen concentrations in Duwamish River estuary at times of low and high tides during August 1968. m- E '5 I I I I I I I I I I I I I I I I I I I I I I I I I I I _ Q :9 I,” I0 - 3 III ,_ <1 II. 5 2 III 5 z '— 2 ~ g I ' v 0 ‘n E Freshwater I8 In 2 IG In _I 3 I4 (I) , $12 — Ism Avenue - Is g South Bridge “1 . ‘3 ‘ — I6 g . . . "I . I4 - I '. ' " I ‘- ' a ' x :z A ~ I; la - ML ."~\,"— I2 5 Spokane Street Observed, 3-foot(I-me1re)depfh —— V §I6 BrIdge Campuied, top sublayer ----- _ I0 I“: M I4 nv/“I ,‘\"I n‘ ‘ A» m I ,. I I II \ \ — « .. I ., w, I Ix; . I ~ \ l2 - ‘w‘ r K. III,’ I0 I I I I I I I I I l I I I IJ L I I I I I I I I I I I I I I 5 I0 IS 20 25 30 SEPTEMBER |97| FIGURE 22.—Water temperatures and tide stages in Duwamish River estuary at times of high and low tides during September 1971. ‘ l2IIIIllljrlIIII|IIlI1ITII1IIII Freshwater Observed , 3—fooI(I-metre)depth — Computed, top sublayer ' leth Avenue 50th Bridge _ Spokane Street Bridge DISSOLVED-OXYGEN CONCENTRATION, IN MILLIGRAMS PER LITRE m I \_ M , A _. I , . . A ,A..-_~, “,9 k I» . pg. .4 V I. «w ' 4 l I I I I I I I I I I I I l I I I I I I I I I I l I I I I I 5 I0 I5 20 25 30 SEPTEMBER |97l FIGURE 23.—Dissolved oxygen concentrations in Duwamish River estuary at times of high and low tides during September 1971. much higher in the upper layer than in the wedge. The agreement between the observed and computed chlorophyll a concentration data is poorer than that for the temperature data. The larger errors in the phyto- plankton model mostly are due to oversimplified mathematical descriptions (equations 26 through 31) of the biological processes and somewhat to the in- accuracies introduced by the insufficient data used to define the boundary conditions for chlorophyll a. The phytoplankton blooms observed at 16th Avenue South, First Avenue South, and Spokane Street Bridges are simulated by the computed data. However, these blooms were not initiated within the modeled region of the estuary; the blooms were inputs to the model in the freshwater inflow and were maintained and advected through the estuary. Chlorophyll a concentration in the upper layer also was modeled without phytoplankton growth in the es- tuary downstream of the wedge toe. Results of this simulation (not shown in the figures) indicated that chlorophyll a and DO concentrations in the upper layer were too low relative to the observed concentrations. Thus, phytoplankton in the estuary downstream of the wedge toe must grow sufficiently to balance the dilut- ing effect of the water entrained from the wedge. Data for samples taken at East Marginal Way Bridge at low tide when the water there was fresh and at Renton Junction where the water is always fresh rarely show chlorophyll (1 concentrations greater than 10ug/l or supersaturated DO concentrations. However, the chlorophyll a concentrations in the freshwater in- flow, which are artificial concentrations computed from observed data in the vicinity of the wedge toe, gener- NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY ally were much higher than 10 ,ug/l during phyto- plankton blooms, and the computed DO concentrations in the freshwater inflow often were above saturation levels (figs. 12, 14, 20, 21) during blooms. These differ- ences between the actual freshwater concentrations and the artificial freshwater input concentrations suggest that the brackish-water reach upstream from the wedge toe is an incubator for the phytoplankton that bloom in the estuary. Thus, the biological process- es in this brackish-water reach are important in de- termining the phytoplankton populations in the upper layer of the salt—wedge reach of the estuary. The agreement between computed and observed chlorophyll a concentrations is better for the 1968 bloom than for either of the blooms in 1967. In 1968, the observed bloom at both First Avenue South and at Spokane Street Bridges lasted as long as at 16th Av- enue South Bridge, and the magnitudes of the observed chlorophyll a concentrations were about the same at the three stations. The input of high chlorophyll a con- centrations in the freshwater inflow to the estuary dur- ing the relatively long 1967 period of blooms generally is confirmed by the observed concentrations at 16th Avenue South Bridge. However, during 1967, observed data show that blooms did not persist as long and chlorophyll (1 concentrations were not as high at First Avenue South and Spokane Street Bridges as at 16th Avenue South Bridge. For example, the August 1967 bloom is barely discernible in the observed data for Spokane Street Bridge (fig. 11). Because the computed chlorophyll a concentrations at the two downstream locations closely follow the concentrations at 16th Av- enue South Bridge, the computed concentrations at Spokane Street and First Avenue South Bridges are in error for the 1967 blooms. The reasons for the lack of downstream persistence of the phytoplankton blooms in 1967 as compared to 1968 are not know. Possible explanations include the following: (1) the species that bloomed in the 2 years could have been different, and those that bloomed in 1967 perhaps could not grow well in the more highly saline water of the downstream part of the estuary; (2) herbivores that graze on phytoplankton could have been more numerous in the estuary in 1967 than in 1968; and (3) a toxic substance that inhibits phyto- plankton growth could have been present in the lower estuary during 1967 but not in 1968. The few available data on the phytoplankton species present in the estuary during blooms appear in table 12. Data for the 1968 bloom exist only for the station at First Avenue South Bridge and only for August 5. On that day, the most abundant taxa were oval flagellates that probably were of marine origin. Oval flagellates were also the most abundant taxa at First Avenue APPLICATION OF THE MODEL TO THE DUWAMISH RIVER ESTUARY South Bridge during the bloom on August 24, 1967, but not 2 days earlier on August 22, 1967. On this earlier date Cyclotella sp., which is of freshwater origin, was most abundant. The available data on the herbivores in estuary samples on a few days during the blooms of August 1967 and August 1968 (table 13) show similar concen- trations of herbivores during both blooms. The phyto- plankton loss rate caused by grazing can be estimated by multiplying the filtering rate of an individual or- ganism by the concentration of the grazing organisms. Hutchinson (1967, p. 528) reports filtering rates ob- served by L. A. Erman for the rotifer Brachionus cal- cyciflorus of 0.002 to 0.014 ml per hour per organism. Thus, the loss rate due to herbivore grazing is approx- imately (0.005 ml hr—1 organisms-1) (10 organisms ml-l) = 0.05 hr—l. Because this value is one-fourth the maximum phytoplankton growth rate used in the model (0.2 hr-l) and is probably nearly as large as the daily average phytoplankton growth rate, grazing by herbivores could be an important mechanism for controlling the phytoplankton populations in the estuary. However, because so few data exist on the concentration of her- bivores in the water flowing into the estuary, modeling herbivore populations and their interactions with the phytoplankton in the Duwamish River estuary is not feasible at present. No data from the Duwamish River estuary are available to support or refute the possibility of toxi- cants affecting the growth of phytoplankton. DISSOLVED OXYGEN Observed and computed DO concentrations are com- pared in figures 12—15, 20, 21, and 23. The data show daily peaks in the DO concentrations during phyto- plankton blooms at both 16th Avenue South and Spokane Street Bridges. During the 1967 period of blooms, the period of computed high DO concentrations lasted a few days longer than the period of high ob- served concentrations, as did the period of the com- puted high chlorophyll a concentrations. Except during blooms the computed and observed DO concentrations in the upper layer usually agree Within about 2 mg/l. The daily means often agree within 1 mg/l. The wedge DO concentrations (not shown) computed with the present model are nearly identical with those computed previously by Stoner, Haushild, and McCon- nell (1974) with the earlier version of the model, and the computed and observed concentrations usually agree within 2 mg/l. 23 PREDICTION OF FUTURE DISSOLVED-OXYGEN CONCENTRATIONS The model was used to estimate changes in D0 con- centrations in the estuary in response to an increase in effluent discharge from RTP, with J une-September 1971 used as the test period. One set of computations was made using historical input data as described in the preceding section. Another set of computations was made using much of this same data but with some changes to account for the effect of an increased flow from RTP. Four variables changed in the predictions were the flow rate and BOD of the RTP effluent, the DO concentration of the freshwater inflow, and the oxygen consumption rate in the wedge. The changes were estimates based on the designed capacity of RTP and on information provided by Metro personnel, who had given consideration to the available preliminary estimates furnished by their consultants. For the predictions, the RTP effluent discharge was increased to 223 ft3/s (6.31 m3/s) as compared to an average of about 37 ft3/s (1.05 m3/s) during June- September 1971. The BOD of the future RTP effluent discharge was assumed to be 5 mg/l as compared to the daily mean of about 3 mg/l (range 1—11 mg/l) during June-September 1971. All freshwater DO concentra- tions, DOf, were decreased by 2 mg/l to account for increased oxygen consumption between RTP and the wedge toe and an increased quantity of effluent with a low DO concentration in the freshwater inflow. Stoner, Haushild, and McConnell (1974) estimate that oxygen in the wedge might be consumed at the rate of 0.27 mg/l per hour for a future RTP effluent discharge of 223 ft3/s (6.31 m3/s). For this study their consumption rate was increased to 0.34 mg/l per hour, a 25-percent increase, to account for the increased en- trainment rate from the wedge as was described ear- lier. Results for both sets of computations for the month of September appear in figure 24. The freshwater DO concentrations shown are all 2 mg/l less than those determined from observed concentrations in Sep- tember 1971 (fig. 23). The differences between DO con- centrations computed with and without increased RTP effluent discharge for Septermber 1971 vary little with time but are less at Spokane Street Bridge than at 16th Avenue South Bridge. Spatial and temporal variations of differences for June-August 1971 (not shown in the figures) were similar to those for September. The estimated monthly average decreases in D0 concentrations in the Duwamish River estuary during June-September 1971 are given in table 8. These de- creases are all less than the constant decrease of 2 mg/l assumed for the concentration in the freshwater inflow to the estuary, DOf. Most of the computed difference is 24 '0 1 I I I I I I I I I I I I I I I I I I I I I I F] IT I 8 II F ‘ uture ‘ l I _ Freshwater It” ,‘II'II' :II/ ,I , I z '1‘" I III I I 6 — ’ A ‘ I I A r ICI’ ,“L_,.."I;IIII’I,I’II'I'I,I“ I ,I ,I/ WI!" L ‘A.IL" ,A. I’V‘ /.'\I‘I’ 17 < 5 My ‘.' I'I’ ‘ ‘ u \"~ ’ ‘J E I; 4 - 5 _I ‘é a: - - I0 8 E Iem Avenue A I97I z 2 _ South Bridgjw I\ \IA’JI'I\I/Furure r a w < UNI] V\ A. \I 2% :I/VVWIIN ,f'I/H AIM/I \‘M‘A’W/‘I\‘MNA¢VV VP \- 6 0:: v ‘-"’v I I,‘ 'I‘ n ,,/\_J“\_ “II r fl fil .55 . m I IA-.. kw! In A , . ‘- [LI _‘I a > 4 _I z 8 _ ‘2 8 -Spokane Street |97I - a Bridge NV‘ Future /\ saws T ~35 Xvwvrwwmw/MN / ”\— A I”. (WW \v' \ , MW" v"_/~I, thIf\\,¢-—\x /"\,I’" 4'IIK‘11111IIIIIIIIIII’lIIIIIIIII I 5 IO IS 20 25 30 SEPTEMBER l97| FIGURE 24.—Dissolved-oxygen concentrations in the top sublayer of the Duwamish River estuary model at high and low tides during September 1971, computed for RTP effluent discharges for 1971 (monthly average=35.7 cubic feet per second or 1.01 cubic meters per second) and for the future (223 cubic feet per second or 6.31 cubic meters per second). caused by the decreased D0 of the freshwater inflow and to the increased oxygen consumption in the wedge water which eventually is entrained into the upper layer. The increased consumption of oxygen by the in- creased BOD in the upper layer accounts for less than 10 percent of the computed difference. At 16th Avenue South Bridge in the upstream part of the estuary, the decreases in D0 concentrations of the top and middle sublayers are influenced mostly by the decrease in D0,», whereas the decreases in D0 con- centrations in the bottom sublayer respond more to the decreases in wedge DO concentrations. The average differences at 16th Avenue South Bridge (table 8) were computed using data at times of the high and low tides (usually a total of four times per day). At Spokane TABLE 8.—Estimated decreases in the monthly averages of computed D0 concentrations in the Duwamish River estuary during June- September 1971 for an increase in the RTP effluent discharge to the probable future maximum Decreases (mg/l) Station June July Aug. Sept. 16th Avenue South Top sublayer ______ 1.9 1.8 1.7 1.6 Bridge, averages Middle sublayer _- 1.7 1.7 1.5 1.4 for high and low Bottom sublayer __ 1,2 1.2 1.3 1.2 tides. Wedge ____________ .7 .8 1.0 1.1 Spokane Street Top sublayer ______ 1.6 1.6 1.5 1.3 Bridge, avera es Middle sublayer __ 1.1 1.3 1.2 1.1 for low ti es Bottom sublayer _- .5 .7 1.0 .7 only. Wedge ____________ .2 .3 .2 .2 NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY Street Bridge near the mouth of the estuary, computed DO concentrations in the wedge and the sublayers at high tide are strongly dependent on the DO concentra- tions in Elliott Bay. Therefore, the average differences at Spokane Street Bridge (table 8) were computed using only data at times of the low tides. Because of a lack of chlorophyll a data for 1971, chlorophyll (1 concentrations were not computed for 1971 or for the future when the RTP effluent discharge rate increases. However, because the nutrient concen- trations are believed to have been sufficiently high in 1971 so as not to limit the growth of phytoplankton (see section "Supplemental Information”), the in- creased RTP effluent should not affect the phyto- plankton concentrations in the estuary or the amount of oxygen produced or consumed by phytoplankton. The predictions which were made using estimates of the probable future values of the input data and parameters indicate that the decreases in D0 concen- trations in the upper layer of the estuary will not be more than the decreases in D0 concentration of the freshwater input at the wedge toe. This indication agrees with the relatively unimportant influence of BOD on DO concentrations in the sublayers noted in the modeling of the historical data. Lastly, the predic- tions were made using preliminary estimates of the effect of future conditions on model inputs and parameters; the confidence in the predictions will in- crease as these estimates improve. SUMMARY AND CONCLUSIONS A numerical model of a salt-wedge estuary de- veloped by Fischer (1974) has been expanded and used to calculate the distributions of salinity, temperature, chlorophyll a concentrations, BOD, and DO concentra- tions in the Duwamish River estuary, King County, Wash. The section of the estuary included in the model extends from the estuary mouth to the toe of the salt- water wedge. The location of the wedge toe, and hence the upstream boundary of the model, is a function of river discharge and tide stage. In the model, the estuary is divided vertically into the wedge and the upper layer, the latter in turn being divided into three sublayers. Longitudinally, the es- tuary is divided into about 35 segments; laterally, the estuary is assumed to be homogeneous. The water- transport processes modeled were longitudinal advec- tion and dispersion in the wedge, entrainment from the wedge to the upper layer, longitudinal and vertical ad- vection in the upper layer, and vertical diffusion in the upper layer. These transport processes were computed using conservation of volume equations, observed data, and predicted tide stages. Additional processes in the model for computing REFERENCES temperature are heat addition to the water by solar radiation and heat transfer at the surface by long-wave radiation, evaporation, and conduction. Phytoplankton are modeled as a chlorophyll a con- centration. The growth of phytoplankton is computed as a function of temperature and light intensity; com- puted respiration is a function of temperature only. Observed nutrient concentrations are believed to be sufficiently high so that nutrients are not limiting the phytoplankton growth in the estuary. Therefore, nu- trients and their effects on phytoplankton were not modeled. The few data on herbivores suggest that their numbers can be sufficiently high to affect the phyto- plankton population; however, the data are insufficient to allow the modeling of herbivore populations and their effects on the phytoplankton. Additional processes in the model for computing DO concentrations in the upper layer are oxygen transfer at the surface, consumption of oxygen by BOD, and the production and consumption of oxygen by phytoplank- ton during photosynthesis and respiration. In the wedge the model uses a constant oxygen-consumption rate. The BOD in the upper layer of the estuary is low, and its effect on the DO concentration is small. The production of oxygen by photosynthesis in the upper layer is sufficient to cause supersaturation levels of DO during phytoplankton blooms but is relatively unim- portant at other times. The effect of phytoplankton respiration on DO concentrations in the Duwamish River estuary is small. The model was calibrated with data for the summer months of 1968 and verified with data for the summer months of 1967, 1969, and 1971. At any given time the observed and computed temperatures in the upper layer at two stations in the estuary agree within about 2°C, and the daily average temperatures usually agree within 1°C. The computed chlorophyll a concentrations increased and decreased with the observed data during a phytoplankton bloom in 1968, but the computed blooms in 1967 persisted farther downstream and lasted a few days longer than did the observed blooms. The blooms are initiated in the brackish waters up- stream from the wedge toe and are advected into the modeled reach through the upstream boundary. Thus, the biological processes in the brackish reach upstream from the wedge toe (outside the model boundaries) are important in determining the phytoplankton popula- tions within the modeled reach of the estuary. Except during bloom periods the computed and ob- served DO concentrations agreed within about 2 mg/l. During blooms both the computed and observed DO concentrations reached supersaturated levels; how- ever, the error in the computations during the blooms is larger than during periods without blooms. 25 The effect on the DO concentration of increasing the discharge of treated RTP sewage effluent to the pro- posed future maximum (223 ft3/s or 6.3 m3/s) was esti- mated with the model. When using the summer data of 1971 as a base, the computed monthly average DO con- centrations in the estuary decreased by 2 mg/l or less. Because the concentrations of nutrients in the Duwamish estuary are presently believed to be suffi- ciently high so as not to limit the growth of phyto- plankton, the increase in nutrient concentrations caused by the increased amounts of sewage in the es- tuary should not affect the phytoplankton growth rates. REFERENCES American Public Health Association, American Water Works As- sociation, and Water Pollution Control Federation, 1971, Stand- ard methods for the examination of water and wastewater [13th ed.]: Washington, DC, 874 p. Barlow, J. P., Lorenzen, C. J ., and Myren, R. T., 1963, Eutrophica- tion of a tidal estuary: Limnology and Oceanography, v. 8, no. 2, p. 251—262. Bella, D. A., and Grenney, W. J ., 1970, Finite difference convection errors: Am. Soc. Civil Engineers Proc., J our. Sanitary Eng. Div., v. 96, no. SAG, p. 1361—1375. Chen, C. W., and Orlob, G. T., 1972, Ecologic simulation for aquatic environments, final report: Walnut Creek, California, Water Resources Engineers, Inc., 156 p. Cupp, E. E., 1943, Marine plankton diatoms of the west coast of North America: Scripps Inst. Oceanography Tech. Ser. Bull. 5 (1), 238 p. Dawson, W. A., and Tilley, L. J ., 1972, Measurement of salt-wedge excursion distance in the Duwamish River estuary, Seattle, Washington, by means of the dissolved-oxygen gradient: U.S. Geol. Survey Water-Supply Paper 1873—D, 27 p. Di Toro, D. M., O’Connor, D. J., and Thomann, R. V., 1971, A dynamic model of the phytoplankton population in the Sacramento—San Joaquin Delta, in Nonequilibrium systems in natural water chemistry: Am. Chem. Soc. Advances in Chemis- try Ser. 106, p. 131—180. Eppley, R. W., Rogers, J.N., and McCarthy, J. J., 1969, Half- saturation constants for uptake of nitrate and ammonium by marine phytoplankton: Limnology and Oceanography, v. 14, no. 6, p. 912—920. Fair, G. M., Geyer, J. C., and Okun, D. A., 1971, Elements of water supply and wastewater disposal: New York, John Wiley & Sons, Inc., 752 p. Fischer, H. B., 1972, A Lagrangian method for predicting pollutant dispersion in Bolinas Lagoon, Marin County, California: US. Geol. Survey Prof. Paper 582—B, 32 p. 1975, Description of the model, pt. 1, in A numerical model of material transport in salt-wedge estuaries: U.S. Geol. Survey Prof. Paper 917, p. 1—8. Gibbs, Martin, 1962, Respiration, chap. 4, in Lewin, R. A., ed., Physiology and biochemistry of algae: New York, Academic Press, p. 61—90. Hustedt, Friedrich, 1930, Bacillariophyta (Diatomeae), in Siisswas- serflora Mitteleuropas: v. 10, chaps. I—VIII, 464 p. Hutchinson, G. E., 1967, Introduction to lake biology and the limno- plankton, v. II of A treatise on limnology: New York, John Wiley & Sons, Inc., 1115 p. Jitts, H. R., McAllister, C. D., Stephens, K., and Strickland, J. D. H., 26 1964, The cell division rates of some marine phytoplankters as a function of light and temperature: Fisheries Research Board Canada Jour., v. 21, no. 1, p. 139—157. Jobson, H. E., and Sayre, W. W., 1970, Vertical transfer in open channel flow: Am. Soc. Civil Engineers Proc., Jour. Hydraulics Div., v. 96, no. HY3, p. 703—724. Koberg, G. E., 1964, Methods tr compute long-wave radiation from the atmosphere and reflected solar radiation from a water sur- face: U.S. Geol. Survey Prof. Paper 272—F, p. 107—136. Kofoid, C. A., and Campbell, A. S., 1929, A conspectus of the marine and fresh-water Ciliata belonging to the sub-order Tintinnoinea, with descriptions of new species principally fromthe Agassiz Expedition to the eastern tropical Pacific 1904—1905: California Univ. Zoology, pub. 34(1), 404 p. Lake Tahoe Area Council, 1968, Bioassay of nutrient sources, first progress report of Eutrophication of surface waters, Lake Tahoe: South Lake Tahoe, Calif, 178 p. 1969, Laboratory and pond studies, second progress report of Eutrophication of surface waters, Lake Tahoe: South Lake Tahoe, Calif, 180 p. 1970a, Pilot pond and field studies, third progress report of Eutrophication of surface waters, Lake Tahoe: South Lake Tahoe, Calif, 180 p. 1970b, First progress report of Eutrophication of surface wa- ters, Indian Creek Reservoir: South Lake Tahoe, Calif, 141 p. Leegaard, Cardine, 1915, Untersuchungen fiber einige Plantoneilli- nien des Neeres: Nyt. Mag. F. Naturv. L. 111 1. 570.5 NV, v. 53—54, 37 p. MacIsaac, J. J ., and Dugdale, R. C., 1969, The kinetics of nitrate and ammonia uptake by natural populations of marine phytoplank- ton: Deep—Sea Research, v. 16, p. 415—422. McKee, J. E., and Wolf, H. W., 1963, Water quality criteria: Califor- nia State Water Quality Control Board Pub. 3—A, 548 p. Megard, R. 0., 1973, Rates of photosynthesis and phytoplankton grth in Shagawa Lake, Minnesota, pub. EPA—R3—73—039 of Ecological Research Series of US. Environmental Protection Agency: Washington, D.C., US. Govt. Printing Office, 70 p. Nelson, J. E., 1972, Vertical turbulent mixing in stratified flow~a comparison of previous experiments: California Univ. (Berke- ley) Dept. Civil Engineering, Hydraulic Eng. Lab., Waste Heat Management Rept. Ser., Rept. No. WHM—3, 33 p. Platt, Trevor, 1969, The concept of energy efficiency in primary production: Limnology and Oceanography, v. 14, no. 5, p. 653— 659. Richards, F. A., and Thompson, T. G., 1952, A spectrophotometric method for the estimation of plankton pigments; pt. 2, The esti- mation and characterization of plankton population by pigment analysis: Jour. Marine Research, v. 11, no. 2, p. 156—172. Ryther, J. H., 1956, Photosynthesis in the ocean as a function of light intensity: Limnology and Oceanography, v. 1, no. 1, p. 61—70. 1963, Geographic variations in productivity, in Ideas and ob- servations on progress in the study of the seas, Volume 2 of Hill, M. N., ed., The Sea: New York, Interscience Publishers, p. 347- 380. Ryther, J. H., and Yentsch, C. S., 1957, The estimation of phyto- plankton production in the ocean from chlorophyll and light data: Limnology and Oceanography, v. 2, no. 3, p. 281—286. Santos, J. F., and Stoner, J. D., 1972, Physical, chemical, and biologi- cal aspects of the Duwamish River estuary, King County, Washington, 1963—67: US. Geol. Survey Water-Supply Paper 1873—C, 74 p. Schiller, Josef, 1930, Coccolithineae, in Rabenhorst’s Krypt- ogamen-Flora von Deutschland, Osterreich und der Schweiz, 10(2): p. 89—267. NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY Shimada, B. M., 1958, Diurnal fluctuation in photosynthetic rate and chlorophyll a content of phytoplankton from eastern Pacific wat- ers: Limnology and Oceanography, v. 3, no. 3, p. 336—339. Slack, K. V., Averett, R. C., Greeson, P. E., and Lipscomb, R. G., 1973, Methods for collection and analysis of aquatic biological and microbiological samples: U.S. Geol. Survey Water- Resources Inv. Techniques, book 5, chap. A—4, 165 p. Steele, J. H., 1962, Environmental control of photosynthesis in the sea: Limnology and Oceanography, v. 7., no. 2, p. 137—150. Stoner, J. D., 1972, Determination of mass balance and entrainment in the stratified Duwamish River estuary, King County, Washington: US. Geol. Survey Water—Supply Paper 1873—F, 17 p. Stoner, J. D., Haushild, W. L., and McConnell, J. B., 1975, Model computation of salinity and salt-wedge dissolved oxygen in the lower Duwamish River estuary, King County, Washington, pt. 11, in A numerical model of material transport in salt-wedge estuaries: U.S. Geol. Survey Prof. Paper 917, p. 11—36. Strickland, J. D. H., 1960, Measuring the production of marine phytoplankton: Fisheries Research Board Canada Bull. 122, p. 5—22 and 95—100. Strickland, J. D. H., and Parsons, T. R., 1968, A practical handbook of seawater analysis: Fisheries Research Board Canada Bull. 167, 311 p. Sverdrup, H. V., Johnson, M. W., and Fleming, R. H., 1961, The oceans, their physics, chemistry, and general biology: En- glewood Cliffs, New Jersey, Prentice-Hall, Inc., 1087 p. Talling, J. F., 1962, Freshwater algae, in Lewin, R. A., ed., Physiol- ogy and biochemistry of algae: New York, Academic Press, p. 743—757. Thames Survey Committee and Water Pollution Research Labora- tory, 1964, Effects of polluting discharges on the Thames es- tuary: Water Pollution Research Technical Paper No. 11, Lon- don, Her Majesty’s Stationery Office, 609 p., 1 pl. Tilley, L. J ., and Dawson, W. A., 1971, Plant nutrients and the es- tuary mechanism in the Duwamish River estuary, Seattle, Washington, in Geological Survey research, 1971: US. Geol. Survey Prof. Paper 750—C, p. C185—0191. [US] Environmental Science Services Administration, 1967—69, Tide tables, high and low water predictions, west coast of North and South America including the Hawaiian Islands: Washington, D.C., US. Govt. Printing Off, pub. annually. 1967—69, Local climatological data: Washington, D.C., US. Govt. Printing Office, pub. monthly. [US] National Oceanic and Atmospheric Administration, 1971, Local climatological data: Washington, D.C., US. Govt. Print- ing Office, pub. monthly. Verduin, Jacob, 1956, Primary production in lakes: Limnology and Oceanography, v. 1, no. 2, p. 85—91. Wailes, G. H., 1939, Canadian Pacific fauna, 1. Protozoa: 1a, lobosa, 1b, regiculoza, 1c, heliozoa, 1d, radiolaria, 1e, mastigophora, 1f, ciliata, lg, suctoria: Fisheries Research Board Canada, Toronto, Univ. Toronto Press, 104 p. Ward, H. B., and Whipple, G. C., 1959, Fresh water biology, edited by W. T. Edmondson 2d edfi New York, John Wiley & Sons, Inc., 1248 p. Welch, E. B., 1969, Factors initiating phytoplankton blooms and resulting effects on dissolved oxygen in Duwamish River es- tuary, Seattle, Washington: US. Geol. Survey Water-Supply Paper 1873—A, 62 p. Yotsukura, Nobuhiro, Jackman, A. P., and Faust, C. R., 1973, Ap- proximation of heat exchange at the air-water interface: Water Resources Research, v. 9, no. 1, p. 118—128. SUPPLEMENTAL INFORMATION 28 PHYTOPLANKTON GENERAL Slack, Averett, Greeson, and Lipscomb (1973) define phytoplank- ton as the community of suspended or floating plant organisms that drift passively with water currents. The Duwamish River estuary contains phytoplankton species of both freshwater and marine ori- gin; the former enter the estuary with the water from the Green- Duwamish River, and the latter enter from Elliott Bay. Only sus- pended phytoplankton are discussed in this report; phytoplankton that float on the surface are not of importance in the Duwamish River estuary at present. During their stay in the estuary, phytoplankters may grow, pro- duce oxygen by photosynthesis, decrease in mass and consume oxy- gen by respiration, be transported by currents, settle, and be eaten by grazers. The eventual fate of the phytoplankton not eaten by grazers in the estuary is either being transported into Elliott Bay or becoming a part of the streambed sediment. Phytoplankton growth and photosynthesis occur mainly in the upper layer of the estuary. Although the top part of the saltwater wedge is within the photic zone, light in most of the wedge is insuffi- cient to support active phytoplankton photosynthesis (Welch, 1969). The growth of phytoplankton and the production of oxygen by photo- synthesis are the important biological processes in the model. Al- though modeled, the use of oxygen and the decrease in biomass at- tributed to respiration are less important biological processes in the estuary. Grazing was not modeled. Phytoplankton photosynthesis increases DO concentrations in the estuary significantly during blooms but has a negligible effect on DO concentration during nonbloom periods. A phytoplankton bloom, as defined in this report, is either 0.5 million or more cells per litre (as defined by Slack and others, 1973) or a chlorophyll a concentration in excess of 4 peg/l (as defined by Welch, 1969) during a continuous period longer than a day. SAMPLE COLLECTION, PREPARATION, AND ANALYSIS The concentrations of chlorophyll a and other pigments were de- termined for water samples collected in the Green-Duwamish River and in Elliott Bay. The plant-pigment data were supplemented by phytoplankton cell counts and taxonomic identification for selected samples. The river samples were collected during 1967—69, usually at six stations starting at Spokane Street Bridge and ending at the Renton Junction monitor (fig. 1). The Elliott Bay samples were collected by Metro personnel (under the direction of R. I. Matsuda) at four sta- tions during the April-June periods of 1967—69. Bay samples were collected from depths of about 3 ft (1 m) and river samples were collected from about 3 ft (1 m) below the water surface (surface samples), or about 3 ft (1 m) above the streambed (bottom samples). Samples at most stations were collected with the Emsworth version of a 4-litre Van Dorn sampler. However, many river samples were obtained from the pumped stream supplying the automatic water- quality monitors. Immediately after sample collection, a 125 m1 aliquot was pre- served with Lugol’s solution for cell counting and identification. Phytoplankton were identified and counted by W. A. Dawson and L. J. Tilley using the inverted-microscope technique (Slack and others, 1973). For analysis of plant pigments, volumes between 0.5 and 4.0 litres were filtered through 0.45-Mm membrane filters. The river samples were filtered at about one-half atmosphere of positive pressure within 1 hour of the time of collection. The bay samples were sub— jected to slightly greater delay before filtration in a vacuum ap- paratus at no more than one-fifth atmosphere negative pressure. The NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY folded filters containing the phytoplankton were stored in desic- cators at 5°C. Acetone extracts from the filter and samples were analyzed for plant pigments by the method of Richards with Thompson (1952), as modified by Strickland and Parsons (1968). The work of Ward and Whipple (1959) was the most useful refer- ence for identification of freshwater organisms. Marine organisms were identified with the aid of the works of Cupp (1943), Wailes (1939), Hustedt (1930), Schiller (1930), Leegaard (1915), and Kofoid and Campbell (1929). ABUNDANT PHYTOPLANKTON TAXA The most abundant organisms found in Elliott Bay samples (table 9) were Skeletonema costatum and Thalassiosira sp. The samples listed in table 9 were selected for taxonomic analysis to represent times of peak chlorophyll a concentrations in the surface waters of Elliott Bay near Seattle during the April-June periods of 1967—69 (R. I. Matsuda, Metro, oral commun., 1971). Although the data in table 9 are for spring months, the two dominant genera probably enter the Duwamish River estuary with the bay water not only in the spring but in summer and early fall as well. Welch (1969) identified Skeletonema as the most abundant taxa in samples from 16th Av- enue South Bridge during a bloom in early August 1965 and specu- lated that an abundant diatom might be Thalassiosira in a later bloom during the same month. Skeletonema was one of the abundant organisms in samples from some estuary stations during blooms in August 1967 (table 12). The list of abundant freshwater organisms in samples from the Green-Duwamish River during March-August 1967 (table 10) shows a diversity of dominant genera. Several genera appear in both list- ings of abundant taxa in samples from the two stations in the fresh- water part of the river (tables 10, 11). Many organisms found in the freshwater samples from the river must have been periphyton which had been dislodged upstream and subsequently carried downstream by the river currents. The abundant organisms in samples from six stations on the es- tuary and river during 2 days of an August 1967 bloom and from one estuary station during 1 day of an August 1968 bloom are listed in table 12. The diversity in the abundant river taxa is again evident by the data for samples from the Renton Junction monitor station and East Marginal Way Bridge. The marine diatom, Skeletonema cos- tatum, was one of the abundant species only in the samples collected from the two farthest downstream stations on August 22, 1967. Cyclotella sp. and oval flagellates, along with "coccoids,” dominate the abundance listing for the estuary during the August 1967 blooms. Cyclotella sp., a freshwater diatom, was not numerous enough to be listed among the abundant freshwater organisms (ta- bles 10, 11); it ranked about ninth in abundance of the freshwater autotrophic organisms found in the Green-Duwamish River. The oval flagellates were probably marine organisms from Elliott Bay; oval flagellates were the 20th most abundant autotrophic organism found in the bay samples of the April-June periods of 1967—69. How- ever, because the taxonomic analyses of the estuary samples may not have distinguished between freshwater and marine oval flagellates, the probable origin of these microorganisms is designated in table 12 as “saltwater(?)” to indicate some uncertainty. Regardless of their origin, both Cyclotella sp. and oval flagellates flourished in brackish estuary water having a wide range in salinity. The longitudinal distributions of Cyclotella sp. (fig. 25) indicate a peak concentration in the brackish surface water between Boeing Bridge (mile 6.5 or kilometre 10.5) and First Avenue South Bridge (mile 3.4 or kilometre 5.5). Downstream of this reach, the consist- ently lower concentrations at Spokane Street Bridge indicate that the prevailing high salinity there may limit the growth of this di- atom. Upstream of the peak-concentration reach, the concentration SUPPLEMENTAL INFORMATION TABLE 9.4bundant taxa in samples collected at four stations in Elliott Bay near Seattle during 1967—69 [Phytoplankton samples collected by R. I. Matsuda, Municipality of Metropolitan Seattle, and counted and identified by W. A. Dawson and L. J. Tilley] 29 TABLE 11.—Five most abundant taxa in freshwater samples obtained at two stations on the Green—Duwamish River during March- A ugust 1 96 7 [Taxa counted, identified, and ranked by W. A. Dawson and L. J. Tilley] Chloro— phyll a Concentration concen- tration Percentage Presence D‘ate (lag/l) Taxon Cells/ml of total of bloom 1967 May 17 5.3 Sheletonema costatum 480 47 No Thalassiosira sp ______ 400 39 No May 17 4.1 Skeletonema costatum ._ 1,500 74 Yes Thalassiosira sp ____________ 430 22 No 1968 Apr. 29 9.8 Skeletonema custatum ______ 610 41 Yes Thalassiosira sp _______ 460 31 No Apr. 29 3.3 Skeletoncma costatum _ 850 75 Yes Flagellate, oval _______ 170 15 No May 27 11.4 Skeletonema costatum . 7,700 88 Yes Thalassiosira sp _.__ 610 7 Yes May 27 9.2 Skeletanema costatu 2,400 68 Yes Thalassiosira sp ”1 840 24 Yes June 3 11.7 ____________ o _____ 1,600 66 Yes Skeletonema costalum ______ 360 15 No 1969 May 14 4.6 12,000 98 Yes May 14 3.6 12,000 98 Yes May 21 2.3 4,700 91 Yes Chaetoceros sp _______ 370 7 No May 21 3.5 Skeletonema costatum A 4,700 86 Yes Chaetoceros sp ______________ 660 12 Yes TABLE 10.—Abundant taxa in 1967 freshwater samples from two stations on the Green-Duwamish River [Taxa counted and identified by W. A. Dawson and L. J, Tilley] Chloro- phyll a Concentration concen- Time tration Percentage Presence Date (P.s.t.) (Hg/1) Taxon Cells/ml of totals of bloom Renton Junction Monitor (mi 13.1, km 21.1 Mat. 14 1020 0.8 Unidentified blue-green 2,500 50 Yes algae. Apr. 18 1035 3.8 Hannaea sp _________ 34 27 No May 12 1200 2.5 Pennate diatoms ,,,,,,,,,, 90 17 No 24 1425 3.6 Unidentified blue-green algae 7,800 71 Yes June 8 1245 1.8 Chrysacoccus sp __________ 350 23 No 27 0805 3.6 "Coccoids" and clusters 2,300 40 Yes July 5 0820 4.0 Oscillatoria s __________ 1,700 34 Yes 11 0800 3.7 “Coccoids” and clusters 1,500 29 Yes 25 1245 4.5 Flagellate spp ________ 2,200 31 Yes Aug. 22 1630 2.7 "Coccoids," solitary ________ 5,600 61 Yes East Marginal Way Bridge] (mi 7.8, km 12.6) Mar. 14 1040 0.5 Crenothrix sp ,,,,,,,,,,,, 700 87 Yes May 12 1245 2.5 Synedra sp in 54 13 No 24 1020 2.2 Oscillatoria sp . 190 36 No June 20 0830 2.2 ____________ do ,,,,,,,,,, 930 38 Yes July 5 0845 3.3 "Coccoids" and clusters .,,, 840 25 Yes 25 1225 6.0 Flagellate Sp ____________ 3,900 36 Yes ‘Samples obtained during low tides when river did not contain saltwater here. of Cyclotella sp. is lower at East Marginal Way Bridge, mile 7.8 or km 12.6 (concentration being tide-dependent here), and decreases rapidly thereafter with distance upstream. Longitudinal distributions of oval flagellates (fig. 26) indicate peak concentrations downstream from Boeing Bridge (mile 6.5 or kilometre 10.5) with the concentration at Spokane Street Bridge (mile 1.2 or kilometre 1.9) either higher or only slightly less than concentrations at adjacent upstream stations. The rapid decrease in Taxon and sampling station Abundance Renton Junction Monitor East Marginal Way Bridge ranking (mi 13.1, km 21.1) (mi 7.8, km 12.6) 1 “Coccoids” and clusters ____________ Oscillatoria sp. 2 Micractinium sp W, ,,-. Pennate diatoms. 3 Scenedesmus sppi -___ Scenedesmus spp. 4 Golenkinia sp __- ____ "Coccoids" and clusters. 5 Oscillatoria sp __________ Chlamydomonas sp. concentration of oval flagellates with distance upstream from Boeing Bridge suggests that they are marine organisms reacting to the in- creasingly fresher water. In general, the concentration of “coccoids” decreases downstream and, therefore, with increasing salinity (fig. 27). PRODUCTION AND CONSUMPTION OF OXYGEN BY PHYTOPLANKTON When blooming, the phytoplankton in the upper layer of the Duwamish River estuary produce more oxygen by photosynthesis than they consume by respiration, frequently resulting in D0 super- saturation. DO supersaturation has not been observed in the wedge. _In the model the production of oxygen during photosynthesis is assumed to be proportional to the increase in chlorophyll a concen- tration caused by phytoplankton growth. Similarly, the consumption of oxygen during respiration is assumed to be proportional to the decrease in chlorophyll a concentration. The coefficient of propor- tionality was assumed to be the same for both processes, which im- plies a photosynthesis-to-respiration ratio of unity for stable popula- tions. The coefficient of proportionality relating the increase of D0 to chlorophyll a, which is denoted by DOC in the main body of this report, was estimated from data obtained in the Duwamish River and estuary during phytoplankton blooms. The procedure used is as follows: 1. The salinity, DO concentration, and the chlorophyll a concentra- tion, S1,, DOb, and CPb, were observed at a point in the river or estuary where there was a bloom. The salinity and DO concen- tration were observed at two other points, one upstream from the first where there was no bloom, S u and DO“, and the other in the wedge, S w and DOW. The chlorophyll a concentrations at the second two points usually were not observed but were known to be a small fraction of the concentration at the first point. 2. The relative volumes of water from the upstream point, V“, and from the wedge point, Vw, required to produce a mixture of salinity, S b, was computed as V = Sw—Sb “ SW—Sll and VW=1—Vu. 3. The DO concentration of this mixture, DOm, was computed as D0," = V,, D0,, + Vw DOW . 4. The excess D0 concentration, defined as ADO=DOb—D0m, was plotted as a function of the chlorophyll a concentration CPI, (fig. 28). 5. If one assumes that the excess DO concentration is the net in- crease in oxygen concentration due only to phytosynthesis and respiration of the phytoplankton during the production of all the chlorophyll a at the bloom point, then, the slope of the line fit to the data in figure 28 is an average measure of the desired coefficient of proportionality, DOC. The available data from the 30 NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY TABLE 12.—Abundant taxa during two blooms in the Green-Duwamish River [Phytoplankton counted and identified by W. A. Dawson] Chloro- Time phyll a Concentration (Post) Depth concen— and tide loca~ Salinity tration Percentage Probable Presence Date Station stage tion‘ (ppt) (all) Taxon Cells/ml of total2 origin of bloom 1967 Aug. 22 Renton 1530 Surface ________ 0 2.6 Micractim‘um sp ___________ 710 42 Freshwater ________ Yes Junction Midflood 510 30 Monitor. (6,700) (398) East Mar— 0930 ,,,,,, do ______ .2 10.5 Cyclotella sp ________________ ,30 88 ginal Way Late ebb Chlamydamonas sp ________ 410 7 Bridge. ("Coccoids" and clusters) __-. (2,800) (47) 1520 ______ do ,,,,,, .4 11.3 Melasim sp _______________ , 00 56 Midflood Chrysococcus sp . 980 35 Boein 0915 ,,,,,, do ______ ___- 41.3 Cyclotella sp _______________ 31,000 98 Bri ge. Late ebb ("Coccoids" and clusters) ____ (1,400) (4) 1425 ...... do ______ .8 21.9 Cyclotella sp ________________ 3,000 93 Midflood Chlamydomonas sp ___ 1,200 3 ("Coccoids" and clusters) (3,900) (11) 16th 0845 ______ do ,,,,,, 16.7 5.6 Cyclotella" sp ___________ 13, 000 98 Avenue Midebb (Coccoids" and cl sters) ,1__ (3, 700) (3) South Bridge. 1400 ______ do ______ 12.9 37.8 Cyclatella sp ________________ 57,000 98 ________ do. - Yes Early Flagellate, oval 5—7 am 510 1 Saltwater(?) __ . Yes flood ("Coccoids” and clusters) (1,800) (3) Freshwater .. . 1405 Bottom ________ 26.4 3.5 Cyclotella sp ____________ _ 690 71 — Early Navicula st __________ _ 110 11 _ floo ("Coccoids’ and clusters) _ (1.200) (123) . First 0825 Surface ,,,,,,,, 20.5 6.6 Cyclotella sp ____________ _ 13,000 92 Avenue Midebb Skeletonema castatum _ . 1,000 7 South (“Coccoids” and clusters) .... (2,400) (18) Bridge. 1340 ______ do ______ 13.0 15.6 Cyclotella sp ________________ 39,000 99 ,,,,,,,, do -. - Yes Eaglyd ("Coccoids" and clusters) _._. (2,000) (5) ________ do ________ Yes 00 1345 Bottom ........ 27.4 1.2 Cyclotella sp ________________ 550 77 ________ do ________ Yes Eaé‘lyd (“Coccoids" and clusters) .-.. (55) (8) ________ do ________ No oo Spokane 0735 Surface ________ 24.0 2.0 Skeletonema costatum ______ 1,400 54 Saltwater __________ Yes Street Early Cyclotella sp ________________ 1,200 46 Freshwater ........ Yes Bridge. ebb 1305 ______ do ______ 23.3 5.5 Cyclotella sp ________________ 2,000 50 ________ d0 ________ Yes Early Skeletonema costatum . 1,100 26 Saltwater ..__ Yes ebb F1age11ate,oval 5—7 am . 550 14 Saltwater(?) _. Yes (“Coccoids” and clusters) (310) (8) Freshwater -. _ No 0755 Bottom ________ 28.5 1.8 Flagellate, oval 6—8 um 200 39 Saltwater(?) .. _ No Early Chilomonas sp __________ . 97 19 Saltwater _--- . No ebb Skeletonema costatum _ - 89 18 ________ do A No 1310 ______ do ,,,,, 28.4 .9 Flagellate, oval 6—8 um . 61 51 Saltwater(?) ._ . No Early Chilomanas sp ______________ 23 19 Saltwater _________ No flo Aug. 24 Renton 0930 Surface ________ 0 6.7 Scenedesmus ______________ 1,100 91 Freshwater ________ Yes Junction Early ("Coccoids" and clusters) __-_ (990) (85) ________ do ________ Yes Monitor. ebb 1540 ______ do ,,,,,, 0 9.6 Scenedesmus ______________ 3,400 90 Early (“Coccoids and clusters) ___- (12,000) (321) flood East Mar- 0190 ,,,,,, do ______ 1.7 21.9 Flagellate, oval 5—7 um .___ 11,000 53 goal Way Early Cyclotellaup sp ________________ 8,900 44 ridge. ebb (“Coccoids’ and clusters) ___. (5,800) (28) 1525 ______ do ______ .2 8.4 Scenedesmus _____________ , 00 46 Early Cyclutella sp ... . 2,100 38 ebb Chlamydomonas _________ _ 470 8 ("Coccoids” and clusters) .... (4,000) (72) Boein 0830 ______ do ______ 11.0 47.8 Flagellate, oval 5—7 pm ..__ 80,000 97 Saltwater(?) ___ Bri ge. Start Cyclotella sp ............... 2,200 3 Freshwater _-- Yes of ebb ("Coccoids” and clu ers) ___. (5,400) (7) Yes 1440 ______ do ______ .8 13.2 Cyclotella sp ............ 4,600 61 Yes Start Scenedesmus .......... . 2,400 32 do Yes of flood Flagellate, oval 5—7 mm __.. 710 10 Saltwater(?) ..- Yes (“Coccoids" and clusters) . (3,300) (45) Freshwater ___ Yes 16th 0805 ______ dc ...... 14.7 33.7 Flagellate, oval 5—7 pm _ 77,000 99 Saltwater(‘?) ._- Yes Avenue Low high Cyclotella sp ____________ _ 150 <1 Freshwater __- No South Scenedesmus __________ _ 110 <1 ........ do _.- No Bridge. ("Coccoids” and clusters) .... (42) (<1) ........ do .. No 0815 Bottom ........ 27.7 3.9 Flagellate, oval 5-7 pun 1,700 87 Saltwater(?) ___ Yes Low high (“Coccoids” and clusters) . (190) (1) Freshwater __. No 1415 Surface ........ 7.4 58.9 Cyclotella sp .......... 66,000 94 ________ do _ Yes Low low Flagellate, oval 7 um 3,000 4 Saltwater(?) __. Yes (“Coccoids" and clusters) - (360) (1) Freshwater __- No 1425 Bottom ........ 26.7 5.6 Fla ellate, oval 3—5 pm 1,100 61 SaltwaterC’) Yes Low low Cyc atella sp ............. _ 250 14 Freshwater __- No (“Coccoids” and clusters) -.._ (91) (5) ________ o ________ No First 0740 Surface ........ 16.7 10.0 Flagellate, oval 5-7 am ..._ 12,000 99 SaltwaterC’) ________ Yes Avenue Low high (“Coccoids" and clusters) _-.. (140) (<1) Freshwater ________ No South Bridge. 0750 Bottom ________ 28.0 2.6 Flagellate, oval 5-7 urn __.. 840 82 Saltwater(?) ........ Yes Low high ( Coccoids“ and clusters) .... (25) (2) Freshwater . N 1345 Surface ________ 11.5 66.8 Cyclotella sp ........... 110,000 99 ........ Low low Flagellate, oval 5—7p.m _ 1,000 <1 Saltwater('?) _ ( Coccoids" and clusters) _ (510) (<1) Freshwater _ 1355 Bottom ________ 28.1 1.6 Flagellate, oval 5—7 um _ 37 83 Saltwatefi?) . Low low ( 'Coccoids" and clusters) ___- (15) (34) Freshwater .‘ ...... N0 SUPPLEMENTAL INFORMATION 3 1 TABLE 12.fiAbunda,nt taxa during two blooms in the Green-Duwamish River—Continued Time Sign Concentration Pst. De th oonoen~ a(nd tide loga- Salinity tration Percentage Probable Presence Date Station stage tionl (ppt) (II/1) Taxon Cells/m1 of total2 origin of bloom S kane 0645 Surface ________ 21.8 9.5 Fla ellate, oval 5—7 am .__. 25.000 97 Saltwater(?) ________ Yes pSatreet Low high Cyjotella sp ________________ 300 1 Freshwater ,,,,,,,, No Brid e. g 0705 Bottom ________ 28.7 .3 Flagellate, oval 5—7 ;.I.m “,1 36 40 SaltwaterI?) ________ No Low high Chilomonas sp ______________ 23 26 Saltwater __________ No ("Coccoids” and clusters) J (6) (7) Freshwater 7. _ N0 1310 Surface ________ 23.4 9.0 Flagellate, oval 5—7 am .c" 18,000 90 Saltwater(?) Yes Low low Cyclotella sp ________________ 1,800 9 Freshwater Yes (“Coccoids" and clusters) .___ (15,000) (74) ________ do Yes 1320 Bottom ________ 28.3 .9 Flagellate, oval 6—8 am h" 88 41 Saltwater(?) __ _ No Low low Peridinium A“. 83 38 Saltwater .__. ._.- No Chilomonas sp ______________ 31 14 ,,,,,,,, do ________ No 1968 . Aug 5 First 0940 Surface ,,,,,,,, 12,9 13.2 Flagellate, oval 6-10 [1.11] ____ 19,000 95 Saltwater(?) ________ Yes Avenue Low low Cyclotella sp ,,,,,,,,,,,,,,,, 720 4 Freshwater ________ Yes South Bridge. 1830 ______ do ______ 27.3 26.0 Flagellate, oval 6—10 am ..._ 16,000 96 Saltwater(?) ________ Yes Early Cyclotella sp ,,,,,,,,,,,,,,,, 480 3 Freshwater ________ No ebb ‘Surface and bottom indicate about 3 feet (1 m) below water surface and about 3 feet (1 m) above streambed, respectively. 2Total is sum of concentrations for the various taxa counted but does not include concentration of "coccoids” and cocc01 clusters. GREEN-DUWAMISH RIVER KILOMETRE 5 IO I5 20 O Duwamish River estuary yield DOC=0.125 mg oxygen per mi- l I I I I , crogram chlorophyll a. Street ESTIMATED INFLUENCE OF NUTRIENTS ON PHYTOPLANKTON GROWTH Spokane Ioo,ooo — East Marginal Way I The effluent from RTP considerably increases concentrations of ammonium and total and soluble phosphate in the Duwamish River - estuary (Welch, 1969, fig. 8; Tilley and Dawson, 1971, fig. 3) but does \ not greatly increase the concentration of nitrate (Welch, 1969, fig. 8; \ AUGUSI ZZI '957 - Santos and Stoner, 1972, p. 62—63). The nutrients contributed by-the \/ F'°°d tide RTP effluent have the potential of increasing the phytoplankton 4 biomass in the estuary. Welch (1969) reported that the addition of I I.‘ _ RTP effluent to Duwamish River estuary samples increased the |\ August 22’ 1967 population of a green-algae population (Scenedesmus sp.) when the I /Ebb "d, samples were incubated in flasks under uniform light of about 7,000 l ‘ lux and a temperature of 20°~21°C, without mixing for 5 or 6 days. In I \ . ' one series of tests, the maximum assimilated carbon-14 (a measure ‘5‘: - of the biomass produced) in samples containing 5-, 10- and 25- \ z. _ percent RTP effluent was about 14, 30, and 36 percent greater than ' in samples without effluent. Welch (1969) also stated that the ob- served data for chlorophyll a concentration in the estuary showed no August 24’ I967 significant increase in the estuary’s phytoplankton biomass follow- _ Low high and ing a 46-percent increase in effluent discharge between 1965 and '._/ earl, ebb me 1966. Later data suggest that chlorophyll a concentrations (figs. ' 9—11, 18, 19) and cell concentrations (table 12) in the 1967—68 blooms 2. were at about the same level as they were in the 1965-66 blooms ‘. (Welch, 1969, figs. 3, 5). As Welch also noted, it is impossible to make definitive comparisons between maximum biomass of blooms be- cause of the difficulty of sampling the maximum biomass for any one bloom. I0,ooo - I000 - August 5, I968 Low low tide _ 0 August 24' I967_ Computed Michaelis-Menton factors were used in estimating the Z / Low low and effects of nutrients on growth rates of phytoplankton in the early flood Iide‘ Duwamish River estuary. These and similar factors, like FL or F7», can be included in equation 29 for reducing the growth rate because of deficiency of a nutrient or the effects of any other parameter. The PHYTOPLANKTON CONCENTRATION, IN CELLS PER MILLILITRE I I I I I I I I , I High high tide ,-' ‘| A : - I I I I I I I I I I I I l o ' 5 ' ' I0 ' l5 GREEN-DUWAMISH RIVER MILE I00 FIGURE 25.—Longitudina1 distributions of Cyclotella sp. during a bloom in 1967 and data at two points during a bloom in 1968. 32 GREEN - DUWAMISH RIVER KILOMETRE 0 ? IP '.5 2? |00,000 I ' r- g E, .......... .: g d s :5 ' 9 _ 8 __ g a l a s — 'l. m -‘ Lu 3 g fAugust 5, [968 '-, :1 ' ,_ 0; Low low tide _ 2' A: High high tide m '3 August 24, I967 E |0,000 - Low high and ' m _ / early ebb tide ' _. . _l '. l.|.l ' ' - o E ~ August 24, l967 '-_ Low low and early flood tide '. - PHYTOPLANKTON CONCENTRATION, I000 - .. - ' August 22, I967 ‘ _ _ /Flood tide . 1 _____ K '. \ u \ . .. \ -. .4 l l l '- °°o ' 5 ' ' IO ‘ :5 GREEN-DUWAMISH RIVER MILE FIGURE 26.—longitudinal distributions of oval flagella‘oes during a bloom in 1967 and data at two points during a bloom in 1968. Michaelis-Menton factor, FN, for a particular nutrient is expressed in the form C F =———, (50) N C+CV2 where C is the concentration of the particular nutrient and Ci)z is C when FN=%. Existing literature does not provide values of 01/2 for all phyto- plankton taxa found in the Duwamish River estuary. However, likely ranges of C V2 for the types and sizes of phytoplankton in the estuary may be 0.01 to 0.1, 0.002 to 0.02, and 0.001 to 0.05 mg/l for nitrate nitrogen, ammonium nitrogen, and phosphate phosphorus, respectively. Especially helpful in providing these data were the reports by Eppley, Rogers, and McCarthy (1969), MacIsaac and Dug- dale (1969), and the Lake Tahoe Area Council (1968—70). The mid- points of the likely ranges of C 1/2 for nitrate and ammonium are 0.055 and 0.011 mg/l, respectively; Maclssac and Dugdale (1969) reported respective values of 0.062 and 0.023 mg/l for natural marine com- munities growing in eutrophic conditions. NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY GREEN-DUWAMISH RIVER KILOMETRE 0 5 IO I5 20 l | | l I I El E El Lu 4‘. :7; 9 u: ‘. u I: (1% = 5 e gIOpOO E 3 '1 :J ' Lu - E ’, 5 h 2 A 22 ' ”/7 _ | .3-” n. , ”as; "a; 967 ’ \ August 22, l967 ‘3 ' ‘: \ Flood tide d //\\ o . , I ; I E 2 ’ . ’- Ir‘7‘r’ _ z . 9 =. ,’ '- '~I < E IOOO - ’f August 24 I967 - z / '. :' ' 3 - I ' ' \ Low low and _ I . g / early flood tide O " I _ I 5 I I; _ I _ I E I ‘1’ August 24, I967 ,9 Low high and E ' early ebb tide - a. loo 1 I \ l I I I I 0 I5 5 IO GREEN-DUWAMISH RIVER MILE FIGURE 27.—Longitudinal distributions of “coccoids” plus “coccoid” clusters during a bloom in 1967. Figure 29 shows the Michaelis-Menton factors for nitrate + nitrite and ammonium nitrogen and for phosphate phosphorus that were computed using equation 50 and midpoint values of the likely ranges of 01/2. Nutrient-concentration data were available for 3-ft (l-m) depth samples collected semimonthly and monthly during most of the probable bloom periods, July-early September of 1967 and 1970—71 at East Marginal Way, 16th Avenue South, and Spokane Street Bridges. Analyses of these data indicate that (1) concentra- tions of ammonium nitrogen, NH4—N; nitrate + nitrite nitrogen (N03+N02)—N; and total phosphate, PO4-P, equaled or exceeded 0.25 mg/l in 79, 92, and 93 percent of the samples, respectively; and (2) variations in nutrient concentrations were unrelated to chlorophyll (1 concentrations. For a concentration of 0.2 mg/l, the curves in figure 29 give factors ranging from 0.79 to 0.95 times maximum growth. Two inferences important to the modeling and the management of the Duwamish River estuary may be implied from the foregoing analysis: 1. Given the present phytoplankton inflow to the Duwamish River estuary, present concentrations of nutrients in the estuary usu- ally are sufficiently high to be considered nonlimiting to the growth of the phytoplankton; 2. Therefore, if the phytoplankton inflow to the estuary does not increase, future increases in nutrient concentrations in the es- tuary from additional RTP effluent or from other sources proba— bly will not greatly increase the phytoplankton biomass in the estuary. SUPPLEMENTAL INFORMATION I4 I I l I (J Lu E / :1 I2 - /_ a; lath Avenue South / if OK Bridge // (n / 5 / 5 I0 - / _ 3’ / E // E P z" 8 - / _ ° / “ B 5 °/ 6‘ E / III / ‘z’ / 8 6 ‘ / ‘ z ' Spokane Street g O.|25 / 5"" Bridge 3 a 4 _ / _ u.I e / 3 ‘3' G / ° 0 '0 / ‘2 o M )6 :n / E 9/ East Marginal Way 6 A/ Bridge 0 0 /. o o A a s L . I O 20 40 60 80 IOO CHLOROPHYLL 0 CONCENTRATION, IN MICROGRAMS PER LITRE FIGURE 28.-—Relation between concentrations of excess dissolved oxygen and chlorophyll a in some 3 ft (l—m) deep samples from three stations during 1967—68. Excess dissolved-oxygen concentra- tion is the difference between sample concentration and a com- puted concentration for wedge and river water mixed in proportion to sample salinity. The study of the effects of nutrients on phytoplankton growth rates in the Duwamish River estuary is continuing. HERBIVORES Herbivores are organisms that obtain their nourishment by con- suming plants. Both freshwater and marine species were found in water samples taken from the Duwamish River estuary during periods of phytoplankton blooms in 1967 and 1968. Concentrations ranged from 0 to 110 herbivores/ml (table 13). These data suggest that concentration of herbivores and number of taxa tend to be higher in the water sampled from the two farthest downstream sta- tions. Out of the seven freshwater taxa identified, rotifers, Didinium nasutum, Vorticella sp., and other ciliates were most often present in the samples. Twelve marine taxa were identified, of which holotrich spp., Laboea conica, and other heterotrich spp. occurred most often in the samples. The freshwater taxa were predominant in samples from the upstream stations, whereas marine taxa were predominant in samples from the downstream stations. 33 [-0 l I I 'i NH3 -N. Cg = 0.0” mg/I 0.8 - - P04 - P, 6% = 0.026 mg/I a: o (N034 N02l- N, Cl = 0.055 rug/l 5 z E z 0.6 _ o I- z m 2 I ‘9 d 0.4 - < I 9 E 0.2 - 0 I I I I 0 0 0.4 0.6 08 Lo NUTRIENT‘ CONCENTRATION, IN MICROGRAMS Pea LITRE FIGURE 29.—Relations between Michaelis-Menton factors and con- centrations of nutrients. TABLE 13.—Concentrati0ns of herbivores in water samples from the Green-Duwamish River for 3 days during phytoplankton bLooms in 1967 and 1968 Sampling Time depth des- Concentration Date (P.s.t.) ignation1 Taxa present2 (herbivores/ml) Renton Junction Monitor 1967 Aug. 22 1530 1.1 24 1930 .2 24 1540 6.8 Aug. 22 0930 .4 22 1520 .1 24 0910 .6 24 1525 .6 Boeing Bridge Aug. 22 0915 Surface __________ F1, 512 .4 22 1425 .3 24 0830 3.2 24 1440 0 16th Avenue South Bridge Aug. 22 0645 Surface __________ F2, 811, 812 2.8 22 1400 ,,,,,, do," M. F1, F6, 812 .6 22 1405 Bottom _____ , F1, F6, F7, 812 3.4 24 0805 Surface _____ S4, 812 1.8 24 0815 Bottom F6, SS, 812 4.0 24 1415 Surface F1, 512 .8 24 1425 Bottom ,,,,, ._ F5, SB, 812 7.3 34 nu. S. GOVERNMENT PRINTING OFFICE: NUMERICAL MODEL OF THE SALT-WEDGE REACH OF THE DUWAMISH RIVER ESTUARY TABLE 13.—Concentrations of herbivores in water samples from the Green-Duwamish River for 3 days during phytoplankton blooms in 1967 and 1968—Continued Sampling Time depth dee- Concentration Date (P.s.t.) ignationI Taxa present2 (herbivores/ml) First'Avenue South Bridge Aug. 22 0825 Surface VVVVVVVVVV F1, S4, 811, $12 5.4 22 1340 ______ do.. ___ F1, F4, 511, S12 2.8 22 1345 Bottom... . _ S7, S12 1.7 24 0740 Surface ... S4 13 24 0750 Bottom-... ... S4, SS, S12 2.1 24 1345 Surface _._ F1, F6, 812 4.3 24 1355 Bottom ____________ 811, S12 .7 Spokane Street Bridge Aug. 22 0735 Surface .......... F2, F7, 31, S6, 511, S12 3.6 22 0755 Bottom-... . F6, SS, 8, $12 16 22 1305 Surface _. - SS, S4, 812 8.0 22 1310 Bottom___. . F3 F6, 81, SS, S6 3.1 24 0645 Surface . 52, S4, S7, 812 110 24 0705 Bottom“ _ F3, 52, S4, S7, 812 2.1 24 1310 Surface . F4, 812 12 24 1320 Bottom._ . $9 6.8 First Avenue South Bridge 1968 Aug. 5 0940 Surface .......... F7, SE, 812 8.9 5 1830 ...... do __________ F4, F6, SS, 312 37 1Surface and bottom indicate samples from about 3 fi (1 m) below the water surface and about 3 fl (1 m) above the streambed, res ctively. 2Herbivm‘es counted and identified by W. A. Dawson are referred to by F (freshwater taxa) or S (saltwater taxa) and a number to indicate a specific taxazFl-Didinium nasutum, F2-Nemata, FS-Rhizopoda and Actinopoda, F4-Vorticella sp., F5-Fllamentous bacteria (decomposers), FG-Other ciliates, and F7-Rotifers; Sl-Appendicularian ("Dikopleura"), 52- Gyrodinium spirale, SB-Laboea strobila, S4-Laboea conica, S5-Laboea sp., SG-Nauplius larvae, S7-Noctiluca scintillans, SS-Parundella sp., 59~Strombidium ep., SlO-Copepodid larvae, Sll-Holotrich app, and SlZ-Other heterotrich app. 1976 - 690-036/94 QE7€ Po, 7 DAYS 1144/ Thermal Loading of “Natural Streams §r \ ENCES ‘ éARV GEOLOGICAL SURVEY PROFESSIONAL PAPER 991 us. DEPOSITORY APR 12 1977 Thermal Loading of Natural Streams By ALAN P. JACKMAN and NOBUHIRO YOTSUKURA GEOLOGICAL SURVEY PROFESSIONAL PAPER 991 UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1977 UNITED STATES DEPARTMENT OF THE INTERIOR CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Dirermr Library of Congress Cataloging in Publication Data Jackman, Alan P. Thermal loading of natural streams. (Geological Survey Professional Paper 991) Bibliography: p. 38-39. Supt. of Docs. no.: I 19.16z991 1. Thermal pollution of rivers, lakes, etc. 1. Yotsukura, Nobuhiro,joint author. 11. Title. III. Series: United States Geological Survey Professional Paper 991. TD427.H4J3 363.6'1 76—608317 For sale by the Superintendent of Documents, US. Government Printing Oflice Washington, DC. 20402 Stock Number 024-001-02959-6 CONTENTS Page Page Abstract ____________________________________________________ 1 One-dimensional model of excess temperature—Continued Introduction ________________________________________________ 1 Dan River near Eden, North Carolina, 1969 ______________ 20 Acknowledgments __________________________________________ 2 Tittabawassee River near Midland, Michigan, 1969 ______ 22 Equations for conservation of thermal energy _____-_, _________ 2 North Platte River near Glenrock, Wyoming, 1970 ________ 23 Heat transfer at the air-water interface ______________________ 4 West branch of the Susquehanna River near Shawville, Analysis of natural temperature ____________________________ 8 Pennsylvania, 1962 __________________________________ 26 Thermally homogeneous streams ________________________ 9 Conclusions regarding one-dimensional modeling __________ 27 Thermal homogeneity of the Potomac River ______________ 10 Two-dimensional model of excess temperature ,,,,,,,,,,,,,,,, 27 Prediction of natural temperature ________________________ 13 Model for a steady uniform channel ______________________ 27 One-dimensional model of excess temperature ________________ 16 Model for a steady natural stream ,,,,,,,,,,,,,,,,,,,,,, 28 Preliminary considerations ______________________________ 16 Applications to field data ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 29 Description of computational procedures __________________ 17 Summary regarding two-dimensional model ______________ 36 Results and discussion __________________________________ 18 Summary and conclusions __________________________________ 37 White River near Centerton, Indiana, 1969 ______________ 18 References cited ____________________________________________ 38 Fenholloway River near Foley, Florida, 1969 ______________ 20 ILLUSTRATIONS Page FIGURE 1. Temporal variations of water temperature at five positions in a hypothetical channel with doubled depth (a=2) ,,,,,,,, 10 2. Temporal variations of water temperature at five positions in a hypothetical channel with halved depth (a=‘/2) ________ 10 3. Sketch of the Potomac River study reach below Dickerson Power Plant, Maryland ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 11 4. Transverse profiles of water temperature at selected sections, the Potomac River below Dickerson Power Plant, Maryland, March 11, 1969 ______________________________________________________________________________________________ 12 5. Longitudinal profile of natural water temperature near right bank, the Potomac River below Dickerson Power Plant, Maryland, March 11, 1969 ____________________________________________________________________________________ 12 6. Longitudinal profile of natural water temperature near right bank, the Potomac River below Dickerson Power Plant, Maryland, May 14, 1969 ______________________________________________________________________________________ 12 7. Transverse profile of natural water temperature at section 1, the Potomac River above Dickerson Power Plant, Maryland, March 11—12, 1969 ___________________________________________________________________________________________ 13 8 Comparison of observed and calculated natural water temperatures at section 4, the Potomac River below Dickerson Power Plant, Maryland, March 1969 __________________________________________________________________________ 13 9. Comparison of observed and calculated natural water temperatures at section 4, the Potomac River below Dickerson Power Plant, Maryland, May 1969 ____________________________________________________________________________ 13 10. Comparison of observed and calculated natural water temperatures, the Potomac River below Dickerson Power Plant, Maryland, 1969, and the Riverside Inlet Canal near Greeley, Colorado, 1970 ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 15 11. Attenuation of remaining excess heat with distance downstream from heat discharge site ,,,,,,,,,,,,,,,,,,,,,,,,,,,, 17 12. Sketch of study reaches, I ________________________________________________________________________________________ 19 13. Sketch of study reaches, II ________________________________________________________________________________________ 19 14. Comparison of observed and calculated water temperatures, the White River near Centerton, Indiana, 1969 ,,,,,,,,,,,, 20 15. Comparison of observed and calculated water temperatures, the White River near Centerton, Indiana, 1969 ,,,,,,,,,,,, 20 16. Comparison of observed and calculated water temperatures, the White River near Centerton, Indiana, 1969 ,,,,,,,,,,,, 20 17. Temporal variation of natural and heated water temperatures, the White River near Centerton, Indiana, 1969 __________ 21 18. Comparison of observed and calculated water temperatures, the Fenholloway River near Foley, Florida, 1969 ,,,,,,,,,, 22 19. Comparison of observed and calculated water temperatures, the Fenholloway River near Foley, Florida, 1969 __________ 23 20. Comparison of observed and calculated water temperatures, the Fenholloway River near Foley, Florida, 1969 ,,,,,,,,,, 24 21. Comparison of observed and calculated water temperatures, the Fenholloway River near Foley, Florida, 1969 ,,,,,,,,,,, 24 22. Comparison of observed and calculated water temperatures, the Fenholloway River near Foley, Florida, 1969 ,,,,,,,,,, 24 23. Comparison of observed and calculated water temperatures, the Fenholloway River near Foley, Florida, 1969 ,,,,,,,,,, 25 24. Comparison of observed and calculated water temperatures, the Dan River near Eden, North Carolina, 1969 ____________ 25 25. Comparison of observed and calculated water temperatures, the Dan River near Eden, North Carolina, 1969 ____________ 25 26. Comparison of observed and calculated water temperatures, the Dan River near Eden, North Carolina, 1969 ,,,,,,,,,,,, 25 27. Comparison of observed and calculated water temperatures, the Tittabawassee River near Midland, Michigan, 1969 “fl 26 28. Comparison of observed and calculated water temperatures, the Tittabawassee River near Midland, Michigan, 1969 ____ 26 Ill IV CONTENTS Page FIGURE 29. Comparison of observed and calculated water temperatures, the North Platte River near Glenrock, Wyoming, 1970 ____ 26 30. Comparison of observed and calculated water temperatures, West Branch of the Susquehanna River near Shawville, Pennsylvania, 1962 __________________________________________________________________________________________ 27 31. Sketch of a two-dimensional natural stream and its stream-tube flow system ________________________________________ 29 32. Concurrent transverse distributions of dye concentration and excess temperature, the Potomac River below Dickerson Power Plant, Maryland, March 1969 __________________________________________________________________________ 30 33. Temporal variations of surface heat dissipation coefficient and source excess temperature, the Potomac River below Dick- erson Power Plant, Maryland, March 1969 ____________________________________________________________________ 30 34. Comparison of observed and calculated transverse temperature distributions, the Potomac River below Dickerson Power Plant, Maryland, March 1969 ________________________________________________________________________________ 31 35. Temporal variations of surface heat dissipation coefficient and source excess temperature, the Potomac River below Dick- erson Power Plant, Maryland, May 1969 ______________________________________________________________________ 32 36. Comparison of observed and calculated transverse temperature distributions, the Potomac River below Dickerson Power Plant, Maryland, May 1969 __________________________________________________________________________________ 33 37. Comparison of observed and calculated transverse temperature distributions, the Potomac River below Dickerson Power Plant, Maryland, May 1969 __________________________________________________________________________________ 34 38. Comparison of observed and calculated transverse temperature distributions, the Potomac River below Dickerson Power Plant, Maryland, October 1969 ________________________________________________________________________________ 34 39. Comparison of observed and calculated transverse temperature distributions, the Potomac River below Dickerson Power Plant, Maryland, October 1969 ________________________________________________________________________________ 35 40. Comparison of observed and calculated transverse temperature distributions, the Potomac River below Dickerson Power Plant, Maryland, October 1969 ________________________________________________________________________________ 35 41. Comparison of observed and calculated transverse temperature distributions, the Dan River near Eden, North Carolina, April 1969 __________________________________________________________________________________________________ 36 42. Comparison of observed and calculated transverse temperature distributions, the North Platte River near Glenrock, Wyoming, January 1970 ______________________________________________________________________________________ 36 TABLES Page TABLE 1. Temporal variation of hypothetical natural temperature under freezing conditions, the North Platte River near Glenrock, 19 Wyoming, January 28—29, 1970 ______________________________________________________________________________ 2. Cross-sectional average hydraulic parameters, the Potomac River below Dickerson Power Plant, Maryland, March 1969 __ 30 THERMAL LOADING OF NATURAL STREAMS By ALAN P. JACKMAN and NOBUHIRO YOTSUKL'RA ABSTRACT The impact of thermal loading on the temperature regime of natural streams is investigated by mathematical models, which de- scribe both transport (convection-diffusion) and decay (surface dissi- pation) of waste heat over l-hour or shorter time intervals. The models are derived from the principle of conservation of thermal energy for application to one- and two-dimensional spaces. The basic concept in these models is to separate water temperature into two parts, ( 1) excess temperature due to thermal loading and (2) natural (ambient) temperature. This separation allows excess tem- perature to be calculated from the models without incoming radia- tion data. Natural temperature may either be measured in pro- totypes or calculated from the model. If use is made of the model, however, incoming radiation is required as input data. In order to formulate a linear decay model for excess temperature, the equations for back radiation, evaporation, and conduction, de- rived from the Lake Hefner study of US. Geological Survey, are linearized with reference to an arbitrary base temperature. It is shown that the resulting surface dissipation coefficient is predomi- nantly influenced by wind speed and the mass-transfer coefficient of evaporation. For one-dimensional problems, the transport of excess tempera- ture is solved by the Lagrangian convection model of traveling with an average heated water packet. In two-dimensional problems, a steady-state diffusion model is combined'with the one-dimensional decay model. Longitudinal dispersion is neglected in both transport models. Comparison of observed and calculated temperatures in seven natural streams shows that the models are capable of predicting transient temperature regimes satisfactorily in most cases. Mass- transfer coefficients of evaporation in streams are much higher than those commonly used for long-term calcualtions in a lake. Factors such as ground-water accretion, abrupt changes in thermal loading, and unstable atmospheric conditions are found to have significant impacts on water temperature regimes. The dissipation of excess heat in natural streams is a gradual process frequently extending over a long downstream distance. For five out of the seven study reaches, the remaining excess heat at a point nearly 30 kilometres downstream was more than 50 percent of the initial thermal load. INTRODUCTION The impact of thermal loading (waste heat dis- charge) on temperature regime of a natural water body is the basic physical information that is required for meaningful assessment of water-quality changes in- duced by such waste discharges. The Geological Survey has been engaged for more than 20 years in the study of thermal loading as an outgrowth of the study of evaporational water losses in lakes. The results of the earlier studies have been published in papers by Harbeck (1953), Harbeck, Koberg, and Hughes (1959), Messinger (1963), Har- beck, Meyers, and Hughes (1966), and Harbeck (1970). These studies used the energy-budget equation derived from the Lake Hefner study (Anderson, 1954) and in- vestigated daily- or weekly-average (long-term) ther- mal regimes of a few western lakes and a couple of natural streams. Toward the end of 1968, the extension of thermal loading studies to natural streams was deemed urgent in View of the mounting concern about thermal pollu- tion caused by power generation (Parker and Krenkel, 1969). At the same time, keen interest was shown in the collection of synoptic data from thermally loaded streams covering a wide range of meterologic and hy- draulic conditions. Accordingly, the Geological Survey, in cooperation with the Atomic Energy Commission (now the Energy Research and Development Administration), under- took a 1-year program of data collection in six natural streams that were thermally loaded and located in dif- ferent regions of the country. Briefly, a stream survey consisted of measuring water temperature, meteoro- logic variables, discharges, and channel geometries at several cross sections downstream from the heat source for a duration of 24 hours. In many streams, two to three surveys were conducted during the year to evaluate seasonal thermal patterns. All survey works were completed between February 1969 and January 1970. In addition, the thermal data of the Susquehana River collected by the Geological Survey in 1962 were considered worthy of reexamination and included in the data set of the present report. Examination of these stream data revealed that, even though the energy-budget equation adequately describes the heat exchange at the air-water interface, it is not suitable for describing the heat transport. In order to predict natural stream phenomena in terms of hourly (short-term) variations, the equation for con— servation of thermal energy must be derived and solved with due considerations given to the transport aspect as shown by recent studies such as Edinger and Geyer (1965), Jaske (1969), Morse (1970), and Harle- man (1972) among others. In solving the thermal energy equation, further- 2 THERMAL LOADING OF NATURAL STREAMS more, some of the customary models and assumptions had to be reexamined. For example, the steady-state exponential decay law predicts that the excess temper- ature decreases exponentially with increasing down- stream distances. In view of highly variable thermal loading patterns and meteorologic conditions in proto- types, such a simple model could not accommodate short-time or equivalently short-distance phenomena (Jackman and Meyer, 1971). Another example is the equilibrium temperature that is used to calculate the surface heat dissipation coefficient. It is an approxima- tion acceptable only for daily or weekly averages but not for hourly averages (J obson, 1973a; Yotsukura and others, 1973). The new approach consists of combining the convec- tive diffusion equation of heat transport with a linear- ized equation for dissipation of excess heat at the air- water interface (Jackman and Meyer, 1971; Jobson, 1973a; Jobson and Yotsukura, 1973). The approach to surface heat dissipation is Lagrangian in the sense of following average water particles downstream. The dissipation coefficient is calculated with reference to the daily average ambient temperature. upstream from the heat source. In addition, a two-dimensional diffu- sion approach was incorporated into the analysis to facilitate the definition of the mixing zone (Yotsukura, 1972). The present report describes the final analysis of the thermal data by the analytical models evolved in the last few years and discussed in the above-mentioned papers. The original thermal data and detailed de— scription of field conditions will be presented by a com- panion paper, which is herein referred to as the Data Report. The following tables describe the measurement units used herein. ACKNOWLEDGMENTS The present model study benefited greatly from con- tributions from a number of specialists in the US. Geological Survey. The authors acknowledge, in par- ticular, E. J. Pluhowski, E. F. Hollyday, F. A. Kilpat- rick, J. F. Bailey, and R. L. Cory for their valuable con- tributions throughout the study. The authors also would like to acknowledge cooperations generously given by W. F. Weeks and C. M. Keeler of the Cold Regions Research and Engineering Laboratory, Han- over, N.H., in conducting the thermal study for the North Platte River. Coordination of the data collection program of this scale was certainly not a simple task. Credit is due E. L. Meyer, whose skill and devotion applied to the task was a major factor in the successful completion of the field work. EQUATIONS FOR CONSERVATION OF THERMAL ENERGY Water temperature expresses the thermal energy of Conversion of units between SI and conventional systems for basic quantities [This table is based on (1) National Bureau of Standards, 1972, The International System of Units (SI): NBS Special Publication 330. (2) National Physical Laboratory, 1966, Changing to the metric system: H.M.S.O., London. (3) Baumeister, T., 1958, Mechanical Engine ‘rs’s Handbook: McGraw-Hill Book C0,, Inc, New York] Conversion _____________ 1 m =3.281 ft 1 ft =0.3048 m _____________ 1 km =0.6214 mi 1 mi =1.609 km _____________ 1 m2 =10.76 fizz . 1 R2 =0.09290 In2 _____________ 1 km2= 0.3861 mi2 1 mi2=2.590 km2 _____________ 1 m3 =35.31 fl." 1 ft3 =0.02832 m3 _____________ 1 kg =2.205 lb 1 1b =0.4536 kg _____________ 1 N =0 1020 kg 1 kg =9.807 N _____________ 1 N =0 2248 lb 1 lb =4.448 N _____________ 1 J =0 2388 cal 1 cal =4.187 J 1 W =0.2388 cal-s‘1 1 cal-s“=4.187 W Conversion of units between SI and conventional systems for derived quantities [Quantities represent those commonly used in thermal studies and conversions are given at the magnitudes observed under normal climatic conditions] Quantity SI units Conventional units Length Metre ________________________ Foot _____________ Kilometre ____________________ Mile _____________ Area Square metre ................ Square foot _____ Square kilometre ______________ Square mile _____ Volume Cubic metre __________________ Cubic foot _______ Mass Kilogram ____________________ Pound ___________ Force Newton ______________________ Kilogram-force __ Pound-force _____ Energy Joule ________________________ Calorie _________ Power Watt ________________________ Calorie per second ____________ Quantity SI units Water density ______________ Wind speed __________________ Atmospheric pressure ________ Specific heat of water ________ Kilogram per cubic metre _____ Metre per sec _________________ Newton per square metre Joule per kilogram per degree kelvin. Joule per kilogram ___________ Watt per square metre ________ Latent heat of vaporization __ Radiation and heat flux ______ Kilogram per metre per newton. Mass transfer coefficient of evaporation. Stefan Boltzman constant of Watt per square metre per black body radiation. (degree kelvin).4 Conventional units Pound per cubic foot Mile per hour ________________ Millibar ____________________ Calorie per gram per degree Celsius. Calorie per gram ____________ Calorie per square centimetre per hour. Gramme per square centimetre per hr per mile-per-hr per millibar. Calorie per square centimetre per min per (degree Kelvin).4 Conversion 1,000 kg-m‘3 =62.43 lb-ft" 1 m‘sec-1=2.237 mile-hr-l 105 N-m"=103 mb 4,187 J-kg'l -K’1 =1ca1'g‘1'C'1 2,495 '103 J'kg'1=596 cal'g'l 1,000 W~m'2=85.97 cal-cm‘2 'hr'1 2-10“3 kg-m‘l-N'1 =3.218‘10"g'cm'2 -mile"-mb'1 5.67'10“ W‘m'2'K‘4 4 =8.14~10'11 cal'cm'2 - min"-K EQUATIONS FOR CONSERVATION OF THERMAL ENERGY 3 a water body. In natural conditions the thermal energy may be considered equivalent to the internal energy. In order to derive an equation for thermal energy bal- ance, one starts with the First Law of Thermodynam- ics, which says that the increase in total energy con- tent of a body is the sum of all heat inputs to the body minus the amount of work that is performed by the body. When one subtracts from the equation of total energy balance those parts that are related to mechan- ical energies, namely, kinetic and potential energies, the differential equation for conservation of thermal energy is obtained as (Pai, 1956) DT pCPDT =KV2 T++S, (1) where p is water density, CI, is the specific heat of water at constant pressure, T is temperature, t is time, K is the thermal conductivity, (1) is the rate of heat produc- tion by viscous dissipation of mechanical energy, and S is the net heat input from sources and sinks. In equation 1, the operator D/Dt is the substantial derivative, or D 6 6 6 6 D—t—6~t+vx6;+vy$+U26—z , (2) where vx, 12,, and v, are velocities in the respective coor- dinate directions. The symbol V2 designates the Lapla- cian operator or a sum of the three second-order partial derivatives with respect to x, y, and 2. Equation 1 as- sumes that water is incompressible even though its density may vary in space and time. Thus, any change in energy content due to volume expansion or compres- sion is neglected. The equation also represents instan- taneous balance of thermal energy within an infinites- imal body of water. Since most applications of the thermal equation in natural water deal with turbulent flows, equation 1 needs to be averaged over a time scale appropriate for a turbulent flow. The procedure, which is based on the Reynolds’ classic averaging method (Hinze, 1959), yields the following equation: DT6 —=—(k Dt 6x (3) 6T 6 ‘67” —yg(ky$+—(k 62 pCp Here T represents time-averaged temperature, S rep- resents time-averaged heat sources and sinks, and the symbols kt, ky, and k, are turbulent thermal diffusion coefficients in the respective directions. In comparing equation 3 with equation 1, note that the molecular conduction term, KV2T, and the viscous dissipation term, (I), are neglected after averaging, because these are quite small relative to other averaged terms. For example, for every l-metre drop in the surface eleva- tion of a uniform stream, this much potential energy is converted through viscous dissipation to thermal energy. This is equal to a heat gain of 9.807 joules per kg of Water and corresponds to a temperature rise of only 0.0023°C. The turbulent diffusion terms, on the other hand, appear in equation 3 as lump-sum expres- sions for the correlation that exists between the devia- tions of instantaneous velocities and temperature from time-averaged values. The symbols vx, vy, and v, in the substantial derivative, D/Dt, are now all time-aver- aged velocities. Owing to the threedimensional nature of equation 3, its solution is often difficult. In order to reduce its dimension, equation 3 may be integrated over the local depth as shown by Leendertse (1970), Prych (1970), and Holley (1971). Assume that the y coordinate is di- rected vertically upward. By means of Leibnitz’s rule of integration and kinematic boundary conditions appro- priate for a natural water body, equation 3 is inte- grated to the following equation: 6YT 6vaT 6szT_ 6t + 6x + 62 _ 6T )+Y_S+ H HW 262 +pCp pC—p 6 6T 6 5;(YK EH62; (YK— +— pCp (4) The symbol Y designates local depth, T is depth- averaged temperature, 12,, and v, are depth-averaged velocities, and K, and K2 are turbulent thermal- dispersion coefficients. The dispersion term in equation 4 consists of two parts. One is the contribution from the correlation between the local deviations of velocities and temperature from depth-averaged values. This correlation is expressed by a gradient-type dispersion term following Taylor’s (1954) model in a steady uni- form flow. An additive part is contributed by depth averaging of the turbulent diffusion term of equation 3. The source-sink term is now defined by YS where S is the depth-averaged value. The symbol H defines bHoundary heat influx normal to the water surface and Hbot defines one normal to the channel bottom. These terms are introduced into equation 4 by the boundary condition which assumes that incoming heat flux is transported inward by turbulent diffusion without any storage at the boundary. . For a natural stream, where the width is small rela- tive to longitudinal distances, it is often convenient to have one-dimensional thermal equations. Equation 4 may now be integrated over the channel width, using the same techniques as used in deriving equation 4. The result is 6_AT +6AVT_ 6 6—t 6x 6x AS +WH +WHbo, p0,. +1on+ In equation 5, A is the cross-sectional area, pT is the ”(KA (5) 4 THERMAL LOADING OF NATURAL STREAMS cross-sectional average temperature, V is the cross- sectional average longitudinal velocity, K is the cross— sectional average thermal dispersion coefficient, and S is the cross-sectional average source sink. The symbols H and H be, are width-averaged surface flux and bottom flux, respectively, while W designates the surface width at the cross section. Even though the details of derivations of the above equations are omitted in the present report, examina- tion of various references shows that these equations are general and most suitable for use in natural water bodies where flows may be transient and boundaries complex. In deriving these equations, no assumptions are made as to the steadiness or uniformity of flow. On the other hand, various diffusion and dispersion terms are added as the averaging progresses from one stage to the next. These terms are introduced as an analogy to gradient—type molecular diffusion. Although some of these diffusive terms have been verified as workable approximations by experiments, notably in a steady uniform flow, one must remember that the expressions for dispersion and diffusion are definitions rather than established physical laws. Fortunately diffusive trans- port is normally small relative to average—velocity (convective) transport and may even be neglected under certain conditions. HEAT TRANSFER AT THE AIR-WATER INTERFACE The mechanics of heat transfer at the flow bound- aries of a natural water course is important in thermal loading problems, because the excess thermal energy possessed by water is ultimately dissipated to the air and the soil through such transfers. It is known that heat transfer at the water surface is one of the major factors in determining water temperature under both natural and thermally loaded conditions. In compari- son, heat transfer through the streambed is normally small, except possibly in extremely shallow streams flowing over solid beds of rock (Brown, 1972). Moreover, there is very little information currently available on this transfer process. For these reasons, the soil-water interface transfer will be neglected hereafter. Heat transfer at the air-water interface may be de- scribed most conveniently with reference to the one- dimensional thermal equation, because the same surface-transfer equations are applied to the more complex two- and three-dimensional equations. The left side of equation 5 may be simplified by substitut— ing the continuity equation of water. Also the source term, S, and the longitudinal dispersion term, KABT/ 6x, are neglected. Equation 5 is then reduced to fl LT: H at 8x pCpY' (6) The depth, Y, is defined by A/W according to equa- tion 5. Four mechanisms have been identified as contribut- ing to the net heat flux, H. They are shortwave radia- tion, longwave radiation, evaporation, and conduction. The largest of these terms is the shortwave radiation, which may assume values as large as 1,200 watts -m'2. It always contributes a positive component to H since essentially no shortwave radiation is emitted by the water. Incoming shortwave and longwave radiation differ significantly from other components in that they do not depend in any way on water temperature. Moreover, the shortwave radiation has a well-defined dependence on time of day through the sun angle, whereas all other terms, while varying in a somewhat random fashion with time in response to variables such as wind speed, have no deterministic dependence on time. Outgoing longwave radiation, evaporation, and con- duction all depend on the water temperature. Because of the complex nature of some of the dependencies, these terms cause equation 6 to be nonlinear. They do not depend directly on the time, although evaporation and conduction depend on such factors as wind veloc- ity, atmospheric stability, and humidity, which vary with time. The maximum values of these components are considerably smaller than that for incoming shortwave radiation. The fact that incoming shortwave and longwave radiation do not depend on water temperature enables one to remove these components from the thermal equation through a transformation of the temperature variable. Suppose there was no waste heat discharged into the stream. The temperature, T,,, which would exist under natural conditions is described by the same equation as equation 6 but with H depending on t and T,,, or 6T" 6T" _ H (Tmt) 7 W W “ To; - (7) It is assumed that the coefficients of equations 6 and 7, such as V and Y, are not affected by thermal condi- tions. If the excess temperature, Te, is defined as the difference between T and T”, then, subtracting equa- tion 7 from equation 6, 6T, +Vfl= H(T,t)-H(T,,,t) ' at 6x pCpY (8) The net flux, H (T,t), may be divided into four compo- nents, HEAT TRANSFER AT THE AIR-WATER INTERFACE 5 H(T,t)=Hr(t)—Hb (T) —He (T,t)—HC(T,t), (9) where H, is the net incoming radiation, both longwave and shortwave, which actually crosses the air-water interface unreflected, Hb is the back radiation (longwave radiation emitted by the water), H, is the heat flux due to evaporation of water, and He is the heat flux due to conduction from the water into the atmosphere. Note that only H, is independent of T; therefore, if equation 9 is substituted into equation 8, the H,(t) from the two terms will cancel, and equation 8 is written as 6T9 6Te _6t_ +V 6x _ —Hb(T)+Hb(Tn)_He(T;t)+He(Tn’t)—HC(T7t)+Hc(Tmt)' pCpY (10) This is a considerable simplification, as equation 10 can be solved without knowing the total incoming radiation. Total incoming radiation, in addition to being a large term of considerable variability, is a quantity that is not now being continuously measured and reported by any agency. For each application it must, therefore, be measured, and this requires ex- pensive equipment that is difficult to maintain and calibrate. In addition to simplifying the modeling problem, ex- cess temperature is a variable of considerable impor- tance in studying the impact of man on the stream environment. In the case of a discharge of waste heat, the heat added to the water will increase the internal energy of the water, thus increasing the water temper- ature. Thermodynamics dictates that the increase in temperature resulting from a waste discharge is a nearly linear function of the amount of heat added to a unit mass of the water. This increase in temperature is the excess temperature, Te. Thus, a waste heat dis- charge may be thought of as a source of excess tem- perature. Just as the excess temperature at the point of dis- charge is a measure of the amount of heat added at that point, the excess temperature at a down-stream point is a measure of the amount of the waste heat that still remains. The solution of equation 10 describes the excess temperature as a function of space and time and, thus, gives an idea of how much of, and to what degree, the stream is affected. Calculations involving excess temperature are not limited to waste heat discharges. For instance, the re- covery of subnormal stream temperatures resulting from cold water releases below impoundments may be estimated using the excess temperature approach. In this case, excess temperatures will be negative owing to the fact that the stream has lower energy content than it would under natural conditions. In order to calculate the excess temperature using equation 10 it will be necessary to know T, at some point as a function of time, for example, a boundary condition. This may be specified at the point of dis- charge, in the case of a waste heat discharge, by divid- ing the rate of discharge of waste heat by the product of the density, the specific heat, and the volumetric rate of flow of waste water into the stream. It is also possi- ble, in those cases where the stream mixes rapidly, to measure the difference between the temperature at a point just upstream of the discharge point and the temperature of the stream once fully mixed below the discharge point. Both of these approaches assume that the temperature just upstream of the discharge is the natural temperature. If this is not the case, it will be necessary to estimate the natural temperature at the upstream point. The solution of equation 10 for excess temperature requires expressions for the heat exchange due to out- going longwave radiation, evaporation, and conduc- tion. In the case of evaporation and conduction mecha- nisms, there are a number of expressions available. Much room for improvement in these expressions exists. The models discussed here are among the most frequently used. The longwave or "thermal” radiation emitted by a body is known to be related to the temperature of a body by the Stefan-Boltzmann equation: H,=ae(T+273.16)4, (11) where o-=5.67-10‘8 watts ~m'2'°K ‘4 is the Stefan- Boltzmann constant, e is the total hemispherical emis- sivity, T is temperature of the body in degrees Celsius, and Hb is given in units of watts-m2. Anderson (1954) reported that for natural waters the emissivity is es- sentially independent of the dissolved-solid concentra- tion and temperature. Emissivity was also found to be independent of the suspended sediment load, but an oil layer on the water surface significantly reduces emis- sivity. Anderson concluded that for oil-free natural waters the emissivity is 0970:0005. In spite of extremely precise formulation provided by equation 11, there are still inaccuracies that may at- tend its use with natural water bodies. These errors arise becauSe of errors in the temperature. Most longwave radiation emitted by a water surface is pro- duced in the first 100 ,um below the water surface (McAlister and McLeish, 1969). Ewing and McAlister (1960) found that the temperatures within this layer 6 THERMAL LOADING OF NATURAL STREAMS may differ significantly from that temperature which would be measured with the smallest conventional temperature measuring device placed as close to the water surface as possible. The true surface tempera- ture is generally found to be lower than the bulk water temperature. Thus, calculations of back radiation based on bulk temperature will overestimate the ac- tual back radiation. The error in calculated back radia— tion will usually be less than 3 percent, because the surface is usually less than 2°C (Celsius) cooler than the bulk. Equation 11, with the temperature measured near the surface using conventional thermometry, should give a satisfactory estimate of back radiation for most purposes. Evaporation contributes to the heat flux at the air— water interface in two ways. First, in order for water in the liquid state to enter the gaseous state an amount of energy, the latent heat of vaporization must be added. This energy, approximately 2.5-106 joules per kg of water, must be supplied from the water body. Thus, the water body loses energy by the evaporation process. In addition, some mass of water is lost, and this repre- sents a decrease in the total energy of the system. This term does not have any effect on the energy content per unit mass and becomes important only in cases where a change in the mass of the system is significant. It will be ignored in this work. The rate of evaporation from a water surface is re- lated to the rate at which water vapor may be transfer- red away from the interface. The transport process in- volved is a turbulent diffusion process. If the air far away from the interface has a lower partial pressure of water vapor than the air very near the interface, then a gradient for diffusion will exist. Note that this gra- dient may be reversed and condensation may occur, causing an influx of energy to the water body. The partial pressure of water vapor at the interface is the saturation vapor pressure for water at the temperature of the interface that will be approximated by the bulk water temperature. The diffusive flux of water vapor away from the interface may be expressed as 6PH20 _1_ _ 12 ”21%., 6y ( > E: where R is the gas constant, T2 is the air temperature, kH2O is the turbulent diffusivity of water vapor in air, and pH20 is the vapor pressure. The diffusivity is known to be a function of wind velocity. If the wind velocity profile follows a logarithmic form, it may be shown (Priestly, 1959) that at a given distance, y2, above the surface K2U2 k :____—_ %0%)nhM%%V (13) where K is the von Karman constant taken to be 0.40, 1) is the turbulent Schmidt number, v2 is the wind speed at elevation y2, and y0 is a constant known as the roughness. This expression is valid only during periods of neutral atmospheric stability, but it shows a direct proportionality of kH20 on wind speed. Among many existing models of evaporation, the most commonly used are the mass transfer models, which are also called Dalton-type equations. These models recognize that evaporation is proportional to the vapor pressure difference and that the coefficient of proportionality is a function of wind speed. Marciano and Harbeck (1954) and Harbeck (1962) studied them carefully and found that the following simple formula gives adequate results for lakes and reservoirs when all quantities are time averaged over a period of 1 day or longer: E=NU2(pS-p2), (14) where N is the mass transfer coefficient, 1)2 is the wind speed at some elevation (usually two metres), p, is the saturation vapor pressure of water at the temperature of the surface, and p2 is the partial pressure of water vapor at the same elevation as used for wind speed. Harbeck (1962) found that the value of N varies with the surface area of a water body according to N=3.42'10'8 - A, "005, (15) where AS is the surface area in square meters andN is given in units of kg-m’l-newton‘l. Taking a typical stream to have an effective surface area for evapora- tion of 1,000 m2, one finds a value of N=2.42-10’8 kg -m’1 'newton‘l. It has been recognized for some time that the value of N may not be a constant but may depend on wind direction and atmospheric stability. While these fac- tors may not be important when using data averaged over long periods of time, they may be very important when using hourly data. Brutsaert and Yeh (1970) showed that rather than N depending on the area of a water body, it should depend on fetch. They find thatN should be proportional to the fetch to the -0.25 power. The average fetch length will depend on wind direction for all but circular bodies, and it is very difficult to assess this factor for streams. There is an increasing interest in the effect of at- mospheric stability on the mass transfer process. It is generally believed that under unstable atmospheric conditions the rate of mass transfer will be increased owing to the existence of buoyancy-driven natural con- vection currents. And under stable conditions, the value of the eddy diffusivity will be reduced, reducing the rate of mass transfer. On the basis of the Lake HEAT TRANSFER AT THE AIR-WATER INTERFACE 7 Mead studies, Harbeck, Kohler, Koberg, and others (1958) recommended a stability dependent form for the mass transfer coefficient: N=1.04-10'8[1—0.03(T2—T)], (16) where T is surface water temperature in °C, T2 is the air temperature in °C, andN is given in units of kg - m‘ 1 -newton'1. The evaporative heat flux may be calculated directly from the flux of water vapor, E, using any of the above expressions. This is done by multiplying the flux by the latent heat of vaporization, A, so that He =>\E. (17) A commonly used value of A is 2495 - 103 joules per kg. The conductive heat flux has generally been calcu- lated from the evaporative heat flux, as was first suggested by Bowen (1926). It may be shown that if the vertical fluxes of water vapor and heat are independent of distance from the surface, and if the turbulent dif- fusivities of heat and mass are identical, then HC=B He and (18) B=6.1-10'4~P(T—T2)(Ps—p2)'l, (19) where B is the Bowen Ratio and P is the atmospheric pressure in newtons m'2. Temperatures are in degrees Celsius, and vapor pressures are expressed in newtons ' m'2. Equations 11, 17, and 18, in conjunction with equa- tions 14 and 19, describe all the fluxes appearing in the one-dimensional excess temperature model, equation 10. The same fluxes are also needed for the two- dimensional excess temperature model. The one-dimensional excess temperature equation can now be solved, provided that necessary initial and boundary conditions are specified and that the wind speed and partial pressure of water vapor at the 2-m level are known. Unfortunately, all the flux terms ex- cept the conductive flux, Ho, are nonlinear functions of water temperature, T. The back radiation depends on the fourth power of water temperature according to equation 11. The evaporative flux depends on the sat- uration vapor pressure of water, which is an exponen- tial function of water temperature, ps=610'exp[19.7071—5383.2'(T+273.16)'1] (20) where p, has units of newtons'm'2 and T is in degrees Celsius. Linearized approximation of Hb and He may be ob- tained by expanding equations 11 and 17 as Taylor series about an arbitrary base temperature, Tb, and truncating all terms of second order and higher. Sub- tracting the expressions for T" from those for T, the following equations are obtained: H,(T) = HAT") + 4047*, + 273.16)3~T,, (21) He(T) = He(Tn) + 5383.2'ANUZpS(Tb) (Tb + 273.16)‘2'Te, and (22) H,(T) = H,(T,,) + 6.1~104~PWU2T, . (23) Substituting equations 21, 22, and 23 into equation 10, the one—dimensional excess temperature equation is 6Te 6T9 6t +V 6x = —__UTe where U, defined as the heat dissipation coefficient, closely resembles the conventional engineering heat transfer coefficient and is given by (24) U = 406(Tb+273.16)3 + Wu, [53832-103019 (Tb +273.16)‘2 + 6.1-104-P] . (25) For most computations, a constant temperature may be used as Tb so long as it does not deviate excessively from natural temperature Tn (Jobson, 1973a, Yotsuk- ura and others, 1973). Two other coefficients for sur- face heat dissipation may be defined in conjunction with U. These are U,= pa and (26) 7d: pCUPY (27) In solving equation 24, note that the left side is iden- tical to the one-dimensional substantial derivative, DTe/Dt. By definition, it describes the time variation of T6 of a particular water packet. Equation 24 can be simplified by interpreting it in the Lagrangian sense of following the water packet moving with velocity V. The travel time, 1', may be defined such that r x = deT' + x0 (28) and 0 t = 1' + to, (29) 8 THERMAL LOADING OF NATURAL STREAMS where x0 and to are given values of x and t at 7:0. Equation 24 is, thus, reduced to an ordinary differen- tial equation with 7- as the single independent variable: dT T e = ———g. 30 d7 7d ( ) The analytical solution to this equation is easily ob— tained as 7. TezTeo exp (_f 1 d7"), 7d (31) where T8,, is an initial excess temperature at 7:0. Note that the time constant, 7d, depends on U, which, in turn, depends on v2 and T, as shown in equation 25. Even if Tb may be assumed constant, 02, the 2-m wind speed, varies rapidly with time. Thus, 7d is a function of time. It is often desirable to predict downstream tem- perature for a travel time longer than the interval for meteorological observation; for example, employ equa- tion 31 for a travel time of 24 hours using Wind-speed data collected at hourly intervals. In cases such as this, the travel time must be segmented into periods, AT, equal in length to the period for a wind speed observa- tion, and the integral in the exponential term of equa- tion 31 becomes a summation of A’T/Td over the travel time 7. This segmentation causes the calculations to have a serial character much as would a direct nu- merical integration of equation 10, although less te- dious and free of convergence problems. Yotsukura, Jackman, and Faust (1973) have shown that the linearization results in appreciable errors when the excess temperature is larger than 10°C. In such cases the numerical integration of equation 10 would be an attractive alternative. Recapitulating on the various heat exchange equa- tions introduced in this section, it is evident that the description of evaporation processes needs further im— provement. Equation 14 was adopted in this report on the basis of proven practical usefulness, but it is an empirical equation. It will be very attractive to develop equations that do not depend on water temperature. The aerodynamic models of estimating evaporation have an advantage over the mass transfer models in that they do not contain empirical coefficients such as the mass transfer coefficient and that they accommo- date more theoretically sound description of evapora- tion processes. Jobson (1973b) and Pierson and J ackman (1975) reported progress made in this course of approach. The total incoming radiation, Hr, was removed from the energy equation when equation 10 was derived for excess temperature. This heat flux is important when the natural temperature needs to be estimated. This quantity is best determined experimentally. When ex- perimental determination is not possible, however, there are models for both the longwave and shortwave incoming radiation. Sellers (1965) described a number of models for both fluxes. Koberg (1964) presented a method for estimating longwave radiation, and Ander- son (1954) presented a model for determining the re- flectivity of a natural water surface. These models should only be employed where direct measurement of incoming radiation is not feasible. It is also worth noting that, in spite of nonlinear relations between heat fluxes and water temperature, the heat exchange equations and their linearized ap- proximations can be used for widely varied time inter— vals ranging from 15 minutes to more than 24 hours, on the condition that they are used within a natural range of temperature variability (Jobson, 1972; Yet- sukura and others, 1973). ANALYSIS OF NATURAL TEMPERATURE As was demonstrated above, the use of excess tem- perature, Te, is a very attractive method of evaluating man’s impact on the thermal regime of a stream. This technique may be used where excess temperature is the only variable of interest, as might be the case in many planning studies. There will, however, always be instances where one must predict actual water tem- peratures. A study of the influence of discharges on the biota of a stream is one example. Another example is where multiple manmade alterations of unknown magnitude affect the thermal regime of a stream in a given reach. In such cases, the excess temperature alone is not enough. It will also be necessary to deter- mine the natural temperature. In general it is not pos— sible to observe Tn downstream of a waste heat dis- charge, since that discharge will have modified the natural temperature. An exception to this occurs when excess heat mixes slowly in the transverse direction, allowing unmodified natural temperature on one side of the stream throughout the reach. Here the natural temperature may be observed. Where this is not possi- ble, the natural temperature must be predicted. In the analysis of natural temperature, a lineariza- tion of surface heat flux term, H (Tn,t), may not offer much advantage, because the incoming radiation, H,, cannot be removed from the equation for T,, as was the case for Te. In solving equation 7, therefore, all four components of H must be taken into consideration. It is also clear that the variation of depth, Y, is of comparable importance to the right side of equation 7 as that of H. Natural streams frequently have signifi- cant depth variations in both longitudinal and trans- verse directions. It is common to think of the stream as made up of an assemblage of shallow, fast-flowing riffles separating deep, slow-flowing pools. Pool and ANALYSIS OF NATURAL TEMPERATURE 9 riffle regimes are frequently encountered in nature. This casts considerable doubt on any assumption of depth uniformity in either longitudinal or transverse direction. Some insight to the effect of longitudinal depth variation on the natural temperature may be gained by resorting to the concept of the thermally homogeneous stream. THERMALLY HOMOGENEOUS STREAMS A thermally homogeneous stream is defined as one for which no longitudinal gradient of temperature exists. The temperature is thus a function of time but not of position. Parts of shallow streams far down- stream of their headwaters may approach homogeneity rather closely. Consider a steady uniform stream where V and Y are constants denoted as V0 and Y0 and where a sinusoidal net heat flux is assumed as follows: H(T,,,t) = H, cos wt. (32) Substituting equation 32 into equation 7, the thermal equation is 6T 7% a 0T, = H, pom If the stream is thermally homogeneous, then, by defi- nition, 8Tn/6x= 0 and equation 33 can be integrated to Tn = TM + L sin wt, pCpYow cos wt. (33) (34) where Tm, is the temperature at t=0. As there are no longitudinal gradients of temperature, equation 34 is valid throughout the homogeneous portions of the stream. Now assume that this homogeneous water suddenly enters a reach where the depth changes to Y=aY0. If a is constant, Y and V are constant. The surface heat flux along the new reach is assumed the same as before. The thermal balance in the new channel is thus ('9T,1 +V 6T” _ H0 at 6x — 7005 wt. (35) pYa The boundary condition at the entrance to the new reach is given by equation 34. Equation 35 can be integrated by the introduction of travel time, T, as defined by equations 28 and 29 in the previous section. The solution that satisfies the bound- ary condition is obtained as _ Ho - _ T,,(t,'r) — T,,0+ m s1n w(t 7') H + apCpYow (36) [sin cut—sin wot—7)]. Equation 36 may be interpreted in two different ways. The travel time, T, is equal to x/V, where x is the downstream distance measured from the entrance to the new channel. Thus, by fixing a value for 1-, equation 36 gives temporal variation of water temperature ob- served at a fixed point. Alternatively, by fixing a value for t, equation 36 gives an instantaneous temperature profile along the channel. The first method of interpre- tation is used here. Figures 1 and 2 show temporal variations of temper- ature, Tum—Tm, as observed at several positions, which are expressed in units of traveltime rather than distance. Note that the reference temperature, Tm, is constant but arbitrary. Figure 1 is for doubling the depth (a=2), and figure 2 is for halving the depth ((2:14). The value of w is 77/ 12 radians per hour, and Ho/pCPYOw is assumed to be unity. In both figures it is clear that the temperature regime in the new channel became nonhomogeneous, as temperature at a given time depends on position. In other words, longitudinal temperature gradients were introduced to the new channel as the result of depth change. According to figure 1, at a position 12 hours downstream of the point at which the depth of the channel is doubled, tempera- ture is constant. Amplitude of the diurnal oscillation decreases continuously as a position approaches the 12 hour point from either direction. This is a somewhat pathological condition that occurs only for a doubling of depth. Figure 1 demonstrates that when the depth in- creases, the downstream diurnal amplitude will be less than that for the homogeneous stream. On the other hand, when the depth decreases, downstream diurnal amplitudes will be greater than that for the homogeneous stream, as seen in figure 2. Another in- teresting effect of the depth change is the shift of the time for maximum and minimum temperatures. For instance, while the temperature in the homogeneous stream reaches a maximum at 30 hours, the maximum temperature will occur at about 28 hours for a point 3.3 hours downstream and about 32 hours at a point 20.7 hours downstream, in the case of halving of depth. When a=1, equation 36 reverts to equation 34, and there is no longitudinal temperature gradient at any location. These results, although highly idealized, suggest some conditions one would expect to find in a natural stream. For a thermally homogeneous regime, the amplitude of the diurnal temperature oscillation should be the same for all positions. Further, the time at which the maximum and minimum temperatures are observed should be the same for all positions. On the other hand, there are a number of factors that could prevent even a very shallow natural stream from 10 THERMAL LOADING OF NATURAL STREAMS 1'0 I ’1- ‘ I I l r I 0.8 — ,// _ . - / / \ Observed at downstream posmon 0.6 which requires traveltime of: - TURE RELATIVE TEMPE I o a l / -10 1 1 I I i‘lhz I 24 27 30 33 36 39 42 45 48 TIME,|N HOURS FIGURE 1.—Temp0ral variations of water temperature at five p0si~ tions in a hypothetical channel with doubled depth (a=). becoming thermally homogeneous. Longitudinal vari- ations in depth can prevent homogeneity in a channel of any size. However, these variations will have an effect only if the depth variations are such that the average depth (averaged with respect to longitudinal position) changes over relatively long distances. A par- ticle of water passing through a short, shallow reach will suffer a more rapid temperature increase at mid- day than would a particle of water of average depth. But owing to shortness of the shallow reach, the actual difference in temperature between the two particles may be quite small, assuming they were initially at the same temperature, and this difference would be eliminated if the shallow reach was followed by a short, deep reach. Variations in the exposure of a 'stream to incoming radiation or wind can also prevent a stream’s attaining thermal homogeneity. Passage through a deep, narrow canyon will significantly reduce the total radiation in- cident on the water surface during a day. In such a case, a longitudinal gradient in temperature can be established even if the stream was homogeneous and free of such gradients at the upstream end of the can— yon. Passage through a reach with appreciably differ- ent exposure to wind would produce a similar effect. Both radiation and wind exposure are affected by bankside vegetation (Pluhowski, 1970). THERMAL HOMOGENEITY OF THE POTOMAC RIVER Very little data on thermal homogeneity of streams exist. The addition of waste heat to a stream induces longitudinal temperature gradients. This renders an otherwise homogeneous stream nonhomogeneous. Thus, it is impossible to determine whether the streams studied here were or were not homogeneous, and this study adds little to what is known about this concept. I I I I I Observed at downstream position which requires traveltime of: _ 3.3 hours 12.0 TEMPERATURE RELATIVE TO A REFERENCE 24 27 30 33 36 39 42 45 48 TIME, IN HOURS FIGURE 2.—Temporal variations of water temperature at five posi- tions in a hypothetical channel with halved depth (a=‘/2). The data on the Potomac River are a notable exception. Two of the three studies at this site were conducted under conditions such that the heated plume, which discharged at the left bank, did not affect the water near the right bank for at least 35 km downstream of the discharge site. The Potomac is 500—1,000 m wide in this reach and 1—2 m deep. The shallow nature of the stream led to greatly inhibited lateral mixing. The third study was at an extremely low river discharge, and the large fraction of river water that was diverted through the powerplant condensers was able to reach the right bank within about 3 km of the discharge point. Figure 3 shows a map of the Potomac study reach and a plot of average depth as determined by discharge measurements versus the longitudinal position along this reach. It reveals considerable variability in depth. The pool at the end of the reach is the result of back- water behind the lip of Great Falls natural dam. The pool at about 8 km is utilized for ferry crossings. The shallow area below the power plant site will not permit operation of small outboard motors during very low flows. During both the March and May studies in 1969, a helicopter equipped with a Barnes PRT—5 radiometer was employed to determine longitudinal and trans- verse temperature patterns in the river. Figure 4 pre- sents data from the overflight of several different cross-sections during the March study. Note that even as far downstream as section 11, 35 km from the pow- erplant, the heat plume is still confined essentially to the left half channel. These temperature determina- tions were all made within a 30-minute period starting at 12:05 pm. on March 11, 1969. ANALYSIS OF NATURAL TEMPERATURE 'I. 8 I I I I I U) Section 7 Section 10 LU _ — D: 1.6 ’— Lu E E 1.4 — I. E 1.2 — Section 9 — Lu D m 0 1.0 '— < '1 (I g 0.8 — _ < Section 4 0.6 I | L I I O 5 10 15 20 25 30 DISTANCE FROM POWER PLANT DISCHARGE SITE, IN KILOMETRES Flow direction ‘ 78° 77c ‘ I I PENNSYLVANIA Section 1 __ .___- (0 km) OHagerstown *% Section 3 . lg) MARYLAND (1.3 km) / chkerson 206 X Power Plant ~ VIRGINIA Section 2 Leesburg Mason Island . (0 km) 390 _ _ ‘v’,.°' Section 4 Study reach § (2.3 km) ‘ .s‘ Section 6 I I Whites Ferry\ (5_1 km) Section 7 (7.7 km) Q S E K \\ O 1 2 3 4 MILES Section 9 I I I I I I (15.5 km) I I I I I I f ' o 1 2 3 4 5 KILOMETRES \\ \ Section 10 Seneca ‘ \\ (28.8 km) 0 NA TURAL DAM FIGURE 3.—Sketch of the Potomac River study reach below Dickerson Power Plant, Maryland. 11 12 11 11 9 Section 6 (5.9 km) 9 Section 7 (7.7 km) CU ._ U _. — .— E 7 g 7 . 5 ._ :9, _ 5 _ ._ Lu kw 5 3 3 )_ LB RB LB RB <1: I ~ ”J 9 9 . 2 Section 8 (12 km) SectIon 9 (15.5 km! 7 — - 7 E 2 \ _ _ 5 _. _. 5 fig— 3 _ 3 LB RB LB NEAR RB 9 U 7 _ Above section 10 (27 km) 0 U "‘ Z 5 —- E _ LLI 3 3N 0: 3 LB RB I- <1 I 9 :1 Section 11 (35 km) 1: 7 — c — 2 to LU 5 73 _ ’_ k—i—k—v 3 LB RB TRANSVERSE DISTANCE FROM LEFT BANK TO RIGHT BANK FIGURE 4.—Transverse profiles of water temperature at selected sections, the Potomac River below Dickerson Power Plant, Maryland, March 11, 1969. Careful comparisons of the radiometer measure- ments with simultaneous ground measurements made using Whitney thermometers show agreement to within O.1°C, except in the hottest part of the plume where appreciable “cold skin” effect is observed. This effect is caused by the evaporation and conduction at the water surface. There is a thin film of water at the surface where transport of heat is solely by molecular conduction, because turbulent transport diminishes as one approaches the water surface from below. As the heat loss is occurring at the upper edge of this film, and the heat must be supplied from below the film, a net flux of heat toward the surface results in the tempera- ture at the surface being somewhat lower than the temperature below the film. Ewing and McAlister (1960) have shown that this film will not be penetrated by turbulence even in the presence of waves unless the waves are breaking. The water temperature near the right bank appar- ently was unaffected by the thermal discharge throughout the reach, and the temperatures observed with the airborne radiometer are, to within the limits of accuracy of our observations, identical to those measured on the ground. Figures 5 and 6 show the radiometer observations of natural temperature as a function of longitudinal posi- tion for the March and May studies, respectively. In THERMAL LOADING OF NATURAL STREAMS 5 I I l TEMPERATURE IN °c A I E 3 I I I O 10 2O 30 4O DISTANCE FROM HEAT DISCHARGE SITE, IN KM FIGURE 5.—Longitudinal profile of natural water temperature near right bank, the Potomac River below Dickerson Power Plant, Maryland, March 11, 1969. ‘9 I I I TEMPERATURE, IN r)C 65 I I I 10 2O 30 40 DISTANCE FROM HEAT DISCHARGE SITE, IN KM .: \I 0 FIGURE 6.—Longitudinal profile of natural water temperature near right bank, the Potomac River below Dickerson Power Plant, Maryland, May 14, 1969. both cases the natural temperature is nearly constant with position. The maximum variability is 0.4°.C. There does appear to have been a cooling trend in the upper part of the reach during the May study, but this is probably not significant. On the basis of figures 5 and 6, it seems reasonable to conclude that the Potomac River near Dickerson, Maryland, is an example of a thermally homogeneous stream. In spite of some known variations in depth, there is no distance-averaged longitudinal gradient of natural temperature, and local gradients are small. Figure 4 reveals that there were some observable transverse gradients in natural temperature, particu- larly at section 10. However, the greatest lateral fluc- tuation of natural temperature was at section 1, just above the thermal discharge site. Figure 7 shows natural temperature versus transverse position at sec- tion 1, observed at several times during the March study. At some times the lateral variations in tempera- ture were quite pronounced. It is difficult to ascertain the source of these varia- tions. Monocacy River joins the Potomac River on the left bank about 2.5 km upstream of the site. The flow of the river was small, about 20 1113 sec'1 compared to 149 ANALYSIS OF NATURAL TEMPERATURE 13 1 l l T T ‘l T b- .1 4 1— — I- '/Z\—/. - U, I' Time: 1800, March 11, 1969 ‘ 2 " -I ‘” I J — 8 z 3 I— _ — 2400, March 11, 1969 _‘ Lu. _ g — . I— < - -4 0: Lu - _ 3 Lil—1 2 — _ " 1 0550, March 12, 1969 _ 41 1 l l L l I 0 4o 80 120 160 200 240 280 320 TRANSVERSE DISTANCE FROM LEFT BANK, IN METRES FIGURE 7.—Transverse profile of natural water temperature at section 1, the Potomac River above Dickerson Power Plant, Maryland, March 11—12, 1969. m3sec‘1 for the Potomac, and the variations do not ap- pear to be substantially related to this inflow. The South Branch of the Potomac River joins at the right bank about 30 km upstream. This is a large flow and the slow transverse mixing could lead to variations, but again the pattern of the variations does not seem to support the idea that variations were caused by un- mixed inflow. Judging from the pattern of cool center- stream conditions during the day and warm center- stream conditions during the night, it seems most probable that there is a long, deep channel in the center of the stream upstream of section 1. The air- borne radiometer results indicate, however, that transverse variations of temperature are less pro- nounced at most downstream sections. PREDICTION OF NATURAL TEMPERATURE When one assumes a thermally homogeneous stream, the longitudinal gradient term, 6T,,/6x, van- ishes, and equation 7 is reduced to dTn _ H . dt — pCpY An explicit finite difference form to solve equation 37 numerically is (37) H(T"(t),t+%) - At my Note that H for an interval At is calculated by using T,(t+At) = Tn(t) + (38) known water temperature, Tn, at t and given meteorologic conditions at t + At /2. Figures 8 and 9 are the results of calculations using equation 38 for the station nearest the right bank of the left channel at section 4 of the Potomac River. As the initial condition, the temperature observed at this station during the first measurement of the study was used. The heat-flux term was adjusted by trying vari- ous combinations of the depth, Y, and the mass trans- fer coefficient, N, in order to investigate an optimum 4-4 I I 1 I ‘l I 1 I 1 I I 4.0 3.6 3.2 Calculated — Observed 2.8 - TEMPERATURE, IN CELSIUS | 2.4 — 2.0 — 1 I 1 I I l 1 l 1 | 1 12 16 20 24 4 8 12 TIME, IN HOURS, MARCH 11—12, 1969 FIGURE 8.—Comparison of observed and calculated natural water temperatures at section 4, the Potomac River below Dickerson Power Plant, Maryland, March 1969. 20.4 I 1 I R I 1 —I I I I I I 20.0 19.6 19.2 TEMPERATURE, IN CELSIUS 18.8 12 16 20 24 4 8 12 TIME, IN HOURS, MAY 14—15,1969 FIGURE 9.—Comparison of observed and calculated natural water temperatures at section 4, the Potomac River below Dickerson Power Plant, Maryland, May 1969. 14 THERMAL LOADING OF NATURAL STREAMS range of N values suitable to stream conditions. The adjustment of parameters was done in steps to improve the fit of calculated temperature at the end of a run with the observed temperature at that time. Based on data from the March study, the best combination was found to be Y=1.83 m and N=2.31-10'8 kg-m‘l -newton'1. This depth value is somewhat disturbing, since all observations of depth above section 4 would suggest a value near, and perhaps slightly less than, 1 m. This error could be due to consistently high radio- meter readings, but in this case we would expect poor prediction during the night when solar radiation was near zero. This is not the case, and the cause of this depth discrepancy remains an unanswered question. It should be noted that 1.83 m is, however, fairly typical of the depth in much of the rest of the river. The calculated temperature at section 4 for the March study is presented as a function of time in figure 8. The calculated values agree quite well with the ob- served values. The maximum discrepancy is O.3°C, and the average of the absolute error over the 24—hour period is about 01°C. The shape of the calculated curve is somewhat sensitive to the value of the mass transfer coefficient, N. Data from the May study on the Potomac was an- alyzed using the same procedure. However, both Y and N were fixed at values found for the March data. The results are presented in figure 9. The agreement is again very good. A maximum error of 0.3°C and an average absolute error of about 02°C is observed. Note that, because depth was already established by the March study, there is no requirement that calculated and observed temperatures agree at the end of the 24 hour period. The second method of calculating natural tempera— ture is to solve equation 7 directly without assuming the state of thermal homogeneity. By introducing the travel time 7, as the single variable, equation 7 is re- duced to an ordinary differential equation. For a uni- form subreach, an explicit difference scheme for the differential equation is A H(Tn(7),t0+7+ 73m Tn(‘r+A7) = Tn(7) + pCpY (39) Equation 39 was applied to the data from the Potomac River. As explained previously, equation 39 is a Lagrangian equation of temperature for a packet of moving water. In order to compare calculated tempera- tures with observed ones at a fixed location, equation 39 was solved repeatedly for different to and Tm, (to) observed at section 1. The value of A7- was chosen to be 1 hour. As the observed depth was used in equation 39, the value of N was smaller than that found in the homogeneous stream calculations. It was fixed at N=1.5-10‘8 kg-m-'1 newton'1 for the March and May data. Some results of this calculation are shown in figure 10. The agreement with observed temperature is gen- erally satisfactory except at section 7 for the March study, where the calculation underestimates the ob- servation by almost 1°C. The observation station, number 9 of section 7, was located at the midstream in the March study, and about 45 percent of the total discharge was flowing to the left side of this station. It was first suspected that there may have been some effects of the heated plume. However, the air-borne radiometer data indicated that this station was clearly outside the heated plume. A more reasonable explana— tion appears to be the variation of T”, at section 1. Between 0150 and 0550 hours, March 12, water tem- perature near the left bank, which was used for the calculation, was cooler than the midstream by 0.7 to 09°C. This initial difference could have been convected downstream without much modification. The water near the left bank side at section 7 could actually have been cooler than midstream for the period between 0720 and 1120 hours, March 12. For the May study, the transverse variation of Tm (to) between the left bank and midstream was between 0.1 to 04°C throughout the survey period. The error of calculation for the May data is less than 0.5°C. Figure 10 also includes the calculations made for the Riverside Inlet Canal near Greeley, Colorado. The data are not suitable for publication as yet, because of some instrumental problems in the radiation and tempera- ture measurements. The calculations were based on the adjusted data, which included a uniform 20 percent reduction of recorded incoming solar radiation values. The canal is 16 km long with little vegetation along both banks. The channel width ranged from 14 to 28 m, and the depth variation was between 0.6 and 1.2 m. The travel time, 7-, for 16 km reach was about 5.5 hours. The temperature variation within a cross sec- tion was negligible except near the entrance to the canal. The large diurnal fluctuation in Tn is caused mostly by that of net incoming radiation. The calcula- tion was based on A7=1fl1 hours and N=1.5-10“8 kg-m‘l' newton'l. The agreement between observed and calculated T” is satisfactory. In summary, the one-dimensional thermal equation, in conjunction with surface exchange equations, ap- pears to provide a workable model for the natural temperature data available in this study. On the other hand, some questions on the transverse variation of temperature as well as the magnitude of mass transfer coefficients are not adequately answered by the pres- ent analysis. Note that the calculation based on the homogeneous stream model, equation 38, is much simpler than that ANALYSIS OF NATURAL TEMPERATURE I l I I I l I I I l I I Section 7, Potomac River \~ ‘ = T..(t.)exp (—I; (31) where to is the time when a particular packet of waste water with excess temperature T,” leaves the effluent site, 7 is the travel time measured relative to to, and t=7+to as defined by equation 29. All the results dis- cussed below were generated using some variation of this solution. Before calculations can be undertaken, the input data must be put in a manageable form, because tem- perature data and meteorologic data collected at var- ious times are quite cumbersome to use in the raw form. The temperatures and meteorologic variables were interpolated to each quarter hour and meteorologic conditions were assumed constant during each 15-minute interval. Estimated travel times were rounded off to the nearest quarter hour. Starting with a packet having a given excess tem- perature at time to, equation 31 is employed to calcu- late the excess temperature of the same packet at tO+Ar where AT=15 minutes. The calculated tempera- ture at t0+Ar may then be used together with equation 31 and meteorologic data to calculate Te (t0+2A1-). This procedure is repeated, starting with another water packet leaving the discharge site 15 minutes later than in the previous claculation. Repeating this procedure as many times as available data permit generates a calculated temperature history at the downstream sec— tion. The Lagrangian solution method indicates that if the first excess temperature data at the discharge site is available at to, the first time at which downstream temperature may be calculated is at to +1- where r is the travel time. Thus, observed downstream temperatures between t, and t0+7 may not be compared with calcu- lated temperatures. As the Data Report describes, however, the decision to impose a uniform data collec- tion period of 24 hours on all thermal surveys was made prior to the choice of predictive models. Accord- 18 THERMAL LOADING OF NATURAL STREAMS ingly, application of the Lagrangian method is re- stricted to a reach with the travel time less than 24 hours. This inconvenience, however, was more than compensated by an advantage of the Lagrangian solu- tion; that is, one is able to assess surface heat dissipa- tion closely by following moving water masses with known initial excess heat. Several approximations, which are required in order to complete the above calculation, should be discussed. First, in computing actual water temperatures at a downstream cross section, the natural temperature must be added to the excess temperature. Because no other data on natural temperature were available, the upstream temperature was taken to be the natural temperature that would exist at all downstream points in the absence of waste heat discharge. The limitations on this assumption have been discussed in the section on the "Analysis of Natural Temperature.” It is also necessary to estimate the travel time and average depth for each reach. The travel time may be estimated by dividing the length of the reach by the average velocity for the reach. Since complete stream discharge measurements were available at most cross sections, the average velocity at each cross section was calculated by dividing total discharge by total cross- sectional area of the reach. The average velocity for the reach may then be calculated as the average of the velocities at the upstream cross section and down- stream cross section. This method can lead to rather serious errors if, as may well be the case, the channel geometries at the upstream and downstream cross sec- tions are not typical of the reach as a whole. Average depth for the reach could also be calculated by taking the average of the average depths calculated from the discharge measurement at the upstream and down- stream cross sections. This would again be of question- able accuracy. In order to alleviate the above errors, use is made of the following relation among discharge, travel time, and reach volume, QT=ASY, (40) where As is the total surface area of the reach and the depth Y is defined such that ASY represents the reach volume. Letting 7=nA7 and substituting these expres- sions into equation 31, AS n T€(to+7)= T8000) exp(— .21 Uu)’ (41) n l: where U.., is a transient value of UK for the i-th A7 period. Note that Q and A, can be determined accu- rately from field data and maps. Equation 41 was the form actually used in calculation. One additional specific approximation should be mentioned. Equation 7 for natural temperature is transparent to the phase transition that actually oc- curs at 0°C. Negative values of H can result in temper- atures less than 0°C were the water not to freeze. It is this hypothetical negative temperature, and not the freezing point of water, which must be used in cases where a stream is partially frozen and at 0°C. To do otherwise would require that H (Tn,t)EO, which is clearly incorrect. This problem arises in the data from the North Platte River where the unaffected stream was indeed frozen and at 0°C. For the North Platte River, it was necessary to calcu- late the hypothetical natural temperature as a func- tion of time. This was done by applying the homogene- ous stream method of natural temperature prediction using equation 37. Field radiometer data were employed, and the value of N , the mass transfer coeffi- cient, used was 2.681018 kg-m-1 newton-1. The initial condition for the integration was chosen to produce a slightly negative average value of H for a 24-hour period since field observations indicated an increased rate of ice formation. The predicted natural tempera- tures appear in table 1. RESULTS AND DISCUSSION One-dimensional temperature prediction was under- taken for all streams presented in the Data Report ex- cept the Holston River, where no discharge data were available, and the Potomac River, where two-dimen- sional analysis appeared to be more appropriate. These calculations employed the one-parameter evaporation and conduction model as shown in equations 14 and 18. In most cases, the value ofN was 2.68-10‘9 kg - m’1 ‘newton’l, hereafter referred to as the standard value for the mass transfer coefficient. In cases where other values were used, an explanation will accompany the discussion of these results. Figures 12 and 13 present sketches of river aline- ment, cross-sectional location, and other pertinent in— formation for all streams analyzed in the following section. WHITE RIVER NEAR CENTERTON, INDIANA, 1969 The temperatures calculated using the Lagrangian one-dimensional model for the various downstream sections at the White River near Centerton, Indiana, are compared with observed temperatures in figures 14—16. The reach length was only 12 km. Therefore, little heat loss occurred, and only the results from sec- tions 4 and 7 are presented to avoid congestion. The standard value of the mass transfer coefficient was employed. Heat loading varied somewhat during all three studies, and excess temperatures just below the D ONE-DIMENSIONAL MODEL OF EXCESS TEMPERATURE 19 TABLE 1.—Temporal variation of hypothetical natural temperature under freezing conditions, the North Platte River near Glenrock, Wyo» ming, January 28—29, 1970 t in hours T in CC 1400 __________________________ 0.00 1500 __________________________ 0.04 1600 __________________________ 0.08 1700 ___________________________ 0.04 1800 __________________________ —0.01 1900 __________________________ —0.09 2000 __________________________ —0.21 2100 __________________________ —0.30 2200 __________________________ —0.38 2300 __________________________ —0.50 2400 __________________________ —0.62 0100 __________________________ —0.74 0200 __________________________ —0.86 0300 __________________________ —0.98 0400 __________________________ —1.10 0500 __________________________ —1.26 0600 __________________________ —1.42 0700 __________________________ —1.10 0800 __________________________ — 1.80 0900 __________________________ —1.67 1000 __________________________ —1.49 1100 __________________________ —1.10 1200 __________________________ —0.64 1300 __________________________ —0.30 1400 __________________________ —0.11 discharge site were 1.7—2.5, 5.5—6.4, and 3.0—4.4°C for the February, June, and October studies, respectively. Figure 14 for the February study shows very good agreement between predicted and observed tempera- ture at cross section 4 and good agreement at cross section 7 except during the immediate predawn period. Because this was a period of very light winds, it is unlikely that errors in the estimation of evaporation and conduction could have caused the deviation. It ap— pears most likely that the deviation was caused by a failure in the assumption of constant natural tempera- tures along the reach. Figure 15 shows the comparison of prediction with observation for the June study. The rate of cooling has been rather substantially underestimated during most of the study. The period of underestimation coincides very closely with the period during which observed winds dropped to less than 1 m-secl. The observed wind velocities possibly were not characteristic of the winds that prevailed over most of the channel. The unexpectedly high dissipation rates could not be due to natural convection in this case, because water temper- atures were nearly equal to air temperatures during some of this period. Figure 16 shows the modeling results for the October study. Winds were high during the first 8 hours of the study although still less than 3 m-secl. Excess tem- peratures were generally in the vicinity of 4°C at all cross sections. The amount of cooling appears to have been overpredicted rather than underpredicted during Section 7 (6.9 km) Section 1 Whlte River near Centerton, lnd. (— 5‘1 km) Powerplant DISCHARGE CANAL Section 5 (3.5 km) . Section 6 .’Sect|on 2 (5,7 km) (~0.2 km / Section 4 0 1 MILE (2.1 km) s ‘5 (0.2 km) 0 1 KILOMETRE Fenholloway River near Foley, Fl. Section 5 Buckeye Cellulose (10.9 km) Corp. plant Section 2 t‘ 1 ' (fgcslil) (0.1 km) Section 3 Section 4 ' (3.4 km) (7.8 km) Section 5A and 6 1 2 3 4 MILES (133 km) l—rl-l—Ll—I'i—r—LH Section 7 o 1 2 3 4 5K|LOMETRES “8-1 km) Dan River near Eden, N. C, Section 2 g . DAfllfl\Sction 1 / (0.1 km) 0 1 2 3 4 MILES Hie—He 01 2 3 4 SKILOMETRES FIGURE 12.—Sketch of study reaches, 1. Tittabawassee River near Midland, Mich. Section 2 and' 2A (0.2 km) Section 4 Section 3 (6‘5 km) Powerplunt (3.0 km) Section 5 t‘ 1 Sec 1cm 1 2 3 MILES \(110 km) (—4.8 km) 0 Section 6 (17.2 km) 0 1 2 3 4K|LOMETRES North Platte River near Glenrock, Wyo. Section 4 Section 5 Section 6 /(10.9 km) (19.5 km) (29.3 km) / \ Section 1 (“2.3 km) / jPowerplant \Section 3 (3.1 km) DAM \ Section 2 (0.2 km) 2 3 4 MILES 012 3 4 5 KILOMETRES O 1 2 3 MILES Section 0 Section 2 (5.8 km) (0.8km)\ o 1 2 3 4K|LOMETRES DAM Section 3 ‘ (8.7 km) Section 5 f (24.1 km) Shawville owerplan _ /Section _1 /Sectlon1 Section 4 (3.2 km) (4.9 km) (15.9 km) West Branch of the Susquehanna River near Shawville, Penn. FIGURE 13.—Sketch of study reaches, 11. 20 THERMAL LOADING OF NATURAL STREAMS 10.6 I u l 1 "fi' ' '1' 26IIIIIIIIIIII " \ Section 4 (I) Calculated U) — _. 3 10 2 Section 7 _“ Observed to ~ U) _, _| u.I u: 24 - — u 0 I z E I _~ . _ I _., LIJ L” I CC 9 Section4 II I 3 ’ a 22 - ’ a '— * < I, at: l: Calculated natural temperature m u] _ —4 n. ~v - a. 2 —— Calculated \ Section _ E W 8 — b d ‘ 7 1' l“ ._ _ ———O serve \\ \l’ ‘ I-20— _I \ I h H—- < 7_4.1.|.1.|.| 191||||11|111 12 16 20 24 4 8 12 8 12 16 20 24 4 8 9 TIME, IN HOURS, FEBRUARY 26—27,1969 FIGURE 14.—Comparison of observed and calculated water tempera- tures, the White River near Centerton, Indiana, 1969. 33 ' l ' l T l ' l ‘ l g 31 — -- —- a ——--- ----- - -‘~~‘~ ----- _I _ Section 7 u.1 _ 0 Z _. 29 '_ Calculated — u.| -__ II Observed D - .. ’— < (I E 27 _ Calculated natural temperature— 2 Lu ,— _ — 25 ’ __ 24 1 I 1 l l I I l I I 12 16 2O 24 4 8 TIME, IN HOURS,JUNE11—12,1969 FIGURE 15.—Comparison of observed and calculated water temper- atures, the White River near Centerton, Indiana, 1969. the morning hours, when wind speeds were again very low. The cause is a decrease in measured natural temperature of 0.5°C between 0500 and 0700 hours. During the same period one would expect a similar decrease in the actual temperature at affected downstream sections, as no unusual transient had been observed moving downstream. This expected de- crease in observed temperature is absent. These data probably hold greater significance from the standpoint of fluctuations in natural temperature than they do from the standpoint of accuracy of the one-dimensional model, because of the short reach and travel time studies at the White River and the small dissipation. It is interesting to compare the fluctua- tions at the downstream sections with the fluctuations TIME, IN HOURS,OCTOBEFI 7»8,1969 FIGURE 16.—Comparison of observed and calculated water temper- atures, the White River near Centerton, Indiana, 1969. at the upstream natural temperature cross section. Figure 17 plots the deviation of the natural tempera- ture from its value of 2400 hours and the same devia- tions at the cross sections 3 and 7. Large deviations from the natural temperature pattern occur at cross section 3 in the morning hours of both days. Large deviations also occur at section 7 on the morning of the second day. These deviations are as great as 12°C. The assumption of the thermal homogeneity in natural temperature may have been the source of the errors in temperature prediction. FENHOLLOWAY RIVER NEAR FOLEY, FLORIDA, 1969 The Fenholloway River near Foley, Florida, is in- teresting for several reasons. It is the smallest of the streams studied, unusually small for a stream receiv— ing a waste heat discharge. Because of its small dis- charge and large heat loading, it had the largest excess temperatures of any stream studied, as high as 17°C. Finally, it had significant flow accretion in the reach studied. Most of the accretion was unmeasured, but there was one measured inflow from Waldo Spring which entered the river just downstream of section 5A. The temperature of the spring was about 21°C for both the March and June studies (it was not observed dur- ing the November study). The fact that the tempera- ture was fairly constant, at about the mean annual temperature for the region, suggests that the spring was ground water outflow and, therefore, is repre- sentative of the temperature of any accretion resulting from ground-water pickup by the channel. Figures 18—20 show comparisons of calculated and observed temperatures at the Fenholloway River. The starting point for the calculations was section 2, just downstream of the discharge. The reach studied was 18.1 km long. The standard value of the mass transfer coefficient was employed, although results indicate a ONE-DIMENSIONAL MODEL OF EXCESS TEMPERATURE 21 1'4 - . Section 1 A Section 3 1.2 _ 0 Section 7 - 0.8 '- 0.4 - I F3 N l DEVIATION FROM MIDNIGHT TEMPERATURE, IN CELSIUS .c'> a. l -0'6 .- —0.8 - 1 I I I I I I I I l I I L 6 8 12 16 20 24 4 8 10 TIME, IN HOURS, OCTOBER 7~8, 1969 —1.0 FIGURE 17.—Temporal variation of natural and heated water temperatures, the White River near Centerton, Indiana, 1969. somewhat higher value might be justified. The water at the upstream end of the reach was generally warmer than the air. All the studies reveal a dramatic overprediction of temperature. Errors as great as 4°C are found at the downstream sections. Much of the error appears to be due to the ground-water accretion. If it were assumed that all the accretion at section 7 occurred just before the section, mixing the cool ground water (computed as the difference between the discharge at section 7 and that at section 2), a cooling of 2°C for the March study and 5°C for the June study would be predicted. For the November study, section 5A replaced section 7 and a cooling of 2°C at this section would be predicted. These predictions compare fairly well with observed errors. Figures 21—23 present modified prediction which as- sumes a complete mixing of ground water and heated water. The agreement between calculated and ob- served temperatures is considerably better. The prob- lems with the shapes of the time-temperature relation- ships for the March and November studies remain since they are not affected appreciably by the ground- water accretion correction. A tendency to underesti- mate cooling is still evident. The cooling effect of the accretion is somewhat over- predicted if the accretion actually enters more or less uniformly along the channel, as is probably the case on the Fenholloway. In this case, the reduced excess temperature resulting from accretion to a heated pack- et near the source will cause a reduction in the heat losses from that packet. Therefore, the packet arriving at the end of the reach which has suffered the same apparent heat loss due to accretion as if the accretion all occurred at the most downstream point has suffered smaller losses to the atmosphere. This indicates a very significant under-estimation of heat transport at the air-water interface has occurred. This must, in large part, be due to the extremely unstable atmospheric conditions that prevailed. The results of the November study in figure 20 re- veal a peculiar and sharp increase in predicted tem- perature during the period from 1700 to 2400 hours at cross section 3. The calculated pattern then propagated to the downstream sections causing similar peaks later at successive sections. This was caused by a sharp de- crease in wind velocity at about 1600 hours, which caused the calculated decrease in excess temperature between sections 2 and 3 to decrease from 1.5°C at 1700 hours to 05°C at 2400 hours. During the same period the excess temperature at section 2 also increased nearly 1°C. The great sensitivity to changes of wind speed is the result of predicting evaporative and conductive heat losses using the one parameter mass transfer formula, equations 22 and 23. The observed temperature shows only a slight increase caused by the increased heat loading. It appears that cooling did not decrease dramatically during this period. This is probably the result of large evaporation and conduction losses caused by free convection. During this period the water was about 15°C warmer than the air. DAN RIVER NEAR EDEN, NORTH CAROLINA, 1969 The comparisons of predicted and observed tempera- tures for the three studies of the Dan River near Eden, North Carolina, are shown in figures 24—26. The start- ing point for calculations was section 1 just down- stream of the discharge. The reach was the longest studied, 31 km. Heat loading was fairly constant dur- ing all studies. Excess temperatures were about 2.8, 4.2, and 6.5°C for the April, May and October studies, 22 THERMAL LOADING OF NATURAL STREAMS 28 I I I I I I I I I I 27 — ll _ Section 3 ‘ // 26 — Section 3 __ / N Section 4 / \\ / w 2 \ ’ u) 25 -- \ / .— d \ \ / o e \ / 2 ~ ~ — _ a \ \ / Iii \ \ / II 24 - \ § —- 3 s \ \ / Section 4 I— s \ ’ . \ < SectIon 5 \ / I I: ‘ x / ’ LU ‘ ‘ \ § _ .— l E 23 _ \ I LLI — — \ / I- \\ ‘ \ I / I \ \ \ I \ \ ‘ N ’ \ u —- - — / Section 5 22 — \ \ / \\ Section 7 I \ Calculated - - — -’ ” ’ Section 7 21— —--- Observed ‘§~~ ”’ ' ‘ --— ’ 20 I I l I I I I I I I 18 20 22 24 2 4 6 8 10 12 14 16 TIME, IN HOURS, MARCH 26—27,1969 FIGURE 18.—Comparison of observed and calculated water temperatures, the Fenholloway River near Foley, Florida, 1969. respectively. The standard value for the mass transfer coefficient was used in all cases. The river flow is regu- lated, and only small changes in flow were observed during the studies. These changes should have little effect on the one-dimensional modeling, changing only the travel time. As the estimates of travel time used here are somewhat crude, it is doubtful that the slight discharge changes would affect this estimate. The predictions for the April study shown in figure 24 are in excellent agreement with measured values. The rapid rise in temperature at section 2 in the after- noon of the first day is correctly convected downstream producing an increase in temperature at section 4 dur- ing the early morning hours. This rather unusual in— crease was borne out by measurements. As in the case of the Fenholloway River, there were rapid changes in wind speed during the April study. The wind decreased from nearly 6 m -sec1 to 1 m -seC'l between 1500 and 1900 hours and then rose less dramatically from 0600 to 1000 hours of the second day. However, unlike the Fenholloway River, excess temperatures were only about 3°C. Thus, excess tem- perature would decrease less than a degree Celsius in any reach, and changes due to wind would be only a fraction of a degree. In addition, since air temperatures exceeded water temperatures, evaporation and conduc- tion should not be enhanced by free convection. The predictions for the May study are also good. There is a sharp decrease in temperature at section 2 at about 2100 hours that does not agree with observa- tion. This was the result of a 0.75°C drop in observed temperature at section 1, from which excess tempera- ture is calculated. The results seem to indicate that this drop may have been due to observational error. There is some underestimation of cooling at sections 3 and 4. Air temperatures were lower than water tem- peratures from 1800 hours until the conclusion of the test, and some free convection enhancement of evap- oration and conduction rates may have occurred. TITTABAWASSEE RIVER NEAR MIDLAND, MICHIGAN, 1969 The comparisons of predicted and observed tempera- ONE-DIMENSIONAL MODEL OF EXCESS TEMPERATURE 23 39| I I I I 38 Section 3 36 —- __ __ _ - Section 4 w 2 m 34 — _l m U - Z ui n: 32 _. D l.— E .. g Calculated LIJ 3° _ — — — Observed '— 28 J l l l I Calculated natural temperature Section 7 l I l l l l 25 13 14 16 18 20 22 24 2 4 6 8 TIME, IN HOURS, JUNE 25~26, 1969 FIGURE 19.—Comparison of observed and calculated water temperatures, the Fenholloway River near Foley, Florida, 1969. tures for the two studies of the Tittabawassee River near Midland, Michigan, are shown in figures 27 and 28. Calculations started at section 2 just downstream of the discharge site. The study reach was 11.0 km long in September and 17.2 km in October. The standard value of the mass transfer coefficient was used. Heat load varied somewhat and excess temperatures ranged from 6 to 8°C for the September study and 2.5—4°C for the October study. Air temperatures were substan- tially less than water temperatures for both studies. The predictions for the September study are gener- ally good. Because the one-dimensional model em- ployed here does not account for longitudinal disper- sion, the sharp peak in initial temperature, which ap- peared at the section just below the discharge point at 1030 hours, was propagated downstream by the com- puter solution with no appreciable spreading. The measured values clearly show that the dispersion ac- tually occurred. There is also a slight tendency to over- estimate cooling in spite of the fact that air tempera- tures were 5—10°C less than water temperatures. The predictions for the October study are less satis- factory. The excess temperatures were much smaller but increased by about 60 percent during the study period. The downstream convection of this change ap- pears to have caused some of the error. However, there was indication of a large thermometry error. The ob- served temperatures at section 2 were about 0.5°C less than those observed at section 3 during the early morn- ing hours when the heat load was relatively constant. If the temperature observed at section 3 is assumed to represent the correct temperature at section 2, this would increase the initial excess temperatures by 0.5°C and effectively raise the calculated results at sec— tions 3, 5, and 6 by that amount. Clearly a correction of this magnitude would cause a considerable improve- ment in the agreement of calculated and observed temperatures. NORTH PLATTE RIVER NEAR GLENROCK, WYOMING, 1970 The comparison of predicted and observed tempera- tures for the North Platte River near Glenrock, Wyom- ing, is shown in figure 29. This study was performed in January and is the only data collected during the winter. Calculations were started at section 2 just below the discharge site. The reach was 29.8 km long. The standard value of the mass transfer coefficient was employed. Significant variations in heat loading were 24 THERMAL LOADING OF NATURAL STREAMS 31 T I‘ 30 #1 Section 3 I 29 _.. U) 2 9 Section 4 8 2 I—- E 8 ui I: :3 ’— g 27 — Section 5 LIJ D. 2 Lu ’— ’ ‘ 26 — Section 7 Calculated 25 — - - - — Observed ,, l I I l l I I I I I 14 16 18 20 22 24 2 4 6 8 10 12 TIME, IN HOURS, NOVEMBER 24—25,1959 FIGURE 20.—Comparison of observed and calculated water temperatures, the Fenholloway River near Foley, Florida, 1969, 27 I I I I I I I I 32 U, I I I I I I D 31 —- — m 26 — — Calculated — a ---- Observed -’ Section 7 Calculated 2 3 30 - \ \ I! W 3 2 ~ ~ _ _ _ _ ’ ’ z 8 _. 29 - Observed _. Lu Z n: . D 23 — .. W I- II. < Calculated natural temperature 3 I 27 —w " 33 < E _. 5 E 26 - (L g ,5 I I I I I I E 23 24 2 4 6 8 1o 11 TIME, IN HOURS, JUNE 25—26,1969 20 I I l J I | I I FIGURE 22.—Comparison of observed and calculated water tem- 22 24 2 4 6 8 10 12 14 16 TIME, IN HOURS, MARCH 26—27,1969 FIGURE 21.—Comparison of observed and calculated water tempera- tures, the Fenholloway River near Foley, Florida, 1969. Calcula- tions are modified to account for ground-water accretion. encountered during this study with excess tempera- tures varying from 7.4 to 90°C. The air temperature remained well below freezing throughout the study, ranging from -3.3 to -10.8°C. The river was frozen peratures, the Fenholloway River near Foley, Florida, 1969. Calculations are modified to account for ground-water accre- tion. above the discharge with the flowing water tempera- ture at 0°C. This required the prediction of the hypothetical negative natural temperature as discuss- ed earlier. The prediction is generally quite good. There is a sharp drop at section 3, starting at about 0100 hours, ONE-DIMENSIONAL MODEL OF EXCESS TEMPERATURE 3° I I I I I I I I <0 3 V) _I u.) U E u.I D: 3 '— <1 I I: Section 5A ‘\ z’ 5 \\~-"/ 25 '— _______ —‘ ’— Calculated "T “\\\ --— Observed ‘~____./ 24 I I I | I I I I 19 20 22 24 2 4 6 8 10 12 TIME, IN HOURS, NOVEMBER 24~25, 1969 FIGURE 23.—Comparison of observed and calculated water temper- atures, the Fenholloway River near Foley, Florida, 1969. Calcu- lations are modified to account for ground-water accretion. 17-2 I I I I I I I I I I I w 17 2 U) .1 LLI U E u; 16 D: D '— < I Section2 LLI a. E 15 — Calculated - I- _ —-— Observed 146 I I I I I I I I I I I .1314 16 ‘18 20 22 24 2 4 6 8 TIME, IN HOURS, APRIL 2—3,1969 1011 FIGURE 24.—Comparison of observed and calculated water tempera- tures, the Dan River near Eden, North Carolina, 1969. which does not appear in the data. This was induced in the solution by a drop of 08°C in the excess tempera- ture at the starting point, section 2, occuring between 2400 and 0200 hours. It is possible that the tempera- ture measurement at 0200 hours, which was consider- ably lower than that at 2400 hours and that at 0355 hours, may have resulted from temporary input condi- tions and was not representative of the temperature at section 2 from 0100 to 0300 hours. The prediction would surely have been better had the temperature at 0200 hours at section 2 been 05°C higher. There was also an increase in wind speed from 2 m‘sec‘1 to 4 m - see1 that occurred at about 0100 hours and caused a sharp reduction in the value of T, at section 3. There is some overestimation of cooling, particularly during the early morning hours. During this period, ice formation occurred at the right bank (discharge was 25 27 IIIIIIIIII Calculated - - — Observed 26L 25 — ‘ ‘ — Section 2 24 TEMPERATURE, IN CE LSIUS 23 I l I I I I I I I ‘ 12 14 16 18 20 22 24 2 4 6 8 10 TIME, IN HOURS, MAY 6—7,1969 FIGURE 25.—Comparison of observed and calculated water temper- atures, the Dan River near Eden, North Carolina, 1969. 28 I I I I I I I I I I I Calculated ——— Observed S 27% :7: _I In 0 Z UT 26 _Sec1tIon II D '— < I]: Lu 0. E 25— m '— 24 | I I | I l I L I I I 24 2 4 6 8 10 12 14 16 18 20 22 24 TIME, IN HOURS,SEPTEMBER1647,1969 FIGURE 26.—Comparison of observed and calculated water tempera- tures, the Dan River near Eden, North Carolina, 1969. concentrated on the left bank). This condition ceased somewhere between sections 3 and 4 where mixing supplied enough heat at the right bank to prevent ice formation. There is no quantitative information on the amount of ice that formed. However, any ice formation has the effect of liberating heat to the remainder of the system and would cause observed temperatures to be higher than expected. The melting of ice, which was complete by 1100 hours, would require energy from the rest of the system. This would lower the temperature or, more precisely, slow the rate of increase of tempera- 26 THERMAL LOADING OF NATURAL STREAMS 25— 244 23— TEMPERATURE, IN CELSIUS — Calculated — __ Observed 21IIIIIII 81012141618202224246 TIME, IN HOURS,SEPTEMBER 18—19, 1969 FIGURE 27.——Comparison of observed and calculated water temper- atures, the Tittabawassee River near Midland, Michigan, 1969. 11 I I l I I I I l I I I (I) 2 U) _I LIJ U 10 Section5 _ Z _. Section 3 u.I ‘1 Section 5 D ‘5: u: 9— LU a. E E -- Calculated “— Observed , SectlonG 8 I I I I I I I I I I I 16 18 20 22 24 2 4 6 8 10 12 14 TIME, IN HOURS, OCTOBER 28—29,1969 FIGURE 28.—Comparison of observed and calculated water tempera- tures, the Tittabawasee River near Midland, Michigan, 1969. 4-5 I I I I I T T I I I I 4 —l o) 2 In _ _I uJ U z 3 “ ui _ D: D ’— < 2 _ 0: Lu 0. _ 2 DJ I— — Calculated 1 ‘ --- Observed \ ,’ ‘ .’ Section 6 0.5 I I I I | I AL I I 18 20 22 24 2 4 6 8 10 12 TIME, IN HOURS, JANUARY 28—29,1970 14 16 FIGURE 29,—Comparison of observed and calculated water tempera- tures, the North Platte River near Glenrock, Wyoming, 1970. ture in the postdawn hours, and thereafter prediction and observation should agree. Excepting this ice for- mation, the estimate of cooling appears to have been quite good. WEST BRANCH OF THE SUSQUEHANNA RIVER NEAR SHAWVILLE, PENNSYLVANIA, 1962 The comparison of predicted and observed tempera- tures for the West Branch of the Susquehanna River near Shawville, Pennsylvania, is shown in figure 30. Calculations were started at section 0 just below the discharge site. The total reach length was 24.1 km. The standard value for the mass transfer coefficient was used. Excess temperatures were quite high, ranging from 12 to 13.5°C at section 0. Air temperatures were less than water temperatures at all times. The prediction is generally quite good and, in par- ticular, the average dissipation seems to have been es— timated quite well. Nevertheless, because of the high excess temperature, a problem similar to that encoun— tered during the November study on the Fenholloway River can be noted. A sharp decrease in wind speed at about 1800 hours resulted in such a great rate of in- crease in the excess temperature that the predicted total temperature started rising in spite of decreasing natural temperature. After the low windspeed had per- sisted for a period equal to a travel time, all the heated water packets arriving at section 1 had experienced only low windspeeds and excess temperature stabiliz- ed, and the total temperature decreased owing to de- creasing natural temperature. Unlike the November study on the Fenholloway River, there is no appearance that evaporation and conduction have been underestimated using standard mass transfer expressions. In light of the highly unsta- ble conditions that existed, this is difficult to understand. CONCLUSIONS REGARDING ONE-DIMENSIONAL MODELING For most of the sites studied, the one-dimensional, nondispersive Lagrangian model would appear to be adequate for downstream temperature prediction. Rapid changes in heat loading rates may induce lon- gitudinal temperature gradients so sharp that disper- sion must be considered. This is evident in the results for the September study on the Tittabawassee River. In most cases the conventional mass transfer formu- lation of equation 14 appears adequate using a mass transfer coefficient of 2.68108 kg-m-l' newton'l. In the case of the Fenholloway River, this value is appar- ently too small. This is not entirely surprising in light of the area dependence of the mass transfer coefficient as given by equation 15 and the fact that the Fenhol- TWO-DIMENSIONAL MODEL OF EXCESS TEMPERATURE w 2 U) _I u.I ’ _ o — Section 2 SectIon 3 _____ Z 22 e uI _ D: 20 — —— Calculated I? _ ---- Observed < ~‘ E /,—‘\ Section 5 ‘ a 16 ~\\ // \\\\ _ I— — ‘-—’ ~__ >\ _ 14 Observed natural temperature \\ _ 12 I I I I | I I I I I I \ 8 10 12 14 16 18 2O 22 24 2 4 6 TIME, IN HOURS, OCTOBER 17—18,1962 FIGURE 30.—Comparison of observed and calculated water tempera- tures, west branch of the Susquehanna River near Shawville, Pennsylvania, 1962. loway River was by far the smallest stream studied. It would also appear that the predicted mass transfer is too sensitive to changes in wind velocity where unsta- ble atmospheric conditions prevail. This is particularly evident in the results from the Fenholloway River and the West Branch of the Susquehanna River, where very large excess temperatures made the sensitivity of heat exchange to changes in wind velocity obvious. The calculations strongly support the conclusion, evident from a consideration of the data, that very long reaches are required to dissipate the majority of the waste heat added to a stream. This is particularly true if the excess temperature at the point of discharge is low—a requirement now imposed by most states on waste heat discharges. Only in the cases of the Fenhol- loway River and the West Branch of the Susquehanna River were heat losses at the end of the study reach greater than 50 percent of the heat added. In both cases the excess temperatures were far above those now permitted by most states. The above conclusion suggests that calculations such as proposed here may be necessary in the operation and siting of power plants downstream of existing waste heat sources. If regulations are designed to per- mit excess temperature no greater than some specified figure, a downstream facility should not be permitted that full figure if an upstream discharge exists. But since neither the natural nor the excess temperature can be measured immediately above the downstream facility, either the natural temperature or the excess temperature at that point will require calculation. If the stream thermal regime is reasonably homogene- ous, the excess temperature is the clear choice. 27 TWO-DIMENSIONAL MODEL OF EXCESS TEMPERATURE MODEL FOR A STEADY UNIFORM CHANNEL The two-dimensional equation for thermal balance in a natural water body was derived previously as equation 4. In order to develop a working model for natural streams, it must be simplified as for the one- dimensional models. The source-sink term, S, and the bottom heat exchange term, Hm, are neglected. The flow is assumed to be steady uniform so that the depth-averaged transverse velocity, 11,, is nonexistent and Y and 1),, are sole functions of 2. The longitudinal dispersion term will be neglected on the grounds dis- cussed previously. All these assumptions reduce equa- tion 4 to BT H 262 ”peg“ In equation 42, T is the depth—averaged water tem- perature, U, the depth-averaged longitudinal velocity, and K, is the depth-averaged thermal dispersion coeffi- cient. The symbol H is the surface heat influx and Y is the depth. Following the same approach as in the one-dimen- sional model, the excess temperature T, is defined as (T—Tn), where T" is the temperature that would occur under the same meteorologic and hydraulic conditions if there was no waste heat discharge. The surface influx will be linearized by means of equations 24, 25, and 26. The resulting equationT for Te is UyT 3T 6T 1 6 6—t+v xii—x _ 7&(YK (42) 6T9 +1) 6—71 = Tt 3‘ 6x 16—(YKZa— g7 )— Y 62 (43) Equation 43 forms a boundary value problem in a steady uniform flow when combined with proper boundary conditions. Because of the dependence of v,, Y, and K, on 2, however, a closed form solution is not obtainable for equation 43, and it must be solved nu- merically. A simpler approach to equation 43 is possible if the boundary condition is simplified such that waste wa- ter discharge and excess temperature are assumed to be steady at the discharge site. It consists of combining the one-dimensional exponential decay model for sur- face heat dissipation with the two-dimensional steady-state diffusion model developed by Yotsukura and Cobb (1972). Notice that if the excess heat were conservative so that U, =0, equation 43 is reduced to a steady-state equation (44) 28 Suppose that T’e, the hypothetical excess temperature without decay, satisfies equation 44. Because the heat decay due to surface dissipation is a slow process rela- tive to the heat transport by convection and diffusion, its parameter U,/Y may be approximated by U,/{Y}, which is a function of time alone if {Y}, the cross— sectional average depth, is constant. Accordingly a so- lution to equation 43 is proposed as Te(x,2,t)=T’e (x,2)exp (— I: L dt'). (45) 0 {Y} Equation 45 can be shown to satisfy equation 43 if the heat decay term is replaced by Ufa/{Y} in equation 43. Even though equation 45 is derived conceptually from the form of product solutions frequently used in the diffusion of reacting solutes (Crank, 1967), it could also be viewed as a combination of an Eulerian solu- tion to diffusion and a Lagrangian solution to dissipa- tion. It is similar to the approach used by Yeh, Verma, and Lai (1973) in the prediction of excess temperature for a cooling pond. Equation 45 should be considered as an approximation to What is physically going on rather than as a rigorous solution to a boundary value prob- lem. The decay term in equation 45 will be calculated in terms of travel time in the same Lagrangian sense as in the one-dimensional model. As for steady—state solutions to equation 44, the ap- proach used by Yotsukura and Cobb (1972) consists of transforming equation 44 into aT'e _ 3 2 (97” ax — 50(21):), w), (46) where q is the partial cumulative discharge measured from one side of a cross section and is defined by q=f:u,Y dz’. (47) Note that q=Q when the integration in equation 47 is carried over the entire cross section. Equation 46, in which q replaces 2 as the independent variable, is sim- pler to handle than equation 44, because the convec- tive velocity is reduced to unity. The new coefficient in the dispersive term of equation 46, szxY2, may be ap- proximated by a constant cross-sectional average value, {KZUXYZ}, so that the following approximate so- lution is given to equation 46: 00 2 ' ’ _ l a(q52'+2n+5jq') T e(a!q )_ 2 Tea 2 2 erf Vi n=0 J=1 a(g '3] + 2n + 6ft] ') V5 —erf THERMAL LOADING OF NATURAL STREAMS 0° 2 a(q’,2—2n+8,q’) + E 2 erf \/§ n=1 j=1 ah]; -2n+5jq’> V2— (48) —erf The symbol Tea is the steady excess temperature at the initial cross section, Where Q5 is the steady excess water discharge rate and pCstTeo represents an inflow rate of excess heat. Summation with respect to index j is defined by 61=+1 and 82=—1. The symbol a is a nondimensional longitudinal distance parameter, a=vQ2/2x{K,u,Y2} , while (1’, fractional cumulative discharge q/Q, repre- sents a transverse position. The symbols q’s1 and q’,2 are fractions qsl/Q and qsg/Q respectively, where qS1 and qs2 represent the two ends of a part of streamflow con- taminated by excess temperature at the initial section and q32 _qsl =Qs- By means of field tracer tests Yotsukura and Cobb (1972) showed that equation 48 is quite satisfactory for a number of straight uniform channels and is also usa- ble in a moderately meandering reach of the Missouri River. (49) MODEL FOR A STEADY NATURAL STREAM In applying equation 45 to a steady natural stream, it is necessary to subdivide a study reach into a set of steady uniform channels. A river system is visualized first as a flow system consisting of a fixed number of stream tubes with equal subdischarge Aq. This is shown in figure 31. Note that all solid lines represent impervious boundaries through which no convection and diffusion take place. The system is then divided into a set of subreaches depending on the uniformity of cross-sectional properties. A subreach may further be divided into subchannels depending on the location of tributaries and islands as illustrated in figure 31. In the stream-tube flow system longitudinal distance re- mains the same as in the real river system. Solutions to “conservative” excess temperature, T 'e, are obtained for successive downstream subreaches starting at the waste heat discharge site. For each un- iform subchannel, an upstream T’e is treated as the initial temperature, Tea, and downstream T’e is calcu— lated by means of equation 48. In most applications, the upstream temperature tends to be different from one stream tube to another so that a superposition of solutions is required to obtain the downstream T’e re- sulting from all upstream temperatures. This proce- dure is justified because all differential equations in- volved are linear equations. TWO-DIMENSIONAL MODEL OF EXCESS TEMPERATURE STREAM-TUBE FLOW SYSTEM Waste ‘ load I X REAL RIVER SYSTEM Waste discharge , , ‘ section 44c: Sections l defining hydraulically +4 . uniform ‘ ' subreaches ‘ l it a l l . is 1‘ +7" 17? _ . l —>AQ<—l ‘ l 77/. Observation section—>7 T4—i i—T" 4* V . l m l O=7Aq Q=7Aq FIGURE 31.—Sketch of a two-dimensional natural stream and its stream-tube flow system. Once the steady-state T ’e is found for the entire reach, the calculation of exponential decay is initiated by choosing a set of times, t, at which excess tempera- ture is desired at a downstream cross section. The travel time, 7, between the discharge site and the de- sired cross section and the initial time, to, are then determined, and the exponential loss factor is calcu- lated, as in the one-dimensional model. The cross- sectional loss factor thus obtained is applied to T'e ac- cording to equation 45 to obtain actual excess tempera- ture T,(t) at various longitudinal and transverse posi- tions. This may be added to natural temperature, T”, to obtain actual temperature T. In field problems, the ex- cess temperature at a discharge site tends to vary with time even if waste water discharge rates remain steady. This time variation in excess temperature at the discharge site may be accounted for in Lagrangian decay calculations for field applications. As stated previously, the above model is not pro- posed as an exact solution to a boundary value problem but as an adequate simulation of physical phenomena without burdening oneself with time-consuming nu- merical solutions to the original set of equations. The validity of this approach can be assessed by comparing the model calculations with observed data. APPLICATIONS TO FIELD DATA Temperature regimes observed near Dickerson Power plant on the Potomac River were ideal for test- ing the two-dimensional model. Figure 3 gives a sketch of the study reach. The surveys conducted in March and May 1969 showed that the heat plume did not 29 reach the right bank for a downstream distance of 35 km and the transverse variation of excess tempera— ture was quite pronounced throughout the study reach. In the November study, on the other hand, the heated effluent discharge was 36 percent of the total discharge and the heat plume was observed to reach the right bank at about 1.3 km downstream. A tremendous in- crease in effective transverse mixing was caused by density stratification, jet effect of effluent water as well as low stream velocity. One of the major tasks in the analysis was to assess the magnitude of dispersion coefficient, {K2}, to be used in thermal diffusion. Two dye tracer studies conducted at the time of the March survey turned out to be useful in this connection. The first study was for the purpose of determining the traveltime in the left channel (3.64 km long) along Mason Island using 2 kg of rhoda- mine WT dye, which was injected as an instantaneous line source. The observed dye velocity was 0.46 m-sec'1 and about 10 percent smaller than 0.51 m-seC’1 esti- mated from gaging data at sections 4, 5, and 6. The purpose of the second dye study was to tag heated water from the power plant. A rhodamine WT 20 percent solution was continuously injected at a rate of 4.2 g-sec‘1 into an upstream area of the effluent canal for about 5.5 hours. The mixing in the canal was intense enough to produce a uniform concentration of 11.2 micrograms per liter in the canal at its outlet (section 2) to the Potomac River. Figure 32 shows transverse distributions of both WT dye concentration and Te, which were concurrently observed at several downstream sections. Both coordinates are nondimen- sionalized by arbitrary scales for easier comparison. Good overall agreement between the diffusion of dye and excess temperature is indicated in figure 32. When the dye data were compared with Te observed at other times, it was found that such Te’s were very much in- fluenced by the magnitude of time-dependent initial excess temperature as well as natural temperature. Therefore, the initial assessment of the dispersion co- efficient was based solely on the distribution of rho- damine WT concentration. For the simulation of March data, the reach between the discharge site (section 2) and White’s Ferry site (section 7) was divided into six subreaches, of which the middle three subreaches had two subchannels separated by Mason Island. The total discharge, Q=148.5 m3sec'l, was divided into 48 stream tubes with Aq=3.09 m3seC‘1. At the effluent site, 4 tubes on the left bank side were taken to be uniformly contami- nated by dye and excess heat (QS=12.36 m3sec'1). After several trial calculations by equation 48, an optimal value for the transverse dispersion coefficient was found to be 30 {KZ}=0.52{Y}{V.}, (50) where {V} is the average shear velocity. The constant 0.52 is larger than 0.08~0.26, a range commonly ac— cepted in straight laboratory fiumes and small chan- nels (Prych, 1970; Fischer, 1973) but is close to 0.72 for the Columbia River (Glover, 1964) or 0.6 for the Mis— souri River (Yotsukura and others, 1970). For a slowly meandering stream such as the Potomac, the constant of 0.52 is a reasonable value. Table 2 contains some hydraulic properties related to the dispersion coeffi- cient. The calculation of the heat decay was discussed in detail in the previous sections and is not repeated here. Table 2 contains some hydraulic properties related to 1.0 I I I Section 3 0-8 _ (1.3 km from source) _ 0') Lu 3 0.6 — _ _J < > 0.4 -— — LIJ O D: 0.2 — 3 8 1.0 O 0 I- Section 4 I: — (2.3 km from source) 0.8 ,2 <[ _ 0.6 _I LIJ “5 _ 0.4 LU II I2 _ / — 0.2 E Lu 1.0 0 o. . 2 Section 6 E 0.8 ._ (6.1 km from source) _ ‘63 LIJ 0.6 — O X L” 0.4 — II 0 z 0.2 ~— 9 '— < 0 E Section 7 Z _ (7.7 km from source) - 0.8 m U 2 _ — . O 0 6 U m — > _ 04 O ._ —- 0.2 —-l’/ I I l 0 0.5 0.6 0.7 0.8 0.9 1.0 RELATIVE TRANSVERSE DISTANCE, Z/W FIGURE 32.—Concurrent transverse distributions of dye con— centration and excess temperature, the Potomac River below Dickerson Power Plant, Maryland, 1969. THERMAL LOADING OF NATURAL STREAMS TABLE 2.—Cross-sectional average hydraulic parameters, the Potomac River below Dickerson Power Plant, Maryland, March 1969 _ x _ [v.1 m [ow] [v.1 [K21 Axum (in m) (m m-sec'” (in m) (in ma-sec'l) (in m-sec") (in m2>sec‘1) (in hours) 2424 0.365 1.07 0.724 0.040 0.0223 1.84 5197 0.58 0.73 0.345 0.033 0.0125 1.33 6061 0.29 1.34 0.765 0.044 0.0304 0.83 7637 0.30 1.74 1.127 0.051 0.0458 1.46 traveltime, 7-, and figure 33 shows the variation of sur- face dissipation coefficient, U.., and initial excess tem- perature Tea with time. In calculating U.., two sets of meteorologic data obtained at section 4 and section 7 were averaged, because the difference in U. by the use of individual set was insignificant. Natural tempera- ture observed near the left bank of section 1 was used as Tb in equation 25. The mass transfer coefficient of evaporation, N, was fixed at 1.9 - 108 kg-m'1 newton-1. As the reach was short and the heat loss was small in the March study, no intensive effort was made to as- sess the sensitivity of the calculations to the mass transfer coefficient. In the March study, the effect of transverse stratifi- cation was not observed even at the first survey sec- tion, section 3, which was 1.3 km downstream from the discharge site. The effluent canal joins the Potomac River at an angle less than 12°, and velocity in the canal was about the same as that of the Potomac in the March survey. Therefore, the jet effect of effluent dis- charge was absent. Large transverse temperature gra- dients and, thus, density gradients near the discharge site must have caused transverse mixing much more intense than that due to turbulence alone. The effect EXCESS TEMPERATURE 0.025 IlIIlIl 0.020 COEFFICIENT, IN M/HR AT SOURCE, IN 0C I I I r SURFACE DISSIPATION IIIIIIII l J | I l | 400 800 March 12, 1969 0.015 ' I 1 1 1200 1600 2000 March 11, 1969 1 2400 1 200 TIM E FIGURE 33.——Temporal variations of surface heat dissipation coeffi- cient and source excess temperature, the Potomac River below Dickerson Power Plant, Maryland, March 1969. TWO-DIMENSIONAL MODEL OF EXCESS TEMPERATURE apparently was limited to a region very close to section 2 and could not be detected at section 3, where the ration, x/{Y}, was on the order of 1300. The calculated T, was added to Tn which was ob— tained by the one-dimensional analysis using equation 39. The calculated temperature, T, is plotted against transverse position expressed by cumulative dis- charge, q’, at sections 3, 4 and 7 at selected times in figure 34. Also plotted on these figures are the observed temperatures. The agreement between the predicted and observed temperature is satisfactory. One obvious reason is that the total heat loss through the air-water interface was small, amounting to only 10 percent of 31 the initial excess heat at section 7, the furthest station. Thus, the transverse distribution was determined pre- dominantly by diffusion. Another reason for good agreement is that high wind speed and, thus, high val- ues of Ur induced rather uniform dissipation in the transverse direction, as assumed in the model. Some of the discrepancies between the observed and calculated temperatures are believed to be due to the transverse variation of T,,, which is not accounted for in the model. As mentioned earlier, the data at section 1 showed that such variations amounted to 1°C occasionally. Analysis of the May data was more involved than the March data, because the length of the study reach TIME TIME 11 TIME 17 1 15 11 l l T T Mar. 11,1969 16 (1,3 km 16 1340 14 (213 km 14 10 — (7.7 km downstream ——- 10 downstream Mar 11 1969 downstream from heat source) . 2040 from heat ' ’ from heat Mar. 11,1969 15 source) 15 13 source) 13 9 14 17 12 15 10 1730 2005 13 1 11 14 2240 12 15 10 13 05 30 M .12,1969 11 14 8' 9 14 0325 Mar. 12, 1969 (010 13 m 8 8 3 3 0605 a 5, Mar. 12,1969 —I 9 12 —I 7 12 LIJ m U) U Q 2 (n E 8 11 3 6 11 ii i ‘ 0735 u.1 Lu [E a: 0 3 D 2 ,_ 7 10 |_ 5 10 _~ < <( ui I n; [I u: m 6 9 if 4 9 a 2 2 4 1150 1‘3 “J I 5 e *— 3 8 E E u: 4 7 5 7 I- 3 6 4 6 5 —— 5 3 — 5 4 — 4 2 4 3 — 3 3 —3 2 2 2 — 2 1 1 I I 1 I I I I 1 1 0.6 0.7 0.8 0.9 1 1.0 RELATIVE CUMULATIVE DISCHARGE, q/Q FIGURE 34,—Comparison of observed and calculated transverse temperature distributions, the Potomac River below Dickerson Power Plant, Maryland, March 1969. 32 THERMAL LOADING OF NATURAL STREAMS was extended to 29 km to cover the Seneca area. One major problem with the May data was to extrapolate meteorologic and thermal-effluent source data for about 11/2 days prior to the start of the actual thermal- survey period, May 14—15, 1969. This was necessary for Lagrangian heat-decay calculations, because the average travel time of a packet of water from the plant site to Seneca was 29.8 hours. The water, whose tem- perature was observed at Seneca at 1300 hours, May 14, had left the plant site at 0710 hours, May 13, and had been undergoing surface heat exchange for an appreci- able period of time, which was not covered by the 24 hour survey starting at 1200 hours, May 14. The extrapolation of T, and T, data was accom- plished by use of the temperature record kept by the Dickerson Power Plant for the condenser units No. 2 and 3. The extrapolation of T, was relatively easy, be- cause T, observed at the effluent canal was very close to an average of exit temperatures recorded at units 2 and 3 at the corresponding time. On the other hand, intake temperatures recorded at condenser units 2 and 3 indicated some stratification effects, because they were about 2°C lower than the observed natural tem- perature in the Potomac at section 1. The extrapolations of T,, and T,0 =Ts—Tn were not very satisfactory, proba- bly containing an error of 115°C. The meteorologic data were extrapolated by compar- ing observed data at the river site with those recorded at the National Weather Service Station at Dulles In- ternational Airport, Virginia, which is about 40 km west of the study site. Partly because of this need for extrapolation, the two sets of meteorologic data ob- tained at sections 2 and 9 were again averaged to pro- duce a single time-series set to be used for the entire reach. The averaging and extrapolation is believed to have a relatively small effect on the values of U,. The river discharges in May were only 10 percent less than that in the March survey period. It was ex- pected that diffusion factors would remain similar for the subreaches between sections 2 and 7. After several trials, it was decided that the transverse diffusion fac- tors for the May regime be kept the same as for the March regime (table 2). Taking account of observed local depth and velocity, this amounted to using the constant in equation 50 ranging from 0.56 for section 3 to 0.65 for section 7. A constant of 0.65 was used for the lower subreaches that included sections 9 and 10. The total discharge, Q = 133.2 m3 -sec-1, was divided into 40 stream tubes, of which 4 tubes near the left bank were assigned for thermal effluents with uniform excess temperature (Q,=13.32 m3'sec '1). The calculation of surface heat dissipation was based on U. values shown in figure 35. These values, ranging from 0.009 to 0.018 m -hour-1, are much lower than the I l—U " I V V E (C 85 _T I Assulmed frdm l T Y , ' EL” 2 8-0 ' extrapolated data a: c 3 W s a 2g 7.5 LIJ O I D 7.0 ' $38 6.5 l l l I l l .1“ l l l | ZlIoozo , ‘ VT" , , , , 9 E > ‘ YAsslimed ffom l; E : extrapolated data E 3 0.015 — in ~ * ‘2; : 0 Lu k g g 0.010 C < LL LL u. CE UJ 0 005 l J__L | l l I l l l mu l l l 3 O ' 600 2 ll) 0 1200 1800 2400 600 1200 1800 2400 1 00 May 13, 1969 May 14,1969 May 15, 1969 TIME FIGURE 35.—Temporal variations of surface heat dissipation coeffi- cient and source excess temperature, the Potomac River below Dickerson Power Plant, Maryland, May 1969. March values and are obviously due to lower wind speed in the May period. The calculated temperature, T, is compared with the observed temperature in figures 36 and 37 in the man- ner similar to figure 34. Despite large amounts of ex- trapolated input data, the agreement in temperature distributions at sections 4, 9, and 10 are satisfactory. Some discrepancies observed at sections 4 and 9 may be attributed to either local weather conditions, which were not represented in the U, values, or T, variations in the transverse direction. These discrepancies are ob- served late at night or early in the morning. The calcu- lated Te at Seneca, section 10, shows that the heated plume, T920.05°C, occupies 70 percent of total dis- charge, leaving 30 percent of the discharge near the right bank uncontaminated. According to the calcula- tions, the total heat loss at Seneca is only 25 percent of the initial excess heat after 29.8 hours of travel from the discharge site. The results at White’s Ferry, section 7, are not very satisfactory. The calculations overestimate left bank temperature and underestimate midstream tempera- ture between q/Q=0.55 and 0.75. The peculiar midstream distribution with a hump is not due to tur- bulent diffusion, which dictates a smoother transverse distribution. A more likely cause is either transverse variation of T, or local transverse convection by sec- ondary flows. Since the March and October data do not show such a pronounced hump, the effect of secondary flow is likely to be less than that of Tn variation. On the other hand, the variation of T,, was small in the May data according to observations at section 1. The source of this error at section 7 could not be assessed in further detail. The October data for the Potomac River illustrate an TWO-DIMENSIONAL MODEL OF EXCESS TEMPERATURE 33 TIME 26 I I 26 TIME 23 23 1205 25 — (2.3 km —- 25 22 — (7.7 km downstream from heat source) 22 M 14 1969 downstream av ' from heat 24 — source) —le4 104° 21 —— 21 May 14,1969 23 20 — 24 1600 22 19 — . 23 . . 21 1800 18 ° 22 Calculated 20 2000 19 18 2400 m tn 2 17 2 m (/1 d d 2400 U 21 U 2 2 Lu‘ Lu~ I 20 II 3 a 1. < < a: 19 I E 0600 fig 2 May 15 1969 1 , 1”.” 2 E 0415 May 15, 1969 20 19 1000 18 0815 20 19 18 P— 13 20 F 20 19 '— 19 18 ' . . . . —‘ 18 18 l I 18 17 l 1 17 0.7 0.8 0.9 1.0 0.0 0.2 0.4 0.6 0.8 1.0 RELATIVE CUMULATIVE DISCHARGE, q/Q RELATIVE CUMULATIVE DISCHARGE, q/Q FIGURE 36.—Comparison of observed and calculated transverse temperature distributions, the Potomac River below Dickerson Power plant, Maryland, May 1969. interesting effect of jet and transverse density currents of heated effluents. The Potomac flow was much lower in October relative to those in March and May. The effluent discharge of 18 m3-sec‘1 was flowing into the ambient Potomac discharge of 32 m3-sec'1. The ratio, Qs/Q, was 36 percent. Furthermore, the effluent veloc- ity of 0.90 m‘seC'1 was high in comparison to the am- bient Potomac velocity of 0.2 m-seC'l. The combined 34 TIME 21 21 I I I I 1104 —(15.5 km downstream 20 May 14' 1969 from heat source) 22 1710 21 22 21 2105 22 to 2 . o 3 calculatEd 21 0300 Lu May 15, 1969 0 20 E o 0 ° . . . . 21 1055 Lu 0: 3 20 '3 a: 19 LL! 0. E .8 I I I I .8 E 22 I I I I 22 2310 21 —(28.8 km downstream 21 May 14’ 1969 from heat source) . . o o 20 . o o ' ' _ 21 0505 20 May 15,1969 2° 0910 19 18 0 0.2 0.4 0.6 0.8 1.0 RELATIVE CUMULATIVE DISCHARGE Q/O FIGURE 37.—Comparison of observed and calculated transverse temperature distributions, the Potomac River below Dickerson Power Plant, Maryland, May 1969. effect of jet and transverse convection due to density currents was observable at section 3, where the verti- cal temperature difference of more than 0.5°C existed between 40 and 20 m from the right bank. The depth in this area was 1.4—2.2 m. The total heated plume width at section 3 was on the order of 200 m in contrast to the width of 70 m in the March study, and the right edge of the plume was characterized by a 20-m-wide band of vertical nonuniformity in temperature. Even though the present model assumes that the depth-averaged transverse velocity, 02, is zero every- where, some transverse convective effects due to sec- ondary flows or density currents are accommodated in the equation in terms of effective dispersion coefficient, K2. The model can not accommodate the convective ef- fects of jets or very strong secondary currents. An effort was made to simulate the distribution at section 3 by increasing the diffusion factor until the calculation matches with the data. The best simulation is shown in figure 38. In this calculation, the constant of the dis- persion coefficient, equation 50, was increased to 11.0 THERMAL LOADING OF NATURAL STREAMS 18 18 TIME I I I I ' 17 >— (1.3 km downstream from heat source) —- 17 16 1505 15 Oct.29, 1969 14 13 2105 12 U, 11 2 (I) 10 _I m U 9 .2. 11511 0310 o: Oct.30, 1969 D 10 3 ’2 D: 9 uJ n. 2 8 15 E 10 14 1105 9 13 8 12 7 11 10 10 9 9 o 8 ~ — s 7 - ' 1 1 1 1 7 0.0 0.2 0.4 0.6 0.8 1.0 RELATIVE CUMULATIVE DISCHARGE q/O FIGURE 38.—Comparison of observed and calculated transverse temperature distributions, the Potomac River below Dickerson Power Plant, Maryland, October 1969. from 0.52. On the other hand, a constant of 0.65 could be used in all other downstream sections as expected. The results of calculations for lower subreaches are shown in figures 39 and 40. In the October study as in the May study, meteor— ologic and effluent temperature data were extrapolated by comparing the survey data with those recorded at Dickerson Power Plant and Dulles weather station. Because the travel time to a downstream section was much longer than that for the May study, being 11.2, 25.7, and 59.7 hours for sections 7, 9, and 10 respec- tively, almost half of the calculations for section 7 and all calculations for section 9 used some parts of ex- trapolated data. The calculation at section 10 was not tried. The heat dissipation coefficient, U“ varied from 0.007 to 0.02 m 'hour'l. This range is rather similar to that of the May study, indicating the predominant ef- fect of wind speed on U.. The simulation results shown in figures 38, 39, and 40 show that the present model TWO-DIMENSIONAL MODEL OF EXCESS TEMPERATURE 35 15 15 TIME I I I l | I I | I 14 — (7.7 km downstream from heat source) _ 14 1200 ‘3 Oct. 29, 1969 12 14 1600 13 15 2005 U) 2 m .1 “J 0 Z 0005 _. Oct. 30, 1969 Lu I D E D: 0405 w a. E u.I I— 0805 1215 8 8. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 RELATIVE CUMULATIVE DISCHARGE q/O. FIGURE 39.—Comparison of observed and calculated transverse temperature distributions, the Potomac River below Dickerson Power Plant, Maryland, October 19 with more adequate weather data and with current knowledge on dispersion coefficient can predict heated plume configurations with acceptable accuracy. As for the effects of transverse density currents, it is worth mentioning the results of flume study reported by Prych (1970). He studied the transverse diffusion of excess heat in a rectangular flume by releasing a verti- cal line heat source at ambient velocity. It is reported that the total transverse variance, 01,2, could be consid- ered as a sum of 0,2 due to regular turbulence diffusiv- ity and 0;.2 due to density currents. At a distance on the order ofx/{Y}=2{vx} {Y}/ {K2}, 0'52 stopped increasing because of reduced density currents, and the increase of 01,2 thereafter was entirely due to 0,2. The prototype distance x corresponding to Prych’s criterion is on the order of 30—50 m for the October data and is much smaller than x=1330 m at section 3. Be- cause the heat source on the Potomac River was a block TIME 14 14 1245 (15.5 km downstream Oct. 29 1969 13 — from heat source) 13 ' 12 1535 11 — 10 12 2035 a) 2 11 3 11 ' LIJ U 2 1o — 0040 _‘ Oct. 30, 1969 m 9 — 0: D I; 10 II 0440 UJ 9 n. E ,_ 0830 1025 —»- 9 o 0.2 0.4 0.6 0.3 1.0 RELATIVE CUMULATIVE DISCHARGE q/Q FIGURE 40.—Comparison observed and calculated transverse tem- perature distributions, the Potomac River below Dickerson Power Plant, Maryland, October 1969. source rather than the line source used in the Prych’s study, the effect of density currents must have per- sisted for a much longer distance than 30 or 50 In. Similar effects must have been existent in the March and May data but were not observed at section 3. The observed pattern of density effects appears to agree with Prych’s laboratory study. The two-dimensional model was also applied to the Dan River and the North Platte River data. In both rivers the transverse nonuniformity of excess tempera- ture was significant only in one downstream subreach. Because the river alinement was quite sinuous in both rivers, the temperature distributions at other sub- reaches were all uniform in the transverse direction. The one-dimensional analysis of these subreaches was discussed previously. Figures 41 and 42 present the simulations for one subreach in the Dan River and the North Platte River, respectively. The heat decay in these calculations was insignificant because the sub- reach length was small. The observed temperature dis— tribution at the first cross section was used as the ini- tial distribution in the calculation, because the effluent discharge condition was not compatible with the 36 THERMAL LOADING OF NATURAL STREAMS TIME 19 l l l 19 18 (4.2 km downstream from heat source) 18 ‘7 1310 April 2, 1969 19 1709 22 00 0040 April 3, 1969 0537 TEMPERATURE, IN CELSIUS 0812 1022 14 l | l l 1 o 0.2 0.4 0.6 0.8 1.0 RELATIVE CUMULATIVE DISCHARGE, q/O. 4 FIGURE 41.——Comparison of observed and calculated transverse temperature distributions, the Dan River near Eden, North Carolina, April 1969. model. Because of these reasons, figures 41 and 42 are considered as verifying the diffusion aspect of the model. A constant of 0.65 in equation 50 was used for both simulations. SUMMARY REGARDING TWO-DIMENSIONAL MODEL Ideally, a two-dimensional model for excess temper- ature should be combined with a two-dimensional model for natural temperature. The latter model was not developed in the present analysis, and thus some discrepancies between observed and calculated tem- peratures could not be resolved for the assessment of the two-dimensional model. A typical example is the low temperature near the banks as observed in figure 40. Nevertheless, the overall performance of the present model appears to confirm that the combination of the steady-state diffusion solution and the Lagrangian decay solution is a feasible approach. In this connec- tion, it is worth noting that an attempt was made in the early stage of analysis to solve equation 43 directly by Eulerian numerical approximations. It was aban- doned because the calculations were much more time consuming and inefficient in terms of solution ac- TIME 9 I I I 8 _.(3.1 km downstream from heat source) 8 1620 Jan. 28, 1970 7 2225 w 2 (I) _I III U 2 _‘ 0405 3:1 — 4 Jan. 29, 1970 D '— < LI LU E u.I 0805 ’— 1215 . | l I l 1 o 0.2 0.4 0.6 0.8 1.0 RELATIVE CUMULATIVE DISCHARGE q/O. FIGURE 42.——Comparison of observed and calculated transverse temperature distributions, the North Platte River near Glen- rock, Wyoming, January 1970. curacies than the above-discussed model. A substantial amount of data extrapolation was re- quired for the two-dimensional analysis. This was caused by the incompatibility between the data collec- tion scheme and the model development as noted pre- viously. If thermal data are collected for the purpose of applying the present Lagrangian model, it is clear that the reach length and the travel time should be of prime considerations in planning such a data collection scheme. It is apparent that verification concerning the trans— port phase is more positive than that for the decay phase. A comparison of the present findings with other studies shows that there is little doubt that the trans- verse heat dispersion coefficient, K2, in the absence of abrupt temperature changes, is of the same order of magnitude as the mass dispersion coefficient for neu- trally buoyant solutes. Even though the formulation of SUMMARY AND CONCLUSIONS equation 42 assumes no depth-averaged transverse velocity, the weak convective effect due to secondary or density currents is accommodated in terms of disper- sion coefficient, K2. Furthermore, by means of empiri- cal adjustment of K2, some pronounced convective ef— fects could also be simulated by the model as shown in figure 38. The analysis of the surface-heat decay in the present section is rather limited, as it was evaluated only for the Potomac reach, where heat losses were not sig- nificant at most cross sections. Comparing a total heat loss of 25 percent, calculated for the section at Seneca by the two-dimensional model, with an observed value of 35 percent of figure 11, the magnitudes of both the dissipation coefficient, U,, and the mass-transfer coeffi- cient for evaporation, N, used in the present section may have been too low. The decay aspect has been discussed more extensively in the section on the one- dimensional model of excess temperature. SUMMARY AND CONCLUSIONS As explained in the “Introduction,” development of the present mathematical models was based on the recognition that earlier models, such as the energy- budget equation, the equilibrium temperature, and the steady-state exponential decay law, are not suitable in analyzing short-term phenomena in a flowing natural stream. In concluding the report, therefore, it is appro- priate to summarize pertinent features of the present mathematical models. In order to approach short-term water temperature problems in a thermally loaded stream, the excess temperature, Te, is defined as a difference between ac- tual water temperature, T, and natural temperature, T", which would exist in the stream under the same hydraulic and meteorologic conditions but without heat discharges. One major feature of the present ap- proach is to solve the heat conservation equation sepa- rately for T, and T". Both temperatures are functions of time and space variables. Starting from a three-dimensional equation for con- servation of thermal energy in a turbulent flow, two- and one-dimensional equations for water temperature, T, are obtained by integrating the equation, first, with respect to depth and, second, with respect to surface width of a stream channel. Two assumptions involved are that the heat exchange between water and sur- rounding media takes place at the interface without storage and that the introduction of dispersion coeffi- cients is justified empirically. No formal assumptions are required as to the steadiness or uniformity of a flow as long as the gradient-type dispersion terms are ac- ceptable approximations. 37 The equation for T, is obtained by subtracting the equation for Tn from that for T. Assuming that the entire heat exchange occurs at the air-water interface, the resulting equation is shown to be independent of incoming solar and atmospheric radiation fluxes, the measurement of which requires expensive instrumen- tation. Three other essential heat fluxes, namely, back radiation, latent heat of evaporation, and conduction are dependent on transient meteorologic conditions and are nonlinear functions of water temperature. However, these fluxes can be linearized with respect to T, provided that an arbitrarily chosen base tempera- ture of linearization, Tb, does not deviate excessively from both T and T”. The surface heat dissipation coeffi- cient, U, is simply the sum of the first derivatives of the three heat fluxes with respect to temperature evaluated at Tb. In this manner, the equation for T, is reduced to a linear equation. The one-dimensional equation for Te may be further reduced to an ordinary differential equation by ne- glecting the longitudinal dispersion term and adopting the Lagrangian approach, which employs mean travel- time as the only independent variable. If time-varying initial excess temperature and wind speed are given as inputs, the excess temperature can be solved as a func- tion of traveltime. For the two-dimensional equation for Te, a steady—state Eulerian solution is obtained for a hypothetical "conservative” excess temperature, as- suming that both stream and waste heat discharges remain steady. The actual excess temperature is ap- proximated by multiplying the “conservative” excess temperature by the Lagrangian heat decay factor simi- lar to the one-dimensional model. Only a one-dimensional model is considered for natural temperature in this analysis. The incoming solar and atmospheric radiations are important input data. Linearization of the other three heat fluxes is not employed in the equation for T", because the nonlinear effect is small in the normal range of variation of T,. and these three fluxes at certain times are adequately approximated by use of the known temperature at an immediately preceding time. The solutions are ob- tained by use of the Lagrangian method. A simpler method based on the "thermally homogeneous stream” assumes that Tn is a function of time alone everywhere in a stream. The method has an advantage of requiring lesser amounts of input data than the Lagrangian method. The second part of the report is devoted to the de- tailed analysis of data by the mathematical models. A summary of the result is given at the end of the sec- tions describing, respectively, one-dimensional natural temperature, one-dimensional excess temperature, and two-dimensional excess temperature. Several con- 38 THERMAL LOADING OF NATURAL STREAMS clusions may be drawn from the analysis. The one— and two-dimensional models presented in this report provide adequate descriptions for the varia- tion of natural as well as thermally loaded tempera- tures in a natural stream reach on the order of 30 km in length. There were some streams in which large discrepancies were noticed between observed and cal- culated temperatures. In some cases these errors can be traced to large amounts of ground-water accretion and longitudinal dispersion, both of which are ne— glected in the model. The majority of errors, however, can be attributed to inadequate input data, as well as uncertainty about the mass-transfer coefficient of evaporation. The transport aspects of the present models are gen- erally satisfactory. Longitudinal dispersion terms may safely be neglected in temperature equations even under a typical diurnal cycle of waste heat discharges such as seen below power plants. Optimum values for the transverse dispersion coefficient of heat were found to be 0.52 to 0.65 ' {Y} - {V, }, corresponding to the range commonly accepted for solute masses. As indicated by the Potomac River data, transverse mixing is a very slow process for diluting excess heat. As for surface heat dissipation aspects, the one- dimensional excess temperature model clearly indi- cates the dominant effect of the wind speed and the mass—transfer coefficient of evaporation. An optimum value for the latter was 2.7-10'8 kg-m'l -newton'1. Further stream studies on forced evaporation are de- sirable, especially under unstable atmospheric condi- tions. No definite conclusions could be obtained from the tests of the one-dimensional natural temperature model and the two-dimensional excess temperature model with regard to the mass-transfer coefficient. The data as well as the analysis of excess tempera- ture strongly indicate that heat dissipation through the air-water interface is also a very slow natural proc- ess in removing excess heat. No more than 50 percent of initial excess heat was dissipated in reaches of most streams under investigation. Therefore, this aspect should be given an adequate consideration in planning or regulating operations of a series of power plants along a river system. ' REFERENCES CITED Anderson, E. R., 1954, Energy-budget studies, water-loss investiga- tions, Lake Hefner studies, Technical Report: U.S. Geol. Survey Prof, Paper 269, p. 71—119. Bowen, 1. S., 1926, The ratio of heat losses by conduction and by evaporation from any water surface: Phys. Rev., Ser. 2, v.27, p. 779—787. Brown, G. W., 1972, An improved temperature prediction model for small streams: Water Resources Research Inst. Report WRRI— 16, Oregon State Univ., Corvallis, Oregon, 20 . Brutsaert, W., and Yeh, G. T., 1970, Implications of a type of empiri- cal evaporation formula for lakes and pans: Water Resources Research, V. 6, no 4, p. 1202—1208. Crank, J., 1967, The mathematics of diffusion: London, Oxford Uni- versity Press, p. 120—132. Edinger, J. E., and Geyer, J. C., 1965, Heat exchange in the envi- ronment: EEI Publ. No. 65—902, Edison Electric Inst, New York, 259 p. Ewing, G., and McAlister, E. D., 1960, On the thermal boundary layer of the ocean: Science, v. 131, p. 1374—1376. Fischer, H. B., 1973, Longitudinal dispersion and turbulent mixing in open channel flow, in Annual Review ofFluid Mechanics, v. 5: Palo Alto, Calif, Annual Reviews Inc., p. 59—78. Glover, R. E., 1964, Dispersion of dissolved or suspended materials in flowing streams: U.S. Geol. Survey Prof. Paper 433—B, 32 p. Harbeck, G. E., Jr., 1953, The use of reservoirs and lakes for the dissipation of heat: U.S. Geol. Survey Circ. 282, 6 p. 1962, A practical field technique for measuring reservoir evaporation utilizing mass-transfer theory: U.S. Geol. Survey Prof. Paper 272—E, p. 101—105. 1970, Discussion on the cooling of riverside thermal-power plants, in Parker, F. L. and Krenkel, P. A., eds, Engineering Aspects of Thermal Pollution: Nashville, Tenn., Vanderbilt Univ. Press, p. 133—143. Harbeck, G. E., Jr., Koberg, G. E., and Hughes, G. H., 1959, The effect of the addition of heat from a power plant on the thermal structure and evaporation of Lake Colorado City, Texas: U.S. Geol. Survey Prof. Paper 272—B, 49 p. Harbeck, G. E., Jr., Kohler, M. A., Koberg, G. E., and others, 1958, Water-loss investigations, Lake Mead studies: U.S. Geol. Survey Prof. Paper 298, 100 p. Harbeck, G. E., Jr., Meyers, J. S., and Hughes, G. H., 1966, Effect of an increased heat load on the thermal structure and evaporation of Lake Colorado City, Texas: Texas Water Development Board report 24. Harleman, D. R. F., 1972, Longitudinal temperature distribution in rivers and estuaries, one-dimensional mathematical models: Engineering Aspects of Heat Disposal from Power Generation, MIT Summer Session, Cambridge, Mass. Hinze, J. 0., 1959, Turbulence: New York, McGraw-Hill Book Co., 586 p. Holley, E. R., 1971, Transverse mixing in rivers: Delft Hydraulics Lab. Rept. 8132, Delft, The Netherlands, 92 p. Jackman, A. P., and Meyer, E. L., 1971, Basic data on heat dissipa— tion downstream from large heat sources: Health Physics. Soc., 5th Ann. Midyear Mtg, a symposium, 8 p. J aske, R. T., 1969, Thermal modification of river quality: Rept. BNWL—719, Pacific Northwest Lab., Richland, Wash, 6 p. Jobson, H. E., 1972, Effect of using averaged data on the computed evaporation: Water Resources Research, v. 8, no. 2, p. 513—518. 1973a, The dissipation of excess heat from water systems: Am. Soc. Civil Engineers, Jour. Power Div., v. 99, no. P01, p. 89—103. 1973b, Evaluation of turbulent transfer laws used in comput- ing evaporation rates: U.S. Geol. Survey open-file report, 169 p. Jobson, H. E., and Yotsukura, N., 1973, Mechanics of heat transfer in nonstratified open-channel flows, in Shen, H. W., Environ- mental Impact on Rivers: For Collins, Colo. p. 8.1—8.67. Koberg, G. E., 1964, Methods to compute long-wave radiation from the atmosphere and reflected solar radiation from a water sur- face: U.S. Geol. Survey Prof. Paper 272—F, p. 107—136. Leendertse, J. J ., 1970, A water-quality simulation model for well- mixed estuaries and coastal seas; Vol. 1, Principles of Computa- tion: Santa Monica, Calif, Rand Corp. Memo. RM—6230—RC. McAlister, E. D., and McLeish, W., 1969, Heat transfer in the top REFERENCES CITED 39 millimeter of the ocean: Jour. Geophys. Research, v. 74, no. 13, p. 3408—3414. Marciano, J. J., and Harbeck, G. E., Jr., 1954, Mass-transfer studies, water-loss investigations, Lake Hefner studies, Technical Re- port: U.S. Geol. Survey Prof. Paper 269, p. 46—70. Messinger, H., 1963, Dissipation of heat from a thermally loaded stream: U.S. Geol. Survey Prof. Paper 475—0, p. C175—C178. Morse, W. L., 1970, Stream temperature prediction model: Water Resources Research, v. 6, no. 1, p. 290—302. Pai, S. 1., 1956, Viscous flow theory, I-laminar flow: Princeton, N.J., D. Van Nostrand Co., Inc., p. 41—43. Parker, F. L., and Krenkel, P. A., 1969, Engineering aspects of thermal pollution: Nashville, Tenn., Vanderbilt Univ. Press, 351 p. Pierson, F. W., and Jackman, A. P., 1975, An investigation of the predictive ability of several evaporation equations: Jour. Appl. Meteorology, v. 14, no. 4, p. 477—487. Pluhowski, E. J., 1970, Urbanization and its effect on the tempera- ture of the streams on Long Island, New York: US. Geol. Survey Prof. Paper 627—D, 110 p. Priestly, C. H. B., 1959, Turbulent transfer in the lower atmosphere: Chicago, 111., Univ. Chicago Press. Prych, E. A., 1970, Effect of density difference in lateral mixing in open-channel flows: Rept. no. KH—R—21, Pasadena, Calif, Calif. Inst. Tech., 225 p. Sayre, W. W., 1973, Natural mixing processes in rivers, in Shen, 1w. 3. GOVERNMENT PRINTING OFFICE: 1977 — 789-025/50 H. W., ed., Environmental Impact on Rivers: Fort Collins, Colo., p. 61—637. Sellers, W. D., 1965, Physical climatology: Chicago, Ill., Univ. Chicago Press, 272 p. Taylor, G. I., 1954, The dispersion of matter in turbulent flow through a pipe: Proc. Royal Soc., London, v. 223 A, p. 446—468. Thomann, R. V., 1973, Effect of longitudinal dispersion on dynamic water quality response of streams and rivers: Water Resources Research, v. 9, no. 2, p. 355—366. Yeh, G. T., Verma, A. P., and Lai, F. H., 1973, Unsteady tempera- ture prediction for cooling ponds: Water Resources Research, v.9, no. 6, p. 1555—1563. Yotsukura, N., 1972, A two-dimensional temperature model for a thermally loaded river with steady discharge: Contribution to the 11th Ann. Environmental and Water Resources Eng. Conf., Vanderbilt Univ., Nashville, Tenn., p. 13—26. Yotsukura, N., and Cobb, E. D., 1972, Transverse diffusion of solutes in natural streams: U.S. Geol. Survey Prof. Paper 582—0, 19 p. Yotsukura, N., Fischer, H. B., and Sayre, W. W., 1970, Measurement of mixing characteristics of the Missouri River between Sioux City, Iowa, and Plattsmouth, Nebraska: US. Geol. Survey Water Supply Paper 1899—G, 29 p. Yotsukura, N., Jackman, A. P., and Faust, C. R., 1973, Approxima- tion of heat exchange at the air-water interface: Water Re- sources Research, v. 9, no. 1, p. 118—128. , LufiB: Otarioid Seals of the Neogene By CHARLES A. REPENNING and RICHARD H. TEDFORD GEOLOGICAL SURVEY PROFESSIONAL PAPER 992 Classification, historical zoogeography, and temporal correlation of the sea lions and walruses from the North Pacific region UNITED STATES GOVERNMENT PRINTING OFFICE,WASHINGTON:1977 UNITED STATES DEPARTMENT OF THE INTERIOR CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Director Library of Congress Cataloging in Publication Data Repenning, Charles Albert, 1922— Otaroid seals of the Neogene. (Geological Survey Professional Paper 992) Bibliography: p. Supt. of Docs, no.: I 19.162992 1. Pinnipedia, Fossil. 2. Pinnipedia. 3. PaleontologyiTertiary, 4. Paleontology—North Pacific region. I. Tedford,Richard H., joint author. II. Title. III. Series: United StatesGeological Survey Professional Paper 992. QE882.C15R46 569'.74 76—608333 For sale by the Superintendent of Documents, US. Government Printing Office Washington, DC. 20402 Stock Number 024-001-02988-0 CONTENTS Abstract Introduction . The terminology of geologic time .......................................... Miocene-Pliocene boundary ........................................... Pliocene-Pleistocene boundary ........................................ Rock-stratigraphic units discussed _. Other terminology ...................................................................... Acknowledgments .................................................................... Suprageneric diagnoses ..... .. Superfamily Otarioidea ............................................................ Characters ............................................................ Remarks ...... __ Family Odobenidae .......................................................... Characters ............................................................ Remarks .................. Subfamily Odobeninae _____ Characters ______________ Remarks _________________________ Subfamily Dusignathinae Characters .......................... Remarks ...... __ Family Otariidae ........... Characters .. ..... Remarks ............................. Subfamily ”Arctocephalinae” Characters ......................... Remarks .................. Subfamily ”Otarinae" Characters ................. Family Desmatophocidae,_ Characters. Remarks ...... Family Enaliarctidae. Characters. Remarks ______ Part I: Walruses ...................................... Family Odobenidae . Subfamily Odobeninae ............. Genus Odobenus Brisson . Type species ___________ Diagnosis ____________ Included species ______________ Genus Alachtherium DuBus .. Type species .................... Diagnosis ............ Included species ..................................... Genus Prorosmarus Berry and Gregory . Type species ....................................... Diagnosis... Distribution ...... Included species ........... Genus Aivukus new genus ._ Type species ................. Etymology ______ Diagnosis .............................................................. Page Page Part I: Walruses—Continued Family Odobenidae—Continued Subfamily Odobeninae—Continued Genus Aiuukus new genus—Continued Distribution ........................................................ 14 Included species 14 Aiuukus cedrosensis new species 14 Holotype ................................................................ 14 Etymology 14 Referred material from Cedros Island __________ 14 Referred material from Baja California Sur ........................ 14 Diagnosis ...... 1 4 Type locality and age ...................................... 14 Description ................. 16 Mandible ...................................................... 16 Skull _ 16 Postcranial skeleton .................................. 20 Discussion .................................. 22 Subfamily Dusignathinae ____________________ 22 Genus Imagotaria Mitchell 22 Type species ................................................ 22 Diagnosis ........................................................ 22 Distribution ................................... 22 Imagotaria downsi Mitchell, 1968 ................ 22 Holotype ........................................................ 22 Referred material from the Santa Margarita Formation, Santa Cruz, California ........................................ 22 Material from other localities . 24 Diagnosis ...................................................... 24 Type locality and age ................................ 24 Age of the referred material from the Santa Cruz area .............................. 24 Referral of specimens to the species ...... 27 Description of material from Santa Cruz and comparison with the type ____________________________________ 27 Female mandible .. 27 Female skull ........................................ 28 Juvenile male skull ____________________________ 33 Scapula .................... 34 Humerus .............................................. 35 Ulna ...................................................... 35 Radius ........... 35 Scapholunar ........................................ 36 Cuneiform ............................................ 36 Pisiform ....... 36 Trapezium ............................................ 37 Trapezoid .............................................. 37 Magnum... 37 Unciform .............................................. 37 Metacarpals ________________________________________ 37 Phalanges _______ 38 Femur .................................................... 38 III IV Page Part I: Walruses—Continued Family Odobenidae—Continued Subfamily Dusignathinae—Continued Genus Imagotaria Mitchell—Continued Imagotaria downsi Mitchell, 1968—Continued Description of material from Santa Cruz and comparison with the type— Continued Patella .................................................. 38 Tibia ______________________________________________________ 39 Calcaneum 39 Astragalus ............................................ 39 Navicular .............................................. 40 Cuboid __________________ 4O Cuneiform bones ................................ 40 Metatarsals .......................................... 40 Vertebrae 41 USNM 13487 ................................................ 41 Discussion .................................................... 42 Genus Pontolis True ....... 42 Type species ........................................ 42 Discussion ____________________________________________________ 42 cf. Pontolis magnus, Lyon, 1941 43 Genus Dusignathus Kellogg __________________________________ 43 Type species ........................................ 43 Diagnosis .......... __ 43 Distribution ________________________________________ 43 Dusignathus santacruzensis Kellogg, 1927 .................................... 43 Holotype ................................................ 43 Referred material .............................. 43 Questionably referred material. .. 44 Type locality ........................................ 44 Discussion of the type ................................ 44 Discussion of referred specimens __________ 45 Discussion of questionably referred specimens ........................................ 47 Concluding discussion .. .. 48 Genus Pliopedia Kellogg .......................................... 49 Type species ........................................ 49 Diagnosis ....... ._ 49 Distribution ........................................ 49 Pliopedia pacifica Kellogg, 1921 .................... 49 Holotype ............................ .. 49 Referred material .............................. 49 Type locality and age ........................ 49 Discussion ............................ 49 Humerus .............................................. 50 Ulna ...................................................... 50 Radius ............. .. 51 Right trapezoid .................................. 51 Right metacarpal I ............................ 51 Right metacarpal II .. .. 51 Right metacarpal III ........................ 51 Metacarpal proportions .................... 5 1 Skull ......................................... 52 Possible rear limb elements ............ 52 Possible synonymy ............................ 52 Genus Valenictus Mitchell .......... i. 53 Type species ........................................ 53 Discussion .................................................... 53 Genus Neotherium Kellogg ._ 54 Type species ........................................ 54 Discussion ____________________________________________________ 54 CONTENTS Part I: Walruses—Continued Family Odobenidae—Continued Classification of the walruses ________________________________________________ Summary of the history of the walruses ______________________________ Part II: Fur seals and sea lions ...................................................... Family Otariidae Genus Arctocephalus F. Cuvier ______________________________________ Type species ....... Diagnosis ...................................................... Included species __________________________________________ Genus Callorhinus Gray _________ Type species ................................................ Diagnosis ...................................................... Included species. Genus Pithanotaria Kellogg .......................................... Type species ________________________________________________ Diagnosis ......... Distribution _____________________________________ Included species ............................... Pithanotaria starri Kellogg, 1925 Holotype _________________________________________ Referred material ....................... Diagnosis _____________________ Type locality and age ________________________________ Discussion ............................................................ Genus Thalassoleon new genus Type species ........... Etymology Diagnosis ______ Distribution Included species ........................ Thalassoleon mexicanus new species Holotype ______________________________________ Etymology .............................. Diagnosis ..................... Type locality and age Referred material ......... Description of skull ........................................... Discussion of the postcranial skeleton ...... Summary and discussion ................................ Thalassoleon macnallyae new species “ Holotype ______________________________________________ Etymology ______________________________________ Diagnosis _____________________ Type locality and age ................................ Questionably referred material .............. Discussion ____________________________________________________________ Other fur seals ______ Specimens from the San Diego Formation __________ Specimen from the Etchegoin Formation Fossil sea lions ...................................................... Pliocene record .................... Early Pleistocene record .. Late Pleistocene record Classification of fur seals and sea lions ................ Summary of the history of fur seals and sea lions ____________ Part III: Desmatophocids, enaliarctids, and faunas .............. Family Desmatophocidae ........................................................ Genus Desmatophoca Condon ______________________________________ Type species ............ Diagnosis ______________________________________________________ Included species __________________________________________ Page 55 55 57 57 57 57 57 57 57 57 57 57 58 58 58 58 58 V Page 79 79 79 79 81 81 81 82 82 CONTENTS Page Part III: Desmatophocids, enaliarctids, and faunas—Continued Part III: Desmatophocids, enaliarctids, and faunas—Continued Family Desmatophocidae—Continued Faunas—Continued Genus Allodesmus Kellogg .............................................. 74 Early early Miocene (about 22.5 to 17 m.y. B.P.) ........ Type species ................................................ 74 Late early Miocene (about 17 to 14.5 m.y. B.P.) __________ Diagnosis ...................................................... 75 Early middle Miocene (about 14.5 to 13 m.y. B.P.) Included species ...... .. 75 Late middle and early late Miocene Classification of the desmatophocids .................................. 76 (about 13 to 9 m.y. B.P.) ................................................ Discussion of the desmatophocids ........................................ 76 Late late Miocene and early Pliocene Family Enaliarctidae 76 (about 0 to about 3.8 m.y. B.P.) .................................... Genus Enaliarctos Mitchell and Tedford .................... 77 Late Pliocene (about 3.8 to 1.8 m.y. B.P.) ...................... Type species ................................................ 77 Early Pleistocene (about 1.8 to almost 0.7 Diagnosis .......... .. 77 m.y. B.P.) .. Included species .......................................... 77 Late Pleistocene (0.7 to about 0.01 m.y. B.P.) ______________ Faunas 77 References cited Oligocene (me-22.5 m.y. B.P.) ......................................... 79 Index ILLUSTRATIONS [Plates follow index] PLATE 1. Aivukus, Odobenus, Neophoca. 2. Aivukus, Prorosmarus, Odobenus. 3. Aiuukus. 4. Aivukus, Pliopedia, Imagotaria, Obodenus. 5. Zalophus, Imagotaria, Dusignathus, Prorosmarus. 6. Imagotaria. 7. Imagotaria. 8. Imagotaria. 9. Imagotaria, ?Allodesmus, Neotherium. 10. Imagotaria, Pontolis, Odobenus, Zalophus. 11. Imagotaria, ?Neotherium. 12. Imagotaria. 13. Imagotaria. 14. Imagotaria, Pliopedia, Aivukus, Obodenus. 15. Imagotaria, Dusignathus, Obodenus. 16. ?Dusignathus, Valenictus, ?Pliopedia, phocid. 17. Pliopedia. 18. Dusignathus, Pontolis. 19. Pithanotaria, Arctocephalus. 20. Thalassoleon. 21. Thalassoleon, Neophoca, Callorhinus, Arctocephalus. 22. Thalassoleon. 23. Thalassoleon, Zalophus, Arctocephalus, Neophoca. 24. Thalassoleon, Pliopedia. . Drawing showing restored skull of Aivukus cedrosensis, type female skull IGCU 901 .......................................................... . Locality map, Santa Cruz area _______ . Section of Santa Margarita Formation in the Santa Cruz Aggregate Co. quarry ............................................ FIGURE 1 2 3 4. Drawing showing restored skull of Imagotaria do wnsi, referred female skull USNM 23858 .................................................. 5 6 . Drawing showing restored skull of Thalassoleon mexicanus, holotype male skull IGCU 902 ............................................ . Phylogenetic diagram of the Otarioidea 87 Page 1 6 25 26 28 62 80 VI TABLE . Number of roots on the upper cheek teeth, Imago taria downsi ......... . Dimensions of three humeri of Imago ta ria downsi from the Santa Margarita Formation ...................................................... . Dimensions of a radius and ulna of Imagotaria downsi ...................................................................................................................... . Dimensions of the metacarpals of Imagotaria downsi, transverse diameters i. . Dimensions of Dusignathus santacruzensis, referred specimen UCR 15244 ....................................................................... . Metacarpal measurements of three dusignathine genera .................................................................................................................. . The skull and humerus length and proportions of mature specimens of Thalassoleon mexicanus, Arctocephalus CONTENTS TABLES . Time correlation chart ______________________________ Age assignments of otarioid fauna and their host rocks .......................... Proportions of the humerus in otarioids ............... Dimensions of the female mandible and teeth of Imagotaria downsi (USNM 23858) ................................................................ Dimensions of two skulls of Imagotaria downsi ......................................................................................... Dimensions of the upper teeth of Imagotaria downsi ........................................................................................................ pusillus, and Arctocephalus forsteri ______________________ _ Dimensions of the skulls of Thalassoleon mexicanus and T halassoleon macnallyae .................................................... . Hind limb proportions of fossil and living otariids .............................................................................................................................. OTARIOID SEALS OF THE NEOGENE 1 By CHARLES A. REPENNING and RICHARD H. TEDFORD ABSTRACT The otarioid seals originated in the North Pacific region and most of their history is centered in this area. They include four families: the extant Odobenidae (walruses) and Otariidae (fur seals and sea lions) and the extinct Desmatophocidae and Enali- arctidae. Early in their development, the odobenids dispersed through the Central American Seaway to the North Atlantic, where the living walrus evolved. By late Pliocene time, the odobenids had become extinct in the North Pacific but the modern walrus spread from the Atlantic to the Pacific by way of the Arctic Ocean in late Pleistocene time. The history of the odobenids is known back to the early middle Miocene, at which time they seem to have evolved out of the ancestral otarioid family, the Enaliarctidae. The odobenids were most diverse in the North Pacific during the late Miocene when six genera belonging to two subfamilies are known. A new genus and species, Aiuukus cedrosensis, from the Almejas Formation of Baja California, is a privitive odobenid of the Subfamily Odobeninae. Three Pliocene and younger genera, including the living genus, are recognized in the Atlantic, but evolution in this area was essentially unidirectional toward modern walrus. The history of the otariid seals began in the late middle or early late Miocene when this family evolved out of the last of the Enaliarctidae. The otariids remained a family with little variety, and, relative to the odobenids, slight evolutionary change until the late Pliocene or early Pleistocene, more or less the time of extinction of the odobenids in the Pacific. During this period of little diversification, one new otariid genus, Thallassoleon, is recognized. This genus contains two new species: T. mexicanus, from the Almejas Formation of Baja California, and T. mac- nallyae, from the Drakes Bay Formation of Point Reyes, Cali- fornia. The otariids are today in their period of greatest diversifi- cation—there are seven living genera. By early Pliocene time, the otariids had dispersed to the South Pacific Ocean, whence during the Pleistocene, they spread to their present circumantarctic dis- tribution. They have never reached the North Atlantic. The extinct desmatophocids evolved out of the enaliarctid group in the early Miocene and became most diverse in the middle Miocene. They are last known from rocks of late Miocene age, at about the time that the odobenids began to diversify. The desmatophocids seem never to have left the North Pacific, but they are known from southern California to Alaska and from Japan. The Enaliarctidae were ancestral to the other three families and are largely unstudied. They were derived from primitive ursid land carnivores, presumably during the Oligocene, and may be described as flippered marine carnivores with heterodont denti- tion in which the premolar, carnassial, and molar teeth are differ- 1The American Museum of Natural History, New York City. entiated in form, reflecting their land-carnivore ancestry. The evolutionary stage at which all cheek teeth become homodont is herein arbitrarily taken as that point where the descendant families are to be taxonomically recognized. The Enaliarctidae were most diverse in the early Miocene. From the beginning of the Miocene to the late Pleistocene, eight otarioid faunal units based on generic composition are recognizable in the eastern North Pacific. Otarioid species, as now recognized, seem to have existed for about 2 my on the'average, although some genera have persisted for 5 or more my INTRODUCTION The seals of the world were divided into two major groups by Allen (1880): the ”walkers” and the "wrigglers." Smirnov (1908) subsequently applied the name Otarioidea to those seals which could flex their hind legs beneath them and walk on land, the name Phocoidea to those seals which could not, but rather wriggled on their bellies in terrestrial loco- motion, their hind limbs permanently extended be- hind them. The two groups have been placed in the mammalian Order Pinnipedia (sensu Scheffer, 1958) or Suborder Pinnipedia (sensu Simpson, 1945) since Illiger (1811) first defined the group in its present context. The pinnipeds have always been considered to be related to the fissiped carnivores, and, the wide- spread belief, since about 1960, that the group is not monophyletic (McLaren, 1960) has led to the rejec- tion of ordinal or subordinal distinction from the fissiped carnivores by some workers (McKenna, 1969, p. 235; Mitchell and Tedford, 1973, p. 278). The question of polyphyly is largely a matter of definition. According to some definitions (Simpson, 1961, for example), the Order Pinnipedia would be monophyletic in that it is derived from a group of equivalent or lower taxonomic rank, in this case the Superfamily Arctoidea (or Canoidea). The debate really centers over whether there was one or two protopinnipeds from which all pinnipeds derive: whether the pinnipeds derive from one, two, or three, invasions of the marine environment by fissiped carnivores. The obviously different centers of origin, the phocoids in the Atlantic area and the otarioids in the North Pacific, certainly suggest two protopin- 1 2 OTARIOID SEALS OF THE NEOGENE nipeds which, nevertheless, may have been closely related. In addition, Mitchell and Tedford (1973) have described a protopinniped from the North Pacific which may have been ancestral to all otarioid seals and which, in their analysis, bears only limited resemblance to what might be expected of a proto- phocoid. This report describes several of new otarioid fossils which, along with other evidence, indicate that the sea lions (Family Otariidae) and the wal- ruses (Family Odobenidae) have had long, different, and separate evolutionary histories. The retrospec- tion permitted by these new records supports the opinion that they must have evolved at different times out of one group of primitive otarioids rather than from different terrestrial carnivores. The fossil record of the pinnipeds is very poor. As much as anything, this seems to be the result of mammalian paleontology stopping at the water’s edge. In the marine environment, with abundant invertebrate fossils available, there has been no stratigraphic need for vertebrate paleontology so pinniped remains frequently have been ignored or described as more or less isolated curiosities. Only in recent years has intensive collecting of fossil pin- nipeds taken place. The stratigraphic usefulness of fossil pinnipeds is potentially great, however. Despite their rarity relative to the abundant invertebrate fossil remains, the pinnipeds have rates of evolution comparable to those of other mammals; thus they are much more significant in temporal correlation than many ma- rine invertebrates. In addition, the geographic range of many pinnipeds is far greater than that of many invertebrates. All that is needed is more specimens and more comprehensive study to fulfill this poten- tial. In the past decade, this need has been greatly lessened by the work of Edward D. Mitchell, J r., both in description of new material and in repeated reevaluation of previously described fossil otarioids. Through Mitchell’s work, it has become evident that in the Neogene of the North Pacific Ocean there existed a great assortment of different types of otarioid seals. Many of these seals belong to extinct groups and have been difficult to recognize as being clearly related to either of the two living otarioid families, the sea lions and the walruses. In some of his publications (1966, 1968) Mitchell has provided a revised classification of the ”walking seals.” In his most recent publication (Mitchell, 1968; reaffirmed without explicit discussion in Mitchell and Tedford, 1973) he names, in the essence of established classification, four extinct families, each represented by only one genus, in addition to the two living families. Complexity of the recogni- tion of so many pinniped lineages is made less conspicuous by lowering all categories one rank from that in the conventional classification system, but this does not eliminate the lack of pattern. For example, although Mitchell (1968, p. 1887) concurs with assumptions that the odobenids originated in the North Pacific region, he lists only the living genus and three extinct genera of walrus from the Atlantic in his Subfamily Odobeninae (equivalent in com- position to the Family Odobenidae of other authors). Equally uninformative, only one extinct genus of the sea lions, Pi thanotaria, is listed under his Subfamily Otariinae (which includes all species listed under the Otariidae by other authors). It will be shown in this report that several of Mitchell’s fossil "sea lions," two of his subfamilies, and both of his subfamilial incertae sedis are wal- ruses in a familial sense, Family Odobenidae. In addition, the lineage of fossil sea lions, Family Otariidae, will be shown as an extremely conserva- tive lineage as far as the present fossil record reveals not as diverse a group as the Odobenidae. Reduced ad absurdum, the theme of this report is that it does not take tusks to make a walrus. It seems appropriate, therefore, to begin by defining what does make a walrus and, by diagnoses, to indicate how other otarioids differ. This is followed by the description of a fossil walrus without tusks. This, then, enables the recognition and description of other fossil odobenids and description of their known stratigraphic and geographic range to complete Part I of this report- Part II of this report includes a description of fossil otariids new to the published record, a sum- mary of published records, and a description of their known stratigraphic and geographic range. Part III contains no description of new material but sum- marizes the diversification of a third and extinct otarioid family, the Desmatophocidae (Hay, 1930); it also discusses the presumed ancestral family, the Enaliarctidae of Mitchell and Tedford described by these authors as a subfamily of the otariidae (=Otari- oidea of this report) . THE TERMINOLOGY OF GEOLOGIC TIME Most of the fossil seals discussed in this report lived during the Miocene and (or) Pliocene Epochs. Confusion has existed for many years in the applica- tion of these terms to rocks in the same or different parts of the world by different paleontologists and stratigraphers. No report now dealing with these terms without definition can make geochronologic -‘-‘ INTRODUCTION 3 sense if it deals with interdisciplinary topics or with more than localized geography. MlOCENE-PLIOCENE BOUNDARY In recognition that typology is the only logical defense of usage, this report follows the 1959 recom- mendations of the Mediterranean Neogene Commit- tee of the International Geological Congress regard- ing the typification of the Miocene-Pliocene bound- ary. Subsequent refinements have been summarized recently by Van Couvering (1972), Van Couvering and Miller (1971), and Berggren (1971, 1972). For the most part, our usage in this report (table 1) follows their most recent evaluation of the correlations involved (Berggren and Van Couvering, 1974). In this usage, the boundary between rocks of Miocene age and those of Pliocene age is placed at the top of the evaporitic deposits in southern Italy assigned to the Messinian Stage. This usage places most of the rocks that have been called Pliocene along the Pacific Coast of North America in the Miocene Series. The Pliocene Epoch now seems to have lasted only about 3.2 m.y., whereas as it has been used on the Pacific Coast of North America it was considered to have lasted as long as 10 my2 It is emphasized here that this difference is an insistence on a typo- logical meaning of the epochs and not a change in estimation of the age of the rocks. As closely as available radiometric dates can be correlated to fossil localities and to the stratotypes of the Miocene and the Pliocene Series, fossils from about 1.8 to about 5.0 my old are herein called Pliocene. Although the fossil record of otarioid seals is poor during this time, there appears to have been significant evolution of the fauna, and the Pliocene is here subdivided into early and late parts at the Blancan-Hemphillian North American land mam- mal age boundary, which we approximate as being between 3.5 and 4 my, following Evernden, Savage, Curtis, and James (1964). This boundary between early and late Pliocene is somewhat older than the approximately 3.3 m.y. boundary used by Berggren and Van Couvering (intending to date the Zanclian- Piacenzian contact). At the present time, however, the faunal differences between the Hemphillian and Blancan Ages are recognizable and no such faunal difference is detectable in the West Coast marine invertebrate section (that is, in the San Joaquin Formation or equivalent). The Miocene Epoch, which in past customary usage in the Pacific Coast area was about 10 my in duration, is now considered to be of nearly 17 my. 2Based upon an approximately 1.8 m.y.-old Pliocene-Pleistocene boundary as used in Berggren (1971, 1972). duration as a result of the acquiescence to typology here expressed. Previously the Miocene Epoch had been subdivided into early, middle, and late parts by more or less arbitrary usage in the Pacific Coast area. These subdivisions were provincial in usage, and it was well understood that their application varied with the paleontologic discipline and was not specifically intended as an exact correlation of some archetypal stratigraphic section in Europe. Their usage, nevertheless, was useful in the Pacific Coast Province. In the chronology of North American vertebrate paleontologists (Wood and others, 1941), as radio- metrically calibrated (Evernden and others, 1964), the late Miocene extended from about 16 my ago to about 12 my ago. For the past 10 years at least, (Evernden and others, 1964, p. 147), there has been a growing realization that by this definition, the con- clusion of the Miocene as used by North American vertebrate paleontologists was considerably earlier than that of the type Miocene, regardless of the indecision on where the Miocene-Pliocene boundary should be placed in the type section. With the adoption of the recommendation of the Mediter- ranean Neogene Committee, it appears that the Miocene Epoch persisted about 7 my longer than the Miocene as defined by Wood and others (1941) and as customarily used by most vertebrate paleon- tologists in North America. A more or less equal change is indicated in the usage of the late Miocene by invertebrate paleontologists in the Pacific Coast area, although historically they have extended the Miocene into rocks slightly younger than those recognized as of this age by vertebrate paleon- tologists. Seven million years is a span of time great enough to include several recognizable faunal changes, and hence a span of time for which tempo— ral subdivisions would be useful; to refer to the time from 16 my ago to about 5 my. ago as ”late Miocene” would result in losing useful precision. On the other hand, to redefine the informal subdivisions early, middle, and late Miocene would be to introduce familiar terms with unfamiliar meaning into the language of Pacific Coast paleontology. Redefining the informal subdivisions of the Miocene seems to be the most reasonable choice though, regrettably, it will also be the most con- fusing. Berggren (1971, p. 755) has suggested a threefold subdivision of the Mediterranean Miocene that seems satisfactory for the most part. This subdivision places the Tortonian and the Messinian Stages in the upper Miocene and the Langhian and the Serravallian Stages in the middle Miocene. The 4 OTARIOID SEALS OF THE NEOGENE TABLE 1.—Time correlation chart NORTH WEST COAST WEST TIME .AGE EUROPEAN AMERIcAN MARINE STAGES COAST NOTES (m.y.) (th's report) STAGES 1 MAMXGAEL'AN MEGAFOSSILSG MICROFOSSILS USAGE ' Rancho- . Late gg Late . labrean 1 m E L“ — 1 _ m UJ Calabrlan 8 <2 E 3 1Berggren and Vancouv~ El 0 Early lrvingtonian : 3 U E ering, 1974. For excep— 2 c g. 0- UJ tions see text. _ 2 _ ‘5‘: :: LL] 0.) E g iii c Blancan ° _ 3 .2 E —' c 9 E, ‘2 : E E 2 a g = a .E a: 3 Addlcott, 1972, 1969. U 1‘; m g 3 : 3 D : OJ 0 >. < 8 m E g ‘2 1" _ 4 _ _ 1. E N 0 o 0: Lu m _l Lcuu 8’ fl 8 _‘ 3 “- § g to Z Naeser, Izett, and Wilcox l— 5 — E Eu : ___‘,»I_ UJ 1971. Johnson, Opdyke, . . E ‘ and Lindsay, 1972: based _- 6 _ Messmian '5, 0 upon the earliest Lepux. cu E O a I ‘ — 4 _ 7 _ In Evernden and others.,1964. .J o ——1 > .‘L‘ _ a.) E 0- L — 8 ‘_ H 5 8 P CO 5 'E 4 :1 DJ Oldest Hemphillian date (0 9 ~ 8.0 m.y. Ash Hollow _ 9 _ _J >_ 5 Formation, J.D. Boellr _ 1- % storf, written commu- L 'E C nication to Tedford . (U ' — 10 —J l: g. ~c E 1975. c m _____ .c a; .‘Z 0 OJ — 11 — ‘5 ‘5 E P 6 0 g “3 Pierce, 1970, Barron, LlJ (D a: E —-J 1976: temporal over- _. 12 _ _ E Serravallian 3 = lap of definitions are Z "O .1 involved. — 13 — LLI U ._ >_ _ . 7 - . C Lunsnan Turner 1970 KA2127 L m . i _ 14 _ U E Haj Langhian ‘; 8 Barstovian mammals O .3 : just below, unpublished. _ 5 § Relizian ‘19 — 15 — m D 8 _ 8 a: E Lu '0 Turner, 1970‘ 16 E E C ‘8 2 U _ _ m - ._ —l '7; 5 E F.” LU l— 17 m — g m 0 >. r: '9 m m — 19 — LU E S 2 3 E 8 L f —— 20 — m LL] 3 : >~ _ 21 _ : é’ _ .9 g L. c _ __ ‘V c 0' <0 22 g g g s LU o‘ E = C <1: m _ 23 _ C.) 9.‘ "g uJ <2 5 9 z l E — 24 — —I “J an o O '1’ middle-late Miocene boundary falls at approxi- North America the change from the Barstovian to mately 11 m.y.,postdating the oldest radiometrically the Clarendonian mammalian ages, marked by the dated record of the horse genus Hipparion in Europe, first record of Hipparion, also was about 12 m.y. ago which has been dated at 12.4 m.y. (Lippolt and (Evernden and others, 1964). others, 1963: Howenegg fauna of Germany). In The beginning of middle Miocene is selected to INTRODUCTION 5 be about 14.5 m.y. ago; this date approximates the beginning of the Langhian Stage in Europe as shown by Berggren and Van Couvering (1974). Berggren (1971, p. 755 and table 52.39, not table 52.40) suggested the base of the Langhian Stage as the base of the middle Miocene, a usage continued here, although Berggren and Van Couvering (1974) place the early-middle Miocene boundary within the Langhian at the base of the planktonic foraminifer zone N9. Following Van Couvering and Miller (1971, p. 562), the late Miocene is believed to have lasted until ”something like 5 m.y.” Berggren and Van Couvering (1974) review the arguments about the time of the Miocene-Pliocene boundary that fur- ther support 5 m.y. This date for the end of the Miocene Epoch falls somewhere within the younger part of the Hemphillian mammalian age in North America and approximates the beginning of the Zanclian Stage in Italy. By this definition of time, the 3.5-m.y.-long middle Miocene and the 6-m.y.-long late Miocene, span three recognizable stages of otarioid evolution along the Pacific Coast of North America. In order to identify these stages, the middle and late Miocene are here arbitrarily broken into lesser parts: early middle Miocene, dominated by the genus Allodes- mus; late middle Miocene and at least part of the early late Miocene, dominated by the genus Imago- taria; and late late Miocene, dominated by the new genus Thalassoleon. As now understood, the new genus Thalassoleon extends throughout the late late Miocene and Pliocene, and some incompletely known genera, among them Valenictus, seem to be of both late Miocene and Pliocene age. It is expected that more discoveries, and hoped that more radio- metric dates, will better define late Miocene and Pliocene otarioid faunas along the Pacific Coast. PLIOCENE-PLEISTOCENE BOUNDARY Probably no chronostratigraphic boundary has received more attention than the Pliocene-Pleisto- cene boundary. Berggren and Van Couvering (1974), in reviewing and evaluating the highlights of the investigation of the Pliocene-Pleistocene boundary, conclude that the base of the Calabrian deposits in southern Italy, which is equivalent to the base of the Pleistocene as recommended by the Pliocene—Pleis- tocene Commission of the 18th [1948] International Geological Congress, is between 1.7 and 1.8 m.y. ago. They show it in their charts as a band between 1.7 and 1.8 m.y. ago or as a line 1.8 m.y. ago. The 1.8 m.y. date is used in the present report. The approximate ages, epochal assignments, and correlation of stages used in this report are summarized in table 1. The antiquity of the fossils discussed will also be stated in years, as an inter- pretation based upon the best information available to the authors. Because the epochal assignments differ drastically from those customarily used along the Pacific Coast of North America, one column in table 1, ”West Coast usage," indicates long-estab- lished provincial usage of epoch terminology based upon West Coast marine invertebrates; that by which all previously described fossil otarioid seals from the Pacific Coast have been dated. Present trends seem to indicate that this provincial usage of epoch assignments will be continued along the Pacific Coast of North America for utilitarian con- siderations, while recognizing that it is not com- patible with European usage. Provincial usage of Neogene epochal terminology along the Atlantic Coast has not differed so greatly from European usage, and it seems at this time that the relatively minor adjustments to establish conformity are being incorporated into provincial usage. ROCK-STRATIGRAPHIC UNITS DISCUSSED Several rock-stratigraphic units are discussed here in terms of their age and their otarioid pinniped fauna. Because confusion will undoubtedly result from the new age assignments based upon accept- ance of the recommendations of the Mediterranean Neogene Committee of the International Geological Congress, it is felt that a tabulation here, in sum- mary, will help the reader retain orientation. Cause for this roughshod revision of formational age as- signments lies in the need to discuss chronology when establishing evolutionary sequence. Among fossil otarioids, this need is best illustrated by prim- itive walruses: Prorosmarus alleni is from a forma- tion customarily dated late Miocene, and Aivukus cedrosensis, new genus and species, is from a forma— tion customarily dated middle or late Pliocene. Available information now indicates that A. cedro- sensis, the more primitive genus, is older by as much as 3 m.y. (see table 2). OTHER TERMINOLOGY Anatomical terminology follows, as feasible, the usage of Miller, Christensen, and Evans (1965) for the dog. There are a few obvious exceptions, such as the names of the carpals and tarsals for which Miller, Christensen, and Evans use a nomenclature more frequently used for reptiles rather than that derived from human anatomical terminology with which most mammalian paleontologists, at least, are more familiar. OTARIOID SEALS OF THE NEOGENE .0020 850 08 .8 52005.52 0.550 05 00.: 2250000005 005 SE02: 22000.: 05 0005 00 0<_ 50:8050N v50 50:00050m 28022 550 :.008=00>: E. .2200 522508205 2.50m 3032. 050022 2500 2500 2500 500202.15 mm I mm .2005 08000205Q 5200050m 002 050 :..~0350E: 0200522050 2002802053 2. 50M050 509502 5200555 2500020. 28022 23:22 28022 3500 0002 50380505 5 I 5 02050002050 00025.20500Q E 02007» 22002 .55 0500 30.802 050022 22:2 050022 3500 202 522505.: 3 I m: .50 0002280502: 28022 m0 .200 00505505. 205m 208502 28022 23:2 22:5 2500 52052 E I n: 5805000 000000025. 28022 250350.23 mp «:00 .2505 2502 52005.52 5050553 050022 22:2 22:5 3500 050 5205..— E I 2 205002000: 005000020» 00550.5 50 505508: :0: 550 :m: 0500550005022 am 28022 2050 52052 50.55 50.202002 3 0:00 522508222 5:25:02 55:02 050022 22:2 22:5 2500 .5052050m E I 2 0.8505802 0050000224. 52005505 050022 mm «:00 2:000m 0:500:02 250m 28022 202 23:5 0002 6505502 2 I S .2000 0.2205055 5220555 28022 :5250302: «N «:00 2:5 50.85. 02500502 250m 28022 000.2 2225 2012 050 5052050520 N: I S .50 0.20.2500: 05000.22 .2500 0200050505 vm .200 005502 520055505 02.00% 28022 0502 22:5 202 525502 N2 I 02 205300 0.0025002 .5 5 02.500502 050m 050022 m0 .vm 0:00 .0550 250m 220 «0 0.505 .8302 050022 0002 23:5 0002 525028520 N2 I 02 .50 005000220 .200 5200585 22030.2. «N 505500 00020m 2: 00 ”E05 8302 28022 20..— 050022 22 3500 25002500502: 2 I m 205300 0.202500: 52005505 005.522 20 :<: 0580550005003 02500502 050m 0:000 0.2205055 wNIvm «:00 .0550 0050m 00 0.505 5253 050022 2500 28022 22 2500 :02:002.: 3 I 0 2053020 0.2000M0EN m: 5m 2500.300 0.20m 6.5505 52005505 0020m 28025 050022 22 2012 :520m220um: w I 0 0205000500 00020220‘ 02050005000500 00205205 2550200 0.20m 0050285 502000225 SI: 00500 0202 52005.55 00.853. 050025 3500 050022 22 202 0502 w I 0 02050002000 000200.20. NV 50M05 $0.5 0000 52005.55 05550 050025 3500 050022 002 202 22002230: 0 I m 03505 0.20505 .2200 52005505 20 02500502 250m 02202 0005 050025 350 28022 22 2012 25082000: 0 I m. 00.2002 0.50025 .2200 050025 .8 mm $220> 228552 52005505 228552 28025 2.500 050022 20A 0502 w I v 020020.225: 008.523» .200 52005505 255502025: 005350200 .20 05205500805 00 60 .225 5052302 5:0m2200m2 28025 23:2 050022 22 000A 050 5052550: 0 I 20 00.2002 0.50225 5200555 28025 .50 0022050005 50200002005 SI: .3 .200 .0085 05:05 >05 002055 050025 3.500 28022 002 00012 :550002. 50w: 0.0 I 20 0205000800500 0050520052.. 28025 2050 :550002. 50m: 002205005 S .50 .50 50200002005 SI: .00 0:00 .550 250m 5200585 05255 050025 302 050022 22 20A 050 :50m2200m: 0.0 I 0 02050005000500 00505225 0200: 00000 305 00 00500 0000 08 52 50 @272 3 0520500 . Z 050 .0> 52005505 530250? 050025 2500 28025 2.500 050 50255585 m I w .2820 00.5050050an .200 52005505 2550005. 50m: 3 2:5 505200002 550002. 50m 28025 20.2 050025 2500 2050 50525505 v 2050201025.. 050.5203 :550005. On. 0:00 .0005 50m 52005505 002m 50m 28025 20.2 050025 202 50m: 050 500505 VI N 0:535:00 .500 0520550000054 0288QO 320002 82585 05005 25055202 20050 505200004» 2.5.50 0000 2505 22.500 wMflm no flomwaEI—onm «mfioo umwg HWEDGEEOeEH wm< I wwgm 35:205de 002005 0002 5.820 050 0500K 20.2000 .20 0050552000 0M<|N 0.55005. SUPRAGENERIC DIAGNOSES 7 The following acronyms or abbreviations are used: Specimen numbers CAS—California Academy of Science, San Francisco HSC—California State University, Hum- boldt, at Arcata LACM—Natural History Museum of Los Angeles County MCZ—Museum of Comparative Zoology, Harvard University, Cambridge SBNHM—Santa Barbara Natural History Museum SU—Stanford University UA—University of Alaska, Fairbanks UCMP—University of California Museum of Paleontology, Berkeley UCR—University of California at Riverside USNM—National Museum of Natural His- tory, Washington, DC. IGCU—Instituto de Geologia, Ciudad Universitaria, Universidad Nacional Autonoma de Mexico. Locality numbers HSC—California State University, Hum- boldt, at Arcata UCMP V—University of California Museum of Paleontology vertebrate locality UCR RV—University of California at Riverside vertebrate locality USGS M—U.S. Geological Survey verte- brate locality, Menlo Park register Others CBL—Condylobasal length of the skull, measured from the posterior surface of the occipital condyles to the anterior tip of the nasal process of the premaxillae m.y.—million years B.P.—Before present ACKNOWLEDGMENTS This study of the otarioid seals of the Neogene was begun in 1965 with the description of the female skull of Imagotaria downsi. The large collection of fossil pinnipeds from Cedros Island, Baja Cali- fornia, collected by Tedford, Frank H. Kilmer, and associates, was added to the study, as were addi- tional materials from the Santa Margarita Forma- tion in the vicinity of Santa Cruz, Calif, and other formations and localities. Several years were spent by Repenning on the study of living seals; this work was aided by Francis H. Fay and the late Richard S. Peterson in particular. Many zoologists specializing in the study of seals, in number about 75, from many countries in both the northern and southern hemispheres, were extremely helpful in providing study specimens, information, and discussion. There is a camaraderie among those who study marine mammals that the senior author has not seen matched in any other group of special- ists. Whether it derives from the smell of salt air or from the surprises of beaching a rubber raft in the surf, he is most grateful for having experienced this bond. Because paleontologists specializing in the study of seals are not so numerous as zoologists of the same bent, they can be thanked by name. To Edward D. Mitchell, Clayton E. Ray, Lawrence G. Barnes, and Q. Bret Hendey, we extend our thanks for years of free exchange of ideas and information. Mr. Rowland Taylor, formerly of Scotts Valley, Calif, kindly gave us permission to collect, for years, from an abandoned quarry on his property, USGS vertebrate locality M1035, from which two skulls and most of the postcranial elements of Imagotaria downsi were collected. Ing. Guillermo P. Salas, former director of the Instituto de Geologia of the Universidad Nacional Autonoma de Mexico, made possible the several expeditions to Cedros Island, Baja California, by the parties of the University of California at Riverside. To the institutions listed with their abbreviations in the preceding section, we extend our thanks for permission to study specimens in their care. Kenji Sakamoto, photographer, U.S. Geologial Survey, made the photographic plates; his considerable labors are greatly appreciated. Janet Brown assisted in preparing the drawings of skull restorations. Rose Trombley typed and retyped the manuscript and compiled the list of References Cited. SUPRAGENERIC DIAGNOSES Superfamily OTARIOIDEA Otarioidea Smirnov, 1908, pp. 1, 14: Gregory and Hellman, 1939, p. 313: Scheffer, 1958, p. 52: King, 1964, p. 7. Otariidae sensu lato: Mitchell, 1968, p. 1897. Characters—"Hind legs capable of being turned forward and used in terrestrial locomotion. Neck lengthened*** Skull with mastoid process large and salient (especially in males [and adults D, and with distinct alisphenoid canals.” (Allen, 1880, p. 3 [for Gressigrada]). Basal whorl of cochlea directed pos- terolaterally. Anterior process of malleus present. No 8 OTARIOID SEALS OF THE NEOGENE head developed on the incus. Internal acoustic meat- us present. Adult auditory bulla composed of about three-fourths ectotympanic ossification of the tym- panic, the entotympanic ossification largely con- fined to the formation of the carotid canal. J ugular process of the exoccipital fused to the mastoid in maturity except in the odobenids, where fusion, if it occurs, takes place in old age. Sexual dimorphism great in all members. Except for the fissipedlike enaliarctids, all otari— oids have flat to little-inflated tympanic bullae, no fossa for the origin of the tensor tympani muscle, homodont cheek teeth, and very poorly developed bony external auditory meatuses. Remarks.——The Phocoidea appear to have ori— ginated and diversified in the ancient North Atlantic basin. Enough of the fossil record is now known to be certain that the Otarioidea originated and experi- enced its major diversification in the North Pacific basin. The record is strongly suggestive that all taxa included in the Otarioidea derive from a common pinniped ancestor. Mitchell and Tedford (1973, p. 278) felt that the pinniped families should be in- cluded with their fissiped relatives in the super- family Canoidea and did not recognize the super— families Otarioidea and Phocoidea. The Otarioidea is retained in the present report in recognition of the probable common ancestry of its four included families. Family ODOBENIDAE Odobaenidae Allen, J ., 1880, p. 5: Smirnov, 1929, p. 242: Ognev, 1935, p. 322. Odobenidae: Palmer, 1904, p. 833 (emended spell- ing): Allen, G., 1930, p. 139: Scheffer, 1958, p. 84: Bobrinsky and others, 1965, p. 168. Odobaeninae: Orlov, 1931, p. 69 (footnote)——Of ques- tionable status as Orlov states, without elucida- tion, that some workers place the walruses, as a subfamily, in the Family Otariidae but does not do so himself. Odobeninae, Mitchell, 1968, p. 1897 (as a subfamily of the Otariidae). Characters—Skull lacking prominent supraor— bital process. Occipital condyles widely flaring and in most species they are low relative to the upper margin of the foramen magnum. J ugular process of exoccipital thin and platelike, remaining unfused to the mastoid at least until late adulthood. Basioc- cipital very broad and roughly pentagonal in form. Tympanic membrane large. Auditory ossicles very large. Internal acoustic meatus very wide and shal- low with almost complete separation of canals for vestibulocochlear and facial nerves. Petrosal apex greatly enlarged, broadly rounded, and flat. Bony Eustachian canal of very large calibre. Palate arched. Hypophyseal fossa broad and shallow. Cheek tooth roots simple and peglike at very early stage of evolution. In primitive tooth forms, upper cheek teeth simple with a posterointernal accessory cusp, lower teeth with weak medial anterior and posterior accessory cusps; in forms with greater tooth specialization, these features are greatly sub- dued (Odobenus) or entirely lacking (Dusignathus). Lateral lower incisor anterior to lower canine, rather than anteromedial; variably present in some genera. Symphysis of mandible deep, strong, narrowly oval in shape and articulating over the entire depth of the strong chin, completely fusing in the adults of the living genus. Angle of the mandible, where the digastricus inserts, is weak and anteriorly located beneath a point between the last cheek tooth and the coronoid process. Humerus long, for a pinniped, and slender; distal termination of the pectoral crest about in line with the medial lip of the trochlea of the distal articula- tion; anteroposterior diameter of the trochlear lip much greater than that of the distal capitulum. Radius with pronator origin at or distal to the midpoint of shaft. Ulna without flat anterior surface of a olecranon above semilunar notch; distal radial articulation distinct and elevated. Scapholunar with deep pit for articulation with magnum. Femoral head distinctly higher than greater trochanter. Proximal end of fibula not usually fused to the tibia, commonly so even in individuals of great age. Calcaneum with most prominent tuberosity on the medial side of the calcaneal tuber. Astragalus with high and essentially vertical fibular articula- tion, not flaring distally onto an extended lateral process. Remarks—Known odobenids are as old as early middle Miocene in the Pacific and early Pliocene in the Atlantic. Much has yet to be discovered regard- ing their earlier history, as it seems that the family already was sharply divided into two groups by late Miocene time. These two groups are diagnosed as follows. Subfamily ODOBENINAE Odobeninae: Mitchell, 1968, p. 1894 (as a subfamily of the Otariidae). See also “Odobaeninae: Orlov, 1931” under Family Odobenidae. Characters—Upper canines elongate. Lower canines reduced. Tympanic membrane: oval window area ratio approximating 20:1, suggesting shallow- water feeding (Repenning, 1972, p. 326). Cheek teeth peglike. SUPRAGENERIC DIAGNOSES 9 Remarks—There is one living genus and species, Odobenus rosmarus (Linnaeus), in boreal Atlantic and Pacific. The odobenines were present in the Atlantic at least by early Pliocene time. However, their record in the Pacific extends back possibly to early late Miocene. The presence of older dusig— nathines only in the Pacific, and the absence of all other otarioids in the North Atlantic, strongly sug- gest that the odobenines originated in the Pacific. Subfamily DUSIGNATHINAE Dusignathinae Mitchell, 1968, p. 1894 (as a sub- family of the Otariidae). Imagotariinae Mitchell, 1968, p. 1895 (as a sub- family of the Otariidae). Characters—Upper and lower canines of about equal size. Tympanic membranezoval window area ,ratio approximating 10:1 in some genera having sea lionlike teeth with or without multiple roots, suggesting deepwater feeding. Cheek teeth with stout peglike roots in other genera. Remarks—The extinct dusignathines seem to have been confined to the North Pacific Ocean and to the Miocene and Pliocene Epochs. Though united by abundant odobenid features of both cranial and postcranial skeletal elements and by the lack of the odobenine character of empha- sizing the upper canines while reducing the lower, the dusignathine odobenids include two distinct types. The existence of these two types was recog— nized by Mitchell in erecting the two subfamilies, one largely or entirely represented by the genus Dusig- nathus, the other largely or entirely represented by the genus Imagotaria. Dusignathus has walruslike stout peglike roots on its cheek teeth and a very large tympanic mem- brane, suggesting shallow-water habits. Imagotaria - has sea lionlike teeth and a smaller tympanic mem- brane with an area ratio to the oval window approx- imating 10:1 (comparable to the sea lions), suggest- ing deep-water habits and a fish-cephalopod diet. .With further knowledge about this subfamily of odobenids, it is reasonable to suppose that these two dusignathine types may prove sufficiently distinct to merit recognition as separate subfamilies, as was done by Mitchell, but of the Odobenidae, not of the Otariidae. Family OTARIIDAE Otariidae Gill, 1866, p. 7: Scheffer, 1958, p. 52: Bobrinsky and others, 1965, p. 166. Otariinae sensu lato: Mitchell, 1968, p. 1896 (as a subfamily of the Otariidae). Characters—Supraorbital processes present, very strong in adult males. Occipital condyles high, close together, and parallel or nearly so. J ugular process of the exoccipital knoblike and usually thick, fused to mastoid in adulthood. Basioccipital moder- ately narrow, parallel sided to trapezoidal in form. Tympanic membrane very small. Auditory ossicles of normal carnivore size. Internal acoustic meatus roundly oval to circular with little or no separation of canals for facial and vestibulocochlear nerves. Petro- sal apex pointed or bluntly pointed and relatively little enlarged for a seal. Bony Eustachian tube not conspicuously enlarged. Palate flat to arched. Hypo- physeal fossa deep and globose. Teeth consistently double-rooted until the late late Miocene; many living species still retain double-rooted upper molars (Repenning and others, 1971, p. 9). Upper cheek teeth lack any persistent posterointernal cusp. Third lower incisor anteromedial to the lower canine. Symphysis of mandible shallow and weak, broadly oval in shape, usually not articulating over the entire depth of the chin, and never fused. Angle of the mandible, where the digastricus inserts, is weak and anteriorly located. Humerus short and stout, tending to be more slender in the fur seals; anteroposterior diameter of the trochlear lip about the same as that of the distal capitulum; distal termination of the pectoral crest about in line with the midpoint of the distal articula- tion except in the fur seals. Radius short with pronator origin proximal to midpoint of shaft. Ulna with flat anterior surface of olecranon above semi- lunar notch (narrower in more ancient genera), distal radial articulation inconspicuous and united with distal articulation. Scapholunar with facet for articulation with magnum little or no deeper than adjacent facet for unciform. Femoral head about as high as greater tro- chanter. Head of the fibula fused to that of the tibia in forms younger than the late Miocene. Calcaneum with medial tuberosity on calcaneal tuber no larger than that on lateral side. Astragalus with low fibular articulation flaring distally and curving toward a paraplantar plane onto an enlarged lateral process. Remarks—Known otariids are as old as late middle Miocene in the North Pacific Ocean. Al- though never entering the North Atlantic, they have, in the past 5 m.y., dispersed around the world in the southern hemisphere. The living Otariidae are subdivided into the sub- families Otariinae (sea lions) and Arctocephalinae (fur seals) by some workers. Repenning, Peterson, and Hubbs (1971, p. 3) have suggested that the subfamilial distinction is unreal because they were unable to find any consistent osteologic differences 10 OTARIOID SEALS OF THE NEOGENE in the skull of these groups, and they regarded the presence of abundant underfur in the fur seals as a retention of a primitive otariid feature, possibly in more than one lineage. In the present study, some differences in the postcranial bones were noted between the fur seals and the sea lions. Together they offer some basis for morphologic separation of living fur seals from sea lions but are of questionable value in diagnosing fossil otariids. All these differences are here taken to represent retained primitive features in the fur seals either because they more nearly resemble the known Mio- cene and Pliocene otariids or because they are logic- ally a feature of shallow-water aquatic carnivores, the obvious source of the pinnipeds. As a conse- quence, the fur seals are presumed to most resemble the ancestral otariids, at least to the extent disclosed by the present fossil record. As discussed in Part II to follow, the fossil record of the otariids strongly suggests that the lineage leading to the Alaskan fur seal Callorhinus diverged from the otariid ancestral group before the sea lions did. Study of their endemic louse fauna and of their bacular morphology supports this suggestion (Kim and others, 1975). It appears most likely that the Subfamily Arctocephalinae, including the two living fur seal genera Callorhinus and Arctocephalus, is polyphyletic, and it is retained in this report only as a convenience in expression. The sea lions presumably evolved into their present form with the loss of abundant underfur, increase in average size—particularly body size rather than head size, increase in the rate of develop- ment of single-rooted cheek teeth, beginning of the reduction of the number of upper molars, and strengthening of the humerus, possibly in relation to larger body size. Subfamily “ARCTOCEPHALINAE” Ouliphocinae Allen, 1870, p. 23. Arctocephalinae von Boetticher, 1934, p. 359. Arctocephalini: Mitchell, 1968, p. 1897 (as a tribe of the Subfamily Otariinae s.1. but not Arctocepha- lina Gray, 1837, p. 582). Characters— Otariid seals with abundant under- fur, usually of small size and relatively large head, third (lateral) upper incisor small and with small root moderately to conspicuously oval in cross Sec- tion, upper molars usually double-rooted, M2 usually present, humerus elongate (greatest length/ least diameter averaging >5.7, see table 3) and with a pectoral crest which terminates distally toward the medial lip of trochlea, anterior part of body and the tip of the 0s penis (or baculum) transversely very narrow (Morejohn, 1975). Remarks—The proportions of the humerus tend to make the fur seals appear somewhat more like the odobenids than do the sea lions. The overall dif- ference between the two otariid types is slight, however, and individual or specific exceptions to some of the characters are known. These characters, with the exception of under- fur, are the basis for assigning all pre-late Pliocene otariids to the fur seals. The fossil record here considered is strongly supported by studies of the louse fauna and the bacula of the living otariids (Kim and others, 1975). It seems evident that the sea lions evolved out of the fur seals in very recent time, less than 3 my ago. Subfamily “OTARIINAE” Trichophocinae Allen, 1870, p. 23. Otariinae: von Boetticher, 1934, p. 359. Otariini: Mitchell, 1968, p. 1897 (as a tribe of the Sub- family Otariinae s.l. but not Otariina Gray, 1825, p. 340). Characters—Otariid seals with very sparse un- derfur, large body size, third upper incisor large with long root of round cross section, cheek teeth single- rooted or nearly so, M2 frequently missing, humerus stout with pectoral crest directed more or less toward the midpoint of the distal articulation, anterior part of the body and the tip of the os penis transversely broad, triangular to circular (Morejohn, 1975). Stirl- ing and Warneke (1971) discuss vocal and behavioral differences between the sea lions and the fur seals. Family DESMATOPHOCIDAE Desmatophocidae Hay, 1930, p. 557. Allodesmidae Kellogg, 1931, p. 227. Desmatophocinae: Mitchell, 1966, p. 39 (including Allodesmus and Dusignathus as well as Desma- tophoca. Desmatophocinae: Mitchell, 1968, p. 1839 (excluding Allodesmus and Dusignathus). ' Desmatophocinae: Barnes, 1972, p. 5 (including Al- lodesmus and Desmatophoca, excluding Dusig- nathus). Characters—Supraorbital processes lacking or very weak (no greater than found in some phocids). J ugular process of exoccipital greatly enlarged into a posterolaterally extending process distinct from, but with maturity fused to, the mastoid process. Basioc- cipital moderately broad and trapezoidal to nearly rectangular in form. Tympanic membrane small. Auditory ossicles large but otherwise not described (Mitchell, 1966, p. 6). Internal acoustic meatus very SUPRAGENERIC DIAGNOSES wide and shallow with almost complete separation of canals for the vestibulocochlear and facial nerves. Petrosal apex very little enlarged. Bony Eustachian canal of very large calibre in Allodesmus but not Desmatophoca. Palate flat and very broad posteri- orly. Mortised jugal-squamosal articulation. Hypo- physeal fossa broad and shallow. Cheek tooth roots double-rooted to peglike, evolving single-rooted cheek teeth much earlier than other otarioid families. Upper cheek teeth simple, main cusps may or may not have posterointernal accessory cusps. Third lower incisor anterior to canine. Symphysis of the mandible deep, strong and very narrowly oval in shape, strongly tending to be narrower ventrally than dorsally and not fusing in adults. Angle of the mandible, where the digastricus inserts, deep, strong, and posteriorly located below the apex of the coronoid process or posterior to this point. At present the postcranial skeleton of the des- matophocid is known only in the genus Allodesmus (Downs, 1956; Mitchell, 1966; Barnes, 1972). Relative to Desmatophoca, Allodesmus is a very specialized ' animal in the features of its skull; it is expected that when the postcranial skeletal elements of Desmato- phoca are known, they will be equally distinct. In general the limb bones of Allodesmus appear odo- benid in those characters just outlined for the Odo- benidae. Although possibly reflecting the specializa- tions of the genus, the limb bones of Allodesmus are distinguished by the following unique or otariid-like features of the otherwise odobenid-like skeletal ele- ments: , Humerus elongate with unusually (for a pin- niped) weak pectoral insertion; anteroposterior dia- meter of the trochlear lip about the same as that of the distal capitulum. Radius very elongate with a pronotor origin conspicuously proximal to the mid- point of the shaft. Metacarpals and phalanges very stout. Femur with an unusually deep trochanteric fossa and a greater trochanter which is not conspicu- ously lower than the head. Tibia short relative to living otarioids. Calcaneum sea lionlike in that the most prominent tuberosity on the calcanear tuber is lateral but marked by overall shortness and broad distal end. Remarks—Although no specimens included in this family, as here understood, are described in this report, a discussion of familiar characteristics and content is required in explanation of the removal from this family of some genera which are described and which have been considered desmatophocids in some published classifications. As here understood, the characters of this family are based entirely upon the genera Desmatophoca and Allodesmus. The 11 contents and characters of the family have been most recently reviewed by Barnes (1972). Morphologically the desmatophocids seem inter- mediate between the odobenids and the otariids in some characters. Like the otariids, they have min- imal development of petrous bone at the apex of the petrosum; thick and knoblike jugular processes of the exoccipital which, however, form distinct pos- terolaterally directed processes different than the condition found either in the otariids or the odo- benids; relatively narrow basioccipital bones; and flat to moderately arched palate. Like the odobenids, the desmatophocids have no supraorbital processes; very greatly enlarged auditory ossicles; a very wide- ly spread internal acoustic meatus (one specimen of Allodesmus, UCMP 83363, has an almost phocidlike separation of the facial canal from the vestibulo- cochlear fossa); a broad shallow hypophyseal fossa; single-rooted teeth early in their history; and deep, strong, and narrowly oval symphysis of the mandible. Evolution of the living lineages of pinnipeds is marked by a pronounced trend in reducing the length of and increasing the strength of the proximal limb elements and in lengthening the more distal ele- ments. This progressive improvement of flipper strength, leverage, and surface area is at a very primitive stage of development in the genus Allo- desmus, the most specialized of the known desmato- phocids. Family ENALIARCTIDAE Enaliarctinae Mitchell and Tedford, 1973, p. 218. Characters—Supraorbital processes lacking or very weak (no greater than found in some phocids). J ugular process of exoccipital greatly enlarged into a posterolaterally extending process distinct from,.but paraoccipital process of the temporal, which, in turn, is connected to the mastoid process by a strong ridge. Basioccipital moderately narrow and nearly parallel sided. Tympanic membrane small. Auditory ossicles of normal carnivore size. Bony Eustachian canal of small calibre. Bullae greatly inflated for an otarioid seal, smooth and flask-shaped because of well- developed external auditory meatus. Tensor tym- pani originating in a fissipedlike fossa lateral to the promontorium. Lacrimal foramen present. Carnas- sials present: P4 three-rooted with protocone, M, with talonid; M1‘2 more or less quadrate, M1 three- rooted. Remarks—Although no specimens included in this family, as here understood, are described in this report, a discussion of the ancestral position of this family is included.To date only one species has been 12 OTARIOID SEALS OF THE NEOGENE described, Enaliarctos mealsi Mitchell and Tedford from the early Miocene of California. PART I: WALRUSES Family ODOBENIDAE Subfamily ODOBENINAE Genus ODOBENUS Brisson Brisson (1972, p. 30) has been designated author of this genus under the Plenary Powers, Opinion 467, although Linnaeus used the name in his first edition of Systemae Naturae (1735, p. 59). In his tenth edition (1758, p. 38), however, Linnaeus used "Phoca rosmarus" and Brisson was the first to use Odobenus following this starting point of the Law of Priority. Opinion 467 designated Rosmarus Briinnich, 1771, a junior objective generic synonym. Type species—Odobenus rosmarus (Linnaeus, 1758). Diagnosis—In the adult, the upper canines are elongate, evergrowing tusks with globular dentine filling the vacated pulp cavity, upper incisors 1 and 2 absent or rudimentary, I3 well developed and lo- cated medial to the tusk and in line with premolars 1- 3 forming a continuous cheek tooth arcade of four peglike teeth having essentially ,equal diameter, upper fourth premolar often present as a greatly reduced and poorly rooted peg, the lower incisors are absent or rudimentary, the lower canine reduced to the size and position of a cheek tooth in line with lower premolars 1—3, mandibular rami firmly fused at the symphysis in adults. Included species—Odobenus rosmarus (Linnae- us): living in the North Atlantic, North Pacific, and Arctic Oceans. Pleistocene records are known in the Atlantic as far south as North Carolina and Paris (Ray, 1960, p. 137), Michigan (Handley, 1953), and, in the Pacific, San Francisco (R. T. Orr, written com- mun., 1968). Odobenus huxleyi (Lankester): Late Pliocene and early Pleistocene of England and Holland and possibly of the United States (Ray, 1960). This species was originally named by Lankester (1865) under the new genus Trichecodon. In 1880 Lankester concluded that the species belonged in the same genus as the living walrus, then called Triu checus. This usage has generally been followed by many European workers (Rutten, 1907; Hasse, 1909; Van Deinse, 1964) but Trichecodon has generally been retained by American workers. Allen (1880, p. 65) stated that he did not consider Trichecodon huxleyi generically separable from the living walrus (in the same year that Lankester published the same opinion) but he retained the generic name (in quotes, p. 14) for the material Van Beneden (1877) described under the name Trichecodon koninckii. Rutten (1907, p. 7) did essentially the same thing by referring to Trichecodon only with reference to Van Beneden’s material, calling huxleyi a species of Trichecus. Kellogg (1922, p. 49), however, somewhat confusingly mentions that ”the figures and descriptions given by Rutten do not warrant his conclusions on the dis- tinctness of the genus Trichecodon” and continues usage of this genus for both species T. huxleyi and T. koninckii. According to both Lankester (1865, 1880) and Rutten (1907), Odobenus huxleyi differs from Odo- benus rosmarus by its markedly greater tusk curva- ture. Rutten also discusses other cranial characters, all of which can be matched by extreme individual variations in the living walrus, as he points out. Ray (1960) further discusses the distinctiveness of the tusks and has pointed out (C. E. Ray, written com- mun., 1975) that Rutten’s material is not necessarily O. huxleyi. Genus ALACHTHERIUM DuBus Type species—Alachtherium cretsii DuBus, 1867, p. 562. Diagnosis—In theadult, the upper canines are elongate, ever-growing tusks (as inferred from the alveoli; thevtusks are unknown, or unrecognized); two well-developed upper incisors are placed medial to the anterior margins of the tusks and do not form part of the cheek tooth row; four upper premolars and one reduced molar are present and are relatively widely spaced in comparison with Odobenus. The occiput is markedly rectangular in posterior View. These features are known from Alachtherium an— twerpiensis of Hasse (1909). From Hasse’s illustra- tions, it would appear that the snout was more elongate and that the tusks were more procumbent than in Odobenus. Two lower incisors are present; the lower canine is positioned as and reduced in size to that of a cheek tooth—in line with the lower premolars 1—4, the mandibular rami are not fused at the symphy- sis but the symphysis is very elongate—extending posteriorly to the anterior margin of P4 as projected normal to the alveolar margin, the lower cheek teeth are widely spaced in comparison with Odobenus. These features are known from the type species. Included species.—Alachtherium cretsii DuBus, 1867, p. 562: Early Pliocene (Scaldisian) of Antwerp (Mourlon, 1877). The type is a mandibular ramus. Van Beneden (1877) assigned a braincase, as well as a number of isolated postcranial elements, to this species, these also ostensibly from the Scaldisian of PART I: WALRUSES Antwerp. Rutten (1907, p. 10) questioned Van Bene- den’s reason for assigning the braincase to A. cretsii, pointing out inconsistencies in the configuration of . the type mandible and of the referred braincase. He concluded that they are not the same and that the braincase should be given a new name, which he ’ provided: Trichecus (=Odobenus) antverpiensis. Alachtherium antverpiensis (Rutten, 1907, p. 12): Early Pliocene (Scaldisian) and late Pliocene (Poe- derlain) of Antwerp. Although he pointed to con- siderable differences between Trichecus huxleyi and Trichecus antverpiensis to distinguish these two species, Rutten (1907, p. 12) did not deem it advisable to propose a new genus on the basis of the cranium which Van Beneden had referred to Alachtherium cretsii. Rather, he elected to put this new species in Trichecus (=Odobenus). Hasse (1909) described a new skull, as well as a great number of postcranial bones, from the Poeder- lian deposits overlying the Scaldisian deposits of Antwerp. He mentioned Rutten’s paper in connec- tion with the features of Trichecus huxleyi, but he failed to mention Rutten’s discussionof Trichecus antuerpiensis. Although the skull described by Hasse is remarkably similar to that described by Van Beneden (from the older Scaldisian deposits) and named by Rutten, these similarities were not mentioned; rather, Hasse (p. 312) assigned his spe- cies to Alachtherium on the basis of similarity of dental formula and tooth form with the type man— dible of Alachtherium cretsii DuBus. Hasse (p. 312—313) then listed the differences be- tween A. cretsii (mandible only) and his species (fewer but not altogether dissimilar to the distin- guishing features noted by Rutten), and he con- cluded that his was a different species of Alach- therium than was the type mandibular ramus de- scribed by DuBus. He named it, rather prominently pointing out that he was the author, ”Alachtherium antwerpiensis G. Hasse, 1909.” From the published record, it is not certain that the type of T. antverpiensis Rutten, 1907, the cranium from the Scaldisian deposits included in A. cretsii by Van Beneden, is the same species as the type of A. antwerpiensis Hasse, 1909, the skull from the Poe- derlian deposits of Antwerp. At present it seems more likely that they are the same species and Alachtherium antverpiensis (Rutten) is here used. A review of the European material is needed to elim- inate the uncertainties of identification, but such a review is beyond the scope of this report. Genus PROROSMARUS Berry and Gregory Plates 2 and 5 13 Type species—Prorosmarus alleni Berry and Gregory, 1906; USNM 9343, the left ramus of a mandible. Diagnosis—"Two well developed incisors in each ramus*** [ lower] canine [reduced] but retains its primitive position [not in line with the cheek teeth] and caniniform shape*** upper jaw must have retained*** functional incisors in the adult [judged by wear on the medial side of the lower canine]* ***symphysial surfaces did not become anchylosed.” [extracted from Berry and Gregory, 1906, 444—446]. Berry and Gregory infer from the position of the lower teeth that the upper canine was not nearly so enlarged as in living Odobenus; this seems unlikely, however, as large isolated walrUs tusks have been found in the Yorktown and equivalent formations since 1906 (C. E. Ray, oral commun., 1969). These tusks strongly resemble those of Odobenus huxleyi and have globular dentine filling the pulp cavity. These characters were listed in contrast to Odobenus. The resemblances to Alachtherium cretsii DuBus, as described by Van Beneden (1877), are greater in some respects but Prorosmarus differs by having a mandible more like that of Odobenus without the extreme upturn of the jaw found in Alachtherium, which, as pointed out by Berry and Gregory (1906, p. 450), ”implies a shorter, more upturned facial region." At least in configuration of the mandible, Prorosmarus much more greatly resembles Odobenus than does Alachtherium. And the cheek teeth of Prorosmarus are closely spaced as in Odobenus. The mandibular symphysis, though unfused and elongate, does not extend nearly as far back as in Alachtherium but terminates beneath P2 . An odobenid humerus from the type area, Yorktown, Va., in the Museum of Comparative Zoology, Harvard, specimen, MCZ 7713, collected by the J. B. Woodworth Expedition, is the size of the humerus of a large male walrus as pointed out by C. E. Ray (oral commun., 1968). Although clearly odobenid, it differs from the humerus of the living walrus in that the deltoid tubercle has not migrated off of the pectoral crest but, rather, is only a prominence on the lateral margin of the crest, as on the sea lion humerus. In this respect, the Yorktown humerus is similar to that of Alachtherium. In all other respects this humerus, like that of Alachtherium, is quite comparable to that of living Odobenus, being elongate for a pinniped, having the medial lip of the trochlea of much greater anteroposterior diameter than the distal capitulum, and having the pectoral crest directed distally toward the medial lip of the trochlea rather than toward the center of the articulation. 14 OTARIOID SEALS OF THE NEOGENE Distribution—The Pliocene of western North Atlantic Ocean. Hazel (written commun., 1971), states that the Yorktown Formation exposed at the type locality belong to his Orionina vaughani ostracode assemblage zone; he considers the zone to be upper Miocene, but he suggests (Hazel, 1971, p. 8) that the upper part may be of Pliocene age. Akers (1972) states that the Orionina vaughani assem- blage zone contains foraminifers indicative of the last half of N18 or N19 zones of Blow (1969) and thus is Pliocene, between 3.5 and 5 my old. Included species.—The genus is monotypic. Genus AIVUKUS new genus Type species.—Aivukus cedrosensis new species. Etymology—The Inupik-speaking Eskimos of the Bering Strait region call the walrus ”Aivuk” and a very similar name is used by all Eskimos according to John J. Burns (written commun., 1972). Diagnosis.—In the adult, the rostrum is elongate and not massive; the upper canines are elongate and long growing but growth stops when the pulp cavity is filled with annular dentine; the canines have no globular dentine centrally; the lower canines are reduced in size but are 50 percent larger in diameter than the first premolar; the mandibular symphysis is unfused and much less sloping than in Proros- marus, terminating approximately beneath P1 and P2 . Probable dental formula 21-lC'4P'1M 21-10-4P-1M x Distribution—Late late Miocene of Baja Cali- fornia. Included species.—-The genus is monotypic. 2:32 Aivukus cedrosensis new species Plates 1-4 and 14; figure 1 Holotype.———A partial skull complete on the right side with associated fragments of the right petro- sum, an isolated I3, fragments of the left mandibular elements, distal part of humerus, parts of radius and ulna, scapholunar, and incomplete metacarpal I, IGCU 901, field No. RHT 1290, collected by David P. Whistler in 1965. Etymology—The species is named for Cedros Island, Baja California. Referred material from Cedros Island—UCR 15260, Field No. "Cedros 4," collected by Frank H. Kilmer, David P. Whistler, and George T. Jefferson in 1964: associated scapholunar, trapezoid, cuneiform, unciform, metacarpal I, metacarpal II, head of meta- carpal IV, and a phalanx. This material is about twice the size of the front limb elements of the holotype and is the basis for assuming that the type is a female. The two sizes are quite comparable to those of male and female of the living Odobenus rosmarus. UCR 15241, Field No. RHT 1312, collected by Richard H. Tedford and Whistler in 1965: male scapholunar fragment, distal articulation of ulna, unciform, and trapezium. UCR 15242, Field No. RHT 1304, collected by Tedford and Whistler in 1965: isolated female unci- form. UCR 15243, Field No. ”Cedros 4, limb C", col- lected by Kilmer, Whistler, and Jefferson in 1964: female humerus. HSC 309, collected by Robert E. Jones in 1969 from ”coffee-colored marine mammal beds,” female metacarpal III. Referred material from Baja California Sun— USNM 184045 and 184046, abraded first metacarpal and proximal portion of upper canine, collected by Sr. Felipe Moreno Aviles, Rancho e1 Refugio, Santi- ago, Baja California Sur. Diagnosis.—Only one species is here recognized in the genus Aivukus. Type locality and age—The type specimen is from UCR locality RV 7309 (Tedford locality 11) approx- imately 68 feet above the base of the upper Miocene and Pliocene(?) Almejas Formation on Cedros Is- land, Baja California, Mexico. As recognized on Cedros Island by F. H. Kilmer, the Almejas Formation (Mina, 1957) is composed of at least 800 feet of yellow and gray sandstone with a few conglomeratic beds, more abundant in the younger half of the section. The basal 200 feet contain fossil marine mammals, and one algal-rich pecten-bearing conglomerate is near the top of this interval. The younger 600 feet of the formation contain a molluscan fauna which is the basis of the following statement by F. H. Kilmer (written com- mun., 1970) on the age of the formation. The precise geologic age of the Almejas Formation on Cedros Island does not appear determinable at the present time, at least on the basis of marine invertebrate fossils. The most refined geologic age interpretation available is that of Jordon and Hertlein (1926), who studied the marine invertebrate fauna of the Almejas in detail and concluded that, with respect to other West Coast marine invertebrate faunas, it bore closest resem- blance to the fauna of the San Diego Formation and that it was assignable to the time interval from middle to early late Pliocene. Their usage of the term ”Pliocene” is essentially that of the marine PART I: WALRUSES 15 Cenozoic chronology of western North America developed by Weaver and others (1944) and modified for the southern California area by Durham (1954). The Almejas fauna may be at least partially correlative with the marine fauna of the Careaga Formation of California with which it has a number of short- ranging species in common. There appears to be no sound basis, with respect to marine invertebrates, for attempting direct age correlation of the Almejas faunas with those of the European section at the present time. On Cedros Island, the Almejas Formation unconformably overlies diatomaceous rocks of the Tortuga Formation (Mina, 1957). Near the top of the Tortuga, molds of an arcid pelecypod closely resembling Anadara obis- poana, were collected and this species is common in rocks of Luisian age in the Miocene sequence of California (Reinhart, 1942). The Almejas is unconformably overlain by terrace deposits containing a marine invertebrate fauna of probable late Pleis- tocene age (Hertlein, 1934). The referred material with field Nos. RHT 1312 and RHT 1304 was collected from approximately the same part of the stratigraphic section as was the type; those specimens with the field designation ”Cedros 4” were collected lower in the formation, from 20 to 30 feet above the basal unconformity. Thus, except for HSC 309, whose locality is un- recorded, all material known of Aivukus cedrosensis from Cedros Island was collected from 20 to 70 feet above the base of the 800-foot-thick Almejas Form- ation; the marine invertebrates from 130 to 730 feet above the base of the formation suggest a middle to late Pliocene age in the usage of Weaver and others (1944) or late late Miocene to Pliocene in the usage of this report; and the unconformably underlying Mio- cene rocks indicate that the fossil odobenid probably is not so old as middle Miocene. Howard (1971) has described the avifauna from the beds in the Cedros Island section which cbntain Aivukus cedrosensis, noting that the fauna is quite similar to that from the San Diego Formation ”al- though in no instance are the species the same.” She refers to the fauna as being of early Pliocene age, or late Miocene in the usage of this report. Barnes (1973) has described a pygmy sperm whale from these beds and infers a similar age. The presence of Dusignathus santacruzensis a- bout 50 feet above the base of the Almejas Forma- tion, as discussed in a following part of this report, suggests an age comparable to that of the Purisima Formation in the Santa Cruz area of California. This would be Pliocene, younger than the glauconite (min- imal) date of 6.7 t 0.5 m.y. from the base of the Purisima (J. D. Obradovich, written commun., 1964). In addition, fossil remains questionably referred to Dusignathus are known from the basal glauconite bed of the Drakes Bay Formation of Galloway (1977), Point Reyes, Calif. This glauconite has been dated at 9.3 i 0.5 m.y. (A. J. Galloway, oral commun., 1970); however, the glauconite sample contained detrital biotite and may be considerably younger than the date indicates. Abundant remains of a primitive fur seal, decribed in Part II of this report, are present in the lower part of the Almejas Formation. Remains of this seal have also been found in the glauconite bed of the Drakes Bay Formation and in the Purisima Formation. However, the species from these northern formations seems more modernized than the species from the Almejas Formation. Lawrence G. Barnes, currently studying the cetacean fauna of the Almejas Forma- tion, reports (oral commun., 1974) other similarities with the Drakes Bay Formation: both faunas contain a balaenopterid assignable to Balaenoptera sp., a stenodelphine, a phocoenid, and a unique delphinap- terine which suggest to him a temporal correlation. He has found the cetacean fauna of the Purisima Formation to be very similar. The single mollusk known from the Drakes Bay Formation was identified by F. S. MacNeil as Nep- tunea colmaensis (Martin), known elsewhere only from the lower part of the Merced Formation (late Pliocene) of the San Francisco area. This evidence again suggests that the Drakes Bay glauconite age is too old. The Purisima Formation of the Santa Cruz area includes, in its upper part, beds apparently equivalent in age to the lower Merced Formation (Addicott, 1969). Axelrod (1971, p. 77) mentions cones of Monterey Pine from the Drakes Bay Formation. He notes that the cones resemble those of the modern population at Monterey more closely than the cones from the population living at Waddell Creek, Santa Cruz County, Calif, a population that is much closer to the fossil locality. In all probability the unit con- taining the cones which Axelrod refers to as the Drakes Bay Formation (without definition or other indication of location) is a surficial Pleistocene deposit and not the Drakes Bay Formation of Gallo- way (J. A. Wolfe, oral commun., 1976). Axelrod’s Drakes Bay Formation is thus interpreted herein as if his wording had been "deposits near Drakes Bay.” Although pinniped remains are abundant in the basal 70 feet of the Almejas Formation, nothing has been found that resembles the primitive otariid Pithanotaria starri nor the often associated odo- benid Imagotaria downsi; these species are from deposits in California believed to be 9—12 my old. The strong fauna] similarity shown by the seals and 16 the cetaceans in the lower Almejas, Purisima, and Drakes Bay Formations certainly suggests approxi- mate contemporaneity although the somewhat more advanced fur seal from the Drakes Bay Formation and at least the upper part of the Purisima Forma- tion does suggest that these two units may be somewhat younger than the basal Almejas fauna. As an approximation based upon these similarities, therefore, the basal 70 feet of the Almejas Formation is here taken to be between 6 and 8 my old, late late Miocene. However, the overlying 700 feet of the Almejas, containing the mollusk fauna discussed by Kilmer, may well extend into the Pliocene; con- ceivably into the late Pliocene as suggested by the similarity of the invertebrate fauna to that of the Careaga and San Diego Formations. It seems that Aivukus cedrosensis is probably at least one million years older than Prorosmarus alleni, as would be expected from its less advanced morphology if the two belong to the same odobenine lineage. USNM 184045 and 184046 are from the Rancho e1 Refugio locality in the Cape region of Baja Cali- fornia Sur (Hertlein, 1925, p. 3, 7; Lindsay, 1965). The locality is about 3 miles north of Sierrita de la Trinidad and about 12 miles east-southeast of the town of Santiago (roughly one mile north of the ranch house of Felipe Moreno). The locality is in a south-facing slope of the Salada Formation and is an estimated 50 feet below the top of the formation as exposed in this slope. Age assignments of the Salada Formation in this area vary from late Miocene to late Pliocene (see Hertlein, 1925), but there appears to be general agreement that it is more or less contempo- raneous with the Almejas Formation. This locality is 8 miles south of the 'h'opic of Cancer, the southern- most record of any odobenid. DESCRIPTION Mandible (pl. 2).—Four fragments of the left man- dibular ramus of the type specimen of A. cedrosensis were collected as float washing downslope from where the skull was found. The roots of the canine and first two postcanine teeth are in place; parts of the alveolae of the third and fourth postcanine teeth also are preserved. That part of the jaw which would have shown the position of a fifth postcanine is totally missing but wear on the upper dentition suggests that a fifth was present. Enough of the ramus is preserved to indicate that if a fifth post- canine was present it was shallowly rooted, and reduction of the last upper postcanine also suggests considerable reduction in size of the lower. The condyloid and coronoid processes are missing as are OTARIOID SEALS OF THE NEOGENE all parts of the mandible anterior to the canine. However, the peglike lateral upper incisors both show two distinct facets of wear (pl. 1): a sloping facet on the lateral side from contact against the lower canine and a facet more or less transverse to the axis of the tooth; this wear clearly indicates that at least one large incisor was present in each ramus of the mandible to occlude end-to-end with the two large upper incisors; in order to occlude in this way, the lower incisor must have been essentially anterior to the lower canine. All lower teeth are distinctly single rooted and peglike: solid columns of dentine with a heavy jacket of cementum and nearly circular in cross section. Unlike Prorosmarus alleni, the lower canine of A. cedrosensis is distinctly larger in cross section than the postcanine teeth, though conspicuously smaller than the upper canine. It is approximately 14.5 mm in anteroposterior diameter at the alveolar margin; .the first postcanine is 9.5 mm, the second 8.8 mm, the third about 10 mm, and the fourth very approxi- mately 7.8 in the same dimension. The first post- canine tooth, though of large diameter at the alveo- lar margin, is the shortest, its root terminating very close to that of the relatively elongate canine. The canine itself extended backward in the jaw to a point‘ between the second and third postcanine roots. What is preserved of the mandibular ramus ex- hibits a remarkable similarity to Prorosmarus alleni. Behind the symphysis the ramus is conspicuously bowed outward and has a straight inferior margin back to the bladelike angular process. The inferior margin is narrow and forms a rather sharp crest which, anteriorly, begins to drop downward toward an inferior genial eminence, most of which is lost on the specimen. What is preserved of the symphysial surface in Aivukus cedrosensis indicates that the articular area was less inclined than in P. alleni, possibly in adjustment to the larger canine tooth. The alveolar distance from the anterior margin of the canine to the posterior margin of postcanine IV is approximately 65 mm, the comparable measurement on P. alleni is 104.5 mm. Except for its smaller size the mandible of A. cedrosensis differs from that of P. alleni by having a relatively much larger canine, presumably by retaining a reduced fifth postcanine tooth, and by having a less procumbent symphysial region. These differences are interpreted as repre- senting a more primitive stage of evolution within the Odobeninae. Skull (pls. 1 and 2, fig. 1).—The skull of the type FIGURE 1.—Restoration of the skull of Aivukus cedrosensis. Type female skull IGCU 901. A, Dorsal view. B, ventral View. C, Lateral view. PART I: WALRUSES 17 ‘ ’l .2777” Milk 1Q“! ~ ‘ Wm 18 OTARIOID SEALS OF THE NEOGENE specimen of Aivukus cedrosensis has been some- what distorted and fragmented by weathering. The occipital and basioccipital regions are lost, the brain case is partly crushed, and separation of the front and back parts of the skull in the interorbital region has left some question about proper orientation. Nevertheless, most of the right side is present and it is remarkably well preserved when compared with most fossil pinniped skulls. Judged by suture closure and tooth wear, the skull is from a young adult animal. In general proportions, the skull is quite sea lion- like but in nearly all details it is immediately recog- nizable as an odobenid. A portion of the occipital crest is preserved indicating a low, broad structure as on living Odobenus; however, no trace of a sagittal crest is preserved on the shattered and depressed roof of the brain case, indicating that the braincase was inflated as in the living genus. The lack of a supra- orbital process is conspicuous; a prominent pre- orbital or lacrimal process is present in front of the large orbit; and, walruslike, the maxillary-frontal suture passes through this process and, in a straight line, upward to the nasals. The nasals themselves are not preserved, although the fractures of the skull suggest that their posterior margins, articulating with the frontals, may have formed a straight line continuous with the maxillary-frontal suture. The infraorbital foramen is immense. The palate is con- spicuously vaulted. The area between the glenoid fossa and the mastoid is completely filled in with thick ossification of the ectotympanic. The mastoid process itself is knoblike, rather than crestlike, and a thin, platelike jugular process from the missing exoc- cipital still adheres to its posterior surface. Consistent with a much more elongate skull than in living Odobenus, the interorbital area is narrow and elongate, the ascending ramus of the premaxilla is elongate, and the zygomatic processes of both the squamosal and the maxilla are elongate, though the jugal itself is quite short. Although the nasal processes of the premaxillae are quite prominent, they protrude anteriorly from just above the alveolar margin rather than being well elevated above this level like those of the square- faced modern Odobenus. The palatine fissures of the Odobenus premaxillae are extremely reduced, but they are large and well developed in Aivukus cedro- sensis. Each premaxillae holds two, possibly three, incisors. The medial one or two are small, and their presence on the skull is indicated only by alveoli whose preservation is not distinct. If two medial incisors did exist, the smallest was directly anterior to the other, a situation which seems improbable judging by the dentition of Odobenus. The larger, or probably only, medial incisor has an alveolar di- ameter of about 3.5 mm. The very large lateral incisor, though more sharp- ly curved and tapering to the root, is about equal to the cheek teeth in size and peglike shape. Its greatest diameter, approximately at the alveolar margin in the type skull, is 14.8 mm; this measurement includes a cementum jacket of about 1 mm thickness. From the nearly closed root to the worn crown, the greatest curvilinear length is 60 mm. N o enamel remains on the worn crown. As has been discussed, the wear surface on the crown has two facets (fig. 1B; pl. 1 fig. 10): one lateral and oblique resulting from wear against the lower canine, the other transverse to the long axis of the tooth resulting from attrition with an unknown but presumably equally peglike lower in- cisor. . The maxilla contains an elongate canine and five postcanine teeth; all are single-rooted pegs except the last, which has a transversely compressed root that might be weakly double-rooted, although the roots are not entirely exposed. The canine is the only tooth in the skull which was still growing at the time of death, and it had not begun to constrict toward root closure. However, it does not appear to have been evergrowing but rather has a shallow conical pulp cavity quite similar to that of the cheek teeth in living Odobenus. No globular dentine is developed. The canine is encased in a heavy cementum jacket that is 2 mm thick where the canines are broken off at the alveolar margins. There is no way to judge how far the canines protruded from the skull; from the alveolar margin to the fully open lip of the pulp cavity, they are about 54 mm long with a uniform anteroposterior diameter of 20 mm. In transverse diameter they are slightly narrower, measuring 18 mm. The maxillae are conspicuously swollen, both externally and particularly internally, nearly to the preorbital processes to accommodate these elongate canines. USNM 184045 is the root of a more mature canine whose pulp cavity is closed. This tooth also is of large diameter, possibly from a male. It has a maximum anteroposterior diameter of 27 mm. Weak longi- tudinal fluting is evident at the cementum-dentine interface and is present, less distinctly, on the ex- terior surface of the cementum. The dentine is com- posed entirely of annular layers and no globular dentine is present. The roots of all postcanine teeth except the second and the fifth are exposed on one side or the other of the holotype; all are essentially closed. The first four cheek teeth are slightly curved, cementum-encased PART I: WALRUSES pegs with an approximately uniform diameter down to the abruptly pinched off root termination, the first three are of approximately equal diameter (about 12.5 mm) and decrease slightly in length to the rear (the first is 43.5 mm long, the third 36.5 mm); the fourth is of noticeably smaller diameter (10 mm). The fifth (last) cheek tooth is not peglike; it has a worn but distinct enamel crown with an almost-doubled shallow root. No cementum is evident on the fifth postcanine (M1). Two distinct wear facets appear on the first three postcanines: the larger is slightly convex and slopes dorsomedially toward the palate and the smaller is concave and is on the posterior surface. The posterior facet is not evident on the fourth postcanine and an appreciable amount of the thin enamel crown remains on the labial surface. Abrasion on the very small fifth postcanine is essentially horizontal, and a thin rim of enamel is present almost entirely around the base of the crown. From the labial side of the crown of the fifth post- canine, an emargination of the enamel edge crosses half of the preserved crown, suggesting that the crown was originally bicuspid, or perhaps bilophid as in the M2 of Imagotaria downsi (see following section). By analogy with the tooth wear in living walrus (F. H. Fay, written commun., 1973), the posterior con- cave wear facet appears to have resulted from end-to- end occlusion with the lower teeth while the lingual convex wear facet seems most likely to have resulted from anterior-posterior motion of the tongue when abrasive material, such as mud, was in the mouth. The similarity of the wear facets and of the vaulted palate to those of living walrus suggest that Aivukus cedrosensis, like the living walrus, utilized a tongue- piston type of sucking to extract mollusks from their shells. The ear region of Aivukus cedrosensis is extremely walruslike. The lateral wall of the right middle ear cavity is preserved on the skull and a large part of the right petrosum was recovered from the fragmented debris. The external acoustic meatus is very large in diameter, larger than in most specimens of Odo- benus, and very closely approximates the size of the tympanic membrane. Only the lateral part of the ectotympanic ossification of the bulla is preserved, but from this it can be seen that the stylomastoid foramen, tympanohyoid fossa, greatly enlarged bony Eustachian canal, and thick bullar wall be- tween the glenoid fossa and the mastoid process are all developed and located very much as in living Odobenus. Within the middle ear cavity, the well- preserved crista tympani shows that the tympanic membrane was very large, measuring 10.8 mm in 19 greatest diameter from the tympanic incisure ven- tromedially to the inward projecting lip of the crista and 6.9 mm in least diameter. No auditory ossicles were recovered but the epitympanic recess is very large, indicating the former presence of ossicles as large as in the living walrus. Although not so enlarged as in living Odobenus, the petrosum of Aivukus cedrosensis is larger than that of any otariid or desmatophocid and is char- acteristically odobenid in the great enlargement of the apex anterior to the promontorium and in the broad internal acoustic meatus, which has wide sep- aration of the passages for the facial and vestibulo- cochlear nerves (pl. 1, figs 6, 7). The region contain- ing the semicircular canals and the floccular fossa they encircle is not preserved. Approximately one- quarter of the oval window of the cochlea is pre- served on the petrosum, but the great uniformity in proportionate shape of this foramen permits an estimate to be made of its size, approximately 2.4 mm and 1.7 mm in maximum and minimum diameters. This size suggests an oval window to tympanic membrane area ratio of 1:28; this ratio is greater than that measured on living Odobenus (Repenning, 1972, p. 321) and approaches the minimum reported for fissiped carnivores. Among the living pinnipeds, only the phocid genus Erignathus has a greater ratio. This ratio suggests quite shallow-water feed- ing habits (Repenning, 1972, p. 322) and, in com- bination with the emphasis on the elongation of the upper canines only, is the basis for the assignment of Aivukus cedrosensis to the subfamily Odobeninae as diagnosed earlier. The elongate facial region, alinement of the canine with the upper cheek teeth, five functional cheek teeth, large external acoustic meatus, great oval window to tympanic membrane area ratio, and less enlarged petrosal apex all are features to be expected in a primitive odobenine. From the apparent greater specialization of Prorosmarus alleni, in greater re- duction of the lower canine and presumed loss of one lower cheek tooth, it could be inferred that this species had greater enlargement of the upper canine, loss of the fifth upper postcanine, and presumably a more Odobenus-like ear region than does A. cedro- sensis. With approximate compensation for lost parts, the type skull of Aivukus cedrosensis has the following measurements: mm Condylobasal length ...................... 2951 Greatest zygomatic width ............ 175: Length of cheek tooth row .......... 77 20 OTARIOID SEALS OF THE NEOGENE TABLE 3.—Proportions of the humerus in otarioids Number of Average ratio A1 Range of Average ratio B2 Range of Specimens Taxon Measured (Trochlea/capitulum) ratio A (Length/least diameter) ratio B Arctocephalus Pusillus ------------------------ 4 1.01 1.05—1.00 5.77 6.10—5.17 Arctocephalus townsendi .. _____ 1 .97 6.46 Callorhinus ursinus ................. 4 .97 1.0—.922 6.41 6.66—5.95 Fur seals ................................................... 9 .98 6.21 Zalophus ---------------------------------------------------- 3 1.03 LOB—.960 5.20 5.47-4.92 Neophoca ........... 2 1.05 1.06—1.03 5.63 5.75—5.51 Phocarctos ...... 4 1.01 1.07—.98 5.63 5.95—5.17 Eumetopias . ______ 3 1.06 1.13-1.01 5.67 5.70—5 62 Sea lions ____________________________________________________ 12 1.03 5.53 Odobenus .................................................. 11 1.18 1 18-139 7.04 8 10—626 Aiuukus cedrosensis type ................................ 1 1.34 ---- Cedros 4 ----- 1 1.21 6.90 ?Prorosmarus alleni ..... 1 1.28 7.32 ALachtherium ............................. 1 1.27 6.15 “Trichecodon” koninckii ..................... 1 1-21 lRatio A : greatest anteroposterior diameter of trochlear lip divided by that of distal capitulum. 2Ratio B : greatest length of humerus divided by least transverse diameter of shaft. Palatal width between canines .. 46 Palatal width between fifth postcanines .......................... 32i Depth of palate from alveolar margins at postcanine 2 .......... 22 Minimum interorbital width ........ 47.+. Greatest diameter of infra- orbital foramen ............................ ’21 Postcranial skeleton.—Except for a few species with obvious peculiarities in some postcranial skele- tal elements, such as the sharply upturned medial epicondyle on the humerus of Neophoca cinerea, it is commonly difficult to identify living otarioid species on the basis of postcranial skeletal elements without knowledge of geographic occurrence and differences in size. For example, several species of the genus Arctocephalus are difficult or impossible to identify on the basis of postcranial elements, and it is even difficult to separate the humeri of male Zalophus, female Eumetopias, and perhaps female Phocarctos: genera which are quite distinct in their skull morph- ology. Specific identity of postcranial skeletal ele- ments becomes even less certain in fossil otarioids because of the added problems of the inability to know individual variability or to interpret ade- quately the stage of evolution. At the familial and subfamilial levels, however, most postcranial skeletal elements are diagnostic. In the following description of the postcranial skele— ton of Aivukus cedrosenis, emphasis is placed on those features that identify the species as an odo— benine odobenid and upon the stage of evolution of the species represents within this family. At present, no features of the postcranial skeleton are recogniz- able as being specifically significant without some knowledge of geologic age and geographic distri- bution‘. In the material recovered of the type specimen of Aivukus cedrosensis is the distal half of the right humerus. At another locality, Cedros 4, a left humer- us (UCR 15243) was collected which is essentially complete and which is identical in all features pre- served to the partial humerus belonging to the type. The crest above the external epicondyle, referred to (in phocids) as the supinator ridge by King (1966, fig. 3), is largely lost in the more nearly complete speci- men and its configuration is better shown on the partial humerus of the type (pl. 3). As in all odobenids, and differing from all otariids, the distal articulation of the humerus of A. cedro- sensis is rotated inward a few degrees such that the antebrachium would be more medially directed than in the otariids. Correlated with this, the greatest anteroposterior diameter of the medial lip of the trochlea is conspicuously greater than that of the distal capitulum, as in the odobenids and differing from the condition in the otariids in which these diameters are about the same (table 3). As in all odobenids, and differing from the otariid sea lions but overlapping the condition found in the otariid fur seals, the shaft of the humerus of A. cedrosensis is relatively slender (table 8), the pec- toral crest is directed distally toward the medial lip of the trochlea rather than toward the midpoint of the distal articulation, and the proximal process of the pectoral crest, the greater tubercle, is narrow in proximal aspect. In the living genus Odobenus, the humerus is as long as the ulna, rather than being PART I: WALRUSES ‘ considerably shorter as in the otariids; the Odobenus condition appears to exist in A. cedrosensis, al- though the incompleteness of the preserved elements casts some doubt on this observation. In contrast to living Odobenus, but similar to other fossil odobenids, the deltoid tubercle is an integral ‘ part of the pectoral crest (as in the otariids) rather than being detached and positioned posterolateral to the crest. The pectoral crest itself is rather unusual in that it drops abruptly to the shaft at its distal termination rather than gradually merging with the shaft as in nearly all odobenids; in this respect the * humerus of A. cedrosensis resembles that of "Tri- checodon” koninckii Van Beneden (1877, pl. 7, fig. 2) and the otariid humerus. In addition, the medial epicondyle is prominently hooked upward as in Odobenus, rather than extending more or less direct- ly medial. Though much larger, the odobenid humerus from the Yorktown Formation of Virginia (MCZ 7713), presumed to be Prorosmarus alleni, is identical in every respect to that of A. cedrosensis except that the pectoral crest merges distally with the shaft in a gradual slope, as in modern walrus. Though some- what less robust, the humerus of Prorosmarus strongly resembles the humerus of Alachtherium cretsii (Van Beneden, 1877, pls. 3 and 4) from the Scaldisian of the Antwerp Basin. Though the humerus of Aivukus cedrosensis is easily separated from that of modern walrus, the partial radius and ulna found with the type specimen , are indistinguishable from these bones of a modern female walrus, at least in those parts preserved. The prominent and elevated distal articular surface for the radius on the ulna and the conspicuously distal position of the insertion for the pronator teres on the radius distinctly mark these bones as odobenid (pl. . 2). The distal end of another ulna (UCR 15241) also shows the distinct radial articulation but in size compares to a small male Odobenus. The female scapholunar (radial carpal, pls.2 and 4) of the type skeleton and two male scapholunars (UCR 15260) and (UCR 15241) are identical to that of Odobenus and are marked as odobenid by the pock- eted articular surface for the magnum. No differ- ences in their structure can be found that are not duplicated in a sample of 16 scapholunars from the living walrus. As in living walrus and ”Tricheco- don” (Van Beneden, 1877, pl. 8, fig. 9), the pitlike articulation for the magnum does not extend nearly as far in a palmar direction as does the articular surface for the unciform. One trapezium (distal carpal I, pl. 4), from a male to judge by its size, is known (UCR 15241). The trap- 21 ezium of A. cedrosensis differs from that of Odo- benus by lesser development of the radial-dorsal rugosity for partial insertion of the abductor pollicis muscle, conspicuous enlargement of an ulnar-dorsal process which extends the articular surface for the trapezoid proximally and which produced a pro- nounced concave articulation for metacarpal I dis- tally. In ulnar view, the facet for articulation with the trapezoid is distinctly more elongate than on the trapezium of Odobenus. One trapezoid (distal carpal II, UCR 15260) was found with other associated male flipper elements. In agreement with the differing articular surface on the trapezium, the trapezium facet on the trapezoid of A. cedrosensis is narrower than is that of living Odo- benus. No magnum (distal carpal III) has been found. Three unciforms are known, one with the large male manus (UCR 15620), one with the small male material (UCR 15241, pl. 4), and an isolated unciform comparable in size to that of a female Odobenus (UCR 15242). Other than in size, these three unci- forms are identical in form and are considerably more elongate in the palmar-dorsal dimension than in Odobenus, although this dimension is quite vari- able in the living walrus. One cuneiform (UCR 15260, pl. 4) is with the large male manus and falls well within the range of form and size variation of this bone in the modern walrus. Metacarpal I is known from the type specimen (pl. 1) and UCR 15260 (pls. 3 and 14) on Cedros Island and from the Rancho el Refugio locality. In size these bones compare closely with the same bone from male and female walrus. In form they differ from this bone of walrus in that the proximal articulation curves onto the dorsal surface, matching the differences noted in the trapezium, and in the relatively slight degree of palmar flattening and broadening of the proximal part of the metacarpal shaft due, in modern walrus, to exaggeration of the insertional area for the functionally important abductor pollicis muscle. These differences are comparable to those noted between the trapezium of A. cedrosensis and that of living Odobenus. As in the modern walrus, the dorsal surface is marked, near its proximal end, by a prominent rugosity for insertion of the extensor pollicis muscle. Metacarpal II and the proximal articulation of metacarpal IV are also present with the flipper elements UCR 15260. Though very similar, these metacarpals differ from those of living Odobenus. The articular surface for metacarpal III on the ulnar side of metacarpal II is a flat and roughly circular surface in A. cedrosensis (pl. 4), whereas it is a 22 convex triangular surface in this bone of modern walrus. As that part of the proximal articular surface of metacarpal IV which contacts metacarpal V is much more salient in A. cedrosensis than in Odo- benus, the head of metacarpal IV in either dorsal or proximal aspect is more nearly an equilateral tri- angle than in Odobenus. HSC 309 is an isolated metacarpal III, distin- guished from that of Odobenus by a flat, elevated, and circular facet for articulation with metacarpal II; this shape is consistent with the difference noted in metacarpal II (pl. 4). Metacarpals II, III, and IV are conspicuously more similar to those of the dusignathine odobenid Imago- taria than to these bones in Odobenus. The inter— metacarpal articulations on these elements in Odo- benus are broad convexly rounded surfaces but are less extensive, flatter, better delimited and, in some cases, protrude from the head on salient platforms in Imagotaria. Odobenus—like broadening and round- ing of the articular surfaces between metacarpals III and IV are evident in Aivukus, but the surfaces between metacarpals II and III retain the presum- ably primitive prominence seen in Imagotaria. Except for an isolated terminal phalanx, no ele- ments of the posterior limb of Aivukus cedrosensis have been recognized in the collection. This phalanx is identical to the second, third, and fourth terminal phalanges of the pes of Odobenus in development of both the attachment for the cartilaginous extension and the bony core for the horny claw. DISCUSSION The skull and forelimb of Aivukus cedrosensis are obviously odobenid; some of the skeletal elements are inseparable from those of living Odobenus ros- marus. Characters that mark A. cedrosensis as being less advanced than modern walrus are retention of two, or possibly three, incisors in each premaxilla and at least one functional incisor in each ramus of the lower jaw, of five upper and probably five lower postcanine cheek teeth, of obviously less specialized upper and lower canines, of the less enlarged petro— sal apex, of the deltoideus insertion on the humerus, and of the insertion of the abductor pollicis on the trapezium and metacarpal I. Characters that mark A. cedrosensis as being more primitive than Proros- marus alleni are the larger lower canine, the canini-- form (rather than tusklike) upper canine lacking globular dentine, the probable retention of a reduced fifth lower post-canine, and the less procumbent symphysial region. The criterion (cited above in ”Suprageneric Diag— noses”) of enlarging the upper canine while reducing OTARIOID SEALS OF THE NEOGENE the lower canine places Aivukus cedrosensis in the odobenid Subfamily Odobeninae by definition. At this stage of discussion, this definition is arbitrary. In the discussion of the odobenid Subfamily Dusig- nathinae, the osteologic characters distinguishing the two subfamilies are further described. It will be seen that many of the features of A. cedrosensis which mark it as less advanced than modern walrus are even more pronounced in the dusignathine odo- benids. Nevertheless some genera of the Dusignath- inae have paralleled in their evolution the obvious specializations of the living walrus to a greater extent than A. cedrosensis while retaining those primitive features which mark them as an extinct side branch of the odobenid evolutionary lineage. The most notable of the primitive features is the maintenance of an unreduced lower canine while enlarging the upper. Subfamily DUSIGNATHINAE Genus IMAGOTARIA Mitchell Type species.—Imagotaria downsi Mitchell, 1968. Diagnosis—A generalized dusignathine odobenid (see section on ”Suprageneric Diagnoses”) without specialization of the tooth crowns but with a strong tendency to fuse the roots of the cheek teeth. Dental formula: 31'1C'4P' 2M 21‘1C'4P'1 or 2M Approximate area ratio of oval window:tympanic membrane = 1:10, comparable to living sea lions and deep-diving phocids. Distribution—The late middle and early late Mio- cene of California; by estimation 9—12 m.y. ago. x 2 = 36—38. Imagotaria downsi Mitchell, 1968 Plates 4-15; figure 4 Holotype.—SBMNH 342, parts of an adult male skull and associated anterior postcranial fragments described by Mitchell (1968). From diatomite of the late middle Miocene part of the Sisquoc Formation, Great Lakes Carbon Co. quarry, about 7 miles south- east of Lompoc, Calif. Referred material from the Santa Margarita F orm- ation, Santa Cruz, Calif.—UCMP 88459, young male mandibular fragments and two small fragments of the skull described by Barnes (1971). Barnes declined assignment of this specimen to species, calling it Imagotaria sp., because the lower premolars 2—4 are distinctly two-rooted, whereas in the type of I. down- si these teeth have single, bilobed roots with a strong lateral sulcus. As it is from the same area and formation as other material here described which PART I: WALRUSES has fused roots, it seems more probable that one species was present which had a high degree of individual variability in root fusion rather than that two very similar species of one genus lived in the same area at the same time. Granite Rock Co. quarry, Olympia, Calif, locality UCMP V-70184. ‘ USNM 23858, female skull and mandible, essen- tially complete except that the nasals are missing. Skull is crushed dorsoventrally. Parts of the maxil- laries, frontals, symphysis of the left mandibular ramus, and some teeth are missing. Four vertebral fragments were found with this specimen and dissec- tion revealed the malleus and incus of both ears. Collected by W. W. Derryberry and C. A. Repenning in the fall of 1963. Abandoned quarry beside Glen Canyon Road near junction with Redwood Drive on the property of Rowland Taylor, formerly of Scotts Valley, Calif, locality USGS M1035. USNM 184060, juvenile male skull without man- dible, essentially complete except that the nasals are missing. The elements of the skull were unfused and are somewhat displaced at time of death; the left side is distorted by crushing. Several teeth are missing. An incomplete scapula was found with the skull. Collected by G. V. Morejohn and students from the Moss Landing Marine Laboratories on April 25, 1973. Same locality as the female skull but about 3 feet lower in the section and 70 feet south along the strike of the beds, locality USGS M1035. USNM 23859, male right front limb without scapula, proximal part of humerus, trapezium, and some phalanges. Found with the proximal phalanx of the first digit of the left manus. All found within 10—15 feet of the female skull, USNM 23858, in the same stratum and excavation. At least four individ- uals are represented by material from this single 15- foot excavation, a fifth from the same horizon was found 100 feet away. This material represents a concentration of fossils comparable to the number of dead animals found on beaches in modern rookeries. The elements of this limb were in articulated position in the matrix. Collected by Repenning and W. W. Chamberlain, 1966. Locality USGS M1035. USNM 23860, immature right metacarpal IV, female(?), found with two phalanges in the same stratum as USNM 23858 but about 100 feet away. Collected by Repenning and J. C. Clark, 1964. Local— ity USGS M1035. USNM 23861, left metacarpal III, male, found about 20 feet lower in the section than USNM 23858 at the same locality. Collected by Richard Baker of Santa Cruz, Calif. 1966. USNM 23862, male right calcaneum from about 10 feet lower in the section than USNM 23858 at the 23 same locality. Collected by Repenning and Cham- berlain, 1966. USN M 23863, left patella and tibia lacking both articulations, large but of uncertain sex, from about 13 feet lower in the section than USNM 28358 at the same locality. Collected by Repenning and Cham- berlain, 1966. USNM 23864, small left tibia, lacking the prox- imal articulation, from about 18 feet lower in the section than USNM 23858. Collected by Repenning and Clark, 1964. USNM 23872, small atlas from 10 feet lower in the section than USNM 23858 and about 5 feet from USNM 23862. Collected by Repenning, 1967. USNM 23865, female right humerus; isolated. Collected by Repenning and Chamberlain, 1966. From palisades along Branciforte Drive, Santa Cruz, Calif, locality USGS M1106. USNM 23866, male(?) left calcaneum; isolated. Collected a few feet from right trapezium USNM 23875 by Repenning, Clark, and L. C. Smith, 1965. Santa Cruz Aggregate Co. quarry north of Bean Creek, Scotts Valley, Calif, Locality USGS M1104. USNM 23867, female left calcaneum and astra- galus; associated. Collected by Repenning and Clark, 1965. Moore Creek, Santa Cruz, Calif, locality USGS M1108. One vertebra of the giant salmon Smilodonichtys was found associated with these ele- ments. USNM 23868, male mandibular fragment show- ing symphysis, canine, and I 3 alveolus; associated with a weathered and incomplete metatarsal and isolated canine tooth. Collected by Repenning and Clark, 1964. East side of Bean Creek north of Scotts Valley, Calif, locality USGS M1037. USNM 184055, very large male left metacarpal V. Collected by Clark, 1968. Same locality as the last mentioned. USNM 23870, female left humerus and left femur; associated with other fragments presumably of the same individual. Collected by Repenning and Clark, 1965. Nelson Road, Mission Springs, Calif, locality USGS M11015. USNM 23875, right trapezium; isolated; found in same gravel bed and a few feet from calcaneum USNM 23866 at USGS vertebrate locality M1104. Collected by Repenning, 1965. USNM 184061, right cuboid, possibly female, found 15 feet lower in the section than the female skull at the same locality, USGS M1035, by Repen— ning and Morejohn, April 24, 1973. UCMP 108066, proximal half of a left metatarsal III, from locality USGS M1035 (=UCMP V—6857) found May 5, 1973, by G. McCafferty. 24 UCMP 102854, head of left metatarsal III lack- ing dorsal articular surface for metatarsal IV from locality UCMP V—71197, 1900 feet south of M1035, found December 3, 1972, by J. Foote. UCMP 107752, right female navicular, from locality USGS M1035 (=UCMP V—6857) found March 11, 1973, by J. Lee. Mr. Gerald Macy of Felton, Calif., has collected four female-sized bones, possibly belonging to one individual, from a locality along Zayante Road, USGS M1243. These are USNM 184084, radius; USNM 184085, astragalus; USNM 184086, trape- zoid; and USNM 184087, proximal half of meta- carpal III. The first three are shown on Plates 13 and 14. Material from other localities.—UCMP 34789, proximal half of right metatarsal III, probably male, from UCMP locality V—3916, White-Seale locality below gray sandstone member of the Santa Marga- rita Formation, about 1/2 mile up Comanche Creek from mouth of canyon in middle of slope on west side of canyon, Kern County, Calif; up section from Comanche Point local fauna, probably early but possibly late Clarendonian mammalian age, late middle Miocene, approximately 11-12 my old but possibly somewhat younger; collected by K. A. Rich- ey in 1939. UCMP 24221, distal end of right tibia described by Kellogg (1925a, p. 93—95), probably male, from UCMP locality 3545, ”on opposite side of canyon from Quarry 9 [type' locality of Pithanotaria starri ]***1.5 miles south and east of Lompoc, Santa Barbara County, California,” in the same formation as the type specimen of Imagotaria downsi but about 5 miles west of the type locality. USNM 13487, immature skull from the Celite Co. Quarry No. 38, 2.6 miles south and east of Lompoc from within 45 feet of the top of the deposit (see Mitchell, 1968, p. 1865). UCMP 24070-82, right hind flipper of early late Miocene age from the lower part of the Towsley Formation, south of Humphreys, Soledad Canyon, Los Angeles County (not San Diego County), Calif. Kellogg (1925b) described this specimen under the name Pontolis cf. magnus. Diagnosis.—Only one species is here recognized in the genus Imagotaria. See Mitchell, 1968, p. 1845. Type locality and age—See Mitchell (1968, p. 1845). The locality is in the Dicalite quarry of the Great Lakes Carbon Co. 7 miles southeast of Lom- poc, Calif. The diatomite in this quarry is the same unit as that in the Celite quarry of the Johns Manville Co. 3 miles south of Lompoc (Dibblee, 1950); the diatomite has been considered to be a OTARIOID SEALS OF THE NEOGENE facies of both the Monterey Formation and the Sisquoc Formation. Because of a regional uncon- formity separating the diatomite from the under- lying Monterey, here of Mohnian age, Dibblee con— siders the diatomite to belong to the Sisquoc. Worn- ardt (1967, p. 11) noted a marked change of the diatom flora in the upper part of the 1,000-foot-thick diatomite and suggested that the Monterey-Sisquoc formational boundary occurs in the upper part of the diatomite. Based upon Foraminifera in the underlying and overlying units, most workers (Bramlette, 1946, p. 212, and Woodring and Bramlette, 1950, p. 101) consider the diatomite in these quarries to be early Delmontian. More recent correlations have indicated that it is Mohnian (Barron, 1976). Both Imagotaria downsi and the primitive otariid Pithanotaria starri have been found in the diatomite of the Lompoc area, and these two seals are known in a number of other areas associated with a ”Margaritan” or ”Jacalitos" invertebrate fauna and with a Clarendonian land mammal fauna. John A. Barron (written commun., 1974) examined the diatomite from the matrix of the type of Pithanotaria starri and from a referred skull of Imagotaria downsi (USNM 13487) and found the diatoms to belong to Schrader’s North Pacific Di- atom Zone XI. The age of these fossils from the Sisquoc Formation of the Lompoc area are here estimated to be 9 or 10 my old, or early late Miocene. Age of the referred material from the Santa Cruz area.—All remains of Imagotaria downsi in the Santa Cruz area have been found in the upper 100 feet of the Santa Margarita Formation, which is as much as 430 feet thick in some parts of the area. Many remains have been found in a widespread conglomerate in the upper part of the Santa Marga- rita in the northern parts of its outcrop area. This bed does not appear to be present to the south at the locality where the two skulls were found. The forma- tion at this locality (USGS M1035) is only about 80 feet thick, and the bones of Imagotaria have been found throughout a 30-foot zone approximately 15— 45 feet below the overlying Santa Cruz Mudstone of Clark (1966). The Santa Margarita Formation here rests on granite and thins to 5 feet against this underlying granite 4,500 feet to the north and to 2 inches 6,500 feet to the north of the locality (J. C. Clark, oral commun., 1973). The implication of the stratigraphic relations in the vicinity of this locality is that the formation was deposited against the south side of an ancient granitic high and that the 30 feet of fossil-bearing deposits represent a facies, thickened because of its proximity to the highland, of the wide- spread conglomerate found in the thicker sections to PART I: WALRUSES 25 the north of this local granitic mass. In most areas the Santa Margarita Formation is of late middle Miocene age in the chronology used here, and is a correlative to the Margaritan Stage of Corey (1954). This is true of the lower part of the formation in the Santa Cruz area (Mitchell and Repenning, 1963, p. 9, 12—14). The upper part of the formation in that area, however, contains several mollusks and echinoids generally considered indicative of an early late Miocene age (Clark, 1966); these fossils are characteristic of beds correlative with the lower part of the J acalitos Formation of former usage (see table 1). Clark (1966, p. 126—131) found his most definitive invertebrate fauna down section from the gravel bed bearing many of the remains of Imagotaria. The desmostylian remains reported from this formation (Mitchell and Repenning, 1963, p. 9, 14—15) have been found only well down section in the basal 100 feet or less of the formation and in those areas where the formation approaches 300—400 feet in thickness. In the Santa Cruz Aggregate Co. quarry on the north side Bean Creek, Scotts Valley (figs. 2 and 3, localities USGS M1104 and UCMP V4004), the San- ta Margarita Formation is nearly 400 feet thick and rests on the Monterey Formation of middle Miocene age in this area. Remains of Imagotaria have been found in the conglomerate bed here, very close to the top of the Santa Margarita Formation (locality USGS M1104). Two hundred and eighty-five feet lower in the formation the type specimen of the sirenian Halianassa vanderhoofi Reinhart was found at locality UCMP V4004. A few feet below this locality, the left M2 of Hipparion cf. H. forcei (USNM 23892) has been found. A short distance to the south, on the opposite side of Bean Creek (Locality USMP V5555), a number of horse teeth have been collected from the Santa Margarita Formation that compare best with Hipparion mohavense. At this locality specimens are usually collected from a pebbly sand (see fig. 3); the oldest part of the formation exposed in this quarry is at least 75 feet higher in the Santa Margarita Formation than the horse tooth from the north side of Bean Creek. Hipparion horses comparable to H. forcei and H. mohdvense occur both below and above the Moraga Formation (informally known as the Grizzly Peak Basalt) in the Berkeley Hills, Calif. This unit has been dated at 10 my (G. H. Curtis, oral commun., 1972); a comparable age is suggested for the lower part of the Santa Margarita Formation north of Bean Creek which contains similar horses. Also one tooth of the primitive horse Archaeohippus has been found at this locality along Bean Creek, the youngest record of this genus. In the Santa Cruz area. the 37°OO' @ .. V7fl184 S’ a. 0' v3“. :‘ a (19“ q . . - vssss Mount Hermon ‘_Wm_ M w \ Ti‘IO-S T.11.S 0 1 2M|LES 0 1 2K|LOMETERS FIGURE 2.—Locality map, Santa Cruz area. 26 OTARIOID SEALS OF THE NEOGENE FEET METHES 100 Well-sorted pebble bed 50 quarry to south Scattered pebbles (UCMP V4004) at arrow Santa Cruz Mudstone of Clark (1966) Locality USGS M1104, Imagotaria downsi, Cetacean,sirenian Well-cemented Astrodapsis bed of early late Miocene age; 2 to 3 feet thick Fine, well-sorted sand occurring in large-scale (to 8 feet) crossbedding dipping south; burrows increase up section, few cetacean and sirenian fossils Well-sorted gravel bed, possible source of Hipparion sp. cf.H.mohavense in Graham quarry to south (locality UCMP V5555) Sand size decreases up section. Burrows. Approximate floor of Graham Crossbedded sands and planar gravel. Type locality of Halianasra vanderhoofi Coarse, poorly sorted gravel with abundant bone fragments including Hipparion sp. cf. cf. H. forcei, desmostylians, cetaceans, sharks, and rays Approximately 50 feet of cover to underlying Monterey Formation Section of the Santa Margarita Formation in the Santa Cruz Aggregate C0. quarry. Measured by Larry Phillips, 1973. FIGURE 3.—Section of Santa Margarita Formation in the Santa Cruz Aggregate C0. quarry. Santa Margarita is overlain by the Santa Cruz Mudstone of Clark (1966); a glauconite overlying the Santa Cruz Mudstone has been dated at 6.7 m.y. (J. D. Obradovich, written commun, 1964). Available information thus suggests that the spe- cimens of Imagotaria downsi and associated otariid remains referred to Pithanotaria are younger than 10 m.y. old, but much older than 6.7 m.y. An early late Miocene age is here assumed, but it is believed that these seals are probably younger than this same pinniped association in the diatomite of the Sisquoc Formation south of Lompoc. The referred specimen (UCMP 34789) from Comanche Creek, Kern County, Calif., appears to be more nearly the age of the type PART I: WALRUSES 27 specimen from Lompoc, whereas the referred speci- men (UCMP 24070—82) from Soledad Canyon, Los Angeles County, Calif., appears to be more nearly the age of the specimens from Santa Cruz. It is believed that the Imagotaria-Pithanotaria pinniped fauna may have existed in the North Pacific from 12 to 9 m.y. ago. Referral of specimens to the species—Association of cranial and postcranial bones of the holotype specimen from Lompoc, SBMNH 342, seems unques- tionable according to the collector, Phil C. Orr (written commun., 1968, quoted by Mitchell, 1968, p. 1845). Those specimens from the most productive locality in the Santa Margarita Formation of the Santa Cruz area (USGS M1035) suggest an accumulation of dead animals of varying ages and sexes; the concen- tration of articulated and disarticulated skeletal elements is very similar to that found on beaches supporting modern rookeries and, as discussed, stra- tigraphic relations in the area favor the interpreta- tion of deposition adjacent to a beach bordering a granitic highland. One partial mandibular ramus and a first metacarpal of an obviously different pinniped, Pithanotaria starri, an otariid smaller than any living genus, was found at this locality, and another mandibular ramus was found 0.4 mile to the south, but no confusion with the walrus-sized Imagotaria downsi is possible. All other pinniped remains from this locality fall into two distinct size groups; these size groups are equivalent to those of male and female living walrus. Of the specimens collected at this locality, indicat- ing a minimum of five individuals, about half are anterior limb elements, which, largely owing to the information provided by the articulated forelimb found here in stratigraphic association with the referred skulls, can be referred to the species by close morphologic similarity with the holotype from Lompoc. The rest, largely posterior limb elements, are referred to the species on the basis of strati- graphic association, size compatibility, and odo- benid characters. Elsewhere in the Santa Margarita Formation of the Santa Cruz area, there are two occurrences of a possible desmatophocid seal (one called ”Desmato- phocine A” by Barnes, 1972, p. 55) and other fossils of the small otariid seal Pithanotaria. These fossils can be eliminated easily from the material assigned to Imagotaria downsi from locality USGS M1035 by morphologic differences. Referral of specimens to Imagotaria downsi from other localities in the Santa Margarita Formation of the Santa Cruz area is by morphologic similarity to, and contemporaneity with, the material from local- ity USGS M1035. The assumption here inVolved in referral, somewhat enlarging upon that of Scheffer (1958, p. 49), is that in the past, as in the present, two species of the same genus of otarioid seals do not live in the same areas. In other areas, the assumption of discrete ranges of very similar pinnipeds cannot be used. Referral of this material is considered questionable even though no morphologic differences may be recognizable in postcranial skeletal elements, as specific identity of many living otarioids, based only upon postcranial skeletal elements, is commonly quite uncertain. DESCRIPTION OF MATERIAL FROM SANTA CRUZ AND COMPARISON WITH THE TYPE Female mandible (pls. 5 and 7).—The mandible of the female skull, USNM 23858, is most distinctive in its sloping, massive chin with a large, oval symphy- seal area over its entire depth. Behind this chin the horizontal ramus appears disproportionately thin and weak; the coronoid process is low with a long anterior margin sloping gradually down to the tooth row; the ramus is markedly bowed outward behind the symphysis; the angular region is narrow and weakly developed; and the pterygoid process is short anteroposteriorly, rounded in form, and does not protrude to the rear much beyond the inferior sig- moid notch. A small medial incisor lies adjacent to the sym- physis in both rami, and a rather large lateral incisor is directly anterior to the medial half of the canine. The canine is fully open at the root and its enamel cap has not completely erupted from the alveolus, indicating some degree of immaturity; it is very long and stout, filling the deep chin and extending behind the symphysis to a point below the third premolar; its "enamel crown is rugose and there is a sharp posterior carina. The cheek teeth are quite small, relative to the jaw and the canine, suggesting a female indi- vidual. They are much smaller than those of the holotype (Mitchell, 1968). Their crowns are formed by a central primary cusp with a prominent lingual cingulum and a small but persistent anterior cingu- lar secondary cusp. A very poorly developed poste- rior cingular cusp is present on some lower cheek teeth. P1 is single-rooted. P2—P4 are also single- rooted, but their roots are bilobed with prominent sulci down their length, suggesting that they were derived from a two-rooted condition (pl. 7). M1 is double-rooted; M2 is present as a rudimentary small single-rooted peg, as in the holotype. Mitchell (1968, p. 1848) mentions an isolated M zfound with the type specimen; this M2 has a small, conical root and a 28 rather globular crown showing a weak central anterior-posterior crest. The mandible of the female skull differs from that of the holotype by its smaller size, less massive structure, and smaller cheek teeth (both actually and relative to the size of the mandible) with more subdued prominence of their crown structures; all these features are quite compatible with differences in sex and maturity. The M] of the type specimen is not as distinctly two-rooted as that of the female skull; rather it has the roots fused nearly to their base (Mitchell, 1968, fig. 9—s, t, u), a variation that is individual in living otarioids (Repenning and others, 1971, p. 9). The female mandible differs in these same ways from the juvenile mandible UCMP 88459 that is from the Santa Margarita Formation in the Santa Cruz area (Barnes, 1971), except that the ontogenetic differences are not present. And, consistent with individual variation in many living seals as noted by both Barnes (1971, p. 6) and Mitchell (1968, p. 1847), the juvenile mandible shows no sign of the rudi- mentary second molar; the M,, P 4 and P3 are clearly double-rooted. This great difference in root fusion led Barnes (1971, p. 9) to feel that his specimen was of "a more primitive species" than I. downsi, but the presence of the female mandible in the same forma- tion of the same area and with P 3 and P 4 root closure comparable to the type but a double-rooted M1 comparable to the juvenile mandible suggests rather strongly that only one species with great variation in root fusion is present. Because of this, UCMP 88459 is here placed in the species I. downsi. The weak angle of the mandible where the digas— tricus inserts in combination with a heavy and deep chin with a long oval symphyseal area over its entire depth, the position of the lateral incisor anterior to the canine, the strong but short pterygoid process, and the strong tendency toward single-rooted cheek teeth in early late Miocene time identifies the man- dible of Imagotaria downsi as odobenid. The pres- ence of a large canine, equal in development to that of the upper canine, identifies the mandible as dusig- nathine odobenid (table 4). No mandible was found with the juvenile male skull (USNM 184060) from locality M1035. Specimen USN M 23868 from locality USGS M1037 is a right lower canine with enough of the mandible adhering to it to show the large symphysis and the base of the alveolus of I 3 anterior to the canine (pl. 7). The canine, which is as large as that of the holotype, is believed to be from a male individual. The tooth is fully mature, providing one indication of the size range of Imagotaria d0 wnsi. This tooth is 17.8 mm in OTARIOID SEALS OF THE NEOGENE TABLE 4.——Dimensi0ns of the female mandible and teeth of Imagotaria downsi (USNM 23858) Mandible (right ramus) Measurements Greatest length .................................................................................. 190 Greatest depth below P2 ..... 47 Least depth at M ............................................................... 34 Height, angular process to coronoid process . ____________ 62 Width of condyle ................. approx. 33 Width at M ............... 13 Width at P2 .......................... 22 Greatest length of symphyseal area ..... 65 Greatest width symphyseal area ................... 29 Length P1 to M 2 .............................................................................. 63 Mandibular teeth (in mm) Width Length Height Root Teeth (transverse at (longitudinal (from crown crown base) at crown base) base) length I 2 ............ 1 2.0+ 1 4.0 13 ............ 1 5.0 1 6.5+ C ............ 13.7 19.3 27.8 47.5+ P 1 ............ 14.0 16.3 P 2 ............ 1 4.2 18.9 P 3 ............ 5.6 8.0 7.5 8.4 P 4 ............ 5.2 7.8 6.5 2.8 M 1 ____________ 5.8 9.4 5.7 M 2 ____________ 1 2.6 1 3.8 ‘Alveolus. transverse diameter and 24.3 mm in anteroposterior diameter at the base of the crown; it has a crown height of 32.7 mm with no correction for moderate tooth wear, and the fully closed root is 69.6 mm long. In crown height it is comparable to an adult male Eumetopias, but the root differs in that it is not so bulbous (hence somewhat thinner), is not so curved, and is much longer. The root has a medial and lateral sulcus running its length such that is has a dumb- bell-shaped cross section, as noted on the holotype (Mitchell, 1968, p. 1848). It was found with the distal fragment of a badly preserved metatarsal, an ex- tremely large metacarpal V, a large humerus, and a few other scraps from a large-sized individual. Female skull (pls. 6 and 7).—The female skull, USNM 23858, is most distinctive in its sea lion-like appearance in combination with its broad basioc- cipital region, vaulted palate, and lack of supra- orbital processes (fig. 4). In palatal aspect, the rostrum is long and broad across the muzzle owing to the large canines. The zygomata are rather thin nd are arched laterally rather than being flat—chee ed as in modern sea lions. The mastoid process is enlarged, but because it is not combined with an FIGURE 4.—Rest0ration of the female skull of Imagotaria downsi. Referred skull USNM 23858. A, Dorsal view. B, Ventral view. C, Lateral view. PART I: WALRUSES 29 \\\\-\\\\\\\\““ \ . ~ A VK .2 v W‘ Mn“ “w“ @5356 u WW ‘ 30 OTARIOID SEALS OF THE NEOGENE enlarged jugular process, appears to be relatively far forward when compared with modern otariids, and more closely resembles the living walrus. The occip- ital condyles are large and widely separated dorsal- ly, as in Odobenus. The margins of the prominently vaulted palate form conspicuous crests which sup- port the cheek teeth and which continue posteriorly beyond the last cheek tooth (M2) nearly to the maxilla-palatine suture. The internal pterygoid pro- cess on the palatine-alisphenoid suture is very large. The basicranium is broad between the glenoid fossae, which are transversely shortened. Because of the shortening of the glenoid fossae, the foramen ovale and posterior opening of the alisphenoid canal face more ventrally than in the otariids and are very similar in their orientation to that of Odobenus. The eustachian foramen is very large. Relative to living walrus, the middle lacerate foramen is small as a result of an extensive bony floor of the carotid canal; in living walrus, there is a notably short bony bridge beneath the canal between the bulla and the basioccipital. The bulla is otherwise Odobenus-like with little inflation apparent on the external surface and essentially no sculpturing so that the style- mastoid foramen opens ventrally with no part of the bulla obscuring it from view in ventral aspect. The hyoid fossa is very large and only slightly concealed in ventral view by a weak lip of the bulla. The wall of the bulla is thick but not so pachyostotic as in Odo- benus of comparable age. The posterior lacerate foramen is of moderate size, not so enlarged as this foramen in the otariids, nor does it merge surficially with the posterior opening of the carotid canal as in otariids. The mastoid process is enlarged as in all otarioids and is backed by a thin plate forming the jugular process of the exoccipital. The jugular pro- cesses of both this female skull and the holotype skull are not fused to the mastoid process, and it is clear that they are not thickened as in living adult otariids but closely resemble the condition in Odo- benus. The basioccipital bone is distinctly odobenid in form. It is very broad between the posterior openings of the carotid canal and is pentagonal, rather than rectangular or trapezoidal. The suture of its articu- lation with the basisphenoid is a straight transverse line medially, but laterally it turns abruptly anteri- orly and then turns again transversely in a some- what irregular pattern to meet the basioccipital- entotympanic suture. The fossae for insertion of the rectus capitis anterior muscles are well developed, as is the sagittal crest separating them, suggesting a fairly mature condition for the animal. In posterior aspect the female skull of Imagotaria TABLE 5.—Dimensions of two skulls of Imagotaria downsi mm Measured parts female male young adult juvenile USNM USNM 23858 184060 Condylobasal length (0)1 ................. .. 287 327: Rostral width across canines (12). . 77 831 Palatal width between P4 _______________ 47 50: Palatal arch from alveolus at P4 . 19 131 Length of P1 to M2 69 92 Length of palate from back of incisors (10) 134 -- Greatest zygomatic breadth (17) .......................... 159 140 Greatest mastoid breadth (20) .............................. 158 143 Greatest basioccipital breadth (between posterior openings of carotid canals) ____________ 54 57: Greatest width of occipital condyles .................. 82 87: Occi ital height, subforamenal notch to ambdoidal crest (not corrected for about 15-mm vertical compression) .......... 96 96: Greatest width foramen magnum ..... 36 43: N arrowest width between orbits ........ 34 30 Greatest width of braincase, above mastoid processes .................................................. 124 131: 1. Numbers in parentheses are those of Sivertsen (1954, figs. 5—7). downsi is marked by widely spaced condyles, more so dorsally than ventrally; a very large foramen mag- num; a low occiput; and flaring mastoid processes joined only by thin jugular processes of the exoc- cipital. In dorsal aspect a very low and short sagittal crest is seen; no supraorbital processes are present on the frontals; and the very large canine teeth are indi- cated by the swollen maxilla, where they articulate with the premaxilla (pl. 7). The size of the female skull is comparable to that of an adult male California sea lion (table 5). Although the individual was not fully mature, it does appear to have reached full size. As the enamel cap of the upper canine was nearly filled with dentine at the time’ of death, the apex of the pulp cavity extended distally only 2 mm into the crown base. Although one upper canine was fractured longitudinally, dental growth lines could not be seen. Comparable measurements of the female skull vary from 65 percent to 72 percent of those of the type male skull. I1 and I2 are not preserved, but their alveoli are distinctly oval. I3 is larger, with a conical crown, circular alveolus, and a very long root, as in some modern sea lions (not fur seals). The crown of the upper canine is roughly conical with a faintly serrate posterior carina and rugose enamel. Its immature root, with fully open pulp cavity, is distinctly longer than in the modern sea lions; it measures 44 mm. The crown is 32 mm long. P1 has a single, roughly circular, and posteriorly curved root nearly three times as long as the crown. The crown is a single cusp with nonserrate anterior and posterior carinae and a lingual cingulum which is strongest posteromedi- PART I: WALRUSES ally, although no distinct cusp is present here (pl. 7). P2 has a similar crown except that a minute "cusp” is present on the posterior part of the lingual cingulum and the cingulum is weaker medial to the apex of the principal cusp. This tooth has a single root that is less than twice the height of the crown and roughly triangular in cross section. Slight grooves on the sides of the root strongly suggest that it was derived from the fusion of three roots; two are labial, the third lies below the rudimentary cingular cusp. P3 is similar to P2 except that the lingual cingulum is lacking medial to the apex of the principal cusp; there is a small anteromedial cingular shelf and a prominent posteromedial shelf and basin supporting a strong ”cusp" which rises out of the medial fused root. The three-rooted nature of P3 is clear, even though the roots are completely fused (pl. 7). P4 has a comparable crown but differs in these respects: there is no trace of an anteromedial cingular shelf; the posteromedial cusp is more prominent; and a small cuspule is present on the posterior carina of the principal cusp. Its roots differ, as it is distinctly two- rooted, the posterior root being heavier. M1 is miss- ing but appears to have been at least weakly two- rooted as judged by its alveolus. M2 is also missing but appears to have been small and single-rooted. The roots of all upper cheek teeth are progressively and rapidly shortened from P1 to M2 (pl. 7). The crown pattern of the upper cheek teeth is similar to that of modern walrus in that it is a simple cusp with a posterointernal cingular cusp. This condition is seen only on fetal or very young walrus because of the rapid destruction of the enamel cap; even when present in modern walrus, the cingular structures cannot properly be called cusps because they are little more than accessory swellings on the side of the swollen enamel cap that forms the tooth crown. Sim- ilar swollen areas are seen on unworn crowns of lower teeth in some modern walrus. These are antero- medial and posterior to the principal cusp and are in the position of the secondary cusps on the lower teeth of Imagotaria. Dimensions of the upper teeth of three specimens are given in table 6. By removal of a section of the roof of the braincase, most of the ventral half of the cranial cavity can be observed. The braincase itself has been partly crush- ed, the parietals telescoping ventrally between the temporals; only the lower half is well preserved. In general aspect the braincase is more elongate than that of modern walrus. Two distinct optic foramina are present, differing both from the flat- tened dumbbell-shaped single foramen of Odobenus and from the single round foramen of living otariids. The optic nerves within the brain cavity lie beside 31 TABLE 6.——Dimensions of the upper teeth of Imagotaria downsi [Dimensions measured in mm at crown base unless otherwise noted] male female Teeth Measurements Juvenile adult young adult USNM SBMNH USNM 184060 342 23858 I‘ : width, transverse... ? 5.2 (root) 3.8 (alveolus) : width, longitudin~~l (See text) 7.4 (root‘- 7.5 (alveolus) 6.9 (root) 4.2 (alveolus) 8.5 (root) 8.3 (alveolus) 11.8 16.4 231 5.71 (alveolus) 7.51 (alveolus) I2 I : width, transverse : width, longitudinal I3 : width, transverse... : width, longitudinal : crown height .. 10:0 14.5 : root length ........... Too immature 451 35.3 C : width, transverse... ..... 25.2 21.5 15.8 : width, longitudinal 26.5 23.1 : crown height.. 49.7 42.1 32.0 : root length ...... Too immature 731 44: (root well open) Pl : width, transve 11.9 (alveolus) 10.5 6.2 : width, longitudinal 12.0 (alveolus) 11.4 7.3 : crown height .. —- 10.7 7.9 : root length ........... Too immature -- 35.3 P2 : width, transverse... ..... 13.0 10.3 6.9 : width, longitudinal 17.0 14.5 8.2 : crown height ............ 16.9 12.2 8.0 : root length ..... Too immature -- 21.0 P3 : width, transverse... ..... 10.2 (alveolus) 12.1 7.1 : width, longitudinal 14.2 (alveolus) 13.9 7.5 : crown height.. -- 10.0 6.2 : root length ...... -- -- 15.5 P‘ : width, transverse... 11.5 11.1 6.0 : width, longitudinal 15.7 13.5 7 5 : crown height ..................... 10.8 9.8 5.7 : root length 15 (anterior and 9.3 10.1 (anterior) posterior, not 9.2 (posterior) mature) M| : width, transverse .............. 9.9 (posterior 4.0 (alveolus) alveolus) : width, longitudinal 14.1 (alveoli) ~ 8.1 (alveolus) M2 : width, transverse... 8.5 ~ 2.9 (alveolus) : width, longitudinal 8.5 -- 4.6 (alveolus) 6.2 -- -- 10.0 (anterior medial) -- 4.9 (anterior lateral) 8.1 (posterior) : crown height ....... : root length ................ each other as in all otarioids, run directly forward from the chiasma, and form what appears as a ped- unculate chiasma, that is low and broad as in wal- rus. Also similar to Odobenus, the nerves diverged within the braincase, rather than within the optic canal, to enter the paired optic canals. Because of crushing, only the floors of the optic canals are preserved, and the dorsal part of the rather small cribriform fossae are superimposed on these. The hypophyseal fossa, which is very Odobenus- like, is a broad, shallow basin with poorly defined lips quite unlike the deeply pocketed fossa of the otariids (pl. 10). The bony tentorium lies directly on the petrosum at the rather large cerebellar fossa, the internal acoustic meatuslis quite broad with almost complete separation of the canals for the facial and vestibulo-cochlear nerves, and the petrosal apex is enlarged and flat; all features are very similar to those of Odobenus. In its relative enlargement, the petrosal apex is actually morelike living walrus than that of Ai vukus cedrosensis, but the internal acoustic meatus is rather far back from the apex, more resembling A. cedrosensis than Odobenus. 32 OTARIOID SEALS OF THE NEOGENE The sylvian sulcus was remarkably deep and long and was supported by an extensive plate of bone from the temporal that unites with the tentorium directly above the oval foramen. The development of this plate to support the anterior ectosylvanian gyrus is greater than in any modern pinniped ex- amined. The sylvian sulcus ran posterodorsally at an angle approximating 45° from the vertical to the floor of the braincase. Within the middle-ear cavity, the structure of the female skull, though somewhat smaller, is essential- ly identical to that of the holotype (pl. 10), described in detail by Mitchell (1968, p. 1854—1863). Most conspicuous are the very large epitympanic recess housing Odobenus-sized ossicles and the very large tympanic membrane. The bony ring of the tympanic membrane of the type skull measures 9.3 mm in its largest diameter, 7.5 mm in its smallest. That of the female skull is nearly as large, measuring 8.3 and 7.2 mm for the same dimensions. The oval window of the holotype has maximum and minimum diameters of 3.2 and 2.1 mm, resulting in an approximate mem- branezwindow-area ratio of 10:1. Comparable mea- surements of the female oval window are 2.8 and 1.5 mm, also resulting in a membranezwindow-area ratio of approximately 10:1. These ratios are quite comparable to those of living sea lions but not to those of living walrus; they suggest that Imagotaria downsi was consistently a deeper diving odobenid than was or is any other known member of the Odobenidae (see Repenning, 1972). This ratio, with such a large tympanic membrane, is the result of a large oval window; this large window, in turn, suggests an enlarged basal whorl of the cochlea. Although the cochlea was not dissected in either the holotype skull nor in the skulls here described, all specimens are characterized by a remarkably high and globular promontorium; this shape also suggesting that the basal whorl is en- larged (pl. 10). Enlargement of the basal whorl appears to have a specific purpose in improvement of underwater hearing and is present in varying de- grees throughout the pinnipeds (Repenning, 1972); this specialization in Imagotaria is greater than in living otarioids and parallels that of the phocids. Though differently formed than in any other pin- niped, the tympanic cavity is basically Odobenus- like (pl. 10). Most distinctive is a posterolateral inward swelling in accommodation of the deep hyoid fossa; Mitchell (1968, p. 1855) noted this and called it the eminentia vagina processus styloidei. This swell- ing into the tympanic cavity is variably developed in Odobenus and partially isolates the posteromedial part of the cavity as a somewhat globose space into which the round window of the cochlea, and its fossa, face. A comparable structure has not been noted in the otariids. The tympanic bulla of the female skull appears to be slightly more inflated than that of the holotype, and it is distinctly less rugose, there being no posterior spur below the hyoid fossa, no spine or fos- sa in the meatal region, nor an underhanging lip to form a canal for the auricular branch of the vagus nerve between the hyoid fossa and the stylomastoid foramen. Such differences are typically sexual and ontogenetic in living otarioids. The head of the malleus is large and swollen, and its articular surfaces are flat, as in all odobenids. The angle between medial and lateral articular facets is about 130° as compared with about 115° in Odo- benus and 100° in otariids and most terrestrial carnivores (pl. 10). The lamina between the head and the anterior process (processus gracilis) is reduced (this process was broken before the plate was made). The neck of the malleus is short, much shorter than in Odobenus. The manubrium is long and is flattened only distally, as in the otariids, rather than along its entire lateral margin from the tip to the short process as in Odobenus. Curvature of the manubrium is slight, as in otariids and terrestrial carnivores; this slight curvature indicates a nearly flat tympanic membrane rather than a somewhat conical mem- brane as in Odobenus. The anterior face of the head of the malleus is somewhat swollen and the anterior process origin- ates from it, as in both odobenids and otariids, rather than having a flat or concave anterior face on the head of the malleus as in phocoids. Vertically, the malleus measures 8.79 mm from the extremity of the manubrium to the top of the head. The proportions of the incus are more Odobenus- like. Although relatively short, the posterior pro- cess (short crus) of the incus is placed very low relative to the articulation with the malleus, and the body is large and inflated as in Odobenus. Except for the wide angle between articulations, the incus greatly resembles that of the phocids. Vertically, the incus measures 6.20 mm from the distal end of the ventral process (long crus) to the top of the body. Horizontally, the incus measures 3.38 mm from the tip of the posterior process to the angle of the mallear articular surface on the lateral side. In summary, the female skull (USNM 23858) from the Santa Margarita Formation of the Santa Cruz area differs from the holotype (SBMN H 342) of Imag- otaria downsi from the Sisquoc Formation of the Lompoc area by its smaller size, smaller and less ornamented cheek teeth, and less rugosity of the PART I: WALRUSES 33 tympanic bulla. The postglenoid lip of the ectotym- panic of the type specimen does not ride forward be- neath the postglenoid process as much as in the female skull, and Mitchell (1968, p. 1850) noted a postglenoid foramen at their union on the holotype that is not present on the female skull. A foramen in this superficial position is sometimes present in modern otarioids, usually unilaterally, and its signif- icance in the present comparison is not known. These differences appear to be either sexual or individual. Juvenile male skull (pls. 8 and 9).—A very imma- ture male skull and scapula (USNM 184060) were found at locality M1035 about 3 feet lower in the section than the female skull. Most of the elements of the skull are unfused and displaced by varying amounts. Two-thirds of the crown of the large lateral incisors and about one-third of the crown of the canines had erupted at the time of depth. The pulp cavity extended to within 10.4 mm of the apex of the 49.7-mm-high enamel crown of the canine. All cheek teeth were fully erupted but had, except for the small M2, widely open pulp cavities in the roots and showed no signs of wear on the crowns (pl. 8). By analogy with the tooth eruption and suture fusion of living otariids, the individual was between 1 and 2 years old. The right canine had been broken off roughly 2.6 cm above the crown apex and about 2.4 cm below the base of the enamel cap. The break exposed 3.9 mm of dentine deposited inside the enamel crown. Mineral deposits on the fracture were cleaned off with a micro sand blaster, and the frac- ture surface was examined under 30 x magnification. Interpreted according to the findings of Kubota, Nagasaki, Matsumoto, and Tsuboi (1961, pl. 7), the following deposits were found across the fracture: Growth layers, male upper canine, Imagotaria downsi, USNM 184060 Enamel ................................................ 0.15 mm Prenatal dentine .............................. .72 Double neonatal line ...................... .13 First year dentine ............................ 3.05 It is concluded from this depositional sequence that the individual died at about 1 year of age. The cause of the double neonatal line is not known, but examination of the upper canine of a one-year-old male Eumetopias jubata revealed the same double structure (deposited entirely within the enamel crown). Comparable growth layer measurements show a few interesting differences. The Eumetopias tooth was considerably smaller than that of the immature male Imagotaria; the measurements were made at a point 1.9 cm above the crown apex and 1.2 TABLE 7.—Number of roots on the upper cheek teeth, Imagotaria downsi Male Female Male Tooth USN M 184060 USNM 23858 SBMNH 342 P1._ one P2 .. one: one lateral sulcus one one: one lateral sulcus one 1 one: two su‘lci suggesting 3- rooted origin one: two sulci one: one lateral P3 .. one: one lateral sulcus suggesting 3- sulcus rooted origin P 4 __ two two one: one lateral sulcus M 1 two: the posterior two: possibly only one much separate broader distally M2 three: two one: small ? anterior, one posterior cm below the base of the enamel. Growth layers, male upper canine, Eumetopias jubata Enamel ................................................ 0.32 mm Prenatal dentine ............................ '.. .32 Double neonatal line ...................... .05 First year dentine ............................ 1.05 Although the enamel crown was nearly twice as large in Imagotaria, the enamel was less than half as thick as in Eumetopias. And dentine deposition during the first year of life was three times greater in Imagotaria. All known odobenids are characterized by thin or no enamel and by excessive dentine deposition in their tusks or upper canines. The juvenile male skull, though of a much younger individual, is considerably larger than the female skull, having a CBL of 327+ mm (40 mm longer than the female skull). The premaxillae were at least partially fused to the maxillae, although their mu- tual sutures appear open, but the two sides of the rostrum were not fused together at the median palatine suture and, as it was buried, were consider- ably displaced; several teeth had fallen out and been lost. Restoration along this suture is not certain; part of the palate obviously was lost from the left maxilla and possibly from the right. And the premaxillae anterior to the incisors are abraded. The CBLlength, therefore, was somewhat greater than measured. The specimen appears to have only two incisors in each premaxilla, but because of the nature of damage during burial, it is not possible to be certain what incisors were present. As with the lower teeth in other specimens, the condition of the roots of the upper teeth appears individually variable (table' 7). Those of the juvenile male skull show a greater development of multiple roots than those of the female skull or the holotype. 34 OTARIOID SEALS OF THE NEOGENE Despite its immaturity, the juvenile male skull shows several features of greater similarity to the type specimen than does the female skull. These similarities strengthen the interpretations of specific identity and female sex of the associated smaller but more mature skull from the Santa Cruz area. Most obvious is the large size of the cheek teeth of the juvenile male, actually larger than those of the holotype, supporting the interpretation that the rela- tively small size of the teeth of the female skull is a sexual feature not to be considered in specific assign- ment. The crowns of the cheek teeth are more ornamented, particularly in the more prominent lingual cingulum, than the teeth of the female indi- vidual, and they more closely resemble the teeth of the type specimen. The M2, known only on the juvenile male skull, has a roundly triangular crown above the three roots with a low transverse loph anteriorly and a slightly higher posterior cusp above the posterior root. Even though clearly not fully developed, the ear region of the juvenile skull has a well-developed inframeatal spine and an apical bullar fossa which are comparable to these structures on the type specimen but lacking on the female skull. In addi- tion, the form and ventral projection of the mastoid process are more similar to that of the holotype. Ossification of the ectotympanic was obviously very incomplete on the juvenile skull, and the floors of the middle-ear cavities must have been very thin and fragile, for they are broken away on both sides and were lost prior to burial. Similarly, both entotym- panic ossifications were lost, and the petrosal on each side was broken from the mastoid, the left being lost and the right being disoriented but present in its approximately correct position at the time of burial. All features of the middle ear otherwise conform to both the holotype and the female skull except that the prominently intruding crista tympanica is not as well developed and measurement of the size of the tympanic membrane is accordingly uncertain. In dorsal aspect (pl. 9), the juvenile male skull is markedly more elongate than the female skull, par- ticularly in the interorbital region. This difference appears to be the most marked of the sexual dimor- phic features in the skull of Imagotaria downsi, al- though the small size of the cheek teeth in the female is nearly as conspicuous. The interorbital elongation of the juvenile male skull is particularly striking in view of its youth, because facial-interorbital elonga- tion correlated with maturation is particularly mark- ed in the living otariids. An approximation of what might have been the size of the juvenile male skull (USNM 184060), had it grown to maturity, can be made from the mandibular length and cheek tooth size of the specimen (UCMP 88459) described by Barnes (1971) and the size of the cheek teeth of the holotype and of the immature male skull. Although the cheek teeth of UCMP 88459, undoubtedly a young male, are essentially the same size as those of the holotype, the length of its cheek tooth row is only 91 percent of that of the holotype because of its immaturity (see Barnes, 1971, table 1). Barnes reconstructed a complete dentary on the basis of the two incomplete rami available (1971, fig. 3); the restoration is about 220 mm in length, which is 30 mm longer than the female mandible. From this, the mandible of the holotype is estimated at 241 mm long, essentially the length of the missing mandible of the juvenile male (USNM 184060) judged by the measurement from the glenoid fossa to the anterior end of the premaxilla. Because the cheek teeth of the juvenile male average 121 percent of the size of those of the holotype, it is assumed that as an adult its mandible would have been 291 mm long and its adult CBL 377 mm, a 50 mm additional elongation in the skull. Therefore an adult male of Imagotaria downsi would seem to have had an extremely elongate skull, particularly in the interorbital region, possibly one- third (or 90 mm) longer than the female skull for essentially the same sized braincase (table 5). Scapula (pl. 8).—Mitchell (1968, p. 1864), in describ- ing two fragments from the glenoid region of the left and right scapulae of the type specimen, noted that the glenoid fossa was shallow and that the coracoid process was very large. These features help distin- guish the scapula of Odobenus from those of the living otariids. A prominent nutrient foramen is present 15 mm posterior to the ventral termination of the base of the scapular spine. An incomplete left scapula was found in associa- tion with the juvenile male skull from the Santa Cruz area. The epiphyseal elements of the glenoid and coracoid had not fused to the body of the scapula and were lost, but the symphyseal surface on the body is curved ventrally in the coracoid region, suggesting a prominent coracoid process. In addition, a prom- inent nutrient foramen is present in the same posi- tion posterior to the termination of the spine base. The position of this foramen is more ventral and posterior than in Odobenus and living otariids (pl. 8). The caudal border, caudal angle, and infraspinous fossa of the body more resemble those of the otariid scapula than they do those of Odobenus; the infra- spinous fossa is distinctly broader, particularly dor- sally, than in Odobenus, and the caudal angle is located more ventral relative to the dorsal extent of the vertebral border. Anteriorly, the scapular notch PART I: WALRUSES is extremely ventral in position and. has a very short radius of curvature; these characteristics resemble neither Odobenus nor the living otariids but are reminiscent of the Antarctic monachine seal, Hyd- rurga. The scapulae of Allodesmus (Mitchell, 1966, pls. 15, 16; Downs, 1956, pl. 26) and that of Pithano- taria starri (Kellogg, 1925a, fig. 1) both appear to share some similarities with that of Imagotaria downsi. All have a low scapular notch with a short radius of curvature, suggesting that this is a primi- tive condition. Humerus (pls. 11 and 12).—The associated male right front limb (USNM 23859, pl. 11) from locality USGS M1035, found a few feet from the female skull, lacks the proximal part of the humerus. In those measurements that can be compared, the humerus of USNM 23859 is 10 percent larger than the humeri of the type (Mitchell, 1968, table III), and is therefore from a very large male (table 8). In other respects, it is identical. The humerus of I. downsi, which is rather robust for an odobenid, is characterized by a pronounced curve of the medial outline from the head to the epicondyle, a transversely short medial epicondyle, an elongate pectoral crest that gradually tapers distally to the shaft almost to the medial lip of the trochlea (toward which it is clearly directed), a very prominent medial lip of the trochlea that has an anteroposterior diam- eter up to 32 percent greater than the greatest anteroposterior diameter of the distal capitulum, and a deltoid tubercle that is on the pectoral crest. The humerus of Imagotaria downsi differs from that of Aiuukus cedrosensis most conspicuously by lacking the abrupt distal termination of the pectoral crest. Ulna (pls. 11 and 13).—This bone, known only from USNM 23859, is characterized by its remark- ably deep olecranon process and short, stout shaft. In other respects, it is odobenid in nature. The anterior margin of the olecranon, dorsal to the greater sigmoid cavity (in the orientation here used the shaft axis is considered to be dorso-ventral), has been damaged, as well as the dorsal lip of the humeral articulation. Enough is preserved, however, to indicate that this margin was narrow and not a broad anterior-facing surface as in the living sea lions. The greater sigmoid cavity for articulation with the humerus has its transverse axis between 15° and 25° from normal to the shaft axis, and in this orientation, as well as in other features, resembles Odobenus. The lesser sigmoid cavity for articulation with the radius is nearly flat and circular in shape. As in all otarioids, the anterior end is heavier than 35 TABLE 8.—Dimensions of three humeri of Imagotaria downsi from the Santa Margarita Formation mm Measured parts USNM USNM USNM 23959 23870 23865 Greatest length, greater tuberosity, to radial capitulum ........................ 265 226 Greatest width across epicondyles 113 88 71 Transverse width at narrowest part of shaft ................................................ 51 42 37 Transverse width across tuberosities ........................................ 85 65 Anteroposterior width midshaft 101 76 62 Greatest anteroposterior diameter of medial edge of trochlea ................ 77 50+ 43 Greatest width of distal articulation ........................................ 80 66 53 Greatest anteroposterior diameter of radial capitulum .............................. 52 42 34 the posterior end of the olecranon crest but there is no lateral process at the approximate midpoint of the crest. The distal articulation for the radius is dis- tinctly separated from the styloid process as in Odo- benus. Radius (pls. 11 and 13).—Two specimens of this bone are known, USNM 23859 (a male) and USNM 184084 (a female). As is the ulna, the radius is marked by its shortness and stoutness, more pro- nounced on the male specimen. Walrus-like, the prominent process for the insertion of the pronator teres lies distal to the midpoint on the anterior margin of the shaft rather than distinctly proximal to the midpoint, as in the otariids. Unlike the sea lion, the anterodistal crest between the pronator teres process and the distal articulation, the radial crest as here used (see pl. 2), is low and not promi- nent. The lowness of this radial crest is comparable to the odobenids, but in Imagotaria the distal termin- ation, the radial process as here used (see pl. 2), is notably medial in position relative to the shaft axis (pls. 12 and 13). In distal view the extreme medial position of the radial process is the most conspicuous feature of the radius of Imagotaria. The grooves for the extensor tendons of the manus are quite shallow; however, the groove for the extensor metacarpi pollicis, like the radial process, is notably medial of the position found in Odobenus, Aivukus, or the otariids; the radius of Imagotaria is comparable in this respect to the radii assigned to Dusignathus and Pliopedia, as will be discussed in a later section. The articular facet for the scapholunar is nearly square but has a convex lateral margin in Imagotaria. Walrus-like, there is no articular facet for the cuneiform on the , ulnar side of the distal termination of the radius. In 36 OTARIOID SEALS OF THE NEOGENE all otariids the cuneiform articulation on the radius is on a slightly elevated platform and visible in distal, as well as ulnar, or posterior, views of the radius. The articular surface on the sides of the head of the radius, for articulation with the lesser sigmoid cav- ity, or radial notch, of the ulna, are much more extensive in Imagotaria than in Odobenus. On the medial side of the head, this articular surface is much like that of Odobenus, but it continues around the posterior side onto the lateral side of the head as a well-developed surface in Imagotaria. Among eight Odobenus radii, only one approached this condition, whereas six clearly had no continuation of this articular surface onto the lateral surface. It appears that pronation of the manus of Imagotaria was about as functional as it is in Odobenus but that supination was much greater in Imagotaria than it is in at least most Odobenus. The dimensions of the male ulna and of male and female radii are given in table 9. Scapholunar (pl. 13).—The scapholunar from limb USNM 23859 articulates with the radius, trapezoid, trapezium, magnum, and unciform, but it has no facet for articulation with the cuneiform. The coun- terpart facet is missing on the cuneiform. Examina- tion of 16 Odobenus scapholunars shows that al- though the size of the cuneiform facet varies, it is always present, suggesting that its lack on the scapholunar of USNM 23859 may be of significance taxonomically. This surface, on the ulnar end of the scapholunar, is continuous with the distal surface for articulation with the unciform in Odobenus and Aivukus, and the opposing surface is large and well developed along the distal margin of the radial side of the cuneiform. In the scapholunar of the otariids, the ulnar termination of the articular facet for the unciform is a sharp lip and there is no cuneiform articular surface. Similarly, the radial side of the sea lion cuneiform shows no articulation with the sca- pholunar along its distal margin, rather it has a large surface along its proximal margin for articula- tion with the radius. This surface is not present in Odobenus or Imagotaria. The pocketed articulation for the magnum on the scapholunar of Imagotaria is about as long as that for the unciform, and this articulation extends about as far in a palmar direction. Sixteen scapholunars of Odobenus, three of Aivukus, and one of ”Triche- codon” (Van Beneden, 1877, pl. 8, fig. 9) have a magnum articulation on the scapholunar that is considerably shorter than the unciform articular surface and it terminates in the palmar direction far short of the termination of the unciform articular surface. TABLE 9.—Dimensions ofa radius and ulna ofImagotaria downsi Measured parts female USNM USNM 23859 184084 Ulna: Length, anterior end of olecranon to styloid process .............................................................. 340 Depth, humeral sigmoid notch to posterior end of olecranon ............................................ 143 Depth, narrowest part of shaft 48 Width, narrowest part of shaft .................... 25 Radius: Length .................................................................. 260 181 Greatest width, proximal articulation ...... 76 47 Greatest width, distal termination .............. 95 66 Least width, proximal articulation ............ 54 37 Depth of shaft at pronator teres origin 58 35 Width of shaft at pronator teres origin 32 17 In all other respects, the scapholunar of Imago- taria downsi is odobenid. In distal aspect the articu- lation for the magnum is deeply pocketed; it does not resemble this articular surface in the otariids, which has nearly the same surface curvature as the adja- cent articular facet for the unciform. Dorsally, in the sense opposite to palmar and as though the flipper were prone on the substrate, there is a round lip terminating the articular surface for the magnum which curves slightly over the dorsal surface; the termination of this facet in otariids is a sharp lip facing distally. From ulnar to radial sides, the common facet for the trapezium and trapezoid is strongly curved, rather than slightly so as in the otariids. Cuneiform (pl. 13).——This bone, from limb USNM 23859, differs from that of Odobenus largely by lacking an articular surface for contact with the scapholunar, as has been discussed. It is otherwise most similar to Odobenus. It lacks the articular sur- face for contact with the radius found in all otariids. The palmar process is long in comparison with either Odobenus or living otariids, and there is a dis- tinct pisiform articulation. Continuous with this articulation, the surface on the ulnar side for articu- lation with the styloid process of the ulna is not so flat as that in Odobenus but not nearly so cupped as is this articulation in the otariids. Odobenus-like, the articulation with metacarpal V is triangular and is present only on the palmar process, whereas in otari- ids it is continuous across most of the body of the cuneiform. Pisiform (pl. 11).—This bone, from limb USNM 23859, is large and has a well-defined articular surface for both the cuneiform and the styloid pro- cess of the ulna. It does not appear clearly odobenid PART I: WALRUSES 37 except that it is long relative to that of the otariids. Trapezium (pl. 11).—This bone was not repre- sented in the articulated limb USNM 23859. An isolated trapezium (USNM 23875) was found a few feet laterally from the left calcaneum (USNM 23866) at locality M1104. These two bones were from the 2- to 3-foot thick gravel bed in the upper part of the Santa Margarita Formation. The isolated trapezium conforms closely to what could be expected for the trapezium missing from the referred articulated limb and it actually articulates well in the limb (pl. 11). It is referred to the species on the basis of both its odobenid morphology and its stratigraphic associ- ation. This trapezium is very distinctive in form. Most conspicuous is a dorsal-ulnar projection of the distal articulation which conforms to the prominent dorsal- palmar convexity of the proximal articulation of metacarpal I. This articulation surface for contact with the metacarpal I is very concave, not moderate- ly so as in sea lions, and very different from 'the flat surface in Odobenus. In living walrus the proximal articulation of the trapezium is concave in a palmar-dorsal direction, a shape reflected in the rounded articular surface on the scapholunar. In otariids this surface is slightly concave to flat in this direction. On the trapezium of Imagotaria, this proximal articulation is convex in a palmar-dorsal direction, a shape reflected in the flat articular surface on the radial process of the scapho- lunar. Comparing these differences, the trapezium of ‘ Aivukus cedrosensis is intermediate between those of Imagotaria and Odobenus. Trapezoid (pl. 13).—Based upon two available specimens, male limb (USN M 23859) and a female specimen (USNM 184086), this bone, in dorsal as- pect, is narrow between the trapezium and scapho- lunar facets, as in Odobenus. As in Aivukus cedro- sensis, this bone has a narrower facet for contact with the trapezium than in Odobenus. The facet for contact with metacarpal II is triangular in outline rather than nearly rectangular as in both Aivukus and Odobenus, and it is more deeply concave than the facet on this bone of those genera. Magnum (pls. 14 and 15).—This bone, available only from the limb USNM 23859, is very Odobenus- like: the surface for articulation with the scapho- lunar is high toward the palmar side, reflecting the pocketed articulation for the magnum on the scapho- lunar, and extends farther toward the palmar pro- jection than in Odobenus. In addition, that part of this articular surface on the dorsal part of the radial face of the magnum is deeply pocketed (more so than in walrus) to receive the rounded lip of the scapho- lunar. The articular area for contact with the scapho- lunar is continuous along the proximal margin to its dorsal termination, rather than being entirely on the radial surface at its dorsal termination as in otariids. In otariids, the surface for articulation with the unciform faces more or less proximally at its dorsal termination, whereas in Odobenus and Imagotaria this surface remains on the ulnar side of the bone. The small articulation for the metacarpal IV on the ulnar side of the magnum in Imagotaria faces more distally than either Odobenus or otariids, narrowing the dorsal half of the distal articulation for meta- carpal III. In dorsal view, using the surface of articulation for metacarpal III for dorso-ventral orientation, there is seen the most marked difference recognizable be- tween the magnum of Imagotaria and that of Odo- benus (pl. 15). In this view, the proximal crest of the Imagotaria magnum appears to have the form of a sigmoid curve; it inclines first in an ulnar direction and curves smoothly toward the radial side as it extends back toward the palmar process. The in- clination of this crest appears to be entirely in a radial direction in Odobenus, and this inclination is consistently so in all modern Odobenus magna. The appearance is largely due to the more distal orienta- tion of the articular facet for metacarpal IV in Imag- otaria; the angle on the magnum of Odobenus, formed by the intersection of the articular surfaces for metacarpals III and IV, varies from 72° to 90°, and on the single known magnum of Imagotaria this angle is 123°. Unciform (pl. I4).—The unciform of Imagotaria downsi, known only from the referred front limb USNM 23859, is essentially identical to those of Aivukus cedrosensis, except that it is somewhat narrower between the radial and ulnar sides and has a sharper proximal crest between the scapholunar and cuneiform articular facets. In these features the unciform of living walrus varies greatly with the individual. Like that of A. cedrosensis, the unciform of Imagotaria downsi differs from modern walrus by being elongate in the palmar-dorsal dimension. Metacarpals (pls. 11 and 14).—As a group, the metacarpals of the male limb of Imagotaria downsi (USNM 23859) are very Odobenus-like but very slender. They closely resemble the metacarpals of Aivukus cedrosensis except that in those features of the latter that differ from the metacarpals of living Odobenus, the metacarpals of Imagotaria downsi differ more extremely. In size the metacarpals of male Imagotaria are very similar to those from males of the two other genera. Compared to that of Aivukus, metacarpal I of 38 Imagotaria downsi is very slender, has virtually no flattening of the palmar surface of the shaft nor broadening of the proximal part of the shaft to form a crest on the anterior (radial) edge for insertion of the abductor pollicis (pl. 14). Although a raised area on the dorsal surface of the shaft is present for insertion of the extensor pollicis, it is separated from the proximal articulation by a distinct depression which is unknown in the Odobeninae and which appears diagnostic of the Dusignathinae. The proximal artic- ulation is saddle-shaped, extends well onto the dor- sal surface (somewhat more so than in Aivukus), and is distinctly different than in Odobenus. Metacarpals II, III, and IV are more slender shaft- ed than the comparable elements of Aivukus cedro- sensis; otherwise they are identical (pls. 4 and 14). Metacarpal V is unknown in Aivukus; in Imago- taria downsi it is known from the limb USNM 23859 and from a very large isolated specimen USNM 184055. It strongly resembles Odobenus, particularly because the shaft is not flattened as in the otariids. The dimensions of the metacarpals are given in table 10. Phalanges (pl. 1 1). —The proximal phalanges of all digits were recovered with the articulated anterior limb, although the distal half of the proximal phalanx for the third digit was lost in collecting. The distal phalanx of the first digit, lacking the proximal articulation, and one complete and two incomplete middle phalanges of the other digits were also recovered. In addition, the proximal phalanx of the first and second digit of the opposite forelimb (left) were found in the same excavation. The phalanges are not so flattened as in the living otariids, but they are distinctly more flattened than in the living walrus. On all phalanges the processes for insertion of the flexor tendons are more pro- nounced than in either of the living otarioids; they more closely resemble those of land carnivores. The only terminal phalanx found, from the first digit, suggests a reduction in nail development compar- able to that otariids, and it has a widened and abrupt termination which clearly supported a carti- TABLE 10.—Dimensions ofthe metacarpals ofImagotaria downsi, trans verse diameters USNM USNM USNM USNM 23859 23861 23860 184055 Measurements (mm) I II III IV V 111 IV V Length ........................ 147 103 94 89 91 86 -- 115 Minimum diameter 21 16 16 16 20 14 13 28 Proximal diameter _. 46 25 26 35 26 26 28 42 Distal diameter ,,,,,, p .. 30 30 27 27 28 25 -- 36 OTARIOID SEALS OF THE NEOGENE laginous extension for flipper elongation in typical otarioid manner. Proximal phalanges of the fourth and fifth digit have a slightly developed palmar curvature comparable to some degree with the curva- ture found in land carnivores. Femur (pl. 15).—Skeletal elements of the hind limb of Imagotaria downsi are poorly represented in the collections from the Santa Margarita Formation near Santa Cruz. Their reference to this species is based upon the following circumstances: 1. Size bimodality of specimens comparable in actual size to front limb elements from the same and other beds as the referred female skull, and also to the holotype and the referred mandible already discussed. 2. Odobenid structure. 3. Close stratigraphic association with anterior limb elements identical to others belonging to the genus, or stratigraphic bracketing of hind limb elements between specimens clearly assignable to Imagotaria downsi, or similarity to other hind limb elements less questionably referred to the genus and from strata from which no other odobenid is known. The material thus referred to Imagotaria includes one poorly preserved femur, three calcanea, two astragali, one patella, and two tibia. The femur was found with a humerus. The three calcanea were found at three localities, one at USGS locality M1035 down section from the female skull and male front limb but up section from an unquestionably referred metacarpal and at the same horizon and a few feet away from an atlas which conforms to the fragments of the atlas found with the female skull. One astra- galus was found with one calcaneum and a second was found with a female radius and a metacarpal III. One patella was found up section from a referred metacarpal III. One tibia was found with the patella, and an isolated tibia was found at the same horizon as referred metacarpal III but about 20 feet away. The single femur referred to Imagotaria (USNM 23870) is very poorly preserved but shows some unique features (pl. 15). Most conspicuous is the extreme flatness of the shaft, not only distally as in Odobenus but also proximally. The lesser trochanter is extremely well developed, more so than in living otariids and contrasting even more strongly with living walrus; the head is short necked but clearly higher than the greater trochanter. Patella (pl. 11).—A single patella and associated immature shaft of a tibia, USNM 23863, were col- lected at the same locality but about 15 feet down sec- tion from the female skull and 5 feet up section from a referred metacarpal III. The patella does not have a distinctly conical shape with the greatly protruded PART I: WALRUSES apex seen in otariids, and is flatter and more like that of Odobenus and land carnivores. Tibia.—The lateral condyle on the average is much larger than the medial, and its supporting crest on the shaft is more extended laterally in Odo benus than it is in the otariids. As a result the fossa for the origin of the tibialis cranialis is broader and in some cases deeper in Odobenus. The tibia found with the patella and referred to Imagotaria (USNM 23863) shows this lateral-cranial crest to be produced equally as much as in any of seven Odobenus tibiae available for comparison and far greater than in any living otariid. The shaft is remarkably straight but within the range of variation in modern walrus. This speci- men lacks both distal and proximal articulations, but it is large and stout, presumably from a young male. This tibia and the associated patella were found in the same bed less than 4 feet from the immature male skull. An isolated tibia, USN M 23864, was found at the same horizon as referred metacarpal III at USGS locality M1035. This bone, though poorly preserved, is recognizably from a mature animal about one- fourth smaller than the immature tibia and patella found 5 feet up section and is assumed to be female. As is the other tibia, this bone is characterized by a prominent lateral-cranial crest and prominent fossa for origin of the tibialis cranialis. The proximal artic- ulation was weathered away. Although somewhat abraded, the distal articulation shows the weak grooves above the medial malleolus for flexor ten- dons that characterize the tibia of walrus; these grooves are uniformly double, long, and strong in otariids. The greater extent of preservation, in com- parison to the tibia described above, emphasizes the straightness of the shaft. Kellogg (1925a, p. 94—95) has described the distal part of a tibia (UCMP 24221) from the Sisquoc Formation south of Lompoc, Calif. He noted odo- benid similarities in this bone, and it is possible that it belongs to Imagotaria downsi. Though larger, it is comparable in form to the smaller tibia from the Santa Cruz area (USNM 23864) by having the weak flexor grooves on the medial malleolus and by the presence of a long narrow facet along the anterior border of the distal articulation. On the tibia from Lompoc, the medial malleolus is separated from the astragalar articular facet, a condition not evident on the tibia from Santa Cruz; the difference may cor- relate with greater size. Calcaneum (pl. 15),—Four calcanea, all mature and with a range in size comparable to this bone in living male and female Odobenus, have been col- lected from the upper part of the Santa Margarita 39 Formation in the Santa Cruz area. The two most nearly complete, USNM 23862 (male) and UCMP 107759 (female), were collected at USGS locality M1035, where the female skull, juvenile male skull, and male limb were found. The male calcaneum was found at the same horizon and a few feet from an isolated atlas which matches the fragments of the atlas found with the female skull and very close to the juvenile male skull. Walrus-like, the calcanea referred to Imagotaria have a very prominent internal tuberosity on their calcanea] tuber. They differ from the calcanea of Odobenus by having a cuboid facet that is between 10° and 15° from normal to the long axis of the bone in dorsal aspect (rather than between 30° and 35° as in walrus) and by the lack of a medial-distal exten- sion of the body beyond the limits of the distal astragalar articulation, the secondary shelf of the sustentaculum of Robinette and Stains (1970, fig. 1), as found in the calcaneum of both Odobenus and otariids. In distal aspect, the cuboid facet is rec- tangular in outline and the calcaneum body is deep on the medial side below this facet. A third calcaneum, USNM 23866, of slightly small- er size than the male calcaneum, displays these characters except that the cuboid articulation and peroneal tubercle have been destroyed. This bone, from locality USGS M1104, was not associated with other elements of Imagotaria, although it was found near the trapezium referred to the genus. A fourth calcaneum, with associated astragalus and partial navicular and other tarsal fragments forming USNM 23867, was found at USGS locality M1108. These bones are of distinctly small size, identical to the female calcaneum from USGS locality M1035. Damage to the calcaneum is similar to that described above, and in addition the internal tuberosity has been broken off. In features that are preserved, in particular the lesser process and distal astragalar articulation, this calcaneum matches the other three and falls within the size variation evident in other specimens of Imagotaria. A right hind flipper, UCMP 24070—82, from the Towsley Formation (Winterer and Durham, 1962) of early late Miocene age was assigned to Pontolis cf. magnus by Kellogg (1925b). The partial astragalus (pl. 14), partial calcaneum (pl. 15), cuboid (pl. 9, fig. 4), and metatarsal III of this specimen are identical to elements of Imagotaria downsi from the Santa Mar- garita Formation. The flipper from the Towsley Formation is here identified as Imagotaria sp. be- cause of these similarities. Astragalus (pl. 14).—One of the two complete astragali referred to Imagotaria downsi was asso- 40 OTARIOID SEALS OF THE NEOGENE ciated with the calcaneum, USNM 23867. It has a nearly vertical fibular articulation which is normal to the trochlear surface for articulation with the tibia and a lateral process which is small, in line with the lateral crest of the trochlea, and does not flare widely in a distal-lateral direction; these two features char- acterize the astragalus of Odobenus. It lacks an astragalar foramen, as do most astragali of Odo- benus. In addition, the process medial to the pos- terior calcanear articulation and extending postero- medial from the plantar side of the body of the as- tragalus, the calcanear process which is greatly enlarged in phocids, is enlarged. Examination of 27 Odobenus astragali suggests that these features are probably quite constant. The configuration of the astragalus would require that the astragalar articu- lation on the fibula would be a nearly vertical sur- face alined with the long axis of the fibula, and dis- tinctly different from that of the otariids. The second known astragalus, USNM 184085, found with a female~sized radius and metacarpal III by Gerald Macy, is identical in size and configura- tion to USNM 23867 except that the lateral process is more bluntly terminated and a well-developed astra- galar foramen is present (pl. 14). The fragment of the astragalus with the pes described by Kellogg (1925b) from the Towsley Formation and here called Imago- taria sp. is identical in those parts preserved, but it is much larger and is considered to be from a male individual. Navicular.—A female navicular, UCMP 107752, is known from USGS locality M1035, and another incomplete one was associated with the pes, USNM 23867, from USGS locality M1 108. N 0 significant dif- ference is evident, although the complete specimen is either abraded or immature as several features are obscure, particularly the entocuneiform facet. A series of 11 naviculars from Odobenus indicates considerable variation in the cuneiform facets and the astragalar fossa, and no consistent differences between the naviculars of Odobenus and Imagotaria have been noted in these articulations. However, among the Odobenus naviculars, the plantar process is consistently on the fibular side of the center of the navicular, and on the two naviculars of Imagotaria, the plantar process is on the tibial side of the center. The same condition is evident on the navicular of Imagotaria sp. from the Towsley Formation (UCMP 24072) (see Kellogg, 1925b, fig. 10). On the tibial side of the dorsal margin of the astragalar fossa of the Imagotaria navicular, there is a conspicuous lip on all three specimens which is developed on only one of the 11 Odobenus naviculars available for comparison. Kellogg (1925b) described how this bone differs from the otariids. It may be noted here that the minute navicular attributed to Neotherium mirum by Kellogg (1931) has the plantar process centrally located. Cuboid (pl. 9).—Qne well—preserved female cuboid, USNM 184061, was recovered from USGS locality M1035. -It compares most favorably with that from the pes described by Kellogg (1925b) from the Tows- ley Formation, and those differences which can be seen can be equaled in the available series of nine cuboids from Odobenus. It also is identical to the cuboid of Neotherium mirum (Kellogg, 1931) except for its larger size; Kellogg noted this similarity in describing Neotherium. In odobenids the angle formed, in dorsal aspect, by the navicular facet and the calcanear facet on the cuboid is never less than 75°. The astragalar facet, near the apex of this angle, may lie in the plane of the navicular facet (in which case it is hard to recognize) or may assume an intermediate orientation. In otari— ids this angle varies from 48° to 70°. By this criterion, Imagotaria downsi, Imagotaria sp. from the Towsley Formation, and Neotherium mirum are odobenids. The cuboid of Imagotaria differs from that of Odobenus by the marked prominence of its plantar process, a fissipedlike feature lost in most living otarioids but present in extinct forms, both otariid and odobenid, but presumably lost in desmatopho- cids judged by Allodesmus kelloggi (Mitchell, 1966). Kellogg (1925b, p. 106) commented on the small size of the astragalar facet on the cuboid of the pes from the Towsley Formation. This facet is usually smaller in odobenids than in otariids, with which Kellogg was comparing the fossil, and it is not uncommonly small (for an odobenid) in Imagotaria. Its limits, however, are usually difficult to judge without having the adjacent navicular in articular position. Kellogg (1925b) described the plantar pro- cess in unfortunate terms; it is present across the entire plantar surface but the pronounced process on the fibular side of this face makes the remainder seem insignificant; hence Kellogg refers to the pro- cess as being reduced and restricted to the fibular side, which is exactly the opposite of its condition. Cuneiform bones.—None of these three elements have been found in the Santa Margarita Formation of the Santa Cruz area. They are described by Kellogg (1925b) from the pes collected in the Towsley Formation. In general Kellogg understated their odobenid nature as his primary comparison was with the otariid Eumetopias. Metatarsals.—No metatarsal I of Imagotaria downsi has been found in the Santa Cruz area. Kellogg (1925b, p. 111—113) decribed a complete first PART I: WALRUSES metatarsal from the Towsley Formation that he compared only with Eumetopias, as he did all meta- tarsals. This was unfortunate, as they all compare more closely with the metatarsals of Odobenus. Metatarsal I from the Towsley Formation is more elongate than that of Odobenus, has a moderate plantar-dorsal curvature not seen in Odobenus, and has a pointed dorsal apex on the proximal articula- tion. When compared with otariid metatarsals, it is conspicuously odobenid in its elongate, slender, and unflattened shaft. The proximal half of metatarsal II is known from both the Santa Cruz area (USNM 23868, found at locality M1037 with a humerus, a mandibular frag- ment, and a fifth metacarpal) and the pes from the Towsley Formation. Both specimens are badly pre- served. Both differ from Odobenus by having a more elliptical cross section at about the midpoint of the shaft, the shaft being compressed in a dorso-tibial and palmar-fibular direction. Metatarsal III is known from four individuals: two (UCMP 108066 and UCMP 102854) from the Santa Margarita Formation in the Santa Cruz area, one from the pes from the Towsley Formation, and one (UCMP 34789) from the Santa Margarita Formation in Tejon Hills, Kern County, Calif. All specimens are only the proximal half or less. All are essentially identical to metatarsal III of Odobenus. Kellogg has compared the specimen from the Towsley Formation with the third metatarsal of Eumetopias. Assign- ment of the specimen from the Tej on Hills to Imago- taria is made only on the basis of its geologic age and geographic proximity, as it is obvious that other odobenids will be similar in the configuration of the proximal part of metatarsal III. Metatarsals IV and V are known only from the pes from the Towsley Formation. As described by Kel- logg (1925b, p. 115—116), they are very poorly pre- served. Vertebrae.—Several fragments of vertebrae were found with the female skull of Imagotaria do wnsi but only one nearly complete vertebra is known. This is an atlas (USNM 23872) found near the right cal- caneum (USNM 23862) and 10 feet lower in the section than the female skull (pl. 7). The atlas is distinctive in several features, including the widely spaced anterior articulations, a very large vertebral foramen, an antero-posteriorly narrow neural arch, the lack of a ventral notch between the anterior articular surfaces, the small transverse processes which are less ventrally directed than in otariids, and, most conspicuously, the position of the canal for the vertebral artery which, in posterior aspect, is seen to be ventral to the lateral part of the articular 41 surface for the axis, rather than lateral to it, and hence to lie on the postero-medial surface of the transverse process. The transverse processes of an atlas were found in direct association with the female skull (pl. 7). In the position of the arterial canal relative to the posterior articulation and the unexpanded condition of the transverse processes, these fragments agree exactly with the nearly complete atlas found 10 feet lower in the section. Because the nearly complete atlas is about one-third smaller than the fragments found with the type skull, it is assumed to be from a young individual. Except for size and features of immatur- ity, the atlas is identical with the fragments found with the holotype of Imagotaria downsi described by Mitchell (1968, p. 1863). USNM 13487 Mitchell (1968, p. 1865—1868) described an unpre— pared skull from the same diatomite quarry as, but about 800 feet farther west of, the referred ”Indivi- dual II” and ”Individual III" of Pithanotaria starri (Kellogg, 1925a, p. 84, 87). This locality is roughly 11/2 miles south of the type locality of P. starri and the locality of the distal end of the tibia discussed and referred to Imagotaria downsi. The locality is about 5 miles west of the type locality of I. downsi. All are from the same stratigraphic unit and are essentially of the same age. The specimen is embedded in a block of diatomite which has been fractured in a way that exposes a frontal section of the skull at the palate. According to Mitchell (1968, table 2), the CBL is 340 mm, 13 mm longer than the juvenile male skull (USNM 184060); other measurements are quite comparable. Elonga- tion of the interorbital region suggests a male indivi- duaL The lateral incisors of this specimen seem very large; Mitchell discusses the possibility that they actually represent still-retained milk canines, but he rejects the possibility. The explanation for what appear to be extremely large lateral incisors is evi- dent upon examination of the juvenile male skull from the Santa Cruz area: the lateral incisors are nearly or completely erupted but the canines are only one-third erupted or less. Mitchell notes that P4 is double-rooted; the posterior root is larger, as in the juvenile male skull from Santa Cruz, but differing from the condition of the P4 roots of the holotype. The resemblance of this specimen to the juvenile male skull and t0 the holotype, as discussed by Mitchell, and its close stratigraphic association with the holotype certainly indicate that Mitchell’s refer- ral of this specimen to Imagotaria downsi is correct. 42 OTARIOID SEALS OF THE NEOGENE DISCUSSION The mandible, skull, and postcranial skeleton of Imagotaria downsi all show decidedly odobenid fea- tures. In general appearance, however, it would seem that the species was adapted for a sea lion-like exist- ence. Sea lions that were contemporary with Imago- taria, as will be discussed, had cheek teeth that were entirely double-rooted and were small animals. The only known genus, Pithanotaria, was smaller than the smallest living otariid, the Galapagos fur seal. By contrast, Imagotaria do wnsi had the dimensions of the living walrus. This great size difference sug- gests that Imagotaria and contemporary otariids were not in direct competition; a suggestion strength- ened by their joint occurrence in several late Miocene deposits of California, including the Santa Marga- rita Formation at Santa Cruz and the Sisquoc Form- ation at Lompoc. From both Odobenus and Aivukus, as well as from other extinct odobenine genera, Imagotaria differs by having more slender metacarpals with a distinct pit on the proximal dorsal surface of metacarpal I, a wide (123°) metacarpal III and IV facet angle on the magnum, a magnum articulation on the scapho- lunar that is extended in the palmar direction, a conspicuously medial location of the radial process of the radius (pls. 12 and 13), a notably short and stout shaft on both radius and ulna, and, dubiously, a relatively weak medial epicondyle on the humerus in addition to the lack of reduction of the lower canines mentioned in the diagnosis of the Dusig- nathinae. The bones of the rear limb of Aivukus are not known. However, Imagotaria also differs from Odobenus in the very prominent lesser trochanter and proximally flattened shaft of the femur and the rectangular articular facet for the cuboid on the calcaneum, which has less inclination from normal to the long axis of this bone. These differences appear to be of use in distinguishing between the odobenid subfamilies Odobeninae and Dusignath- inae. Genus PONTOLIS True Plates 10 and 18 Type species.—Pontolis magnus True (1909); USNM 3792, the basicranium and occiput of a skull that was badly shattered and incompletely repaired and prepared at the time of True’s description (1905 and 1909). DISCUSSION This genus was based upon the above specimen collected before the turn of the century in beds of late Miocene age that are exposed near Empire, Oreg. Based upon its invertebrate fossils (Glen, 1959, fig. 5), the Empire Formation appears to be similar in age to the Purisima Formation, possibly 4—6.7 my. old. In June of 1973, C. E. Ray and D. R. Emlong collected the premaxillary ”Sabertooth” breeding tooth of Smilodonichthys from the Coos Conglomerate Mem- ber of the Empire Formation, at or very near to the type locality of Pontolis magnus. This giant salmon is also known from the Drakes Bay Formation of Galloway and the Purisima and the Santa Marga- rita Formations of the Santa Cruz area. Shotwell (1951) mentioned a mandibular ramus from the type locality of Pontolis magnus; he be- lieved it was most likely that this ramus represented this species. Unstudied material from this locality at the Smithsonian Institution indicates that at least two types of large odobenids are present in this form- ation at this locality. True believed that there were similarities between Pontolis magnus and the living otariid genus Eu- metopias, but incomplete preparation misled him into believing that the auditory bullae were ”com- pletely obliterated” (1909, p. 144) and "crushed and splintered off down to the level of the basioccipital and s0 mingled with the matrix that their form is lost” (1905). Consequently, he derived almost no comparative information from the ventral surface of the basicranium. The floor of the brain cavity was quite well ex- posed and here True noted considerable difference between Pontolis magnus and Eumetopias jubata (1909, p. 147). However, he did not compare these structures with any other pinniped. The floor of the braincase of Pontolis magnus (pl. 10) is marked by the broad and shallow hypophyseal fossa of the odobenids, which contrasts markedly with the deep globular fossa of the otariids. Much of the right petrosum is preserved on the specimen. It has a broad and flat apex, and the internal acoustic meat- us is very broad with almost complete separation of the canals for the facial and vestibulocochlear nerves. The cerebellar fossa is small, and the bony tentorium is closely appressed to the petrosum in this region. These features are diagnostic of the odo- benids. There appears to have been formed a rather strong bone plate to conform to the sylvian sulcus, much as in the female skull of Imagotaria downsi. Further preparation of the type skull (pl. 18) shows that both middle ear cavities were shattered and had been repaired with much fragmentation and disloca- tion of bone prior to True’s examination; the nature of the middle ear cavity, tympanic membrane, epi- tympanic recess, and promontorium of Pontolis magnus remains unknown. Externally, however, the basicranial area is not badly damaged. Major breaks PART I: WALRUSES 43 cut across the mastoid-bulla region of both sides and were partly offset in repairing. However, most ex- ternal features are discernible on one or the other side of the specimen. Contrary to True’s impression, the bulla and basioccipital are largely intact. How- eVer, they are not at all typical of Eumetopias nor any otariid seal. The bullae are very flat and smooth. Associated foramina and canals are very large. The basioccipital is extremely broad with an equidimen- sional pentagonal form. These all are features of the Odobenidae, but none at present clearly distinguish the Odobeninae from the Dusignathinae. As with Imagotaria, development of the petrosal apex is more advanced toward the Odobenus condition than in Aivukus cedrosensis, suggesting greater parallel adaptation in this specialization than in contempo- rary odobenines and more resembling the condition in Imagotaria. The resemblance of Pontolis magnus to Imago- taria downsi is exceedingly great. In the material available for comparison, the most dissimilar fea- ture is the smoothness of the fusion of the posterior parts of the bulla to the mastoid process in Pontolis; in the type specimen of Imagotaria, this region is marked by a prominent groove (for the auricular branch of thevagus) between the stylomastoid fora- men and the hyoid fossa, as well as other under- hanging irregularities such as the small posterior bullar projection. However, these structures are not present on the female or juvenile male skull of Imag- otaria downsi. In addition, these structures are not present on the fragment of the temporal of the holotype of Dusignathus santacruzensis, although virtually nothing else of this specimen can be com- pared with Pontolis magnus. Without further knowl- edge about Pontolis magnus or Dusignathus santa- cruzensis, it does not seem advisable to consider synonymy. Mitchell (1968, p. 1877) concluded, from differences in proportions, that the two species, Imagotaria downsi and Pontolis magnus (and gen- era, by inference), are distinct. Pontolis magnus is here considered a dusignathine odobenid, but gener- ically and specifically it is a nomen dubium. cf. Pontolis magnus, Lyon, 1941 A badly distorted skeleton from the Valmonte Diatomite Member of the Monterey Shale, of early late Miocene age (upper Mohnian, Kleinpell, 1988) was compared to this species by Lyon (1941). The locality is near Lomita, on the north side of the Palos Verdes Hills, Los Angeles County, Calif. Mitchell (1968, p. 1879—1800) discussed it in some detail and concluded that it is ”clearly related” to Imagotaria downsi, remarking that the specimen should be further prepared and studied. The specimen was not further examined for the present report. We concur with Mitchell’s conclusion and consider "cf. Pontolis magnus” of Lyon a dusignathine odobenid of un- known generic affinity. Genus DUSIGNATHUS Kellogg Type species.—Dusignathus santacruzensis Kel— logg, 1927. Diagnosis.—A specialized dusignathine odobenid with elongate upper and lower canines: the upper canines apparently did not occlude with any lower teeth because they show little or no wear from this cause; the lower canines are very close together because of an extremely narrow symphyseal region and apparently occluded only with the elongate lateral upper incisors—they exhibit considerable anteromedial wear from this occlusion. Cheek teeth have stout peglike roots and simple almost conical crowns including a fully developed and peg-rooted P 4 and M1, crowns are capped by thin smooth enamel and show wear entirely on their anterior and pos- terior surfaces. Palate not greatly vaulted; infra- orbital foramina of moderate size. Braincase has both sagittal and lambdoidal crests but they are low. Mandibular rami very deep, narrow, and upturned. Probable dental formula: 11 '1C-4P-1M(?) 0-1C‘4P-1M Distribution—Late late Miocene and Pliocene of California and Baja California, by estimation 4—8 m.y. ago. x 2 = 24-26 Dusignathus santacruzensis Kellogg, 1927 Plates 5, 15, 16, and 18 Holotype.—UCMP 27131, left and right dentaries, ‘part of the right maxilla, dorsal fragment of the occiput, incomplete right temporal and isolated teeth including two upper incisors. From UCMP locality V—2701 in the late late Miocene or early Pliocene part of the Purisima Formation, Santa Cruz, Calif. The specimen is apparently a young adult individual to judge from moderate development of the sagittal and lambdoidal crests, tooth wear, and incomplete fusion between maxilla and premaxilla. Referred ma terial.—LACM 3011, ”parts of an asso- ciated right antebrachium and manus including the following: ulna, radius, cuneiform, unciform, trape- zoid, metacarpals 4 and 5, and the proximal ends of metacarpals 1 and 3” (Mitchell, 1962, p. 4, under the name ”Odobenid, possibly new genus and species”). From LACM locality 1181 in the Purisima Forma- tion, within 20 feet of the stratigraphic horizon of the type, and about 3,000 feet S. 25° W. of the point herein 44 OTARIOID SEALS OF THE NEOGENE presumed to be the type locality. LACM 4342, a complete right fibula. From LACM locality 1666 (fig. 2) which is the same or nearly the same locality as LACM 1181 (Mitchell, 1962, fig. 1). UCR 15244, the right half of a snout bearing lateral I, C, P1, P2, and half of P3. From UCR locality RV—7312 in the late late Miocene part of the Almejas Formation, Cedros Island, Baja Calif. Collected by R. H. Tedford (field number RHT 1294) about 50 feet above the base of the formation and in the same beds as Aivukus cedrosensis. USNM 23869, associated left scapholunar and magnum from the base of the Purisima Formation 5,500 feet southwest of the presumed type locality and possibly 40 feet lower in the section. Collected by C. A. Repenning and J. C. Clark in 1965. USGS local- ity M1 109 on seacliff west of Point Santa Cruz (fig. 2). These specimens are from a glauconite bed which has been dated at 6.7 i 0.5 m.y. (J. D. Obradovich, written commun., 1965, KA 396). Questionably referred material.—UCR 15245, a fragment of a large right mandibular ramus having the pterygoid process, articular process, and part of the coronoid process. Collected by R. H. Tedford (field number RHT 1309), at UCR locality RV—7313, about 15 feet above the base of the Almejas Forma- tion. UCMP 83370, associated elements of a left fore- limb from the Drakes Bay Formation of Galloway (1977) (pl. 15). Collected by J. H. Hutchison, D. P. Domning, and L. G. Barnes, 1968. Drakes Beach, Point Reyes, Calif., UCMP locality V—6930. This material is from a glauconite bed which has been dated by Geochron Laboratories at 9.3 i: 0.5 m.y. (A. J. Galloway, oral commun., 1970). The magnum of this specimen is the only bone well enough preserved to recognize. USNM 23891, an immature right radius lacking the distal articulation from the Purisima Formation (Glen, 1959, p. 160) at Moss Beach, San Mateo County, Calif. Collected by Evelina Dunton of San Mateo, Calif., in March 1970, at USGS locality M1245, a point about 100 feet west of the bluff at the mouth of San Vincent‘e Creek in rocks exposed only at low tide. This locality is about 50 miles northwest of Santa Cruz, and the strata may be somewhat older than those included in the Purisima Formation at Santa Cruz for Glen (1959, p. 164) suggests that they are "probably middle or perhaps even early Plio- cene"; late late Miocene in the usage of the present report and possibly from 6 to 8 m.y. old. UCMP 65318, an, adult right humerus from the same locality as the preceding specimen at Moss Beach, Calif., UCMP locality V—6531. Type locality—Considerable question exists as to the exact locality from which the type specimen was collected (see Mitchell, 1962, p. 20—21). The locality map in UCMP shows the locality, V—2701, to be about 1,350 feet north of Point Santa Cruz lighthouse along the seacliff. On the other hand, the original description is rather explicit in latitude and longi- tude and in its description of the lithology at the locality. These facts seem to designate the seacliff along East Cliff Drive in the Seabright District of the City of Santa Cruz as the locality. This would be 1.3 miles east-northeast of the locality shown on the UCMP locality map and higher in the Purisima Formation where the sandstone "becomes quite soft when weathered.” At some points in this locality, the seacliff has receded as much as 200 feet since 1925 and East Sea Cliff Drive is no longer a continuous street. Remains of cetaceans are relatively abundant in the seacliff. Kellogg (1927, p. 28) describes the type locality as being between Seabright and ”the lighthouse.” The only lighthouse that appears on old maps is at Point Santa Cruz, and, as pointed out by Mitchell (1962, p. 21), this description covers about 1% miles of coast adjacent to the City of Santa Cruz. A further compli- cation, but one which greatly reduces the extent of coastline covered by the words ”between Sea- bright***and the lighthouse” was discovered by Jane Knapp, of the Remington Kellogg Library of Marine Mammalogy in the National Museum of Natural History. This complication was a sketch map of the locality in a manuscript of Kellogg’s marked "Geo- logical Correlations.” On this map Kellogg mis- located ”Seabright” at the headland above Cowell Beach just west of the present Municipal Pier (fig. 2), 3,700 feet north of the lighthouse on Point Santa Cruz, and indicated the type locality of Dusigna thus santacruzensis as being along the seacliff, just below West Cliff Drive, approximately 2,800 feet north of the lighthouse, roughly 1,500 feet north of the locality on the map filed at UCMP (fig. 2). There appears, at present, to be no way to deter- mine which of the three possible localities is more nearly correct. Arbitrarily, the middle one, shown on Kellogg’s sketch, is here considered most likely. DISCUSSION OF THE TYPE In describing the type specimen, Kellogg (1927) failed to note that the very long and relatively slender canines show a remarkable difference in wear: the upper ones show virtually no wear and the one lower canine whose crown is preserved is worn completely beyond the base of the enamel on its anteromedial side. The extreme tips of the upper PART I: WALRUSES 45 canines are truncated, whether from wear or fracture is difficult to determine, and there is a slight wear facet, not cutting through the thin enamel, on the anteromedial base of each upper crown which is matched by a slight wear facet on the posterolateral base of the crown of the lower canine. The great wear of the lower canine was obviously against one or more upper incisors. Kellogg does not mention the extreme narrowness of the mandible at the symphysis: when the two rami are placed together, there is barely 5 mm between the two lower canines at the alveolar margin. If lower incisors were present, they were anterior to the canines, certainly not between them. The combina- tion of a very narrow chin, occlusion of the lower canines against the upper incisors, and lack of normal wear of the upper canines against the lower indicates that Dusignathus santacruzensis had an unusual dentition in which the upper canines were laterally displaced from the lower dental arcade and hung, tusk-like, beside the lower jaw. Despite the equally elongate lower canines, the muzzle of Dusig- nathus must have been very reminiscent of a walrus. The mandibular rami (pl. 5) are deep and narrow and are bent dorsally at the region between the anterior termination of the coronoid process and the digastricus insertion. The condyloid processes are missing, but a sharp crest extends from the lateral termination of the condyle anteriorly along the inferior margin of the masseteric fossa forming a prominent lateral shelf for insertion of the masseter. A comparable but less prominent shelf is present on the mandible of Imagotaria downsi. Kellogg (1927, p. 32) notes that the enamel on all teeth is thin and very smooth. Most teeth show evidence of corrosion or imperfect deposition of enamel, mainly near the base of the crowns. The roots of all cheek teeth are short, curved, swollen pegs whose greatest diameter exceeds that of the crown. The poor deposition or the corrosion of the enamel near the crown bases makes it difficult to be certain, but there appears to be no indication of lateral or medial cheek tooth wear; all clear evidence is on the anterior and posterior faces of the cheek teeth. Although the crowns are nearly conical and roundly lanceolate, their sharpness is increased by this wear, which contrasts markedly with the wear seen on Aivukus cedrosensis and Odobenus. It ap- pears that the lower cheek teeth of Dusignathus occluded between their upper counterparts as in some phocids, particularly Halichoerus, without any placement medial to their upper counterparts as is common in the otarioids. The temporal of the type of Dusignathus santa- cruzensis is very walruslike. Kellogg compared the fossil with Eumetopias, Zalophus, Mirounga, and Cystophora. In view of the walruslike temporal and peglike teeth, it is rather remarkable that possible odobenid relationships did not occur to him. In those parts preserved, the temporal of Dusig- na thus is nearly identical to Imagotaria, particular- ly the female skull. This similarity has been noted and described by Mitchell (1968, p. 1885—1886), along with some dissimilarities. Odobenid features include an extremely large tympanic membrane (measuring nearly 12 mm in its greatest diameter), a very large epitympanic recess (indicating the former presence of walrus-sized ossicles), a tentorium that must have been closely appressed to the now missing capsule of the semicircular canals, and a large bony eustachian canal. The hyoid fossa was very deep, extending upward nearly to the epitympanic recess, and there must have been a well-formed eminentia vagina processus styloidei protruding into the middle ear cavity although this wall of the fossa has been destroyed and all that remains is the trace of the fossa on the medial side of the mastoid process. These features, the deep symphysis of the mandible, and single-rooted cheek teeth in late late Miocene time, when all known otariids had double-rooted teeth, indicate that Dusignathus santacruzensis is an odobenid. Failure to reduce the lower canine while enlarging the upper is characteristic of the dusig- nathine odobenids. The dorsal fragment of the occiput and adjacent parts of the parietals show a weak sagittal crest and a strong but low occipital crest which, except for larger size, is quite similar to this area on the female skull of Imagotaria downsi. DISCUSSION OF REFERRED SPECXMENS LACM 301 1, which is an incomplete anterior limb, was described by Mitchell (1962). He fully recognized the odobenid nature of the limb bones but, following Kellogg, did not then recognize the odobenid nature of Dusignathus and hence did not suggest an associ- ation between the two, even though they appear to have come from the same stratigraphic interval and from localities about 1,800 feet apart as he judged the position of the type locality, or about 3,000 feet apart as the position of the type locality is herein arbitrar- ily selected. Except for general odobenid features, which Mit- chell clearly defines, these bones are significant in the following features: radius and ulna short and stout; radius with the radial process (distal termina- tion of the radial crest) radial and medial to the distal articular surface—conspicuously medial in location 46 in distal aspect; ulna without the prominent ole- cranon of Imagotaria and the proximal radial articu- lation long and narrow; metacarpal I elongate, not flattened proximally, with proximal articulation saddleshaped and extending onto the dorsal surface, and with a distinct basin on the dorsal surface proxi- mal to the insertion for the pollical extensor. These limb bones bear a far greater resemblance to those of the dusignathine Imagotaria downsi from older strata than they do to those of the odobenine Aivu- kus cedrosensis of more comparable age or to those of living Odobenus. They are, however, not identical to those elements of Imagotaria. The most conspicuous differences from the com- parable elements of Imagotaria are the lesser devel- opment of the olecranon and the long radial notch of the ulna, distal narrowness of the trapezoid, distal broadness of the unciform, presence of an articular surface for the scapholunar on the cuneiform, and shorter and more robust metacarpals IV and V; all of which enhance the similarity of these elements to those of Odobenus. In view of the stratigraphic association with the type of Dusignathus and their great similarity to other dusignathine limb elements, this material is here referred to Dusignathus santacruzensis. LACM 4342, described by Mitchell in the same paper as an odobenid fibula, cannot be compared to any dusignathine odobenid as no unquestionably dusignathine fibula is known. Its stratigraphic asso- ciation, however, at or very near to the point of discovery of LACM 3011, and its odobenid nature certainly suggest referral to the same animal, as was done by Mitchell (1962, p. 12). This fibula has the proportions of living Odobenus, however, and does not show the short, stocky nature of the anterior limb elements. A fibula from the Etchegoin Formation, questionably referred to Pliopedia in the following pages, does show an extremely short and stocky nature. USN M 23869 is an associated left scapholunar and magnum from USGS locality M1109 in the glaucon- ite bed at the base of the Purisima Formation at a point 3,300 feet farther west than the anterior limb LACM 3011. These two bones are also referred to Dusignathus santacruzensis on the basis of dusig- nathine characters in. combination with some fea- tures more similar to Odobenus than Imagotaria and because of stratigraphic and geographic association. This locality may be as much as 40 feet stratigraphic- ally lower in the Purisima Formation and is about 1 mile from where Kellogg indicated the type locality on his manuscript map previously referred to. We believe that this association is close enough to rely OTARIOID SEALS OF THE NEOGENE upon the assumption that ”similar species are allo- patric” and that one species is represented. This scapholunar appears dusignathine in that the pocketed articulation for the magnum is long and extends about as far in a palmar direction as does the articular surface for the unciform. This condition matches that of the scapholunar of Imagotaria downsi, and, as there noted, is consistently different from 16 Odobenus, 3 Aivukus, and 1 ”Trichecodon” scapholunars. And this scapholunar has a pro- nounced facet for articulation with the cuneiform, matching the presence of the opposing facet noted on the cuneiform of the anterior limb discussed above and referred to Dusignathus santacruzensis. The associated magnum of USNM 23869 has an angle of 109° between the two articular facets for metacarpals III and IV (pl. 15), greater than the range or 72° to 90° found in the magnum of Odobenus and less than the 123° found in Imago- taria. The dorsal face of the Odobenus magnum is triangular in outline, that of Imagotaria narrowly trapezoidal because the dorsal edge of the scapho- lunar articulation parallels and is close to the dorsal edge of the metacarpal IV articulation. Although the metacarpal IV articular surface on the magnum of USNM 23869 faces less distally than does that of Imagotaria, the dorsal part of the articular surface for the scapholunar also faces less proximally and the two surfaces remain approximately parallel. The surfaces are farther apart than in Imagotaria, how- ever, and the dorsal face of the magnum of USNM 23869 is broadly trapezoidal and quite intermediate between Odobenus and Imagotaria. In other re- spects, this magnum is identical to that is Imago- taria: most notably the scapholunar articular sur- face extends completely to the palmar surface, re- flecting the elongate articular pocket on the scapho- lunar. UCR 15244, a portion of a pinniped snout from the lower part of the Almejas Formation of Cedros Island, is from one of the largest known otarioid pinnipeds (pl. 18). It is distinguished by an extremely long and slender canine (to judge by the long and slender root) whose alveolar diameter is no greater than that of P1; by smallcrowned cheek teeth with very thick and swollen, curved, peglike roots; by very thin and smooth enamel crowns on the cheek teeth of essentially conical form; by cheek tooth wear only on the anterior and posterior surfaces; by a lateral incisor that, while not equaling the canine in diam- eter, is stout and extremely long-rooted; and by an extremely small incisor-canine diastema, suggesting that the lower canine occluded with this large incisor rather than fitted into the diastema between the PART I: WALRUSES incisor and canine. The palate is not markedly‘ arched nor is the infraorbital foramen greatly en- larged for a pinniped. Although generally quite similar to the type of Dusignathus santacruzensis, this specimen from Cedros Island does differ by having much greater size, stouter cheek tooth roots, more nearly conical crown on the cheek teeth, and longitudinal fluting on the very dense cementum and dentine of both the canine and lateral incisor, similar to that on the tusks of "Trichecodon." These differences are here believed to be the expression of sexual dimorphism; the larger specimen from Cedros Island is believed to be a male and the holotype is believed to be a female. The dimensions of UCR 15244 are given in table 1 1. DISCUSSION OF QUESTIONABLY REFERRED SPECIMENs Those specimens questionably referred to Dusig- nathus santacruzensis are, for the most part, so questioned because they are sufficiently removed from the type, both in distance and in stratigraphic correlation, that the assumption of ”different species being allopatric" cannot be used in defense of spe- cific identity. The single exception is UCR 15245, which will be discussed first. All specimens are odobenid, some are demonstrably dusignathine, but none can be assigned to the other odobenid species previously discussed. UCR 15245 is a gigantic mandibular fragment from UCR locality RV-7313 in the lower part of Almejas Formation on Cedros Island. The condyle measures 107 mm in width, and the largest modern male Eumetopias jubata available for comparison measures 57 mm in the same dimension. This spe- cimen is questionably referred to Dusignathus san- tacruzensis because of general stratigraphic associa- tion, very large size, and the suggestion of dusig- nathine affinity in the very broad and sharp-edged shelf (where the superficial masseter inserts) that extends anteriorly from the lateral end of the con- dyloid process to the horizontal ramus, and the anteroposteriorly short pterygoid process that ex- tends only slightly posterior to the inferior sigmoid notch. UCMP 83370, a partial anterior limb from the glauconite at the base of the Drakes Bay Formation at Point Reyes, Calif. (pl. 15), contains a rather well- preserved magnum. In those features considered to be dusignathine characters and in those considered different from the magnum of Imagotaria downsi, as well as other features except size, this magnum is identical to the magnum of USNM 23869 that was discussed and referred to Dusignathus. The angle of 47 TABLE 11.—Dimensions of Dusignathus santacruzensis, referred specimen UCR 15244 Measured parts mm Canine root, apex to alveolus ........................................................ 104 Canine diameter at alveolus, anteroposterior .. _____ 30.0 Canine diameter at alveolus, transverse ................. 22.8 Laminated cementum ofcanine, thickness... ..... 1.5, Canine-incisor diastema at alveoli ......................... 5.2 Lateral incisor alveolar diameter .................. 18.9 'P1 alveolar diameter, anteroposterior .......... 22.8 P1 alveolar diameter, transverse .................... 24.0 P1 diameter at base of crown .......................... 16.2 P1 height, lateral alveolar lip to apex .......... 24.4 Rostral width across canines .......................... 118.2: Rostral width across P3 ........................... 120.6i Root length P3, alveolus to apex .................................................. 26.8 the dorsal face formed by the facets for metacarpals III and IV is 116°; it is somewhat more similar to that of the Imagotaria magnum, but the dorsal face is broadly trapezoidal. Judged by the other bones of this limb, the specimen appears to be from an adult animal and because of its smaller size is believed to be female. The glauconite from which this specimen was collected has had a K/A age determination of 9.3 i 0.5 m.y. according to A. J. Galloway (oral commun., 1970), and thus the rock unit appears to be older than the Purisima Formation in the Santa Cruz area 100 miles farther south. However, the identity of the fur seal found in this same unit, associated fossil ceta- ceans being studied by L. G. Barnes, and the avail- able fossil invertebrate evidence all suggest a con— siderably younger age than indicated by the K/A date, as will be discussed in connection with the description of the fur seal from the same locality (see also the section "Type locality and age" of Aivukus cedrosensis). USNM 23891 is a small immature odobenid radius lacking the distal epiphysis; the distal position of the pronator origin on the anterior surface marks it as an odobenid (pl. 16). The articular surface on the sides of the head, for articulation with the radial notch of the ulna, is continuous from medial to lateral sides, indicating a high degree of supination relative to the great majority of known odobenines and equaling the condition found in Imagotaria. This character- istic suggests that the specimen is that of a dusig- nathine odobenid. The occurrence of this specimen in the Purisima Formation at Moss Beach, San Mateo County, Calif, suggests the possibility that it repre- sents a juvenile Dusignathus santacruzensis and it is questionably referred to this species. The specimen was found 50 miles northwest of the type locality, and about an equal distance southwest of the ante- rior limb from the Drakes Bay Formation discussed. 48 OTARIOID SEALS OF THE NEOGENE UCMP 65318 is a stout, adult, odobenid humerus also from the Purisima Formation at Moss Beach, Calif. (pl. 16). The combination of a pectoral crest directed distally toward the medial lip of the trochlea and an anteroposterior diameter of this lip far exceeding that of the distal capitulum marks this specimen as an odobenid. It is further distinguished by a distinct lateral bowing of the shaft, by a moderately broad U—shaped bicipital groove, by a greater tubercle which is about the same height as the head, by a very prominent medial epicondyle, by remarkably weak development of the insertional area for the deltoideus muscle on the lateral side of the pectoral crest, and by a conspicuously short shaft relative to the size of the head and the distal struc- tures. These features distinguish this humerus from other odobenid humeri thus far considered. The greatest length of this specimen is 271.6 mm, the greatest transverse width across the epicondyles 122.8 mm, and the greatest anteroposterior width across the head to greater tubercle 130.0 mm. Al- though the distal termination of the pectoral crest has been lost because of the midshaft fracture, it would appear to have dropped rather abruptly, dis- tally, to the shaft. Because of its odobenid nature, oc- currence in the Purisima Formation, dissimilarity to other known fossil odobenids so far discussed, and shortness of the shaft (which might well parallel the shortness of more distal elements referred to Dusig- nathus), this specimen is also questionably referred to Dusignathus santacruzensis. Though not iden- tical, this humerus is similar to the humerus of Plio- pedia and Valenictus in several characters, dis- cussed under these genera, in particular the ex- tremely low greater tubercle. CONCLUDING DISCUSSION Stratigraphic and geographic association in com- bination with uniqueness of structure have here been relied upon to refer, with varying degrees of doubt, several dusignathine postcranial skeletal elements to Dusignathus santacruzensis, the holotype of which consists entirely of cranial elements. As with the cranial elements, the postcranial elements show some specializations paralleling those of the odobenines and some unique features which sharply separate them from the older dusignathine genus Imagotaria. Those features of. the holotype and referred spe- cimens which appear to characterize the Subfamily Dusignathinae, insofar as it presently is known, are: (1) lower canine not reduced relative to upper canine, (2) radius and ulna short and stout with a conspicu- ously medial position of the radial process at the distal end of the radial crest of the radius, (3) a pock- eted articular surface for the magnum on the scapho- lunar that is elongate and extends about as far in a palmar direction as the articular surface for the unci- form, (4) in agreement with the last, a scapholunar articular crest on the magnum which continues in a palmar direction to the palmar process without break, a metacarpal III to IV articular facet angle of 110°—130° on the magnum, and a dorsal surface on the magnum that is trapezoidal rather than roughly triangular, and (5) slender metacarpal I with pro- nounced basin on the dorsal surface between the head and the insertional spur for the pollical ex- tensor. Those features of the holotype and referred speci- mens of Dusignathus santacruzensis which appear to represent specializations paralleling those of the odobenine odobenids which are not so conspicuously developed in the older dusignathine genus Imago- taria are (1) cheek teeth with stoutly peglike roots, (2) tympanic membrane very large, (3) medial epicon- dyle of the humerus prominently pointed and up- turned, (4) olecranon of ulna not greatly enlarged, (5) cuneiform articular facet on the scapholunar and scapholunar articular facet on the cuneiform, and (6) proportions of the distal articulations of the trapezoid and unciform. Those features of the holotype and referred speci- mens which at present appear unique to Dusignathus santacruzensis are: ( 1) enlarged upper lateral incisor which occludes with the lower canine, (2) very narrow mandibular symphysis, (3) distinctive cheek tooth occlusion and wear that is restricted to anterior and posterior surfaces, and (4) short stocky humerus with very conspicuous lateral bow to the shaft and very weak anconeal crest and deltoideus insertion. Dusignathus santacruzensis lived in association with Aivukus cedrosensis in the southern waters of Baja California, but nothing has been found in the more northerly deposits of central and northern California which could be assigned to the primitive odobenine Aivukus. In Baja California, the Santa Cruz area, and the Point Reyes area, Dusignathus is found associated with a primitive otariid genus to be described in Part II of this report. If all referrals to this species are correct, there appears to have been only one dusignathine odobenid living along the California and Mexican open coast of the North Pacific Basin from questionably 8 to less than 5 my ago. It replaced the earlier dusignathine genus Imagotaria. Two possible exceptions are the genera Pliopedia and Valenictus. N orthward along the Oregon coast, Dusignathus is either replaced by or equal to the genus Pontolis. PART I: WALRUSES 49 Genus PLIOPEDIA Kellogg Type species.—Pliopedia pacifica Kellogg, 1921. Diagnosis—The dentition and facial portion of the skull are unknown; sagittal and lambdoidal crest lacking and supraoccipital area for insertion of neck muscles low and broad; carotid canal very large, and petrosum and bulla as in Odobenus. Humerus with greater tubercle no higher than head; lesser tubercle large, curved medially, and sloping distally; bicipital groove deep and narrow; deltoid tubercle not on pectoral crest but on lateral surface of shaft; internal epicondyle directed posteriorly as well as internally; shaft straight and not shortened or curved. Ulna short and stout with moderate depth of olecranon process. Radius short and stout with radial process medial in position. Metacarpal I with dorsal depres- sion distal to proximal articulation; more slender than in Imagotaria. Distribution—Known only from the late late Mio- cene inland sea of central California and possibly restricted to this body of water as the contempora- neous genus Dusignathus appears to have occupied the open coasts of California and northern Mexico at this time. Pliopedia pacifica Kellogg, 1921 Plates 4, 14, 16, 17, and 24 Holotype.—USNM 13627 (formerly SU537): por~ tions of both left and right forelimbs collected by Robert Anderson in 1909. Referred material.—USNM 187328, portions of left. and right forelimbs, braincase of the skull and one rib collected by W. P. Woodring in 1932 and C. A. Repenning in 1974 (Woodring and others, 1940, p. 98, their locality 350). USNM 187337, a fibula, and USNM 187338, a phalanx, are questionably referred to Pliopedia pacifica because they are odobenid elements found in the same stratigraphic zone as was USNM 187328. Type locality and age—From the basal conglom- eratic member of the Paso Robles Formation "on summit of hill, one mile southeast of the town of Santa Margarita” (Kellogg, 1921), San Luis Obispo County, Calif. Marine mollusks from the same unit 8 miles to the north suggest an early Pliocene age according to Addicott and Galehouse (1973), late late Miocene in the usage of this report. Addicott and Galehouse (1973) found the inverte- brate fauna of the Paso Robles Formation unique in that is has some seemingly late Pliocene forms as well as forms known from the lower part of the Plio- cene of the Kettleman Hills, Calif. Because of a previously established extreme variability of the apparent late Pliocene form, they placed more re- liance on the early Pliocene form. Woodring, Stew- art, and Richards (1940, chart following p. 78) show this early Pliocene form, Ostrea atwoodi Gabb, as occurring from the Upper Pseudocardium zone to the Macoma zone of the Etchegoin Formation in the Kettleman Hills. Within this stratigraphic range of Ostrea atwoodi, Woodring collected, in 1932, an odobenid radius and fragments of its humerus which were identified by Remington Kellogg as an "eared seal***wholly unlike any known extinct and living otarid***” (Woodring and others, 1940, p. 98, their locality 350). Excavation of this old locality in 1974 produced a complete humerus, two ulnae, and a brain case. Largely on the strength of unique features of the humerus, these are identifiable as Pliopedia pacifica Kellogg. The fragments of the radius and ulna of the holotype are also identical to those of the specimen from the Etchegoin Formation. A tooth of the horse Pliohippus sp. is known from this zone (J. W. Durham, oral commun.), and other teeth have been found in other parts of the Etchegoin of the Kettleman Hills (Durham and others, 1954). The upper Pseudocardium zone of the Etchegoin Formation lies roughly 400 feet down section from the Neverita zone of the overlying San Joaquin Formation in which Woodring collected a different odobenid (Woodring and others, 1940, table on page 46). This odobenid is identifiable as Valenictus imperialensis Mitchell, 1961. Midway between the Neverita zone and the younger Pecten zone of the San Joaquin Formation, a tuff bed occurs along Arroy Doblegado on the east side of North Dome of the Kettleman Hills. This tuff has been dated as 4.3 m.y. old (J. D. Obradovich, oral commun., 1975). The Pecten zone contains a Blancan mammalian fauna. Valenictus is thus older than 4.3 m.y., and Plio- pedia is considerably older than Valenictus, though distinctly younger than the Imagotaria-Pithano— taria seal fauna which is found in rocks that appear to be as young as 9 m.y. As will be discussed, the similarities between Pliopedia and Valenictus are such that the possibility exists that_ Valenictus evolved directly from Pliopedia in the inland sea of California. The presence of Pliopedia in the Etchegoin and the suggestion of a rather short temporal range for this genus in this stratigraphic succession strongly supports the age assignment of the basal Paso Robles Formation by Addicott and Galehouse (1973). The type specimen of Pliopedia pacifica is here considered to be from 5 to 6 m.y. old, late late Miocene in the usage of this report. DISCUSSION Kellogg (1921) noted odobenid similarities in this 50 OTARIOID SEALS OF THE NEOGENE fossil, although he tentatively referred it to the Otariidae. Mitchell (1962, p. 22) also recognized odobenid relationships and later regarded it as an odobenid (1966, p. 38-39). In 1968 Mitchell (p. 1880- 1881) pointed out differences between the humerus of Pliopedia pacifica and Imagotaria downsi. These differences showed that "the two species are clearly distinct” although he did not mention why they were generically distinct. Humerus (pl. 1 7).—Kellogg (1921)failed to note that the preserved distal half of the left humerus of the type, as well as the fragments of the radius and the ulna, are badly crushed; in fact he stated (p. 219) that the ”most striking general characteristic [of the humerus] is the antero-posterior compression of the distal end***" This compression is obviously a result of the overall crushing and distortion of the speci- men whereby the distal articulation and the medial epicondyle have been rotated forward and upward approximately 30° such that the medial lip of the trochlea is actually pressed against the distal end of the pectoral crest. The distance from the region of the pectoral insertion to the trochlear crest has been halved by this rotation. The supinator crest, leading distally to the medial epicondyle, seems remarkably well developed and flaring, as noted by Kellogg, and it appears that this feature has been accentuated by the anteroposterior crushing. Kellogg did note a ”lateral compression of the shaft in the deltoid region." The anteroposterior diameter of the shaft near the most prominent part of the pectoral crest is 86 mm, whereas the lateral diameter at the same point is about 50 mm, without compensation for the anteroposterior compression of the proximal root of the supinator crest, which somewhat exaggerates these measurements. The insertion for the deltoideus muscle is not evident. Correcting for the distortion of the specimen, the distal half of the humerus of Pliopedia pacifica is essentially identical in size and form to UCMP 65318 from the Purisima Formation at Moss Beach which has been questionably referred to Dusignathus san- tacruzensis with one conspicuous exception: the medial epicondyle is directed medially and posteri- orly, as noted by Kellogg (1921, p. 220 and fig. 1d). This posterior orientation of the internal epicondyle is typical of Odobenus, but not of most odobenids including the humerus referred to Dusignathus. Though weathered and leached by root action, the right humerus of USNM 187328 from the Etchegoin Formation of Middle Dome, Kettleman Hills, Calif, is complete and undistorted. It is a long and moder- ately stout humerus not showing the shortening of the humeri of Dusignathus and Valenictus, and the shaft is not curved as is the humerus of Dusig- nathus. The humeri of all three genera are distin- guishecéJ by an exceptionally low greater tubercle which oes not protrude dorsally beyond the hum- eral head; this feature is unique among the humeri of odobenids, although it is present in at least some desmatophocids. The lesser tubercle of Pliopedia is unique in that its dorsal crest slopes distally and the tubercle is curved prominently medially. The bicip- ital groove is very narrow and has a nearly V-shaped cross section, comparable only to Valenictus. The insertion of the deltoid muscle is not obviously marked on the humerus of Dusigna thus, but, because of the rather sharp lateral margin of the pectoral crest, it appears most likely to have inserted on that crest. Except for the genus Odobenus (including Trichecodon), Pliopedia and Valenictus are unique among odobenids in that the tubercle for insertion of the deltoid muscle is clearly removed from the pee- toral crest and is on the lateral surface of the shaft. Except for those features obviously caused by distortion of the humerus of the type specimen, the humerus of the specimen from the Etchegoin Forma- tion is identical. Most distinctive is the posterior curve of the medial epicondyle. This condyle other- wise looks quite usual, being triangular with its lower margin directed dorsomedial, rather than like the swollen knob seen on Valenictus. This humerus measures 306 mm in greatest length from head to distal articulation, 80 mm in head diameter, 133 mm in greatest transverse diameter from the posterior margin of the head to the anterior border of the pectoral crest, and 129 mm across the epicondyles. Ulna (pl. 24).—A fragment of the left ulna of the type specimen, preserving much of the proximal articulation, appears to have been buried beneath the humerus in a position such that the ulnar shaft posterior to the anconeal process lay beneath the olecranon fossa of the humerus; compaction of the deposits forced the supinator crest of the humerus downward onto the trochlear notch of the ulna and the entire proximal articulation of the ulna is dis- placed medially relative to its shaft. Possibly the missing olecranon process of the ulna was respons- ible for the displacement of the distal articulation of the humerus. The depth, or anteroposterior diameter of the ulna shaft below the trochlear notch, indicates that the olecranon was not the deep, hatchetlike process as in Imagotaria, instead it must have been much like Odobenus and Dusignathus. Also Dusignathus-like, the radial notch is narrow, rather than being nearly circular as in Imagotaria. In regard to these features, the ulnae of the specimen from the Etchegoin Forma- PART I: WALRUSES 51 tion match the holotype and, in addition, the ulnae are seen to be short and stout as are those of Imagotaria. The distal tip of the right ulna of the type and those of the ulnae from the Etchegoin Formation show the distinctly odobenid separation of the distal radial articulations and also a greater development of a pisiform articular surface than has been seen in Odobenus. Radius (pl. 24).—The fractured and distorted head of the left radius of the holotype has the ulnar articulation very well developed on the lateral side of the head; it is quite comparable to the radius of both Imagotaria and Dusignathus. This articular sur- face is less developed on the radius from the Etche- goin Formation, a variation common in the radii of Odobenus. The complete radius from the Etchegoin Forma-' tion is 22.5 cm long, has a greatest proximal diameter of 6.95 cm, and an anteroposterior distal diameter of 8.73 cm. This bone is clearly that of a dusignathine odobenid, but it cannot be separated definitely from the radii of either Imagotaria or Dusignathus, al- though the distal articulation for the scapholunar is more oval, as in Dusignathus, rather than roundly rectangular as in Imagotaria. The radii of Pontolis and Valenictus are unknown. Right trapezoid—This nearly complete bone from the type specimen, called the fifth metatarsal by Kellogg (1921, p. 226), lacks the palmar process but appears more similar to those of Imagotaria than Dusignathus (judged by the published sketch, Mitchell, 1962, fig. 4) in that the distal articulation is broad dorsally but narrows in a palmar direction. It is less similar to that of Odobenus than to either of these, but it is quite comparable to the trapezoid of Aivukus cedrosensis. A reasonably large sample of trapezoids of Odobenus was not available to eval- uate individual variation. Right metacarpal I (pl. 14).—This bone of the type specimen was broken, and the distal part was con- sidered questionably the right metacarpal IV while the proximal part was recognized as the right meta- carpal I by Kellogg (1921, figs. 7 and 10). It bears a marked similarity to both Imagotaria and Dusig— nathus in being slender, having a saddle-shaped proximal articulation, and in having a basin on the dorsal surface proximal to the insertion of the polli- cal extensor. This bone is somewhat smaller and much more delicate than the first metacarpal of a male Imagotaria. Because the first metacarpal of Dusignathus is not complete, its relative slenderness cannot be compared. Right metacarpal II.—Called ”right metatarsal II?” by Kellogg (1921, fig. 9): this bone of the type specimen is represented only by its head, which is quite similar to this bone in Imagotaria, narrower in proximal aspect than this bone in Aivukus cedro- sensis, and very narrow in comparison with meta- carpal II of Odobenus. The bone is not present in the specimens referred to Dusignathus santacruzensis. Right metacarpal III (pl. 4).——Called ”left” by Kel- logg (1921, fig. 8) and differing from this bone in Imagotaria in that the articular surface for meta- carpal II is smaller, though largely broken on the specimen and best judged by the counterpart on the head of metacarpal II. It differs in the same way from metacarpal III of Aivukus cedrosensis, and it cannot be compared with the weathered metacarpal III here referred to Dusignathus from the published descrip- tion (Mitchell, 1962, p. 1 1). This bone has a head that is much more rounded and indistinct in structure in Odobenus, but this element of Pliopedia, Imagotaria, and Aivukus cannot be differentiated between these genera. Metacarpal proportions (table 12).—Pli0pedia pa- cifica has a smaller and more delicate metacarpal I than Imagotaria downsi. Although metacarpal III of Pliopedia measures 90 mm in greatest length and is only 4 mm (4 percent) shorter than the same element of Imagotaria (male articulated limb, USNM 23859), the greatest length of the Pliopedia metacarpal I is only 123.5 mm and is 24 mm (16 percent) shorter than this element in the articulated limb of Imagotaria. Except for this difference in metacarpal proportions, the metacarpus of Pliopedia falls well within the range of variability of Imagotaria, in both size and morphology. This difference, in combination with the strong differences in the skull, humerus, radius, and ulna, appears to separate Pliopedia from Imago- taria. Because metacarpal I is not complete in the limb referred to Dusignathus santacruzensis (LACM 3011), it is not possible to judge whether its meta- carpal proportions are more like those of Imagotaria or Pliopedia. It is to be noted, however, that the differences in relative lengths of metacarpals I and III between Pliopedia and Imagotaria are exactly those which exist between Eumetopias and Odo- benus. It would appear that the relatively longer metacarpal I of Imagotaria indicates the sea lionlike adaptations already suggested for Imagotaria by such features as the dentition and tympanic mem- branezoval Window area ratio. Conversely, the pro- portionately shorter first metacarpal of Pliopedia pacifica would suggest more walrus-like adapta— tions, theoretically including dentition and ear ratio, such as are present in the type of Dusignathus 52 OTARIOID SEALS OF THE NEOGENE TABLE 12,—Metacarpal measurements of three dusignathine genera Species Imagotaria downsi Dusignathus santacruzensis Pliopedia pacified, referred (males) referred type Specimens USNM 23859“ USNM USNM USNM LACM 30112 USNM 13627 23860 23861 184055 Greatest measurement MCI MC 11 MC III MC IV MC V MC IV MC III MC V MCI MC II MC IV MC V MC I MC II MC III Length ................................................................ 147 104 94 89 91 86 114 71 78 122 89 Proximal width 46 26 26 35 26 Proximal height rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr 32 35 30 31 4:1 Distal width 30 30 26 27 28 Distal height ................................................... 25 21 20 21 23 Minimum shaft width 21 18 16 19 ‘ Limb. santacruzensis. Skull (pl. 24).—A very badly weathered and frag- mented braincase was discovered with the limb bones from the Etchegoin Formation which, to- gether with one rib, make up USNM 187328. Only the dorsal part of the braincase is intact, but recovered fragments of the basicranial area include part of the right auditory bulla with the petrosal in place. The braincase is low and rounded, strongly resem- bling that of Odobenus. No sagittal crest is present; elongate parallel grooves, presumably leading poste- riorly to nutrient foramina, are present on either side of the midsagittal suture approximately at the junc- ture of the frontals and parietals. The occipital crest is a broad rugose area very much like Odobenus except that it is straighter and more transverse because of the less swollen braincase. The auditory bulla is low and rugose and has a broad surface for articulation with the basioccipital, and the carotid canal is very large, measuring 7.5 mm in diameter. The petrosal is also very similar to that of Odobenus, the internal acoustic meatus is very wide with almost complete separation of the facial and vestibuloco- chlear foramina. Possible rear limb elements (pl. 16).——From their Macoma zone in the lower part of the Etchegoin Formation, Woodring, Stewart, and Richards (1940) collected a walrus fibula and phalanx (USNM 187337 and 187338, from their locality 302a). These specimens were not mentioned in their report. The fibula is most remarkable in its stoutness; its length is 26.8 cm, the greatest proximal diameter 4.8, and the greatest distal diameter 5.2 cm. Fibulae from living Odobenus with comparably large distal and 28 14 25 40 38 2O 28 21 41 26 31 50 28 24 26 34 29 23 30 24 25 24 18 28 21 23 14 27 <18 JMitchell, 1962, p. 14. proximal diameters have total lengths that are about twice as great. This extreme massiveness of the pelvic limbs seems to parallel the massiveness of the humerus and radius noted for both Dusignathus and especial- ly Valenictus. However, the fibula from the Purisima Formation at Santa Cruz, herein referred to Dusig- nathus santacruzensis because of close stratigraphic and geographic proximity to the type specimen, is of Odobenus-like proportions and could not be the same species as that from the Etchegoin Formation. Be- cause the fibula (USNM 187337) is within the re- corded range of Ostrea atwoodi, from which the second specimen (USNM 187328) of Pliopedia pacif- ica was collected, it is possible that it represents this species. Possible synonymy.—Pliopedia pacifica seems to be of the same approximate age as Dusignathus santacruzensis, and in the Central Valley of Cali- fornia, Valenictus imperialensis seems to be only slightly younger than Pliopedia pacifica. Pliopedia pacifica differs from Dusignathus santa- cruzensis by having a more elongate and straighter humerus which has the deltoid insertion on the shaft rather than on the pectoral crest, a narrow bicipital groove, and in a backward bend of the medial ‘ epicondyle. In addition, the skull of Pliopedia lacks a sagittal crest, which is moderately developed in Dusignathus, and its occipital crest is a low broad rugose area as in Odobenus and not the actual crest of Dusignathus. Other known elements are very similar, with the possible exception of the fibula. Valenictus imperialensis, as will be discussed, is a genus and species known only from its humerus. Of PART I: WALRUSES 53 all known odobenid humeri, only the humeri of Valenictus, Pliopedia, and Odobenus have the tuber- cle for insertion of the deltoid muscle on the lateral surface of the shaft; in all other genera, the insertion is either obscure or clearly on the lateral margin of the pectoral crest. Of all known odobenid humeri, only those of Valenictus, Pliopedia, and Dusigna- thus do not have a greater tubercle that is distinctly higher than the humeral head. The unique features shared between Valenictus and Pliopedia and, to a lesser extent, Dusignathus strongly suggest affinity between these dusignathine odobenids. When they are better known, it may be reasonable to classify all three species, which are distinct, under the senior generic name Pliopedia. Genus VALENICTUS Mitchell Plate 16 Type species.—Valenictus imperialensis Mitchell (1961); LACM (CIT) 3926, a left humerus from the Imperial Formation on the east side of Coyote Moun- tains, Imperial County, Calif. DISCUSSION The age of the Imperial Formation was considered early Pliocene by Durham (1954, p. 27); it is con- sidered late Miocene in the usage here used. As discussed under the section ”Type locality and age" of Pliopedia pacifica, the referred incomplete hum- erus of Valenictus, USNM 13643, is from the Nev- erita zone of the San Joaquin Formation, Kettleman Hills, Calif, and is a short distance down section from a tuff bed in the San Joaquin Formation which was dated 4.3 my and up section from Hemphillian land mammals. This would suggest an age of from 5 to 6 m.y., late late Miocene or early Pliocene as used herein. Mitchell (1961) compared the type humerus in detail with the humeri of living and fossil otarioids. His report omitted comparison with UCMP 65318 from the Purisima Formation here questionably referred to Dusignathus, USNM 187328 from the Etchegoin Formation here referred to Pliopedia, USNM 13643 from the San Joaquin Formation here referred to Valenictus, and MCZ 7713 from the Yorktown Formation presumably referable to Pro- rosmarus. These specimens were unknown to Mit- chell at the time of his study. The odobenid affinities of Valenictus imperialensis were clearly defined by Mitchell. The humerus of Valenictus imperialensis is as markedly distinct from that of Prorosmarus as it is from the odobenid humeri considered by Mitchell. Most conspicuous of its features are the distinct lateral bowing of the shaft, an extremely prominent and distally positioned medial epicondyle, a very low greater tubercle, and a very short shaft relative to the size of the head and distal structures (compare Mitchell’s illustrations). In some of these features, it is similar to the humerus of Dusignathus from the Purisima Formation, and the two are easily separ- able from all other known odobenid humeri except that of Pliopedia. However, the humerus of Valenictus is not identi- cal to that questionably referred to Dusignathus (pl. 16). Although of equal length, the shaft of the humerus of Valenictus is thicker, the pectoral crest has an indistinct lateral margin, the deltoid inser- tion is marked with a prominent swollen area on the lateral side of the shaft, and the medial epicondyle is a distally located large round knoblike structure dissimilar to that known on any other odobenid humerus. The lesser tubercle is more massive on the humerus of Valenictus, and the bicipital groove is relatively narrower, though as deep as on the humer- us of ?Dusignathus. In addition, the distal articula- tion is relatively small, having a transverse diam- eter about 12 percent less than that of the distal articulation on the humerus questionably referred to Dusignathus. Because of this small distal articula- tion and the very large medial epicondyle, the distal articulation of the humerus of Valenictus has a transverse diameter that is only about half the diameter across the epicondyles. The lateral epi- condyle is notably small and inconspicuous on the humerus of Valenictus, the entire distal end of the bone giving the impression that the articular surface has been shifted laterally. A close relationship between Valenictus and ?Dusignathus seems obvious, but the differences between the two humeri are much greater than those known to be a result of individual or sexual varia- tion. In its distal structures, the humerus of Valen- ictus differs from that of Pliopedia (pl. 1 7) in the same ways that it differs from that of ?Dusignathus. In addition, Pliopedia lacks the shortening and bowing of the shaft seen in Valenictus and Dusignathus, but it does have the deltoid insertion on its shaft as in the humerus of Valenictus, and the humeri of the two genera have equally narrow bicipital grooves. It is to be noted that Dusignathus, Pliopedia, Pontolis, and Valenictus are all of the same approx- imate age, late late Miocene and (or) early Pliocene. These facts suggest either an appreciable variety of dusignathine odobenids in the North Pacific at this time or a considerable synonymy resulting from unfortunate application of new names to specifically indeterminate specimens. 54 USN M 13643 (pl. 16), the distal part of a humerus, was collected by Woodring, Stewart, and Richards (1940, p. 98) in the Kettleman Hills, Calif., from their Neverita zone of the San Joaquin Formation. The locality was approximately 700 feet down section from their Pecten zone, which contains land mam- mals of early Blancan age [comparable to the Hager- man fauna of Idaho, radiometrically dated at 3.5 m.y. (Evernden and others, 1964)] and approximately 300 feet down section from a tuff bed dated 4.3 m.y. Woodring, Stewart, and Richards quote Remington Kellogg’s comments about the fossil in which he noted the narrowness of the distal articulation and the distal position of the knoblike medial epicondyle, features known only in Valenictus. The lateral epi- condyle, while larger than that of the holotype, is noticeably less produced than that on the humerus referred to Dusignathus. The Purisima Formation of the Santa Cruz area includes deposits equivalent in age to both the Etche- goin and San Joaquin Formations of the Kettleman Hills area. Not only are they of the same age but they were also depositionally continuous (Addicott and Galehouse, 1973). The inland Etchegoin-San Joa- quin sea extended northward in a valley following the San Andreas fault, a valley now occupied by the San Benito River, and left this valley in the vicinity of Santa Cruz to merge with the open ocean. In the San Benito River area, the Etchegoin Forma- tion contained odobenid remains which were collect- ed by C. J. Bleifus and deposited now in UCMP (UCMP 112806—proximal end of a large femur, and UCMP 112806—proximal epiphysis of a large tibia). Southward the Etchegoin sea was connected by a second passage to the open ocean in the Santa Maria area by way of the region of Santa Margarita, where the type specimen of Pliopedia pacifica was found in the basal Paso Robles Formation. The overlying beds of the Paso Robles appear to be entirely contin- ental in origin, however, and there is no known record of the inland sea connecting to the open ocean by this route during the later time that the San Joaquin Formation was deposited. We are forced to conclude that three large odoben- ids of markedly different limb proportions lived dur- ing the late late Miocene and Pliocene in central Cali- fornia, one known from the open coast and the other two in chronologic succession from the warmer inland sea then present in the San Joaquin Valley. The younger of the two inland sea forms (Valenictus) must have frequented the open coast, as its type specimen is known from deposits of the ancient Gulf of California. OTARIOID SEALS OF THE NEOGENE Genus NEOTHERIUM Kellogg Plates 9 and 11 Type species.——Neotherium mirum Kellogg, 1931; USN M 11542, a right calcaneum which Mitchell and Tedford (1973, p. 266) have selected as the lectotype. Kellogg originally included this calcaneum and an astragalus, a cuboid, and a navicular as the hypo- digm, but Mitchell and Tedford have restricted the lectotype to the calcaneum on the grounds that the original material was not necessarily associated. The lectotype and referred material are all from the Sharktooth Hill bone bed in the Round Mountain Silt, Kern County, Calif. The deposits are of the Luisian Stage (Beck, 1952) and are estimated to be about 13-14 my old; early middle Miocene. DISCUSSION Kellogg clearly recognized the pinniped identity of the ”type material” and compared it with skeletal elements from some living otariids as well as with the sympatric Allodesmus kernensis and his ”Pon— tolis cf. magnus” (Kellogg, 1925b, pes from Towsley Formation). The latter is referred to Imagotaria sp. in the present report. Kellogg (1931, p. 302) noted that the cuboid of Neotherium was more like that of his ”Pontolis cf. magnus" than that of any living otariid. The lectotype calcaneum is more similar to that of Imagotaria downsi than it is to any other known pinniped, fossil or living. From this species it differs by being much smaller and more slender and the distal astragalar, or sustentacular, articular facet projects more medially. Most typically odobenid is the prominent internal tuberosity on the calcanear tuber. Typical of all primitive otarioids so far known, no sustentacular shelf is present, as was noted by Kellogg. Compared with fissipeds, the calcaneum is more bearlike than otterlike in that the sustentacular articulation is very near the distal end of the bone, at the cuboid articulation. It is otarioid in that the sustentacular articulation meets and merges, over a sharp crest, with the cuboid articulation, and the calcanear tuber is quite short relative to bears and is transversely broad at the heal. The calcanear pro- cess of the astragalus, extending posteriorly from the trochlea and so greatly enlarged in the phocids, is typically otarioid in form and characteristically is much larger than in the canoid fissipeds. The astragalus of the hypodigm also is odobenid character in its nearly vertical fibular articulation and relatively small lateral process that is alined with the lateral trochlear crest, rather than flaring outward in a distal-lateral direction. The head is transversely narrow relative to that of Imagotaria PART I: WALRUSES 55 downsi, otherwise the astragalus appears to be a miniature replica of that of Imagotaria. Mitchell and Tedford (1973) illustrate other limb elements from the type locality of Neotherium mir- um; they suggest that these elements may belong to Neotherium. A navicular, LACM 4733 (their fig. 20), strongly resembles this element in Kellogg’s ”type material” of Neotherium in that the plantar process is centrally located. However, because it is about twice the size of the navicular described by Kellogg, it thus suggests sexual bimodality and that the type material is from a female individual. Mitchell and Tedford illustrate a left metacarpal IV, LACM 4360 (their fig. 21), which is a miniature of that from Imagotaria downsi, as are some other elements referred to Neotherium mirum. An isolated right radius lacking the distal epi- physis, USNM 187377 (pl. 11), found in the Round Mountain Silt in the bed of Kelley Canyon, 3 miles north of Sharktooth Hill, by M. N. Alling in 1927, was associated with a fifth metatarsal of Allodesmus kernensis. This small radius, measuring 11 cm in length without correction for the missing epiphysis, is distinctly odobenid in that the pronator teres insertion is distal to the middle of the shaft. In addition the radial process at the distal termination of the anterior or radial crest is located conspicu- ously medial to the center of the shaft. Like other skeletal elements that presumably belong to Neo- therium, this radius is a miniature replica of that of Imagotaria, differing only by being relatively less massive and one-half the size of the radius of a male Imagotaria downsi. That Neotherium mirum is a pinniped seems certain, and the features of the few known bones, especially those of the lectotype calcaneum, all ap- pear to be odobenid. The resemblances are greatest when compared with Imagotaria downsi, and it is here considered a dusignathine odobenid. However, the small size, the relatively elongate calcaneum, and the relatively small head on the astragalus all suggest a more primitive form than that represented by Imagotaria. It may be that it is an enaliarctid ancestor to the odobenids. Mitchell and Tedford (1973, p. 279) referred Neotherium to their Enaliarc- tinae. The discovery of the rest of the animal prom- ises to reveal much about the unknown earlier history of the odobenids. CLASSIFICATION OF THE WALRUSES As herein defined, the walruses now known are classified as follows: Family ODOBENIDAE Subfamily ODOBENINAE Genus Aivukus new genus Aivukus cedrosensis new species Late late Miocene, Baja California Genus Prorosmarus Berry and Gregory Prorosmarus alleni Berry and Gregory Early Pliocene, Virginia Genus Alachtherium DuBus Alachtherium cretsii DuBus Early Pliocene, Europe Alachtherium antverpiensis (Rutten) Early and late Pliocene, Europe Genus Odobenus Brisson Odobenus huxleyi (Lankester) Late Pliocene and Pleistocene, Europe and possibly the USA Odobenus koninckii (Van Beneden), nomen dubium Late Pliocene, Europe Odobenus rosmarus (Linnaeus) Pleistocene and living, N. Atlantic and N. Pacific Subfamily DUSIGNATHINAE Genus Neotherium Kellogg Neotherium mirum Kellogg Early middle Miocene, California Genus Imagotaria Mitchell Imagotaria downsi Mitchell Late middle and early late Miocene, California Genus Pontolis True Pontolis magnus (True), nomen dubium Late late Miocene, Oregon Genus Dusignathus Kellogg Dusignathus santacruzensis Kellogg Late late Miocene and Pliocene, California and Baja California Genus Pliopedia Kellogg Pliopedia pacifica Kellogg Late late Miocene, California Genus Valenictus Mitchell Valenictus imperialensis Mitchell Late late Miocene and Pliocene, California SUMMARY OF THE HISTORY OF THE WALRUSES Insofar as they are presently recognizable, the earliest known walruses are of early middle Miocene age in the North Pacific basin and of Pliocene age in the North Atlantic basin. At least by late Miocene time, the family was clearly divided into two groups; one seems to have become extinct by the end of the Pliocene; the other leads to the living walrus. Both of these subfamilies were well along the road of adap- tive specialization in the late Miocene. 56 OTARIOID SEALS OF THE NEOGENE With the exception of the poorly known Neother- ium mirum, which may well be the enaliarctid ancestor of the odobenids, the dusignathine genus Imagotaria is the most generalized genus included in the family. However, it already had a strong tend- ency, in the late middle and early late Miocene, toward single-rooted cheek teeth, had enlarged ear ossicles and other auditory features which today characterize the walrus, and had other peculiarities of the cranial and postcranial skeleton which clearly mark it as an odobenid. There is very little about Imagotaria which pre- cludes the supposition that it is prototypic of the stem odobenid from which both subfamilies evolved. Per- haps the feature of Imagotaria most objectionable to such a supposition is the development of a tympanic membranezoval window area ratio of around 10:1, a far greater divergence from the range of ratios found in other carnivores than that present in the living walrus. From this specialization, it seems possible that so primitive a walrus as Imagotaria survived until approximately 9 my. ago by occupying the realm now occupied by the sea lions and extending its range of feeding to depths not regularly frequent- ed by living walrus. During the temporal range of Imagotaria, the known otariids were small- to medi- um-sized animals; the earliest large-sized otariid is from the late late Miocene, as will be discussed, perhaps 8 my. ago. This otariid is found with the Odobenus-like genus Dusignathus in deposits from northern California to Mexico and, in the southern part of this range, with the primitive odobenine genus Aivukus. The late late Miocene genus Dusignathus, while retaining several primitive features characterizing its subfamily, was clearly advanced in other features which suggest parallel specialization with the con- temporary odobenines. In the Almejas Formation of Cedros Island, Baja California, Dusignathus occurs with the odobenine Aivukus; this fact suggests that this parallel specialization was not enough to cause a great amount of ecologic competition. It is difficult to judge whether the distinctively different cheek-tooth wear shown by these genera was caused by a dif- ference in diet or simply resulted from the unusual geometry of the mandible of Dusignathus. The lack of a greatly arched palate and of lingual tooth wear in Dusignathus suggests that it was not a tongue- piston sucking feeder as was Aivukus and as is Odobenus. Pliopedia pacifica was a contemporary of Dusig— nathus, but present records suggest that it was an inhabitant only of the warmer inland sea then occupying the Central Valley of California. In the Pliocene Valenictus inhabited both the Central Val- ley and Gulf of California inland seas, but there is no clear evidence that Dusignathus survived until so recent a time as the only records from the Purisima Formation in the Santa Cruz area are from lowin the formation, a short distance stratigraphically above a 6.7-m.y. glauconite date. Only otariids more ad- vanced than those from the Almejas Formation have been found in the upper Pliocene, part of the Puris- ima Formation in the Santa Cruz area, as discussed under the consideration of the otariid seals to follow. Aivukus cedrosensis is the most primitive known odobenine. Some postcranial elements are insepa- rable from living Odobenus, whereas others are intermediate between the dusignathines and Odo- benus. Although the skull is intermediate between the dusignathines and Odobenus in the loss of the posterior cheek teeth and reduction of the lower canines, the completely odobenine nature of the cheek teeth indicates a major step toward Odobenus from the condition of Imagotaria. From this, one may surmise that the odobenine odobenids were a distinct lineage at least as long ago as the early late Miocene. Prorosmarus alleni is more advanced toward Odo- benus than is A. cedrosensis by the loss of an additional posterior lower cheek tooth, further reduc- tion of the lower canines, and enlargement of the upper canines. This species may be as much as 3 my. younger than A. cedrosensis. It seems quite plausible that a species of Aivukus comparable to A. cedrosensis, or perhaps somewhat more similar to Imagotaria, was the emigrant from the North Pacific, perhaps as long ago as 9 my, and that the rate of odobenine evolution accelerated in the North Atlantic in the absence of the dusig- nathine odobenids and the otariids. The failure of the dusignathine odobenids and of other otarioids to enter the Atlantic may be related to the southerly distribution of Aivukus cedrosensis in the late Miocene; possibly it was then the only otarioid whose range extended as far south as the Central American Seaway. Paradoxically, it seems to be the northerly distribution of Odobenus which accounts for its reentry into the Pacific during the Pleistocene without permitting the otariids access to the Atlantic. The absence of any otariids in the North Atlantic suggests that this group, to be discussed next, did not expand their southerly distribution until after the effective closure of the Central American Seaway and after the Atlantic invasion by the odobenine odobenids. Most critical to establishing the date of this event, therefore, is the report of late Miocene or PART II: FUR SEALS AND SEA LIONS early Pliocene otariids from Sacaco, Peru (Robert Hoffstetter, written commun., 1973) and it is here presumed, on the basis of this tentative age, that the Central American Seaway was effectively closed to otariid crossing by about 5 my ago. This closure could have been an ecologic barrier as easily as it could have been a land bridge; Savage (1974, p. 26) suggests strong currents through the narrowing straits although very warm waters in the shallowing straits seem a more likely barrier. The presence of the varied late Miocene and Plio— cene phocid seals of the North Atlantic (Van Bene- den, 1877) would have an orienting effect on the North Atlantic odobenines toward benthonic feed- ing. This would explain the progressive morphology of P. alleni and the longer and more varied record of odobenine odobenids in younger deposits of the North Atlantic. In the odobenine record of the North Pacific, A. cedrosensis is succeeded only by late Pleistocene Odobenus, largely from Alaskan lati— tudes, and it is highly probable that these actually derive from the North Atlantic as no morpho- logically intermediate forms are known from the North Pacific region. PART II: FUR SEALS AND SEA LIONS Family OTARIIDAE Genus ARCTOCEPHALUS F. Cuvier Type species.—Phoca ursina Linnaeus [fide F. Cu- vier, 1826] = Arctophalus antarcticus (Thunberg, 1811) [fide Allen, 1905, there being no doubt in his mind that Cuvier’s description and illustration were of a South African specimen] = Arctocephalus pusil— lus (Schreber, 1776), Peters, 1877. Diagnosis.—Fur of the forelimb extends distally past the wrist onto the dorsum of the foreflipper, and facial angle is always greater than 125° (Repenning and others, 1971, p. 3). Premolars single-rooted, molars usually retain double roots; basioccipital parallel-sided or nearly so. Dental formula 31'1C'4P'2M mx 2 - 36- Os penis transversely narrow at apex and in anterior part of shaft (Morejohn, 1975). Included species.—Arctocephalus pusillus (Schre- ber): living along shores of southern Africa and southeastern Australia; late Pleistocene, South Af- rica (Hendey, 1974, p. 189). Arctocephalus townsendi Merriam: living along Pacific shores of Mexico and California, late Pleis- tocene, San Miguel Island, Calif, (Repenning un- publ. data, 1969). 57 Arctocephalus gazelle (Peters): living on oceanic islands in and south of the Antarctic Convergence. No fossil record. Arctocephalus forsteri (Lesson): living on islands of southern New Zealand waters and along coast of South Australia and southern Western Australia. No fossil record although known from aboriginal sites. Arctocephalus tropicalis (Gray): living on islands north of Antarctic Convergence. No fossil record. Arctocephalus australis (Zimmerman): living a- long Atlantic and Pacific shores of South America from Brazil to Peru. No fossil record although Arcto- cephalus fischeri (Gervais and Ameghino, 1880) may belong to this species; the published description (Ameghino, 1889, p. 342—343) is not adequate for identification and the specimen is now lost (D. E. Russell, written commun., 1972). Arctocephalus galapagoensis Heller: living in the Galapagos Islands. No fossil record. Arctocephalus philippii (Peters): living in the Juan Fernandez Islands. No fossil record. Genus CALLORHINUS Gray Type species.—Phoca ursina Linnaeus, based up- on Steller’s Ursus marinus from Bering Island (Gray, 1850, p. 359). Diagnosis—Fur of the forelimb extends only to the wrist, where it terminates in a sharp, straight line,3 facial angle always less than 125° (Repenning and others, 1971, p. 3). Premolars single-rooted, M1 usu- ally double—rooted, M2 single- or double-rooted, and basioccipital parallel-sided or nearly so. Dental for- mula 3I'1C'4P‘2M 21~1C-4P-1M x 2 36‘ Ventral processes of bacular apex very broad rela- tive to dorsal process and base flattened dorsoven- trally (Morejohn, 1975). Included species.—Callorhinus ursinus (Linn- aeus): living along shores of North Pacific from southern California, to central Japan. Late Pleis- tocene, Seward Peninsula, Alaska (Repenning un- publ. data, 1968). As here conceived, Arctocephalus is the genus most like the ancestral otariids and both the sea lions and the genus Callorhinus diverged from this line- age leading to Arctocephalus. By reason of some Callorhinus-like features of late late Miocene and “R. L. DeLong (written commun., 1973) has recently shown us photographs of a pup born (in 1972) on San Miguel Island, Calif, with fur extending down the flipper below the wrist in Arctocephalus fashion. This example is the only known exception to this most conspicuous character of Callorhinus. As Arctocephalus townsendi also frequents San Miguel Island, the possibility of a natural hybrid exists. 58 OTARIOID SEALS OF THE NEOGENE Pliocene otariids from California, described below, and because the earliest recognized sea lions are of late Pliocene age, it is herein suggested that Callor- hinus diverged from the Arctocephalus lineage be- fore the time of divergence of the sea lions. A recent study of the endemic louse fauna of living sea lions, of Arctocephalus, and of Callorhinus, in combination with a study of the bacula of these living forms and of their fossil history (Kim and others, 1975) strongly supports this interpretation of the greater antiquity of the Callorhinus lineage. Genus PITHANOTARIA Kellogg Type species.—Pithanotaria starri Kellogg, 1925a. Diagnosis—A small genus of fur seal with M2 missing, M1 to P4 diastema, double-rooted cheek teeth except P1 and presumably P1, cheek teeth simple with a sharply pointed crown having very weak internal cingula and no secondary cuspules, third (lateral) upper incisor with a distinct postero- internal cuspule rather than being simple and nearly conical as in most modern otariids. Dental formula 31'1C'4P'1M 21'10'4P'1 or 2M Os penis unknown. Distribution—Known only from the late middle and early late Miocene of California, perhaps from 12 to 9 m.y. ago. Included species.—The genus is monotypic. ‘x2=34or36. Pithanotaria starri Kellogg, 1925 Plate 19 Holotype.—SU Museum No. 11, now CAS 13665, the impression of a nearly complete, somewhat crushed young individual from the late middle Mio- cene Sisquoc Formation 1.4 miles south of Lompoc, Calif.; in the same formation but about 5 miles west of the type locality of Imagotaria downsi. Referred material—UCMP 74813, impression of the inferior surface of a skull showing primarily the palate and one zygomatic arch with one mandibular ramus and one vertebra from the Santa Margarita Formation about 6 miles east of Seaside, Calif, locality UCMP V-6627. Collected by Keith Stafford, Piedmont, Calif, in 1964. UCMP 26785, ”Individual III” of Kellogg (1925a, p. 87—93), a cast of right and left hind limbs from the Celite C0. quarry No. 15, 2.5 miles south of Lompoc, Calif. Same formation and age as the type. The present condition of this specimen is such that much of the original detail appears to have been lost through spalling of the diatomite, although enough is still preserved to unquestionably identify it as Kellogg’s specimen. UCMP 26784, ”Individual II” of Kellogg (1925a, p. 84—87), a cast of a left forelimb from the same locality as ”Individual III.” In its present condition, this specimen is nearly completely lost through frac- turing and spalling of the diatomite. The only dis- tinctive feature remaining is the somewhat unu- sually oriented lesser tubercle of the humerus noted in Kellogg’s description. UCMP 108069, a portion of a right mandibular ramus bearing P 3 from UCMP locality V71197, 1,900 feet south of locality M1035 along Glen Canyon Road, Santa Cruz, Calif. Santa Margarita Forma- tion. Collected by D. P. Domning, May, 1973. USNM 184062, left metacarpal I from USGS locality M1035, 12—15 feet south and 8 feet lower in the section than the juvenile male skull of Imago- taria downsi, along Glen Canyon Road, Santa Cruz, Calif, Santa Margarita Formation. Collected by G. V. Morejohn, April 1973. USN M 184056, the proximal three-quarters of an adult humerus from the upper part of the Santa Mar- garita Formation, Santa Cruz area, locality USGS M1036 on the west side of Bean Creek along fence marking north boundary of Canada del Rincon en el Rio land grant in NE% sec. 13, T. 10 S., R. 2 W. Collected by C. A Repenning and J. C. Clark in 1965. Diagnosis.—Only one species of the genus Pithan- otaria is recognized at this time (1976). See diagnosis of the genus. Type locality and age—”No. 9 quarry of the Celite Co. [purchased by Johns Manville Corp. in the late 1920’s] 1.4 miles south of the intersection of Ocean Ave. and ‘C’ Street, Lompoc, San Bernardino [actu- ally Santa Barbara] County, California" (Kellogg, 1925a, p. 74). Although no other mention of a quarry No. 9 that is 1.4 miles south of Lompoc has been found in the literature, the same diatomite unit being mined at the modern Celite Co. quarries of the Johns Manville Corp., which are 1 to 2 miles farther south, is present at this locality. There seems to be no reason to suspect that the mileage was incorrect in the description of the type locality nor that the type was from the main quarries of the modern Celite Co. The type description mentions that the specimen was found 200 feet above the base of the 1,400-ft-thick diatomite unit. Bramlette (1946, p. 212) places the Mohnian-Delmontian boundary somewhere in the lower part of the diatomite unit on the basis of Mohnian foraminifers in the underlying unit. The estimated age is 10—12 m.y., late middle or early late Miocene. John A. Barron has examined the diatoms in the matrix of the type specimen and reports that PART II: FUR SEALS AND SEA LIONS 59 they are equivalent to Schrader’s North Pacific Diatom Zone XI. DISCUSSION Kellogg’s description of the holotype and of his two referred specimens leaves little need for additional comment. The type is an immature individual and possibly also a female, for there is no indication of an os penis although the skeleton is nearly complete and nearly completely articulated; such bones as the sternebrae and a hyoid element are preserved in approximately the normal position. As Kellogg notes, there is very little in the postcranial skeleton of this most ancient of now known otariids to separate it from the living eared seals. From a comparison of the humerus of the type and that of the articulated forelimb UCMP 26784, which show impressions of lateral and medial sides, respec- tively, it appears that the humerus is relatively more slender than that of the sea lions and that the distal termination of the pectoral crest is directed toward the medial lip of the trochlea for articulation with the ulna, as in living fur seals. Relative to the size of the canine, the lateral upper incisor is distinctly smaller than that of living sea lions. To the extent that the skeletal elements of the living fur seals can be separated from those of sea lions, it appears that Pithanotaria starri should be considered a fur seal. Considering the primitive nature of comparably ancient odobenids, it is surprising to see so modern- looking an otariid contemporary with them. All that separates the skull of Pithanotaria starri from the modern otariids are double-rooted cheek teeth and an ursine-like cingular cusp on the medial side of the lateral upper incisor; both features are not men- tioned by Kellogg and are not clearly shown on the type specimen but are evident on the referred skull. Double-rooted cheek teeth, except P1, are indicated on the type specimen, however, and though the medial side of the lateral upper incisor is not shown, the posterior broadening of the crown near its base suggests that this is a weak development of a cingulum and not a more nearly conical crown as found in modern otariids. The posteromedial cingu- lar cusps on the lateral incisors are well shown on referred skull UCMP 74813 (pl. 19) from the Santa Margarita Formation east of Seaside, Calif, a de- posit of an age presumably similar to that in which the type specimen was found.4 ‘No associated fossils are known from the Santa Margarita Formation east of Seaside, but the unit at this location represents only the lower part of the formation recognized elsewhere and intertongues with the underlying Monterey Formation (J. C. Clark, oral commun., 1973). Regarding the skull of the type specimen, Kellogg indicates that the braincase is deeper than that of Callorhinus of comparable age and that the rostrum is relatively short and low. This description may be correct but these features are certainly not recog- nizable on the type specimen which, owing to crush- ing, shows largely the dorsal aspect of the braincase, rather than the lateral, and which shows so much crushing and distortion of the rostrum that no judgment of the height of the rostrum can be made. At the generic level, Pithanotaria starri is distin- guished by small size, apparently five postcanine cheek teeth in each upper tooth row rather than six as in modern fur seals, a small diastema between the upper premolars and the single molar, and by a simple cheek-tooth crown consisting of a single narrowly pointed cusp with very weak internal cin- gulum. Although a small genus, it is not unusually small and is roughly comparable in size with the smallest of living fur seals, Arctocephalus galapago- ensis, assuming that the type specimen is a female. The type clearly shows five upper postcanine teeth with a small diastema between the molars and premolars. The crowns are very simple with no external cingula, and the impression of the lingual surface of the last cheek tooth on the left mandibular ramus of the holotype clearly indicates a very weak internal cingulum; the crowns are identical to those of some living species of Arctocephalus, particularly A. galapagoensis, A. tropicalis, and the more simple variation of the cheek teeth of A. forsteri. It is not possible to maintain that there was no sixth postcanine (second molar) in the upper jaw of the holotype; it may have been present and lost, for at least two of the lower cheek teeth have fallen out of the right mandibular ramus and P1 is preserved in a position half out of its alveolus. Because the type is an impression, the possible presence of an alveolus cannot be checked. However, there is no indication that the second upper molar was present on the type and it is clearly not present on referred specimen UCMP 74813. Because the lack of a second upper molar is very unusual, it is the strongest basis for generic distinction. In their review of the living fur seals Repenning, Peterson, and Hubbs (1971, p. 22) noted only one individual in all the specimens exam- ined in which the second upper molar was missing; this was CAS 1185, a female Galapagos fur seal (pl. 19). The loss of the second upper molar is common in living sea lions. The impression of the left mandibular ramus is present with the palate of UCMP 74813. No indica- tion of the lower teeth is preserved. The ventral margin shows a rather sharp crest in the region of 60 OTARIOID SEALS OF THE NEOGENE the digastricus insertion. This crest is formed by the lateral surface curving ventrally and medially to meet the nearly planar medial surface. Anterior to the digastricus insertion, both surfaces are curved and the ventral margin is rounded. In addition, the inferior pterygoid process beneath the lower sigmoid notch, though damaged, is shown to be notably elongate. Although elongation of this process is not visible on the lateral impression of the mandible of the type, the process can be seen to protrude posteri~ orly beyond the notch; preservation of the type is such that the inferior margin in the region of the digastricus insertion is not well defined. UCMP 108069, from the Santa Margarita Form- ation of the Santa Cruz area, does not show the posterior mandibular structures but does show the sharp ventral margin in the region of the digastricus insertion and the rounding of the margin anteriorly. In addition, it shows a deepening of the jaw toward the symphysis beneath P3 and the development of a similarly sharp ventral margin from this point anterior to the symphysis. The deepening of the ramus below P 3 is evident in the mandible of the type specimen. All cheek teeth of this specimen except P1 were two-rooted, and the crown of P3 is identical to that of the type and has a very weak and rounded lingual cingulum (pl. 19). USNM 184056 is the proximal three-fourths of a right humerus comparable in size to a small living fur seal, though larger than that of the type of Pithanotaria starri; this size suggests that the type specimen may be a female. It is from the Santa Margarita Formation of the Santa Cruz area and is referred to Pi thano taria because of stratigraphic and geographic association. It is inseparable from the humerus of any similarly sized otariid except by a single criterion. As on the type specimen, the crest for insertion of the deltoideus muscle on the lateral side of the pectoral crest is very weakly developed, con- spicuously less prominent than in any other known otariid, fossil or living. An isolated first metacarpal, USNM 184062 (pl. 19), was collected at USGS locality M1035; it has the same proportions as that of the holotype but it is somewhat larger. Its length is 62 mm, that of the metacarpal I of the type specimen about 54 mm. The bone is from a fully mature individual, differing in this respect from the holotype. The articular surface for contact with metacarpal II, if present, is very inconspicuous, suggesting an otariid; the strati- graphic associationin the area with both Imagotaria and Pithanotaria, as well as general similarity to this bone of the type, indicate assignment to P. starri. The bone is remarkably dissimilar to the first metacarpal of living otariids in that it shows no flattening of the shaft. And a prominent fossa is present on the proximal dorsal surface for insertion of the extensor. In fact, its general appearance is that of a miniature metacarpal I of Imagotaria, it differs from it largely by lacking a prominent articulation for the second metacarpal and by having a flatter proximal articulation. In summary, Pithanotaria starri is the oldest known otariid, but it is remarkably like living otari- ids except that the cheek teeth are double rooted and primitive structures characterize at least some of the limb bones. It is generically distinct from the living fur seals in the above features and in having lost M2 (if this is not a variation in the individual specimen showing no M2), and in having a cingular cusp on the anteromedial side of the I3. It is also characterized by a P4 to M1 diastema, present in some living fur seals. The loss of M2 would seem to preclude Pithanotaria starri being ancestral to the living fur seals, and it may be assumed, therefore, that there were other otariids living 12—10 m.y. ago. Nearly all known records of Pithanotaria are stratigraphically asso- ciated with specimens of the primitive odobenid Imagotaria downsi. Genus THALASSOLEON new genus Type species.—Thalassoleon mexicanus new species. Etymology.—Gr. Thalassa, the sea; and Leon, lion, masculine; hence, ”sea lion.” Diagnosis.—A large otariid genus with short ros- trum, exceedingly broad nasals, a facial angle (see Repenning and others, 1971, fig. 1) of about 150° in adult males, M2 lost, M2 present, all cheek teeth (except P1) double-rooted, persistent diastema be- tween M1 and M2 and a lesser and variably present diastema between P4 and M1, cheek tooth crowns form a single lanceolate cusp with moderate and rounded internal cingulum and no accessory cusps, trapezoidal basioccipital that is very broad posteri- orly (for an otariid), coronoid process of the mandible very broad and its posterior margin not undercut, pterygoid process of the mandible is shallow with very little shelflike medial protrusion, vertebral fora- men of all vertebrae small relative to living species, specifically variable primitive structures of limb elements. Dental formula: 31:1C:4P:2M 21:1C:4P:1M Os penis very long without forked apex, dorsal process of apex small, base flattened dorsoventrally. Distribution—Late late Miocene and Pliocene, Pacific coast of Baja California and California. x2=36. PART II: FUR SEALS AND SEA LIONS 61 Included species—Thalassoleon mexicanus n. sp., late late Miocene, Baja California. Thalassoleon macnallyae n. sp., late late Miocene and Pliocene, California. Thalassoleon mexicanus new species Plates 20--23; figure 5 Holotype.—IGCU902, adult male skull with some po'stcranial elements, collected from UCR locality RV—7301, about 15 feet above the base of the upper Miocene Almejas Formation, Cedros Island, Baja California, by R. H. Tedford and D. P. Whistler on August 4, 1965; field No. RHT 1273. Etymology.—The specific name is given in appre- ciation of the cooperation shown by the Government of Mexico and particularly by the Instituto de Geol- ogia, Universidad Nacional Autonoma de Mexico. Diagnosis. —Ectotympanic ossification of the tym- panic bulla lacking conspicuous medial ornament- ation in mature males, fibula not fused to tibia with maturity, metatarsal I is stout and short, distance from lambdoidal crest to midpoint of frontal-parietal suture (in young individuals) more than half the distance to the anterior limits of the braincase. Type locality and age—From about 8 feet to about 104 feet above the base of the Almejas Formation, Cedros Island, Baja California, Mexico. By estima- tion, between 6 and 8 my old. This material was found in stratigraphic association with both the odobenine odobenid Aivukus cedrosensis and the dusignathine odobenid Dusignathus santacruzen- sis. The age of this part of the Almejas Formation is more fully discussed under Aivukus cedrosensis n. sp. and Thalassoleon macnallyae n. sp. Referred material.—Thalassoleon mexicanus is the most abundant pinniped in the UCR collection from the Almejas Formation. This collection in- cludes 2 complete skulls, 5 shattered and/ or signifi— cantly incomplete skulls, 5 skull fragments of sig- nificance to the understanding of individual and sexual variation, 25 complete or incomplete mandi- bular rami, 10 isolated cheek teeth, and about 300 postcranial elements. Of these, the following are con- sidered most informative. UCR 15252, crushed juvenile male skull collected from UCR locality RV-7302, by G. T. Jefferson in 1964 from 20—30 feet above the base of the formation; field No. Cedros 4E (pl. 22). UCR 15253, rostrum and temporal of an adult female skull, collected from UCR locality RV—7303 by R. H. Tedford on August 9, 1965, from about 25 feet above the base of the formation; field No. RHT 1295 (pl. 22). UCR 15251, partial adult male skull collected from field locality Cedros 4A by Jefferson in July 1964, from 20—30 feet above the base of formation. UCR 15258, hind quarters, including the os penis, of an adult male skeleton, collected from UCR locality RV—7307 by D. P. Whistler on August 14, 1965, about 70 feet above the base of the formation; field No. RHT 1321 (pls. 21 and 23). UCR 15254, most of two anterior limbs, verte- brae and ribs, collected from UCR locality RV—7304 by Juan Felix and Tedford, August 15, 1965, about 104 feet above the base of the formation; field No. RHT 1324 (pl. 22). UCR 15255, fragments of skull, mandible, two anterior limbs, and vertebrae collected from UCR locality RV—7305 by Whistler on August 11, 1965, about 70 feet above the base of the formation; field No. RHT 1307. UCR 15256, female humerus collected from UCR locality 7306 by Juan Felix and Tedford on August 11, 1965, between 40 and 60 feet above the base of the formation; field No. RHT 1322 (pl. 22). UCR 15249, astragalus collected about 60—70 feet above the base of the formation by Whistler on August 8, 1965; field No. RHT 1288 (pl. 23). UCR 15250, radius collected about 40 feet above the base of the formation by the UCR party; field No. RHT 1320 (pl. 22). UCR 15257, male humerus, ulna, and distal termination of radius collected from UCR locality 7306 by Juan Felix and Tedford on August 11, 1965, between 40 and 60 feet above the base of the forma- tion; field No. RHT 1312 (pl. 22). DESCRIPTION OI“ Skill. Thalassoleon mexicanus is equal in size to the largest of the living fur seals, A. pusillus from South Africa and southeastern Australia. The skull of the holotype, an old male,5 has a condylobasal length of 272 mm; that of the referred adulttmale skull UCR 15251 is 259 mm in the same dimension. King (1969, p. 844) found a range of condylobasal length in 42 male skulls of A. pusillus (her A. doriferus) from 254 to 304 mm. The largest skull of all other living species of Arctocephalus seen by Repenning, Peterson, and Hubbs (1971, p. 23) had a condylobasal length of 268 mm. Although no complete female skull of T. mexi- canus is known, incomplete female skull UCR 15253 is comparable in size to female skulls of A. pusillus. 5The left upper canine was removed from the skull of the type and was found to have a nearly closed root. In addition, cementum annuli are present on the root, as on modern otariids, indicating that the individual was at least 8 years old (pl. 21). These cementum annuli and the marked sexual dimorphism suggest that intraspecific behavior patterns of Thalassoleon mexicanus were very similar to those of living otariids. Repenning (1976) discusses this and other aspects of the behavioral evolution of the otarioid seals. 62 TABLE 13,—The skull and humerus length and proportions of mature specimens of Thalassoleon mexicanus. Arctocephalus pusillus, and Arctocephalus forsteri Species Specimen CBL Humeriis lengthl B/ A = A (mm) = B (mm) (percent) A. pusillus. . ,,,.....USGS 7008 287 234 81.5 T. mexicanus ,,,IGCU 902 + 272 215 79.3 UCR 15257 2 A. forsteri ................ USGS 7107 246 188 76.5 1. Most distal part of capitulum to most proximal part of greater tubercle. 2. Largest available male specimens but not from same individual. Possibly of even greater significance is the simi- larity in size of the postcranial elements of T. mexi- canus and A. pusillus, for A. pusillus, unlike other living fur seal species, has sea lion-like proportions in that its body is large relative to its head. As an approximation of this similarity, the condylobasal length of the skull is compared (table 13) with the greatest length of the humerus for A. pusillus, A. forsteri, and T. mexicanus. The proportions suggest that T. mexicanus is intermediate in this feature. Judged by the size of postcranial elements, males of T. mexicanus attained a weight of at least 650—700 lbs (295—318 kg). Dorsally, the adult male skull is distinctive in its extreme development of the sagittal crest (fig. 5a), which extends anteriorly to a point about midway between the brain case and the supraorbital process (seen on two male specimens), very widely flaring nasal bones, located far to the rear such that they extend behind the anterior limits of the supraorbital processes (evident on three male skulls and the female rostrum), premaxillae whose ascending or nasal processes form essentially no part of the lateral surface of the rostrum in the region just below the termination of the nasal bones (shown on three adult male skulls, one juvenile male skull, and one adult female rostrum) but rather he entirely within the large piriform aperture (juvenile male skull) or form broad anterior-facing surfaces on the lateral margins of the aperture (four adults), and extremely large preorbital processes (both sexes). On the juven- ile male skull, the distance from the rudimentary lambdoidal crest to the parietal-frontal suture at its midline is over half the distance to the front of the brain case as determined by the squared anterior corners housing the sigmoid gyri. In lateral aspect (fig. 50) the skull is most distinc- tive in the posterior position of the nasals such that a line from the tip of the nasals to the gnathion makes an angle of 45° or less with the line of the alveolar margins of the cheek teeth. Measured by the method employed by Repenning, Peterson, and Hubbs (1971, fig. 1), the facial angle of the three adult male skulls OTARIOID SEALS OF THE NEOGENE is between 150° and 151°. In anterior aspect, the prominent preorbital (”lacrimal”) processes and very broadly oval infraorbital foramina are distinctive. In ventral aspect (fig. 5b) the basioccipital is trape- zoidal in form and unusually broad posteriorly com- pared with Arctocephalus and most living otariids. This bone is essentially rectangular and has parallel lateral margins in living fur seals, but comparably broad and trapezoidal basioccipital bones are pre- sent in some of the living sea lions, notably Neo- phoca and Zalophus. The double-rooted cheek teeth and the Ml-M2 diastema also are distinctive in ventral aspect. The auditory region of T. mexicanus is distinctive- ly otariid. Externally the bulla is possibly more rugose than those of the average living fur seals, and it is moderately inflated but certainly within the range of variation in these features. Within the middle ear, the elongate promontorium, the very small tympanic membrane, the very small epitym- panicum, and the lack of a prominence on the posterior wall (caused by the deep hyoid fossa char- acteristic of the odobenids) are all clearly otariid features and differ in no way from living species. Within the brain cavity (pl. 22), the apex of the petrosum has minimal enlargement, the internal acoustic meatus is essentially circular in cross sec- tion with no separation of the passages for the facial and vestibulocochlear nerves, and the prominent floccular fossa is detached from the bony tentorium, which arches distinctly above the petrosum rather than being closely appressed and fused to it. In all features the auditory region of T. mexicanus is clearly otariid, remarkably so for such an ancient species, and it has no differences which could sep- arate it from that of the living otariids. No auditory ossicles have been found. The mandible closely resembles that of living A. pusillus in that the coronoid processes are very broad and not undercut and overhanging along their pos- terior margins. (See fig. 7 in Repenning and others, 1971.) The pterygoid process, also similar to A. pusillus, is long and shallow and projects only slightly as a medial shelf. Skull dimensions are given in table 14. Although larger, the general proportions of the skull and particularly the short broad rostrum and flaring nasal bones bear a strong resemblance to the living A. forsteri and A. australis. In size and in the morphology of the mandible, T. mexicanus is most FIGURE 5.——Restoration ofthe male skull Thalassoleon mexicanus. Holotype IGCU 902. A, Dorsal view. B, Ventral view. C, Lateral view. PART II: FUR SEALS AND SEA LIONS 63 64 OTARIOID SEALS OF THE NEOGENE TABLE 14.—Dimensions of the skulls of Thalassoleon mexicanus and Thalassoleon macnallyae E: E 4: ‘3 5 5 g c m 8 g . 3 a E: E 'é E w .s . 2% ‘55 E s a s E a“ z 7., 3: i E E 3:; 3E an; as .55 f :5 3:5 :3 '_ h 3 3 .r: a 5 a w c G 8 a g 8 x 3 5‘ 3 8 a“; 3 s s a 8 § 8 as 6? :2 a”: 0 <3 <5 3 3 z D o :> T. mexicanus: IGCU 902 (Type) ...................................... M 36 272 167 147 69 29 44 70 51 14.5 UCR 15251 ..... M 36 259 3156 3144 62 29— 43 64 -- 14.5 HSC 310 ___________ M 334 3267 3170 --- 61 30 44 65 -- 14.6 UCR 15252 ..... M 315 207 104 89 33 332 318 54 40 210.0 UCR 15253 ................................ F 336 38 -- 330 353 - 8.5 T. macnallyae: UCMP 112809 .............................. M 3 36 134 -- - -- ~- 47 ‘ From Sivertsen, 1954. comparable to the living A. pusillus. The cheek tooth crowns most resemble the living A. philippii and A. townsendi. The double-rooted premolars, very large preorbital processes, \very broad infraorbital for- amina, and very broad basioccipital bone are not known in the living fur seals. DISCUSSION OF THE POSTCRANIAL SKELETON The vertebrae of Thalassoleon mexicanus are very similar to those of the living otariids. The principal difference appears to relate to a lesser specialization of the venous system. With the type specimen, IGCU 902, all cervical and most thoracic vertebrae were collected. Although the last cervical and more pos- terior vertebrae are distorted, the other cervical vertebrae exhibit essentially no distortion, and from these, it is apparent that the vertebral foramina are small relative to that of living otariids (pl. 23). The small size is primarily expressed in the narrowness of the ventral part of each foramen. From this relation, it is presumed that the right and left vertebral sinuses of T. mexicanus were smaller than those of living otariids, a condition indicating a more restricted ability to remain underwater (Harrison and Kooyman, 1968, p. 240). Four complete male scapulae from two individuals (UCR 15254 and UCR 15255) are in the collection from Cedros Island. They are basically otariid with two—thirds of the external surface occupied by the supraspinous fossa, ‘but, unlike nearly all living otariids, they have no straight anterior or cranial border with a more or less distinct cranial angle. Instead, the cranial border and the vertebral border form one arc continuous posteriorly to the caudal angle. Further, the entire blade is less elongate, measuring about 1:1 in height from the glenoid 2Crown only about half erupted. aApproximation because of incomplete specimen. cavity to the vertebral border relative to the length from the cranial border to the caudal angle. The difference in the T. mexicanus scapula would seem to indicate the lack or rudimentary development of the distinctive otariid muscle, the episubscapularis (Howell, 1929, p. 70). Two female and six male humeri in reasonably complete condition from Cedros Island are in the UCR collection and a third female humerus is in the HSC collection. In all features except one, these are identical to the humeri of living A. pusillus doriferus from southeastern Australia. None of the specimens is quite as large as the largest of the humeri of the living species of comparable sex, but they are larger than any known humerus of other living species of Arctocephalus. On all specimens, the insertion for the deltoideus is less prominent on the fossil than on the humerus of this large living fur seal but this feature is comparable to that of other living species. The humeri are distinctly like those of living fur seals in that they are more elongate than those of the sea lions, and the pectoral crest is directed distally toward the medial lip of the trochlea. One reasonably complete left ulna from the Alme- jas Formation is in the UCR collection, UCR 15257. The epiphysis of the olecranon is much less inclined posteroventrally than in living fur seal species so that its posterior process for the triceps is notably higher than the trochlear notch. In addition, the cranial or anterior surface of the olecranon, dorsal to the trochlear notch, is very narrow, and the lateral crest of this surface, for insertion of the anconeus muscle and marking the anterior limit of the dorsal part of the origin of the long pollical abductor, is positioned posteriorly such that there is consider- ably more relief to the dorsal margin of the lateral PART II: FUR SEALS AND SEA LIONS 65 TABLE 15.—Hind limb proportions of fossil and living otariids h: I “5 3 5 a E E a a m” 8 1. M be :1 A: an 1.2 .2 '53 3 5 .2 3 5 “‘ 3 11 5 r 5 §°8 Te: 7&5 TE TE T-‘S m 0 e m m ,5 s < a. s a: 8 a o u n a: m 8 3: 2 < 2 Adult males: E. jubata CBL 385 ........................ 355.0 150.5 101.5 192.5 314.0 2.36 3.50 1.84 1.13 P. hookeri CBL 333 ________________________ 264.0 116.5 74.3 126.0 281.0 2.26 3.55 .08 .94 A. pusillus CBL 287 ________________________ 230.0 96.0 71.8 123.5 247.5 2.40 3.21 1.87 .93 Z. californianus CBL 276 ........................ 239.5 95.0 74.0 126.0 257.5 2.51 3.24 1.89 .87 T. mexicanus CBL:270 ________________________ 269.5 107.5 83.5 127.0 239.0 2.52 3.22 2.11 1.12 A. forsteri CBL 244 ________________________ 204.0 85.0 58.5 99.5 212.5 2.40 3.49 2.05 .96 Young males: C. ursinus CBL 184 ________________________ 143.5 59.0 34.2 83.5 165.0 2.52 4.19 1.78 .90 N. cinerea CBL 263 ........................ 217.5 90.0 61.2 109.0 208.5 2.42 3.55 2.00 1.04 llschial depth is measured from ischiatic tuberosity to ventral margin of pubis normal to this margin. half of the trochlear notch than is found in living otariids. These features suggest that extension lever- age for the triceps and possibly the anconeus was not so well developed as in living otariids, an important feature because the otariids swim primarily with their front flippers. Four essentially complete male and two female radii are in the UCR collection from Cedros Island. Their proportions are similar to the radius of living Eumetopias, though smaller. From those radii of the living species of Arctocephalus that have been ex- amined, the radius of T. mexicanus differs by being broader over its distal half because of a much more prominent radial crest. The radial crest itself arises at the pronator origin (pronator teres process of Howell, 1929, p. 32) somewhat more distally than in the living fur seals and most living sea lions, though still proximal to midshaft. As with living species of Arctocephalus, the extensor grooves at the distal extremity of the radius are deeper and more distinct than they are on the radii of living sea lions; this distinctness applies particularly to those extensor grooves other than the most prominent and anterior one for extensor metacarpi pollicis which lies on the lateral side of the radial crest and process. No difference considered significant was observed in the carpals and metacarpals between T. mexi— canus and these bones in the living species of Arctocephalus. It should be noted, however, that the carpals differ in detail from those of most living genera of sea lion; the sea lion carpals most similar to those of Thalassoleon and Arctocephalus are those of the genus Zalophus. In many features, including form of the baculum (Kim and others, 1975), the genus Zalophus, of all sea lion genera, is the most like the fur seals. The pelvis of Thalassoleon mexicanus is propor- tionately very large; it is longer than the tibia(pl. 23), a condition seen only in Eumetopias in the living otariids (table 15). However, the other bones of the posterior limb are roughly proportional in length to those of the living otariids. There appear to be no significant differences be- tween the femora of T. mexicanus and those of the living otariids. Similarly the tibia and fibula show no distinguishing features with one exception. On each of the nine tibiae from Cedros Island in the UCR collection, there is a prominent proximal articu- lar surface for the fibula and a comparable articular surface on the head of the three known fibula. This condition is unknown in the living otariids in which the fibula is firmly fused to the proximal end of the tibia. The patella is typically otariid in its conical form. There are very subtle differences between the form of the tarsals of T. mexicanus and those of the living otariids which appear to be distinctive. On the astragalus, the tibial articulation surface does not flare so widely nor extend as far onto the lateral process; on the plantar surface, the posterior cal- caneal articular surface is much more extensive, extending nearly to the end of the lateral process. These conditions appear constant on six astragali from Cedros Island. The astragalar foramen is well 66 OTARIOID SEALS OF THE NEOGENE developed on all known astragali; this feature differs from those of living otariids, where it is variably developed. On the calcaneum, the posterior articular surface for the astragalus is much longer than in most living genera except Zalophus (which has less curvature of this surface) and Eumetopias. In all living genera (minimal on Callorhinus as noted by Robinette and Stains, 1970, p. 583), the anterior or distal articular surface for the astragalus on the medial process of the calcaneum, or sustentaculum, is smaller than the process itself and the process protrudes farther distally than the articular surface, forming what Robinette and Stains (1970, fig. 1) call the secondary shelf of the sustentaculum; this sec- ondary shelf is essentially lacking in the 11 calcanea from Cedros Island. Four cuboids, one navicular, one entocuneiform, and one ectocuneiform of T. mexicanus from Cedros Island seem to show no distinctive characters which would separate them from the living otariids. All metatarsals are represented in the collection; they are uniformly short and stout. With one exception, the metatarsals of the living otariids are at least one- third longer, proportionate to basal and shaft diam- eters, than those of T. mexicanus (pl. 23). The single exception is the sea lion genus NeOphoca that has remarkably short metatarsals in comparison with other living otariids. A single os penis, from specimen UCR 15258, was collected on Cedros Island. This specimen is less like this bone of any known otariid than any other skeletal element. Although measuring 162.2 mm in length, nearly twice as long as that of an adult Arctocephalus pusillus, it has the appearance of a juvenile in that the ventral knob of the apex projects forward as a continuation of the shaft and the dorsal knob is expressed only as a small spur (see Morej ohn, 1975). The base is equally distinct; it is dorsally depressed and flattened and ends caudally in two lateral knobs. Among living otariids such dorsal- ventral flattening of the bacular base is known only in Callorhinus. The shaft, though nearly oval in cross section, does have a slight ventral flattening; no urethral groove is present. The entire shaft is quite narrow except near the base. SUMMARY AND DISCUSSION Thalassoleon mexica'nus is very well represented by fossil remains from the late late Miocene of Cedros Island. The species is a very large, but otherwise primitive otariid. Compared with living fur seals, the genus Thalas- soleon is characterized by the combination of very large size, short rostrum, broad and posteriorly placed nasals, double roots on all cheek teeth except the first premolars, M1-M2 diastema, trapezoidal basioccipital that is broad posteriorly, very large preorbital (”lacrimal") processes, broadly oval in- fraorbital foramina, broad coronoid processes on the mandible, small size of the vertebral foramina, lack of a distinctly straight cranial border on the scapula, weaker deltoideus insertion on the humerus, weaker triceps and other extensor insertions on the ole- cranon of the ulna, relatively broad distal half of the radius, long innominate, minor differences in the articulation of the tarsus, notable shortness of the metatarsus, relatively large postcranial skeletal ele- ments, and a unique os penis. Some of these features, most presumed to be primitive, are retained in living otariid genera, most notably the genus Eumetopias, although the short- ness of the metatarsus is found only in Neophoca and the posteriorly broad basioccipital is found in Neophoca and Zalophus. The species Thalassoleon mexicanus is distin- guished from the other known species, to be dis- cussed next, on the basis of the combination of a lack of fusion of the fibula to the tibia, notably lesser ornamentation of the medial edge of the ectotym- panic, the anterior position of the parietal-frontal suture, somewhat less metatarsal elongation, and possibly smaller vertebral foramina. In their review of the living species of Arcto- cephalus, Repenning, Peterson, and Hubbs (1971, p. 32) suggested that most insular species may have evolved from a mainland form similar to A. australis and appear to have one feature in common—cheek tooth simplification. The simple cheek tooth pattern of the late late Miocene T. mexicanus and of the late middle Miocene Pithanotaria starri suggest, to the contrary, that cheek tooth complication, best seen in the living A. australis and A. pusillus in the form of well-developed anterior and posterior cingular cusps, . is the advanced condition. Those living species associated with continental shores, South America, Africa, and Australia, now appear to be the most advanced of the living fur seals, and insular species, with simple cheek tooth crowns, appear to retain a primitive condition. This conclusion is in agreement with several sea lion-like features of A. pusillus, referred to in the review of Repenning and coworkers 1971, (p. 9 and 10). Thalassoleon macnallyae new species Plates 23 and 24 Holotype.—UCMP 112809, fragments of an adult male skeleton including basicranium, maxillary fragment, vertebral and costal fragments, and ante- PART II: FUR SEALS AND SEA LIONS 67 rior and posterior limb elements collected from UCMP locality V66128 by Kathleen McNally, then of San Francisco State College, in the fall of 1965. Etymology.—The species is named in honor of Kathleen McNally Martin of Fremont, Calif., who collected the type specimen. Diagnosis.—Ectotympanic ossification of the tym- panic bulla with conspicuous medial ornamentation in the form of an elevated ornate ridge as generally found on adult male Callorhinus ursinus, fibula fused to the tibia proximally, metatarsal I slender and somewhat elongate, possibly larger vertebral foramina than present in T. mexicanus, and parietal- frontal suture posteriorly located. Type locality and age—UCMP locality V66128; from the basal glauconite bed of the Drakes Bay Formation of Galloway (1977), 1,350 feet east of the end of the road to Drakes Beach, Point Reyes Nation- al Seashore, Calif., below high-tide level. This glau- conite has been dated at 9.3105 m.y., as discussed under Aivukus cedrosensis n. sp., and contains an odobenid questionably referred to Dusignathus san- tacruzensis. As discussed below, it seems improbable that this locality could be as old as 9.3 m.y. Questionably referred material.—LACM 4343 is the proximal end of a right ulna from the Purisima Formation near Point Santa Cruz, Santa Cruz, Calif. (Mitchell, 1962, p. 18—20). This specimen is from the lower part of the formation as exposed in the area and is no more than 20 feet stratigraphically above the basal glauconite bed, which has been dated at 6.7 m.y. (J. D. Obradovich, written commun., 1964). Santa Cruz City Museum specimen No. 9975.1 is the partial cranium and frontal region of an immature fur seal from the Purisima Formation near Soquel Point, City of Santa Cruz. It was collected by Gerald Macy. The deposits exposed in the Soquel Point area are stratigraphically considerably higher in the section of rocks included in the Purisima Formation than those at Point Santa Cruz. USNM 184076 (pl. 24) is the posterior half of the skull of a juvenile from the Purisima Formation along Capitola State Beach. It was collected in 1974 by L. J. Macdonald and Steve Turner, then of Foothill College, Santa Clara County, Calif. The deposits exposed at Capitola State Beach are strati- graphically the highest known rocks included in the Purisma Formation. This locality (USGS M1241) is 1.9 miles northeast of Point Soquel. The molluscan fauna from this part of the formation is considered by W. O. Addicott (1969, fig. 2, locality 4) to be of late Pliocene age and to be correlative to the lower part of the Merced Formation of the San Francisco Penin- sula. The lower part of the Merced Formation has been recognized by marine molluscan fossils be- neath the continental Santa Clara Formation in the vicinity of Stanford University, Santa Clara County (Addicott, 1969), and 11 miles to the southeast of there a Blancan mammalian locality (C. A. Repenn- ing, unpubl. data, 1973, USGS vertebrate locality M1219) occurs approximately in the middle of the Santa Clara Formation. By these extended correla- tions, it is judged that the upper part of the Purisima Formation, as exposed at Capitola State Beach, is of latest Hemphillian or Blancan age, and latest early Pliocene or late Pliocene. DISCUSSION The fragmentary specimen UCMP 112809 from the Drakes Bay Formation of Galloway is that of a male individual as large as male specimens of T. mexicanus; it includes the basicranial region of the skull, showing a posteriorly broad, trapezoidal basi- occipital bone, a maxilla fragment showing the alveoli of a double—rooted P4, M1, and M2, and an M1 to M2 diastema, and a mandibular fragment bearing a double-rooted P4 with a simple crown having a moderately strong internal cingulum and the pos- terior root of ,P 3. The P4 preserved in the mandibular fragment measures 8.40 mm in antero-posterior di- ameter at the crown base and is slightly smaller than the P 4 of six male individuals of T. mexicanus which average 9.26 mm in this measurement with a range of 8.45-9.82 mm. The mandible, however, is as massive as any from T. mexicanus. In these pre- served parts, T. macnallyae differs from T. mexi- canus only by possibly smaller cheek teeth and the greater ornamentation of the medial lip of the ecto- tympanic bone. These particular features are quite similar to the living genus Callorhinus, and it seems possible that a larger sample from Point Reyes might provide evidence for the origin of the genus Callo- rhinus out of Thalassoleon macnallyae. The ques- tionably referred specimens from the Purisima Formation at Santa Cruz also strongly resemble Callorhinus. In addition to the greater ornamentation of the ectotympanic, the specimen from Point Reyes differs from those from Cedros Island in that the fibula is fused proximally to the head of the tibia in both hind limbs of the specimen, as in all living otariids. All tibiae from Cedros Island show an articular surface for the fibula on their heads and none show a fusion with the fibula. The holotype includes two fragmental vertebrae, one of the more anterior thoracics, and a lumbar vertebra. Although their incomplete preservation precludes confident assignment to a position in the 68 OTARIOID SEALS OF THE NEOGENE vertebral series and the size of the foramen varies greatly with position in otariids, these specimens appear to have larger vertebral foramina than T. mexicanus. A final difference is seen in the single preserved first metatarsal of the holotype, which is identical to male first metatarsals from Cedros Island in dimensions of the base but is somewhat more slender distally and 6.5 mm (8 percent) longer (pl. 23). The larger sizes of the vertebral foramina and first metatarsal suggest that this species is somewhat more advanced than T. mexicanus from Cedros Island. Questionably referred specimen LACM 4343, the proximal end of a right ulna from the Purisima Formation near Point Santa Cruz, was described by Mitchell (1962, p. 18-20). He noted that this ulna differs from those of living otariids by the narrow anterior margin of the olecranon, and, further, the epiphysis of the olecranon is much less inclined posteroventrally than in living fur seal species and many sea lion species, the conspicuous exception again being Eumetopias. In these features, the ulna from the Purisima Formation agrees with that of Thalassoleon mexicanus. Questionably referred specimen Santa Cruz City Museum No. 9975.1, the partial cranium and frontal region of an immature individual from the Purisima Formation near Soquel Point, is distinctive in the posterior position of the parietal-frontal suture at the midline of the skull roof. The suture is less than half the distance that the front of the braincase is from the lambdoidal crest. This position greatly resembles the position of the suture in immature Callorhinus ursinus and differs from that of T. mexicanus and living Arctocephalus. The presence of this Callor— hinus-like feature in Thalassoleon sp. from the Puris- ima Formation of Santa Cruz suggests a closer affinity to T. macnallyae from Point Reyes than to T. mexicanus from Cedros Island. In the material from this locality at the Santa Cruz City Museum is a very immature left metatarsal IV noteworthy for the relative shortness of its shaft in comparison with this element of living fur seals. USNM 184076, the posterior half of the skull of a juvenile individual from Capitola State Beach, is moderately crushed dorsoventrally. The bone was removed from the right side of the skull, revealing a well-preserved endocranial cast (pl. 24) which shows good detail except in its ventrolateral area, where the vertical crushing has confused or destroyed the detail of the major neopallial convolutions. The parietal-frontal suture crosses the midline of the skull roof at a point which is less than half the distance that the front of the braincase is from the lambdoidal crest and the stylomastoid foramen opens widely anteroventrally as in young Callor- hinus, rather than being partially obscured in ven- tral view by an underhanging lip of the bulla. Brain casts of Zalophus and Callorhinus, prepared by W. I Welker, and endocranial casts of Zalophus, Eumetopias, Callorhinus, and Arctocephalus for- steri were available for comparison. In addition, skulls of juvenile individuals of four other species of Arctocephalus were available for checking the vari- ability of some features. From the comparison pos- sible with these specimens, it appears that the cere- brum of the fur seals is characterized by many fewer, convolutions of the neopallium than that of the sea lions, a situation in agreement with the inferred primitive status of the fur seals. However, the basic pattern of sulci and gyri is the same in all otariids and is very similar to that of the bears except that the olfactory bulbs are greatly reduced and the anterior region of the cerebral hemispheres has been pushed backward, forcing the sigmoid gyri outward and downward and the coronal gyri, pseudosylvian sulci, and the posterior ectosylvian gyri into a near- vertical position. Among the fur seals, Callorhinus appears distinct by having an unconvoluted and notably narrow gyrus ectosylvius posterior relative to the width of the adjacent gyrus supersylvius posterior. Dorsally, a prominent sulcus, the postcruciate, connects the longitudinal fissure to the sulcus lateralis, separ- ating the postcruciatus and lateralis gyri, which appears to be unique to the genus. In these features of Callorhinus, the juvenile endocranial cast from Cap- itola State Beach is very similar. Although the pos- terior ectosylvian gyrus is broader than in three specimens of Callorhinus, it is relatively narrow for an otariid and distinctly less convoluted than this gyrus in available specimens of Arctocephalus. The dorsal aspect (pl. 24) shows a very Callorhinus-like postcruciate sulcus running transversely from the longitudinal fissure to the lateral sulcus. Greater ornamentation of the ectotympanic, fu- sion of the fibula to the head of the tibia, elongation of the metatarsals, and enlargement of the venous sinuses all increase the resemblance of T halassol- eon macnallyae to living otariids and are presumed to indicate that the species was more advanced, at least in these respects, than T. mexicanus. The glauconite in which the type specimen of T. mac- nallyae was found has been dated at 9.3 m.y. and is associated with foraminifers believed indicative of the Delmontian Stage (Galloway, 1977). The only mollusk from this unit was identified by F. S. Mac- Neil as Neptunea colmaensis (Martin), known else- PART II: FUR SEALS AND SEA LIONS 69 where only from the lower (Pliocene) part of the Merced Formation. Material questionably referred to T. macnallyae from the Purisima Formation at Santa Cruz is younger than a glauconite bed dated at 6.7 m.y. One referred specimen of T. macnallyae is from the youngest part of the Purisima in beds correlative to the lower Merced Formation. No direct dating is available for the lower part of the Almejas Formation of Cedros Island containing T. mexi- canus, but its age is here judged to be between 6 and 8 m.y. As mentioned in the discussion of the age of Aivukus cedrosensis, the dated glauconite from the Drakes Bay Formation of Galloway (1977) contain— ed detrital biotite and may be younger than the date indicates. Also mentioned were the cetacean faunas of the Drakes Bay Formation, the Purisima Forma- tion, and the Almejas Formation, which have a high degree of similarity at the generic level according to Lawrence G. Barnes. With the differences already noted, the pinniped faunas are equally similar. The only mollusk from the Drakes Bay Formation ap- pears to be Pliocene, probably late Pliocene. Finally, the glauconite date of 9.3 m.y. is extremely close to the youngest age inferred for the pinniped fauna characterized by Imagotaria downsi and Pithano- taria starri, seals considerably more primitive than those from the Almejas, Purisima, and Drakes Bay Formations. It is therefore concluded that the Drakes Bay Formation of Galloway must be younger than the glauconite date indicates and that it probably is about the age of the Purisima Formation of the Santa Cruz area, between 4 and 6.7 m.y. old. The lower part of the Almejas Formation of Cedros Island may or may not be older than the Drakes Bay Formation of Galloway, depending upon which of two interpretations of the species of Thalassoleon is correct. Either the two species lived at the same time along more than 1,000 miles of the Pacific Coast, the present range of the California sea lion, and the northern species was more advanced in several respects, or the modernization of Thalassoleon mac- nallyae actually represents a later historic event and T. mexicanus and the lower part of the Almejas Formation are older than the Drakes Bay Formation of Galloway and the Purisima Formation near Santa Cruz. OTHER FUR SEALS A few other fragmentary otariid fossils are known from late late Miocene and Pliocene deposits that are generically indefinable but appear to record a signifi- cant stage in otariid evolution: the first development of single-rooted cheek teeth. This development takes place first in the most anterior cheek teeth. Although the holotype of Thal- assoleon macnallyae is relatively complete in com- parison with many types of named fossil seals, un- fortunately the more anterior cheek teeth are not known. Whether the following specimens belong to this species, to other unnamed species of the genus, or should be recognized as distinct genera cannot be judged at least until the anterior cheek teeth of T. macnallyae are known. Specimens from the San Diego Formation.— LACM 4323 (formerly UCLA 2282), a mandibular ramus described by Burleson (1948); LACM 16062, proximal epiphysis of a humerus; LACM 16063, an astragalus, and LACM 16064, a single-rooted cheek tooth, possibly P2 from a somewhat larger individ- ual. Burleson noted that the mandibular fragment matches closely a female specimen of Callorhinus ursinus in size, slenderness, position of the P 4 and M1 on the dorsal crest of the dentary rather than on the medial side of the crest as in Arctocephalus townsendi and in tooth crown form. She suggested that the mandible is that of a fur seal intermediate between Pithanotaria starri and living Callorhinus, and she named it Pithanotaria sp. She did not men- tion what now appears to be the most significant feature of the specimen—that P 3, P 4, and M1 are double rooted. P2 is single rooted with a lateral sul- cus on the root and is set in the ramus slightly out of alinement with the tooth row, being directed antero- laterally. The crown of the isolated tooth, LACM 16064, is better preserved than the three present in the man- dibular fragment, and in form it is identical to C. ursinus in that it is a simple cusp with a fairly prominent internal cingulum. The crown of the tooth is larger than those of the mandible and those of male C. ursinus, which do not appear to have cheek teeth noticeably larger than females. The presence of the tooth from the same (or nearly the same, see below) locality suggests that the mandible is that of a female. The mandible differs from that of Pithanotaria starri by having a more rounded inferior margin in the region of the digastricus insertion and beneath P2, in addition to having a single root on P2. The. singlerooted second premolar and small size of the teeth clearly separate this fur seal from Thalasso- leon mexicanus. The P 20f T. macnallyae is unknown. LACM 16062, the humerus head, provides no useful information. LACM 16063, the astragalus, is identical to those of female ’ T. mexicanus from Cedros Island and, except in-size, to that of the type of T. macnallyae; it differs from, the astragali of 70 living otariids by the lesser degree of lateral-distal flaring of the tibial articular surface onto the lateral process and, on the plantar side, by the much greater extent of the posterior articular surface for the calcaneum, which extends well out toward the end of the lateral process. Though distinguishable from living otariids, this astragalus could not be separ- ated from those of female T. mexicanus. The locality of the mandible is LACM Ice. 1072 and that of the other remains is questionably LACM 1073 along Reynard Way between Eagle Street and Red- wood Street, San Diego, Calif. This locality is known primarily for its fossil birds (Howard, 1949, p. 180) and the question about the locality seems to be whether the material was collected on the east side of Reynard Way (LACM Ice. 1072) or on the west side (LACM loc. 1073). The LACM records indicate that the specimens of questionable locality were entered into the museum on May 25, 1947; that is the same date when George P. Kanakoff, of the LACM staff, collected the localities for the bird remains described by Howard, but catalog records in the museum do not include the collector’s name, only that they were from the “bird locality.” UCLA 2282 (now LACM 4323) was collected on the east side of Reynard Way by C. W. Kennell (Burleson, 1948, p. 247). This locality is a short distance from a locality where an undescribed fossil tapir was found in strata of the San Diego Formation, which E. C. Allison believed to be “probably below rocks contain- ing Patinopecten healeyi and other characteristic San Diego formation marine fossils***” which may “correspond to the Plesippus-bearing rocks between similar marine Pliocene faunas on Soledad Moun- tain.” (in Leffler, 1964, p. 58—59). The age almost cer- tainly would be Blancan, late Pliocene, comparable in age to, or only slightly younger than, the specimen from Capitola State Beach in the youngest part of the Purisima Formation. Significantly, Howard has found the avifauna associated with the pinnipeds of the Almej as Formation of Cedros Island to be similar in composition to that associated with the fossil otariid of the San Diego Formation “but in no instance are the species the same” (Howard, 1971, p. 14—15). The avian fauna thus is evidence permissive of the conclusion that the pinniped bearing strata of the Almejas Formation are older than the San Diego Formation and the younger part of the Purisima Formation. Specimen from the Etchegoin F ormation.—USNM 184065 is the anterior part of a mandibular ramus collected by J. R. Macdonald and Kent Krispin at USGS locality M1220, about 50 feet below the Pseu- docardium bed of the Etchegoin Formation exposed OTARIOID SEALS OF THE NEOGENE in North Dome of the Kettleman Hills, Kings County, Calif. This locality is in the same approx- imate part of the formation as the locality in which Pliopedia pacifica was found, possibly 5—6 my old. This specimen is larger and more robust than that from the San Diego Formation. While not so robusta jaw as that of a male Callorhinus ursinus, the length of the tooth row alveoli from the anterior surface of the canine to the midpoint of P 4 is well within the range of living male Callorhinus. P4 was distinctly two-rooted, P 3 was two-rooted for at least two-thirds of the root length, although the separating alveolar septum was broken out during preparation, and P 2 was single-rooted and was oriented out of alinement with the tooth row, being directed anterolateral. The only tooth present is the canine; it has a distinct tre- foil cross section of the root near the base of the crown produced by prominent posterior and lateral sulci; such a lower canine is very typical of male C. ursinus. The mandibular fragment, however, is much less massive than that of the holotype of T. macnallyae (obviously a male) and somewhat more massive than the known mandibles of female speci- mens of T. mexicanus. This mandibular fragment was found about 30 feet above the Siphonalia zone of Woodring, Stewart, and Richards (1940) from which these workers report (p. 98) a Hemphillian-age horse tooth and about 1,290 feet down section from the Blancan mammalian fauna of their Pecten zone of the San Joaquin Form- ation. This fossil was roughly contemporaneous with the older late late Miocene part of the Purisima Formation in the Santa Cruz area. However, if the specimen is a fragment of a male individual, it is of an individual smaller than any fur seal yet dis- covered in the Purisima. FOSSIL SEA LIONS Earlier it was stated that certain features of Pith- anotaria starri, from the late middle Miocene, sug- gested that this species and genus was not directly ancestral to the living fur seals. As small size is herein considered a primitive condition of the Otari- idae, the large size of Thalassoleon mexicanus sug- gests that this late late Miocene genus might already have evolved in the direction toward the sea lions to a point beyond that from which most living fur seals could be derived. Both of these suggestions imply the possible existence of other middle and late Miocene and Pliocene otariids as yet unknown. This possi- bility, in combination with the scarcity of fossil sea lions, hinders interpretation of the origin of the sea lions. The development of modern sea lions out of the fur seals would appear to have taken place in the late PART II: FUR SEALS AND SEA LIONS 71 Pliocene or more recently, an interpretation sup- ported by the endemic louse fauna on the living otari- ids and by the otariid bacular forms, as well as the present fossil record (Kim and others, 1975). In the section on “Suprageneric diagnoses” it was stated that loss of abundant underfur, increase in body size, increase in rate of development of single- rooted cheek teeth, and beginning of the reduction in the number of upper molars were the principal criteria whereby the divergence of the sea lions from the fur seal lineage might be recognized. The first, loss of abundant underfur, is not recognizable in the fossil record; the rest are qualitative, particularly because Thalassoleon mexicanus is a large fur seal. Along the Pacific coast of North America, the oldest known otariid with any single-rooted cheek tooth other than the first premolars is a small form resembling Callorhinus of late late Miocene age, and the oldest known otariids of large size and morpho- logically comparable to the living sea lion genera are from beds containing land mammals of Irving- tonian (early Pleistocene) age. From late Pleistocene deposits, a number of fossil otariids have been recovered along the west coast of North America, in Japan, and in the western South Pacific which rather clearly belong to extant species. On the whole, Pleistocene records are few and very incomplete. Pliocene record.—“Allodesmus(?)” sp. of Kaseno (1951) = “Eumetopias sp. cf. E. jubata” of Shikama (1953, p. 11) = “Eumetopias sp.?” of Mitchell (1968, p. 1882); this specimen is a very large and massive mandibular ramus with unusual teeth from the Omma Formation of Ishikawa Prefecture, Japan. Ikebe, Takayanagi, Chiji, and Chinzei (1972, fig. 2) consider this formation as being roughly between 2.5 and 3.5 my old (late Pliocene as here used). Nirei (1969) has discussed the age of the Omma in more detail and believes that the Pliocene-Pleisto- cene temporal boundary is to be found within it, prob- ably within its upper half (personal-communication to Kiyotaka Chinzei). As Kaseno’s specimen came from near the middle of the Omma Formation, it would appear to be very near that boundary or about 2 my: old, and probably somewhat younger than the fur seal from the San Diego Formation. Because of its large size and advanced root fusion, this specimen is a good candidate for a very early sea lion. The cheek teeth all have fused roots except for M1, which retains a two-rooted condition over half the root length. This clearly represents a greater rate of development of single-rooted cheek teeth than is exhibited by the San Diego fossil if their geologic ages are comparable. The cusp of the cheek tooth crown, though pointed, is notably low, and the base of the crown is somewhat larger than in living Eumetopias. As with all otariids younger than Pith- anotaria, but not all odobenids or desmatophocids, only one molar is present in the lower dentition. Kaseno’s Eumetopias sp. from the Omma Forma- tion is the only sea lion described in publication that may be as old as latest Pliocene. In the eastern Pacific a single femur, now in the University of California Museum of Paleontology, from deposits at Capitola, California (D. Domning, oral commun., 1975), may represent a similarly ancient large sea lion. The femur is very large, is between 4 and 3 mil- lion years old, and was found in association with the ancestors of the Alaskan fur seal. Early Pleistocene record.—Leffler (1964) has de- scribed a large isolated otariid tooth, evidently a lower cheek tooth judged by the length and straight- ness of the root, from the Elk River Formation of Cape Blanco, Oreg. He compared this tooth with those of living Zalophus and Eumetopias, noting greater similarity in structure to Zalophus and in size to Eumetopias. Packard (1974a) has described an otariid radius from the uppermost unit of the underlying Port Orford Formation of Baldwin (1945)6 of the same approximate area which he felt was most similar to Eumetopias. An immature otariid radius (USNM 184057), lacking both epiphyses, was recently found by R. J. Janda in the same area in the Clinocar- dium-bearing pebble and cobble gravel of the Port Orford Formation; this radius also is most similar to Eumetopias in size, distal broadness, and relatively distal position of the pronator teres process. In addi- tion, a nearly complete skull (USNM 187108) and left mandibular ramus (USNM 187109) are known from the Port Orford Formation. The specimens appear to represent an extinct species of the genus Eumetopias (C. E. Ray, written commun., 1975). The Clinocardium beds of the Port Orford Forma- tion have been considered to be of about the same age as deposits in Humboldt County, Calif, known as the “Moonstone Beach” locality (Allison and others, 1962); here a number of postcranial bones similar to those of living Eumetopias jubata have been found. The specimens, which are in the Geology Depart- ment, Humboldt State College, catalogued under locality 205, consist of a proximal phalanx of the first digit of the manus, a scapholunar, and a metacarpal III. This material, though long considered middle to late Pliocene in age, was found associated with a scaphoid from a large species of Mammuthus, also in 6Formational nomenclature for this section is currently in a state of flux. Most investigators agree that there is little difference in age between what is here called the Elk River Formation and the underlying Port Orford Formation. 72 OTARIOID SEALS OF THE NEOGENE the collection in Humboldt State College. This sca- phoid indicates that the deposit is at least as young as early Pleistocene. A picture comparable to that of the Pacific coast of North America is found in Japan. A few early Pleistocene specimens are known which have been assigned to extinct species of living genera including Eumetopias watasei Matsumoto (1925), part of a rostrum, and Zalophus kimitsensis Matsumoto (1939), fragment of a mandible. These two were considered to be of ’Calabrian (early Pleistocene) age by Matsumoto, coeval with his Paralephas proto- mammonteus zones. Teilhard de Chardin and Leroy (1942, p. 55) consider this species of mammoth to be of post-Villafranchian Pleistocene age, probably early Pleistocene as used herein. Shikama and Ta- kayasu (1971) indicate that Z. kimitsensis is of this (“Giinz-Mindelian”) age and state that it cannot be separated from the living species. Mitchell (1968, p. 1881—1883) questioned the specific distinction of Z. kimitsensis from living Zalophus californianus and was unable to recognize E. watasei as any living otariid species; accordingly, he placed the specimen assigned to E. watasei in a new genus as Orien- sarctos watasei (Matsumoto). Admittedly the spe- cimen differs from known specimens of the living species, but it is hardly complete enough to dismiss the possibility that it is an extinct species of the genus Eumetopias. Eumetopias(?) kishidae Shikama (1953) = Zalo- phus californianus of Mitchell (1968, p. 1882); it is a very Zalophus—like rostrum of uncertain provenance, but reportedly is from Pleistocene deposits of Tokyo. The specimen has a single-rooted P2; the more pos- terior teeth are lost. More recently Shikama and Takayasu ( 1971) have described a fragmentary man— dible from the Shibikawa Formation of the Oga Peninsula under the name of the living species, Zalophus lobatus (= Z. californianus japonicus). They state that these beds are about 2.2 my old, which, in the usage of the present report, is close to the temporal limit between late Pliocene and early Pleistocene; all teeth appear to be single-rooted. It appears that Zalophus kimitsensis of Matsu- moto from the early Pleistocene and Zalophus loba- tus of Shikama and Takayasu from the late Pliocene or early Pleistocene of Japan are indistinguishable from the living (or recently extinct) Zalophus cali- fornianus japonicus. This suggests that the evolu- tion of at least one living species took place earlier in Japan than along the eastern North Pacific shores. Both specimens, however, are mandibular fragments and may not be specifically identifiable. Further evidence from the early Pleistocene of both western United States and Japan would be most useful. Late Pleistocene record.-—A number of published and unpublished records of late Pleistocene sea lions are known from North America all of which appear to represent the living species Eumetopias jubata and Zalophus californianus. In the South Pacific the Ohope Skull from the late Pleistocene (Castlecliff- ian) deposits near Whakatane, New Zealand, was described in manuscript as an extinct species of Neophoca by J. A. Berry (Fleming, 1968, p. 1185). However, most late Pleistocene records around the margins of both North and South Pacific oceans appear to represent living species of sea lions. Arctocephalus caninus Berry (1928) from North Is- land, New Zealand, reported to be of Pliocene age (Fleming, 1968), is now recognized to be Phocarctos hookeri (Berry and King, 1970) and to be less than 1,000 years old (Weston and others, 1973). Arcto- cephalus williamsi McCoy (1877) from the mouth of Melbourne harbor, originally described as of Plio- cene age, has long been known to be Neophoca cinerea (Allen, 1880, p. 770; King, 1964, p. 129) and seems to be of late Pleistocene age (Gill, 1968) CLASSIFICATION OF FUR SEALS AND SEA LIONS As herein defined, the otariid seals now known are classified as follows: Family OTARIIDAE Genus Pithanotaria Kellogg Pithanotaria starri Kellogg Late middle and early late Miocene, California Genus Thalassoleon new genus Thalassoleon mexicanus new species Late late Miocene, Baja California Thalassoleon macnallyae new species Late late Miocene and Pliocene, California Genus Arctocephalus F. Cuvier Arctocephalus pusillus (Schreber) Late Pleistocene, South Africa; historic, South Africa and southeastern Australia Arctocephalus gazelle (Peters) Historic, subantarctic islands of Atlantic and Indian Oceans Arctocephalus forsteri (Lesson) Historic and pre-historic, southern New Zea- land and Australia Arctocephalus tropicalis (Gray) Historic, islands of South Atlantic and southern Indian Oceans Arctocephalus australis (Zimmermann) Historic, Atlantic and Pacific shores of South America, roughly south of lat 15° S. Arctocephalus galapagoensis Heller Historic, Galapagos Islands PART II: FUR SEALS AND SEA LIONS 73 Arctocephalus philippii (Peters) Historic, Juan Fernandez Islands Arctocephalus townsendi Merriam (?)Early Pleistocene, California (J. Firby, oral commun., 1972); late Pleistocene, California; historic, Pacific coast of North America from about lat 20° N. northward to Point Conception, California Genus Callorhinus Gray cf. Callorhinus ursinus but retaining two roots on most teeth Late late Miocene and Pliocene, California Callorhinus ursinus (Linnaeus) Late Pleistocene, Alaska; historic, circum- North Pacific north of about lat 33° N. Genus Phocarctos Peters Phocarctos hookeri (Gray) Historic, New Zealand region Genus Neophoca Gray Neophoca sp. Late Pleistocene, New Zealand (Fleming, 1968, p. 1185) Neophoca cinerea (Peron) Late Pleistocene and historic, Australia (in- cluding Arctocephalus williamsi McCoy) Genus Otaria Peron Otaria byronia (Blainville) Late Pleistocene, Argentina (Ameghino, 1889, p. 343); historic, Atlantic and Pacific shores of South America south of about lat 10° S. Genus Zalophus Gill Zalophus kimitsensis Matsumoto, nomen dubium Early Pleistocene, Japan Zalophus californianus (Lesson) Possibly early Pleistocene, Japan; late Pleisto- cene and historic, eastern and western North Pacific shores between lats 20° and 45° N. Historic, Galapagos Islands Genus Eumetopias Gill Eumetopias sp. Latest Pliocene or earliest Pleistocene, Japan; early Pleistocene, Oregon and California Eumetopias jubata (Schreber) Late Pleistocene and Holocene, circum-North Pacific north of about lat 33° N. ?Otariid, incertae sedis Genus Oriensarctos Mitchell Oriensarctos watasei (Matsumoto) Early Pleistocene, Japan SUMMARY OF THE HISTORY OF FUR SEALS AND SEA LIONS The earliest known otariids, insofar as they are recognizable at the present time, appear to have been small fur seals of late middle Miocene age (Pithano- taria). The lineage leading to modern Callorhinus appears to have diverged from the main otariid line by late late Miocene time. The sea lions appear, on the basis of a meager fossil record, to have evolved from the Arctocephalus lineage in the later Pliocene of earliest Pleistocene in the North Pacific Ocean, possibly the western North Pacific. The wider distribution of the fur seals in the southern hemisphere, in comparison with sea lions, seems to suggest an earlier crossing of the equator. However, the total absence of otariid seals in the North Atlantic suggest that both must have crossed the equator after effective (for pinnipeds) closure of the Central American Seaway; this closure may have occurred in late late Miocene time—5—6 m.y. ago (Repenning and others, in press). Robert Hoff- stetter reports (written commun., 1973) rare otariid remains in the Miocene or Pliocene deposits (about 5 my ago) of Sacaco, Peru. The Central American Seaway appears to have been the avenue of intro- duction of the monachine phocid seals to the Pacific (Hendey, 1972) and of the sirenian genus Hallianasa, found in association with Pithanotaria in the Cali- fornia area in late middle and early late Miocene deposits. It therefore is suggested that the fur seals origin- ated in the North Pacific and dispersed into the South Pacific at a late date, probably late late Mio- cene time, and that the development of the sea lions out of the fur seals in late Pliocene and early Pleisto- cene time was either shortly followed by their dis- persal to the southern hemisphere, before the late Pleistocene when living genera and species are known in both hemispheres, or was an event which took place independently in both northern and south- ern hemispheres from the already endemic fur seals. Contradicting the latter possibility is the evidence of the endemic lice found on living sea lions of both northern and southern hemispheres (Kim and others, 1975). Except for the late appearance of the sea lions, the fossil history of the otariid seals is remarkably simple when compared with that of the walruses. All that seems to separate the earliest known members of the family from the living species are such minor features as double-rooted cheek teeth, plainer crowns on the cheek teeth, broader basioccipital bones, smaller vertebral foramina, unfused fibula, and minor differences in muscular attachments, articu— lar patterns, and proportions of the limbs. Many of these differences slightly increase the similarity of these earlier otariids to the walruses, but the increase is slight and the oldest known otariid, Pithanotaria, 74 OTARIOID SEALS OF THE NEOGENE is clearly an otariid. There obviously is a consider- able history of the pre-late middle Miocene Otariidae that is yet unknown. PART III: DESMATOPHOCIDS, ENALIARCTIDS, AND FAUNAS Family DESMATOPHOCIDAE As indicated in the introductory section, “Su- prageneric Diagnoses,” the desmatophocid otarioid seals are distinguished by several features; the more conspicuous are: lack of supraorbital processes, nasals penetrating the frontals, posterolaterally pro- jecting jugular processes of the exoccipital, moder- ately broad basioccipital widening posteriorly, pos- teriorly very broad and flat palate, mortised jugal- squamosal articulation, and the development in some of single-rooted cheek teeth very early in the known history of the otarioid seals. In the genus Allodesmus there is minimal en- largement of the petrosal apex, Wide separation of passages for the vestibulocochlear and facial nerves on the medial surface of the petrosal, a broad, shallow hypophyseal fossa, rather small tympanic membrane, and large ossicles. These features are either not known or not described in the genus Desmatophoca. As noted in the original description of Desmato- phoca, and in its name (Condon, 1906, p. 13), the des- matophocid seals are phocidlike in a number of fea- tures, though clearly otarioid in most. In addition to characters unique to this family such as the dis- tinctive jugular process, they possess features found in the odobenids, as the lack of supraorbital pro- cesses, and in the otarioids, as the narrow basioccip- ital bone. Only two genera are currently included in the Desmatophocidae. Mitchell has variously placed these genera, Desmatophoca and Allodesmus, in either the same subfamily (1966, p. 40; Subfamily Desmatophocinae of the Family Otariidae—in the sense here used) or in separate subfamilies (1968, p. 1897; Subfamilies Desmatophocinae and Allodes- minae of the Family Otariidae—in the sense of Otarioidea as here used). Barnes (1972, p. 61), in a recent review, placed Desmatophoca and Allodes- mus in the Subfamily Desmatophocinae of the Family Otariidae—in the sense of Otarioidea as here used. This grouping is followed here, and the group is given familial status, equivalent in morphologic and phylogenetic distinctiveness to the Otariidae and Odobenidae. Genus DESMATOPHOCA Condon Type species.——Desmatophoca oregonensis Con- don, 1906. . Diagnosis—A large desmatophocid with double- rooted cheek teeth, well-developed internal cingulum on the cheek teeth, incisive foramina large for a desmatophocid, mortising of the jugal-squamosal articulation weak, orbits relatively small for known desmatophocids, about 17 percent of the CBL. Dental formula: 3I:1M:4P:2M 2I:1C:4P:1M x 2 z 36 Os penis unknown. Included species.—Desmatophoca oregonensis Condon. Known only from the late early Miocene Astoria Formation of coastal Oregon (Saucesian and Relizian Stages or more likely only Saucesian ac- cording to Snavely, Rau, and Wagner, 1964, esti- mated age about 15—16 m.y.). Packard (1974b) has described a humerus assigned to this species, and a number of unstudied specimens from the type area are in the collection of the National Museum of Natural History. An otarioid rib, University of Alaska Department of Geology N0. UA 2420, may belong to an individual of the genus Desmatophoca. The rib was found in Astoria-equivalent deposits of the Narrow Cape For- mation on Kodiak Island and is characterized by a large head and tubercle and by a more elongate neck than known from otariid seals. The head, however, is not so swollen as those known for Allodesmus (Mitchell, 1966, pl. 13). Desmatophocine A of Barnes (197 2, p. 55) from the upper part of the Santa Margarita Formation of the Santa Cruz area, California, is part of a mandible that bears a strong resemblance to that of Desmato- phoca oregonensis. The specimen was collected high- er in the section than some remains of Imagotaria downsi. Similarities to Desmatophoca oregonensis include a transversely compressed canine and doubly rooted P 2 and P3, as well as other desmato- phocid characters mentioned by Barnes. The speci- men is somewhat smaller than the mandible of the type of Desmatophoca oregonensis, but it could well equal in size the mandible of a female individual if the type is a male. Its more recent geologic age and more nearly coalesced cheek-tooth roots suggest that it may be an unknown species, possibly assignable to the genus Desmatophoca. Genus ALLODESMUS Kellogg Type species.——Allodesmus kernensis Kellogg, 1922. PART III: DESMATOPHOCIDS, ENALIARCTIDS, AND FAUNAS 75 Diagnosis—Large to small desmatophocids “***with crowns of teeth bulbous and smooth; lin- gual cingulum of cheek teeth reduced and smooth***,” (quotes from Barnes, 1972, p. 5), incisive foramina very reduced, mortising of the jugal- squamosal articulation greatly expanded, orbits very large relative to Desmatophoca, between 20 and 25 percent of the CBL. Dental formula: 31'1C'4P'2M 2I'1C'4P‘1M Os penis (of A. kernensis) recurved as in Odobenus, circular in cross section except for slight ventral flat- tening, one ventral and one larger and slightly bilobed dorsal process on the apex (Barnes, 1972, p. 34). Included species.—Allodesmus kernensis Kellogg, 1922 (including Allodesmus kelloggi Mitchell, 1966, following Barnes, 1972): very large species with “***premaxillae expanded into prenarial shelf"‘** premolars deep rooted with single bilobed root; ven- tral margin of dentary concave dorsally***” (ex- tracted from Barnes, 1972, p. 6). See Mitchell (1966) and Barnes (1970 and 1972) for more details. This species is known entirely, or almost entirely, from early middle Miocene deposits (Luisian Stage according to Beck, 1952, estimated age about 13—14 m.y.) near Bakersfield, Calif. Mitchell (1966, p. 25 and 26) lists some specimens from other sites in southern California of questionable specific assign- ment. x2=36 Mr. and Mrs. Martin R. Sorenson have collected an isolated cheek tooth, USNM 184058, from the Santa Margarita Formation of the Santa Cruz area which greatly resembles those of Allodesmus kernensis. This locality (UCMP V5555), an active sand quarry where most fossil material is found after slumping down the cut face, is difficult to assign to a strati- graphic horizon. However, the fauna found thus far includes Desmostylus, Paleoparadoxia, and Hip- parion, genera which elsewhere in the area char- acterize the older part of the Santa Margarita Forma- tion. It is therefore believed that Allodesmus sp. cf. A. kernensis from UCMP V 5555 is from strata older than local records of Imagotaria downsi, Pithano- taria starri, and “Desmatophocine A” of Barnes. Mitchell (1968, p. 1881). has assigned Eumetopias sinanoensis Nagao (1941) to Allodesmus kernensis. Although the Japanese specimen is of near record size and very robust, particularly in the size of the cheek teeth, this generic assignment certainly ap- pears to be correct. Without additional information, it is questionable whether specific synonymy can be demonstrated. The specimen is from middle Miocene rocks in Nagano Prefecture and indicates the extent of the former range of this genus. Allodesmus courseni (Downs, 1956): a small spe- cies lacking a prominent prenarial shelf, with double-rooted cheek teeth, and lacking a dorsally concave inferior margin of the mandible. For more details see Downs (1956) and Barnes (1972, p. 39—40). The type specimen was collected from early middle Miocene (Luisian) deposits in Los Angeles County, Calif. In the National Museum of Natural History there is a cast of a Japanese fossil skull (USNM 24915) questionably referable to Allodesmus courseni (pl. 9). According to Tokio Shikama (written commun., 1967), the specimen was probably destroyed during World War II; it “was found in late 19th century and stored in a shrine of Utsunomiya***The formation north of Utsunomiya (area of Hachimanyama) is middle Miocene Kanomata-zawa formation.” Shi- kama also states that the matrix, according to Dr. J. Suzuki, who purchased the specimen from the shrine in 1927, was a green tuff containing Tertiary mol- lusks. From the cast, the specimen measured 23.9 cm from the anterior tip of the rostrum to the most posterior part of the backward-projecting lambdoid- al crest (comparable to a CBL of 27.7 cm for the type of Allodesmus courseni). The cast shows a very low sagittal crest, narrow interorbital area with no supraorbital process, a somewhat procumbent canine and greatly enlarged third incisor, and one cheek tooth with bulbous and smooth crown. Al- though it appears to have a better developed pre- narial shelf than A. courseni, the shelf is not so accentuated as in A. kernensis. Allodesmus packardi Barnes, 1972: a medium- sized (or large, as Barnes suggests that the type may be a female individual) species with broad skull, apparently with reduced prenarial shelf, single- rooted cheek teeth with marked anterolateral orien- tation, palate very broad with widely diverging cheek tooth rows. For more details see Barnes (1972). The type and only specimen is from early middle Miocene deposits in Menlo Park, Calif. A femur (USNM 23881) from the same formation and same general area may belong to this species; it is rela- tively more elongate but otherwise identical to the femur of Allodesmus kernensis as described by Mitchell (1966, p. 16 and pl. 20). Barnes (1972) has described, as Desmatophocine B and Desmatophocine C, two additional partial man- dibles which differ from known mandibles of named species, but which are not adequate for definition of a new species. Desmatophocine B is questionably and 76 OTARIOID SEALS OF THE NEOGENE Desmatophocine C certainly from the middle Mio- cene rocks from which Allodesmus kernensis is known. For further discussion see the report by Barnes (1972). CLASSIFICATION OF THE DESMATOPHOCIDS Only two desmatophocid genera are currently recognized; one Desmatophoca, is monospecific. Fol- lowing Barnes (1972), except that the group is re- tained as a distinct family, they are here classified as follows: Family DESMATOPHOCIDAE Genus Desmatophoca Condon Desmatophoca oregonensis Condon Late early Miocene, Oregon and Alaska(?) (rib) ?Desmatophocine A of Barnes Early late Miocene, California Genus Allodesmus Kellogg Allodesmus kernensis Kellogg Early and late(?) middle Miocene, California Allodesmus sinanoensis (Nagao) Early middle Miocene, Japan Allodesmus courseni (Downs) Early middle Miocene, California, ?Japan (Utsunomiya shrine) Allodesmus packardi Barnes Early middle Miocene, California Desmatophocid incertae sedis ?Desmat0phocine B of Barnes Early(?) middle Miocene, California Desmatophocine C of Barnes Early middle Miocene, California DISCUSSION OF THE DESMATOPHOCIDS From approximately 16 m.y. ago to possibly 9 m.y. ago, there seems to have been a variety of desmato- phocids in all the coastal waters of the North Pacific Ocean. Although the variety is not large, most species now known show extremely high special- ization compared with the contemporary primitive otariids (Pithanotaria) and odobenids (Neotherium and Imagotaria). Unlike the odobenids and the otariids, the desma. tophocids seem to have appeared rather abruptly in the late early Miocene, to have acquired full diversifi- cation by early middle Miocene, and to be quite rare in the late Miocene. There are no younger records. At present the record appears to be insufficient to attempt any phylogenetic interpretation. By reason of its somewhat greater age and less specialized skull, however, Desmatophoca oregonensis may be assumed to approximate the ancestral desmato- phocid from which Allodesmus evolved. Mitchell (1968, p. 1888 [lines 6, 7, 12—14] and fig. 16) considered Desmatophoca an approximation of the ancestral form for all otarioids (his “otariids”), but Barnes ,(1972, p. 62) rejects this suggestion on the grounds that the genus is far too specialized, a rejection with which we are in agreement. The familial characters of the Desmatophocidae, as out- lined in the introductory section of the present report, are much too distinctive and specialized even to suggest that the very different odobenids and otari- ids could have been derived from this family. Mit- chell (1968, p. 1888 [lines 9, 10, 14—17]) seems in- clined to agree with this. Moreover, Mitchell and Tedford (1973) have recently documented an ances- tral group from which all otarioid families could have been derived. Family ENALIARCTIDAE Mitchell and Tedford (1973) have erected the Enal- iarctidae (their subfamily Enaliarctinae) to include otarioid (their otariid) pinnipeds of hemicyonine ursid derivation which, because of their primitive structure, cannot be assigned to other otarioid families. Although they strongly favored the inter- pretation that Enaliarctos mealsi, or a closely re- lated member of the same family, was ancestral to some of the otarioids (p. 278), they expressed strong reservations about the derivation of the desmato- phocids from the enaliarctids, specifically stating that E. mealsi was not ancestral to Desmatophoca oregonensis (p. 254). In terms of the characters adopted in this paper for diagnosis of the Desmatophocidae, most, perhaps, represent conditions theoretically derivable from the more primitive (that is, more canoidlike) features of Enaliarctos. Few of the derived features that char- acterize the Desmatophocidae are possessed or hint- ed at in Enaliarctos mealsi. On the other hand, as Mitchell and Tedford had concluded, the observable cranial features of E. mealsi do agree well with those displayed by the Otariidae, as seen by comparing the family diagnoses presented herein (see section on “Suprageneric Diagnoses”). Other features such as the shape of the auditory bullae, presence of a tensor tympani fossa, and carnassial cheek teeth are clear- ly arctoid characters retained in Enaliarctos that emphasize the primitive nature of this genus. In sum total, it cannot be fairly argued that the close resemblance of Enaliarctos to the otariids rules out relationship of other unknown members of the Enaliarctidae with other otarioid families. The lack of derived characters typical of specific otarioid families only reinforces the morphologically central position of Enaliarctos mealsi. PART III: DESMATOPHOCIDS, ENALIARCTIDS, AND FAUNAS 77 Although the geologically younger Neotherium mirum is here considered a primitive odobenid, as discussed, morphologic data are not yet available to debate the stand that it was not an enaliarctid greatly advanced over the condition of E. mealsi toward the odobenids. There is no doubt that other enaliarctids existed which more closely resemble the Otariidae and Desmatophocidae as these are recog- nizable in unstudied material in the National Muse- um of Natural History. Therefore, it is here main- tained, in the absence of contradictory evidence, that the enaliarctids are an ideal group from which all other otarioid families, Desmatophocidae, Otariidae, and Odobenidae, could easily/have been derived. The enaliarctids are characterized as a group dis- tinct from the other known otarioids by the com- bination of the following conspicuous features: lack of supraorbital processes, nasals penetrating the frontals, ursidlike mastoid and jugular processes, narrow basioccipital, unspecialized j ugal-squamosal articulation, inflated and flask-shaped bullae, fissi- pedlike tensor tympani muscle, and hemicyoninelike dentition. Mitchell and Tedford enumerate the pinni- pedlike features that distinguish this group from contemporary hemicyonine ursids. Genus ENALIARCTOS Mitchell and Tedford Type species.—Enaliarctos mealsi Mitchell and Tedford, 1973. Diagnosis—At present only the genotypic species has been described and generic and specific diag- noses necessarily follow the familial diagnosis (see section on “Suprageneric Diagnoses”). For details see Mitchell and Tedford (1973, p. 218). Included species.—Enaliarctos mealsi Mitchell and Tedford, from early Miocene deposits on Pyra- mid Hill, Kern County, Calif. The deposits are believed to be about at the temporal limit separating the Zemorrian and Saucesian Stages (Beck, 1952) and are approximately 22.5 m.y. old (Turner, 1970, p. 101). This age is early early Miocene; in fact, Berg- gren (1972) would place this age at the Oligocene- Miocene boundary. It is evident that a considerable record of otarioid evolution is still unknown between 22 and 16 m.y. ago. This species was described on the basis of two skulls, a braincase, and some isolated teeth, includ- ing lower carnassials associated with E. mealsi on the basis of their occlusal relations with the corre- sponding upper teeth. The lower jaw, anterior cheek teeth, and the postcranial skeleton are unknown. Mitchell and Tedford (1973, p. 272—275) illustrate and discuss the pinniped and limb axial elements in the L. E. Wilson collection at Yale University. These remains, obtained near Woody in the Kern River district, constitute part of the Woody local fauna and were collected from the outcrops of the Pyramid Hill Sand Member of the J ewett Sand at that locality. In Mitchell and Tedford’s (1973, p. 275) view, this local fauna is only slightly younger than the Pyra- mid Hill local fauna, which includes the type of E. mealsi. The importance of the fragmentary Woody pinniped fauna lies in its indication that more than one kind of pinniped was in existence in early Miocene time. Knowledge of the exact nature of these pinnipeds will have to await further evidence from the Pyramid Hill Sand Member or other contempora- neous deposits, but limbs of appropriate size for association with Enaliarctos are present along with specimens of larger forms attributed by Wilson to the desmatophocid Allodesmus. If these identifications are correct (the latter doubted by Mitchell, 1966, p. 20), stocks ancestral to the enaliarctids and desma- tophocids may already have diverged from a com- mon ancestor prior to early Miocene time. FAUNAS The Neogene record of the otarioid pinnipeds, though far from complete, is sufficiently well known to recognize lineages and to specify the general composition of faunas during different temporal in- tervals. Some suggestion of latitudinal ranges in the North Pacific Ocean is evident as well as the approx- imate times of dispersal to the southern hemisphere by the otariids and to the Atlantic by the odobenids. To this time (1977), the desmatophocids and the enaliarctids are known only from the North Pacific. In the fauna] record of the North Pacific basin, there are two major gaps. The longest, just men- tioned, is that between Enaliarctos mealsi of about 22 m.y. age and Desmatophoca oregonensis of about 16 m.y. age. Within this time span, or possibly earlier, evolution of the enaliarctids into the desma— tophocids and possibly the odobenids took place. Further information regarding otarioid evolution during the interval between 22 and 16 m.y. ago will be most rewarding. The second major gap, that between 5 and about 2 m.y. ago, clouds the history of the Otariidae between Thalassoleon spp. and the early Pleistocene otariids apparently assignable to extinct species of living genera. At the present time, only the fragmentary remains from the San Diego and Purisima Forma- tions and the mandibular ramus, Eumetopias sp., described by Kaseno (1951), are known during this time span. Kim, Repenning, and Morejohn (1975) have suggested that divergence of the lineage lead- ing to Callorhinus may be of greater antiquity than 78 OTARIOID SEALS OF THE NEOGENE that leading to the living sea lions because of species differentiation in the sucking lice endemic on mod- ern otariids and because of bacular morphology. Callorhinus-like features of Thalassoleon macnal- lyae and of the fragmentary late late Miocene and Pliocene specimens here discussed may record the beginnings of this early differentiation, but the major features of this history, yet to be discovered, must lie in this second major gap. In generalized outline, the desmatophocids, odo- benids, and otariids all experienced three major stages in their evolution but at distinctly different times in each lineage. The first, which is not yet docu- mented in any lineage, consists of the evolution of homodont dentition out of the heterodont, fissiped- like dentition of the enaliarctids. This first stage appears to be the most reasonable criterion for future separation of advanced enaliarctids from the most primitive members of the three derived otarioid families. The second major evolutionary stage is the evolu- tion of single-rooted cheek teeth, presumably in response to simplification of crown pattern. In the desmatophocids, this stage was achieved in some species of Allodesmus 13—14 m.y. ago. In the odo- benids, the oldest genus in which the dentition is known, Imagotaria, appears to have such variation in the presence of double-rooted or single-rooted cheek teeth that it seems reasonable to suppose that the single-rooted stage was achieved in this odobenid about 9 m.y. ago; all younger odobenids have single- rooted cheek teeth. In the more slowly evolving otariids, some single-rooted cheek teeth are not known before about 5 m.y. ago; all cheek teeth were not single—rooted in any form before 3 m.y. ago; and living fur seals, as well as some sea lions, still retain some posterior cheek teeth having double roots. The third major evolutionary stage of the otarioids is best called diversification. This stage seems to follow shortly upon the development of single-rooted cheek teeth and shortly follows the extinction of the preceding diverse family. While only one species of Desmatophoca, with double-rooted cheek teeth, is known about 16 m.y. ago, one or two million years later three species of Allodesmus and “Desmatopho- cine B” of Barnes had evolved. All of these had rather distinct specializations of the head, three had single-rooted cheek teeth, and two appear to have dispersed around the North Pacific to Japan. In one to two more million years, 11—12 m.y. ago, the supremacy of the desmatophocids appears to have come to an end, and undiversified but more abun- dant odobenids (Imagotaria) and otariids (Pithano- taria) are found in the same deposits as the last known remains of Allodesmus and possibly Desma- tophoca (“Desmatophocine A” of Barnes). Geographic distribution and ecologic restriction of the several desmatophocids of the middle Miocene are not clearly evident in the present record. Al- though many or all seem to have lived at the same time in various parts of California, in general, only one species is found at one locality. It seems evident that Allodesmus kernensis and Desmatophocine “B” of Barnes lived in the inland Temblor Sea of the southern San Joaquin Valley, as no specifically identifiable specimens have been found outside this area. Allodesmus packardi and Allodesmus cour- seni seem to have lived along the open coast, possibly with latitudinal differences in ranges. Both A. cour- seni? and A. sinanoensis are represented in Japan, indicating that the genus was widespread in the North Pacific. Although the otariids remain undiversified, all later odobenids from possibly 8 to about 4 m.y. ago have single-rooted cheek teeth and are diverse. Two subfamilies are recognizable during this time inter- val, the Odobeninae and the Dusignathinae; the genera Aivukus, Dusignathus, Pliopedia, and Valen- ictus and the problematical Pontolis are known from the North Pacific. The odobenids seem to have been quite abundant; they are known from south of the Tropic of Cancer (Aivukus) to almost 45° N. lat (Pontolis). However, the present record suggests that the odobenines, at this time, were distributed more southerly than the dusignathines because the odo- binines are not certainly known north of 29° N. lat while the dusignathines are not known south of this latitude. Before 5 m.y. ago, the odobenines succeeded in invading the North Atlantic, and, because of their southerly distribution, it seems most probable that they did so by the Central American Seaway. Failure of contemporary otariids to invade the Atlantic suggests that they had a more northerly distribution than the odobenines, comparable to that of the dusignathine odobenids. Within the Dusignathinae, the poorly known genus Pliopedia may have been coeval with the genus Dusignathus for 1 or 2 m.y., between 5—7 m.y. ago. Pliopedia seems to have preferred the warmer waters of the then existing inland sea of central California. From at least 5 to possibly 8 m.y., Dusig- nathus inhabited the waters of the open coast at least from lat 38° N. to 28° N. Depending upon the correct age of the Drakes Bay Formation, the youngest record of Dusignathus may be the type specimen, something less than 6.7 m.y. In the central Cali- fornian sea, Pliopedia is succeeded by Valenictus whose age is more than 4.3 m.y. Valenictus is also PART III: DESMATOPHOCIDS, ENALIARCTIDS, AND FAUNAS 79 known in the Imperial Valley of California in de- posits of the ancient Gulf of California, and it may be that this poorly known genus gained access between these two inland seas by occupying the open coast environment vacated by Dusignathus. Thus part of the Imperial Formation may be only 4—6 m.y. old. As mentioned, single-rooted cheek teeth (other than the first premolar) are not known in the Otari- idae until about 5 m.y. ago. A separate lineage leading to living Callorhinus may have been estab- lished at this time, and the fur seals seem to have dispersed south of the equator shortly thereafter. However, it seems evident that major diversification of the otariids began about three million years ago and that the diversity of the living fur seals and par— ticularly the living sea lions represents the same sort of diversity seen in the odobenids from 6—9 m.y. ago and in the desmatophocids from 13—14 m.y. ago. In summary (fig. 6), the Neogene pinniped faunas of the Pacific coast of North America have, or can be expected to have, the following composition: Oligocene (pm-22.5 m. y. B.P.).—Published record: None. Undoubtedly the enaliarctids were present in the Oligocene (following Berrgren’s usage of this time term), but judged by the cranial features of Enaliarctos mealsi, without evidence of the nature of their feet or flippers, it may be difficult to decide whether they are fissipeds or otarioid pinnipeds. At this time, it seems most reasonable to separate the Enaliarctidae from the hemicyonine ursids at that point where they become obviously aquatic in habit. Early early Miocene (about 22.5-17 m.y. B.P.).— Published record: Enaliarctos mealsi from the begin- ning of this time. This species shows greatest simi- larities to the Otariidae or, conversely, the Otariidae show the least modification of the features of this primitive type of otariioid. It is to be expected that other early Miocene enaliarctids, some already known but unstudied, will show features indicating evolution toward the other two otarioid families. Because of its complete homodonty in the late early Miocene, Desmatophoca suggests that the enaliarc- tine~desmatophocine transition took place in the early early Miocene or before; this suggestion is corroborated by the postcranial elements from the early early Miocene Woody local fauna, as pointed out by Mitchell and Tedford (1973, p. 274). The only described enaliarctid is associated with land mammals believed to indicate a late Arikareean age and with mollusks of the “Vaqueros Stage” (Mitchell and Tedford, 1973). Late early Miocene (about 17—14.5 m.y. B.P.).— Published record: Desmatophoca oregonensis from Oregon and probably Alaska. It is probable that early forms of Allodesmus, resembling A. courseni, also evolved in this period because of the high degree of specialization of other species known from the early middle Miocene. Though poorly known, the primitive nature of the few known bones of the early middle Miocene odobenid Neotherium suggests that there were enaliarctids in the late early Miocene immediately ancestral to the odobenids, unless Neo- therium itself proves to be such an enaliarctid. The complete homodonty, loss of the upper second molar, and general otariid appearance of the late middle Miocene Pithanotaria suggests that a considerable gap separates it from the immediate enaliarctid ancestor of otariids. Hence, such an ancestor may also have lived during the late early Miocene, and in fact such forms are already recognizable in un- studied specimens in the National Museum of Natu- ral History. From near the type locality of Desmatophoca oregonensis, as nearly as it can be determined, a relatively large rhinoceros maxilla has been col- lected which is tentatively assigned to the genus Aphelops and appears to be identical to specimens from Mascall-equivalent beds in southeastern Ore- gon. The specimen is now in the National Museum of Natural History (USNM 187123). A Barstovian land mammal age, possibly early Barstovian, is indi- cated. Megainvertebrate and microinvertebrate faunal control, which is excellent, indicates that “Temblor” and Saucesian Stages are represented. Early middle Miocene (about 14.5—13 m.y. B.P.).— Record: Allodesmus packardi, Allodesmus courseni, Allodesmus kernensis, “desmatophocine B” and “C” of Barnes, and Neotherium mirum, all from California but also two Allodesmus records from Japan. If Neotherium does not prove to be an enaliarctid, the enaliarctids may be extinct by this time except that there is no known record of an otariid from this time interval, so the enaliarctid immediately ancestral to the otariids may have been of this age. The renowned Sharktooth Hill locality of Kern County, Calif, is the most productive of marine mammals of this age. Allodesmus kernensis, Neo- therium mirum, “desmatophocine C” of Barnes, and presumably “desmatophocine B” of Barnes are known from this locality. A few land mammals have been found in this locality (listed by Mitchell, 1966, p. 29); they are of Barstovian age. The fossil beds are closely tied into the late “Temblor” megainverte— brate stage and the Luisian microinvertebrate stage. Late middle and early late Miocene (about 13-8.5 m.y. B.P.).—Record: Imagotaria downsi, Pithano- taria starri, ?Allodesmus sp., and ?Desmatophoca 80 OTARIOID SEALS OF THE NEOGENE NORTH AGE ““45 SOUTH PACIFIC NO RTH P AC | F I C ATLANTIC (this report) (m.y.) < o O :I: O. 0W l|.l.l z (Living species) ARETOS Late j ._. 1 _ Early OTARIA (Living genera) PLEISTO- CENE EUME’K ALLO HINQ €1\ (OPHUS OPIAs§ PHOC 0. huxl? ' Eumetapias sp., Japan Late San Diego specimen ? AHCTO EPHALUS T. macnallyae fix/AL lew .n “PLIOPEDIA on PLIOCENE Early Vsp. “- Not studied, Peru ‘2; P. magnu: aha D. ran ta , cmzensis 5 _ T, macmleyae o on “at: AATHUS: a . V. imperialensis on can ,0 MGM — 6 _ Etchegoin specimen n a M I. baoaosm‘ARus' °;, " u ,a Late K 7 _. T. mexicanus -— v‘o nAIny‘us“ Late Desmato- phocine ”A" I. dawns:— THALASSOLEON no.00ao' Early IMAGOTARIA°~ . - ,,. .4.“ o.“ a . ‘7 .‘Du'o". \ —13 Early l Late OTHERl U O C E N E MIddle N .nflna “a° l I.._ In D. oreganenxi: Late Not studied, Oregon 17 N at studied, Oregon Early EXPLANATION VIII/Ill OTAHIIDAE Early _21_ ODOBENIDAE DESMATOPHOCIDAE _24_ :1 ENALlARCTIDAE ENALIAHCTOS OLIGO' CENE FIGURE 6.—Phylogenetic diagram of the Otarioidea. sp. (“desmatophocine A” of Barnes). The primitive phocids entered the North Pacific through the Cen- odobenid and otariid seem to have been fairly com- tral American Seaway at this time, although there is mon seals in California, but the two questionably no conclusive fossil evidence as yet known. Repen- assigned desmatophocids are represented by only ning and Ray (1977) give reasons for believing that ‘ one record each. this happened at an earlier date. Hendey (1972) has suggested that the monachine As nearly as available correlations indicate, this PART III: DESMATOPHOCIDS, ENALIARCTIDS, AND FAUNAS 81 otarioid fauna seems to correlate closely with the Clarendonian land mammal age and with the “Mar- garitan” and earlier “Jacalitos” megainvertebrate stages. The association of Imagotaria with Pithano- taria is remarkably persistent. Records from the inland sea area in the southern San Joaquin Valley do not include Pithanotaria, however, and thus they offer a very weak suggestion that this primitive fur seal may have preferred the cooler waters of the open sea. There exists at present considerable doubt about the temporal extent of this fauna into the late Miocene. The very large sample from the Santa Margarita Formation of the Santa Cruz area is at least largely, and possibly entirely, from rocks con- taining invertebrates of the “Jacalitos” stage and thus the fauna is presumably younger than 10 or 11 m.y. In contrast, the fossil pinnipeds from Point Reyes, Calif., represents the next younger fauna, which is distinctly different than that containing Imagotaria and Pithanotaria. This fauna from Point Reyes has been directly dated, however, at 9.3 m.y. As discussed, the pinnipeds, the cetaceans, and the only mollusk known from the Drakes Bay Formation of Galloway at Point Reyes all suggest a consider- ably younger age than the radiometric date seems to indicate. In the present report, the pinniped fauna from Point Reyes is considered as being late late Miocene or Pliocene, possibly no older than 6 m.y. and possibly as young as 4 m.y. Late late Miocene and early Pliocene (about 8.5- about 3.8 m.y. B.P.).—Record: Thalassoleon mexi— canus, Thalassoleon macnallyae, Aivukus cedrosen- sis, Dusignathus santacruzensis, Valenictus imperi- alensis, Pontolis magnus, Pliopedia pacifica, and a poorly known fur seal possibly representing the lineage leading to Callorhinus. Both latitudinal and environmental differentia- tion of faunas seems to be apparent during this time. The only Tertiary odobenine odobenid of the Pacific, Aivukus, seems to have been restricted to southern latitudes from Cedros Island southward at least to the Tropic of Cancer. The dusignathine Dusignathus extended its range from Cedros Island northward to at least Point Reyes, and the dusignathine Pontolis may have represented still another faunal differen- tiate from Oregon. The warm inland seas seem to have been favored by the dusignathine genera Plio- pedia and Valenictus over the latitudinal range of Dusignathus. The two species of Thalassoleon seem to have been either separated on a north-south basis or the northern species represented a somewhat later and more advanced species (contrary to the radio- metric date of the Drakes Bay Formation of Gallo- way). Within the northern of the two then existing interior seas, the San Joaquin-Etchegoin sea, the otariids are represented by a single mandibular fragment of a small fur seal which may represent the beginning of the Callorhinus lineage. Toward the end the early Pliocene, Prorosmarus alleni is known from the Atlantic coast of North America, but the history of the odobenines of the North Pacific largely began and ended in the late Miocene; Dusignathus and Pliopedia may have lived until the beginning of the Pliocene, and Valenictus may have survived until the late Pliocene. According to Robert Hoffstetter (written commun., 1973), late Miocene or early Pliocene collections from Sacaco, southern Peru, contain a few remains pro- visionally identified as postcranial elements of an otariid. From this it would appear that the otariids first dispersed to the South Pacific Ocean about 5 m.y. ago, presumably of the Thalassoleon stage of evolution. Late Pliocene (about 3.8—1.8 m.y. B.P.).—Record: a small fur seal from the San Diego Formation, resembling Callorhinus but with a primitive astra- galus and some teeth still double rooted, probably “Allodesmus sp.” of Kaseno (1951) from Japan, which seems to be the earliest record of the sea lion stage of otariid evolution. The San Diego specimen could be the same species as that known from the late late Miocene of the Etchegoin Formation, and if itis, shows little difference and a slow rate of evolution in what is presumed to be the early stages of the lineage leading to modern Callorhinus. On the other hand, the Japanese sea lion described by Kaseno and currently included in the genus Eumetopias shows a comparatively sudden development of single-rooted cheek teeth and of large size, both features con- sidered to be characteristic of the sea lions. The late Pliocene history of the otarioid seals is obviously very incomplete. There is no record of any odobenid, and it is reasonable to suspect that this ancient, abundant, and diversified family became extinct in the North Pacific. Also from rocks of late Pliocene age, 5,000 feet below the top of the Yakataga Formation in the Malaspina District of Alaska, is the oldest record of a phOcoid seal in the North Pacific (Repenning and others, in press), a phocine radius of very modern aspect (USNM 23876, pl. 16), most similar to that of Pusa sibirica among the radii of living phocids. Early Pleistocene (about 1.8-0.7 m. y. B.P.).—Early Pleistocene records consist largely, but not entirely, of postcranial elements from California, Oregon, and Japan, which are, in contrast to the San Diego fossil, indistinguishable from living genera; how- 82 OTARIOID SEALS OF THE NEOGENE ever, a few specimens of dentition and one un- described skull from North America indicate extinct species. Mitchell (1968) has placed a partial rostrum from Japan in an extinct genus, Oriensarctos. Very limited, but quite indicative, terrestrial mammals are associated with these early Pleistocene otarioids which indicate this age but, in North America, these mammals do not exclude a younger age. Although there are no early Pleistocene records in the Southern Hemisphere, the presence of the living species of the three southern sea lion genera in the late Pleistocene of Australia, New Zealand, and Argentina suggests that they had evolved in place for some time. As the earliest known sea lion seems to be only about 2 m.y. old, it is assumed that southern dispersal was in the early Pleistocene. The pos- sibility that southern and northern sea lion genera evolved independently, each in its own hemisphere, from the then native fur seals seems to be ruled out by the nature of their endemic sucking lice; both north- ern and southern sea lions are host to the same endemic species of louse, a species not known from any other host, including the northern and southern fur seals (Kim and others, 1975). Late Pleistocene (about 0. 7-0.0] m.y. B.P.).-Late Pleistocene records also are fragmentary but are known from California, Oregon, Alaska, and Japan. Many records, particularly in Alaska, are associated with terrestrial mammals. Only living otariid spe- cies are known; the oldest North Pacific record of Odobenus rosmarus is a humerus (USNM 184059) dug from a fossil beach ridge of the Pelukian (Sangamon) transgression 5-6 miles up the Kokolik River northeast of Point Lay, Alaska. Because of the lack of any record of an odobenine odobenid in the North Pacific between the late late Miocene and the late Pleistocene, it is presumed that Odobenus entered the North Pacific at this time from the North Atlantic. As presently known, the late Pleistocene record of otarioid seals is identical to that of today except for changes of ranges associated with glaciation and the influence of man. By late Pleistocene time the fur seals had crossed the South Atlantic and are known from South Africa; presumably they had already established their present circumpolar distribution. In addition, the fossil record of Otaria from Argentina would indicate that this sea lion had entered the western South Atlantic, but the southern distribution of the sea lions has not enlarged since that time, and they still are basically animals of the North and South Pacific. REFERENCES CITED Addicott, W. 0., 1969, Late Pliocene mollusks from San Francisco Peninsula, California, and their paleogeographic signifi- cance: California Acad. Sci. Proc., v. 37, no. 3, p. 57—93. 1972, Provincial middle and late Tertiary molluscan stages, Temblor Range, California, in Pacific Coast Miocene Biostratigraphic Symposium: Soc. Econ. Paleontologists and Mineralogists Proc., Pacific Sec., March 1972, Bakersfield, Calif, p. 1—26. Addicott, W. 0., and Galehouse, J ., 1973, Pliocene marine fossils in the Paso Robles Formation, California: U.S. Geol. Survey Jour. Research, v. 1, no. 5, p. 509-514. Akers, W. H., 1972, Planktonic Foraminifera and biostratigraphy of Neogene formations, northern Florida and Atlantic coastal plain: Tulane Studies Geology and Paleontology, v.9, p. 1—139. Allen, G. M., 1930, The walrus in New England: J our. Mammalogy, v. 11, p. 139—145. Allen, J. A., 1870, On the eared seals (Otariadae), with detailed description of the North Pacific species: Harvard Univ. Mus. Comp. Zoology Bull., v. 2, p. 1—108. 1880, History of North American Pinnipeds: U.S. Geol. and Geog. Survey Terr., Misc. Pub. 12, xvi + 785 p. Allison, E. 0., Durham, J. W., and Zullo, V. A., 1962, Cold-water late Cenozoic faunas of northern California [abs.]: Geol. Soc. America Spec. Paper 68, p. 2. Ameghino, F., 1889, Contribucidn al conocimiento de los mami- feros fosiles de la Repdblica Argentina: Buenos Aires, Pablo E. Coni E Hijos, Actas Acad. Nacion. Ciencias en Cordoba, 1027 p. Axelrod, D. I., 1971, Fossil plants from the San Francisco Bay Region: Sacramento Geol. Soc. Guidebook, p. 74—86. Baldwin, E. M., 1945, Some revisions of the late Cenozoic strati- graphy of the southern Oregon coast: Jour. Geology, v. 53, no. 1, p. 35—46. Barnes, L. G., 1970, A re-evaluation of mandibles of Allodesmus (Otariidae, Carnivora) from the Round Mountain Silt, Kern County, California: California Univ. Mus. Paleontology Contr., PaleoBios no. 10, 24 p. —1971, Imagotaria (Mammalia, Otariidae) from the late Miocene Santa Margarita Formation near Santa Cruz, Calif- ornia: California Univ. Mus. Paleontology Contr., PaleoBios no. 11, 10 p. 1972, Miocene Desmatophocinae (Mammalia: Carnivora) from California: California Univ. Pubs. Geol. Sci., v. 89, p. 1—68. 1973, Praekogia cedrosensis, a new genus and species of fossil pygmy sperm whale from Isla Cedros, Baja California, Mexico: Los Angeles County Mus. Contr. Sci., no. 247, 20 p. Barron, J. A., 1976, Marine diatom and silicoflagellate biostrati- graphy of the type Delmontian Stage and the type Bolivina obliqua Zone, California. U.S. Geological Survey, Jour. Research, v. 4, no. 2, p. 339—351. Beck, R. S., 1952, Correlation chart of Oligocene, Miocene, Plio- cene, and Pleistocene in San Joaquin Valley and Cuyama Val- ley areas, in Field trip routes, geology, oil fields: Am. Assoc. Petroleum Geologists-Soc. Econ. Paleontologists and Mineralogists-Soc. Explor. Geophysicists Guidebook, Joint Ann. Mtg., Los Angeles, Calif, 1952, p. 104. Berggren, W. A., 1971, Tertiary boundaries and correlations, in Funnell, B. M., and Riedel, W. R., eds., The micropaleontology of Oceans: Cambridge, University Press, p. 693—809. REFERENCES CITED 83 —1972, A Cenozoic time-scale—some implications for regional geology and paleobiogeography: Lethaia, v. 5, no. 2, p. 195—215. Berggren, W. A., and Van Couvering, J. A., 1974, Late Neogene biostratigraphy, geochronology and paleoclimatology of the last 15 million years in marine and continental sequences: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 16, nos. 1-2, p. 1-216. Berry, E. W., and Gregory, W. K., 1906, Prorosmarus alleni, a genus and species of walrus from the upper Miocene of Yorktown, Virginia: Am. Jour. Sci., ser. 4, v. 21, art. 40, p. 444—450. Berry, J. A., 1928, A new species of fossil Arctocephalus from Cape Kidnappers: New Zealand Inst. Proc. Trans, v. 59, p. 208—211. Berry, J. A., and King, J. E., 1970, The identity ofthe Pliocene seal from Cape Kidnappers, New Zealand, previously known as Arctocephalus caninus: Tuatara, v. 18, no. 1, p. 13—18. Blow, W. H., 1969, Late middle Eocene to Recent planktonic fora- miniferal biostratigraphy, in Bronnimann, P., and Renz, H. H., eds.: Internat. Conf. Planktonic Microfossils, 1st, Geneva 1967, Proc., v. 1, p. 199-421. Bobrinsky, N. A., Kuznetsov, B. A., and Kuzyakin, A. P., 1965, [Guide to the mammals of U.S.S.R.] [2d ed.]: Izdatel’stvo, Prosveshchenie, Moscow, 382 p. [in Russian]. Bramlette, M. N., 1946, The Monterey formation of California and the origin of its siliceous rocks: U.S. Geol. Survey Prof. Paper 212, 57 p. Brisson, M. J., 1762, Regnum animalein classes IXdistributum Edito altera auctior, 2d ed.: Lugduni Batavorum, 8 + 296 p. Br'unnich, M. T., 1771, Zoologiae fundamenta: Hafnia et Lipsiae, 254 p. Burleson, G. L., 1948, A Pliocene pinniped from the San Diego Formation of southern California: California Univ. Pub. Zoology, v. 47, no. 10, p. 247—254. Clark, J. C., 1966, Tertiary stratigraphy of the Felton-Santa Cruz area, Santa Cruz Mountains, California: Stanford, Calif, Stanford Univ., Ph.D. thesis. Condon, T., 1906, A new fossil pinniped (Desmatophoca ore- gonensis) from the Miocene of the Oregon coast: Oregon Univ. Bull., Suppl. v. 3, no. 3, p. 1—14. Corey, W. H., 1954, Tertiary basins of southern California, Chap. 3 of Pt. 8, in Jahns, R. H., ed., Geology of southern California: California Div. Mines Bull., 170, p. 73—83. Cuvier, F., 1826, Phoques, v. 39, in Dictionnaire des sciences naturelles...: Strasbourg and Paris, (60 vols), p. 540—559. Dibblee, T. W., Jr., 1950, Geology of southwestern Santa Barbara County, California: California Div. Mines Bull. 105, p. 1—95. Downs, T., 1956, A new pinniped from the Miocene of southern California: with remarks on the Otariidae: J our. Paleontology, v. 30, no. 1, p. 115-131. ' DuBus, B., 1867, Sur quelques mammiferes du Crag d’Anvers: Acad. Sci. Belgique Bull., ser. 2, v. 24, p. 562—577. Durham, J. W., 1954, The marine Cenozoic of southern California, Chap. 3 of Pt. 4, in Jahns, R. H., ed., Geology of southern California: California Div. Mines Bull., 170, p. 23—31. Durham, J. W., Jahns, R. H., and Savage, D. E., 1954, Marine- nonmarine relationships in the Cenozoic section of California, pt. 7 in Chap. 3 of Jahns, R. H., ed., Geology of southern California: California Div. Mines Bull. 170, p. 59-71. Evernden, J. F., Savage, D. E., Curtis, G. H., and James, G. T., 1964, Potassium-Argon dates and the Cenozoic mammalian chronology of North America: Am. Jour. Sci., v. 262, p. 145— 198. Fleming, C. A., 1968, New Zealand fossil seals: New Zealand Jour. Geology and Geophysics, v. 11, no. 5, p. 1184—1187. Galloway, A. J., 1977, Geology of the Point Reyes Peninsula, Marin County, California: California Div. Mines and Geology Bull. 202, (in press). Gervais, H., and Ameghino, F., 1880, Les mammiferes fossiles de l’Amerique du Sud: Buenos Aires and Paris, xi + 225 p. Gill, E. D., 1968, Fossil sea lion as a palaeoclimatologic indicator: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 5, p. 235-239. Gill, T., 1866, Prodrome of a monograph of the pinnipedes: Essex Inst. Proc., Salem, Commun., no. 5, p. 3-13. Glen, W., 1959, Pliocene and lower Pleistocene of the western part of the San Francisco peninsula: California Univ. Pub. Geol. Sci., v. 36, no. 2, p. 147—198. Gray, J. E., 1825, An outline of an attempt at the disposition of Mammalia into tribes and families with a list of genera apparently appertaining to each tribe: Annals Philosophy (London), n.s., v. 10 (26 of entire series), p. 337—344. _1837, Description of some new or little known Mammalia, principally in the British Museum collection: Mat. Nat. History, n.s., v. 1, p. 577—587. __1850, Seals, in Catalog ofthe specimens ofMammalia in the collection of the British Museum: London, British Museum, pt. 2, 53 p. Gregory, W. K., and Hellman, M., 1939, On the evolution and major classification of the citets (Viverridae) and allied fossil and recent Carnivora: a phylogenetic study of the skull and dentition: Am. Philos. Soc. Proc., v. 81, p. 309—392. Handley, C. 0., Jr., 1953, Marine mammals in Michigan Pleis- tocene beaches: Jour. Mammalogy, v. 34, no. 2, p. 252—253. Harrison, R. J., and Kooyman, G. L., 1968, General physiology of the Pinnipedia, Chap. 7 in The behavior and physiology of pinnipeds: New York, Appleton-Century-Crofts, p. 211—296. Hasse, G., 1909, Les morses du Pliocbne Poederlien a Anvers: Soc. Belge Géologie Bull., Mém., v. 23, p. 293—322. Hay, O. P., 1930, Second bibliography and catalogue of the fossil Vertebrata of North America: Carnegie Inst. Washington Pub. 390, pt. 2, xiv + 1074 p. Hazel, J. E., 1971, Ostracode biostratigraphy of the Yorktown Formation (upper Miocene and lower Pliocene) of Virginia and North Carolina: U.S. Geol. Survey Prof. Paper 704, 13 p. Hendey, Q. B., 1972, The evolution and dispersal of the Mona- chinae (Mammalia: Pinnipedia): South African Mus. Ann., v. 59, pt. 5, p. 99—113. __1974, Late Cenozoic Carnivora of the southwestern Cape Province: South African Mus. Ann., v. 63, p. 1—369. Hertlein, L. G., 1925, Pectens from the Tertiary of Lower Califor- nia: California Acad. Sci. Proc., ser. 4, v. 14, no. 1, p. 1—35. ____1934, Pleistocene mollusks from the Tres Marias Islands, Cedros Island, and San Ignacio Lagoon, Mexico: Southern California Acad. Sci. Bull., v. 33, pt. 2, p. 59—73. Howard, H., 1949, New avian records for the Pliocene of Calif- ornia: Carnegie Inst. Washington Pub. 584, pt. 6, p. 177—199. 1971, Pliocene avian remains from Baja California: Los Angeles County Mus. Contr. Sci., no. 217, 17 p. Howell, A. B., 1929, Contributions to the comparative anatomy of the eared and earless seals (genera Zalophus and Phocu): U.S. Natl. Mus. Proc. 273, v. 73, art. 15, 142 p. 84 OTARIOID SEALS OF THE NEOGENE Ikebe, N., Takayanagi, Y., Chiji, M., and Chinzei, K., 1972, Neogene biostratigraphy and radiometric time scale of' Japan—an attempt at international correlation: Pacific Geology, v. 4, p. 39—71. Illiger, J. C. W., 1811, Prodromus systematis mammalium et avium; additis terminis zoographicis utriusque classis, corumgue versione germanica: Berlin, C. Salfeld, xviii + 301 p. Johnson, N. M., Opdyke, N. D., and Lindsay, E. H., 1972, Paleo- magnetic zonation of Blancan and Irvingtonian faunas from the upper San Pedro Valley, Arizona [abs]: Geol. Soc. America, Abstracts with Programs, 1972 Ann. Mtg., v. 4, no. 7, p. 555. Jordan, E. K., and Hertlein, L. G., 1926, A Pliocene fauna from Maria Madre Island, Mexico: California Acad. Sci. Proc., ser. 4, v. 15, no. 4, p. 209—216. Kaseno, Y., 1951, Pliocene pinniped remains from Kanazawa, Ishikawa Prefecture, Japan: Paleont. Soc. Japan Trans. and Proc., n.s., v. 2, p. 57—64. Kellogg, R., 1921, A new pinniped from the upper Pliocene of California: Jour. Mammalogy, v. 2, no. 4, p. 212-226. 1922, Pinnipeds from Miocene and Pleistocene deposits of California: California Univ. Pubs, Dept. Geol. Sci. Bull., v. 13, no. 4, p. 23—132. 1925a, New pinnipeds from the Miocene diatomaceous earth near Lompoc, California, Part IV in Additions to the Tertiary history of the pelagic mammals of the Pacific coast of North America: Carnegie Inst. Washington, Contr. Palae— ontology, Pub. 348, p. 71—96. 1925b, Structure of the flipper of a Pliocene pinniped from San Diego County, California, Part V in Additions to the Tertiary history of the pelagic mammals of the Pacific coast of North America: Carnegie Inst. Washington, Contr. Paleontology, Pub. 348, p. 97-116. 1927, Fossil pinnipeds from California, in Additions to the palaeontology of the Pacific Coast and Great Basin regions of North America: Carnegie Inst. Washington, Contr. Palaeon- tology, Pub. 346, p. 25—37. 1931, Pelagic mammals from the Temblor formation of the Kern River region, California: California Acad. Sci. Proc., ser. 4, v. 19, no. 12, p. 217—397. Kim, K. C., Repenning, C. A., and Morejohn, G. V., 1975, Specific antiquity of the sucking lice and evolution of otariid seals: Symposium on the Biology of the Seals at the University of Guelph, Guelph, Canada, 1972, Proc., Rapports Proc\es Verbaux des Re'un., v. 169, p. 544—549. King, J. E., 1964, Seals of the World: London, British Mus. (Nat. History), 154 p. 1966, Relationships of the hooded and elephant seals (genera Cystophora and Mirounga): Jour. Zoology, v. 148, p. 385-398. 1969, The identity of the fur seals of Australia: Australian Jour. Zoology, v. 17, p. 841—853. Kleinpell, R. M., 1938, Miocene stratigraphy of California: Tulsa, Okla., Am. Assoc. Petroleum Geologists, iv + 450 p. Kubota, K., Nagasaki, F., Matsumoto, K., and Tsuboi, M., 1961, Histological studies on the growth layers in the maxillary canines of fur seals as an indicator of age. Part I, The teeth at the age of one to nine years: Tokyo Med. Dent. Univ. Bull., v. 10, no. 3, p. 261—285. Lankester, E. R., 1865, On the sources ofthe mammalian fossils of the Red Crag, and on the discovery ofa new mammal in that deposit, allied to the walrus: Geol. Soc. London Quart. Jour., v. 21, pt. 3, no. 83, p. 221—232. 1880, On the tusks of the fossil walrus, found in the Red Crag of Suffolk: Linnean Soc. London Trans, ser. 2, Zoology, v0. 2, no. 6, p. 213—221. Leffler, S. R., 1964, Fossil mammals from the Elk River Forma- tion, Cape Blanco, Oregon: Jour. Mammalogy, v. 45, no. 1, p. 53—61. Lindsay, G. E., 1965, A fossil hunt: Pacific Discovery, v. 18, no. 6, p. 18—24. Linnaeus, C., 1758, Systemae naturae..., 10th ed., rev.: Stockholm, Laurentii Salvii, v. 1, pt. 1, 532 p. Lippolt, H.—J., Gentner, W., and Wimmenauer, W., 1963, Alters- bestimmungen nach der Kalium-Argon-Methode an tertiaren Eruptivgesteinen Siidwestdeutschlands: Baden-Wiirttem- burg Geol. Landesamt. Jour., v. 6, p. 507—538. Lyon, G. M., 1941, A Miocene sea lion from Lomita, California: California Univ. Pub. Zoology, v. 47, p. 23—42. McCoy, Fredrick, 1877, Prodromus of the Palaeontology of Victoria, Decade 5: Melbourne, Victoria Geol. Survey, p. 7-9. McKenna, M. C., 1969, The origin and early differentiation of therian mammals: New York Acad. Sci. Annals, v. 187, no. 1, p. 217-240. McLaren, I. A., 1960, Are the pinnipeds biphyletic: Systematic Zoology, v. 9, no. 1, p. 18—28. Matsumoto, H., 1925, [On a new species of fossil sea lion from Umegase, Province of Kazusa]: Geol. Soc. Tokyo Jour., v. 32, p. 45—49 (in Japanese). 1939, [On two species of fossil mammals from Kazusa]: Zool. Mag. Tokyo, v. 51, no. 5, p. 257—266 (in Japanese, English summary). Miller, M. E., Christensen, G. C., and Evans, H. E., 1965, Anatomy of the dog: New York, W. B. Saunders, x + 941 p. Mina, U. F., 1957, Bosquejo geoldgico del Territorio Sur de la Baja California: Asoc. Mexicana Geolo’gos Petroleros Bol., v. 9, nos. 3-4, p. 139—269. Mitchell, E. D., J r., 1961, A new walrus from the Imperial Pliocene of southern California: with notes on odobenid and otariid humeri: Los Angeles County Mus. Contr. Sci., no. 44, 28 p. 1962, A walrus and a sea lion from the Pliocene Purisima Formation at Santa Cruz, California: with remarks on the type locality and geologic age of the sea lion Dusignathus santacruzensis Kellogg: Los Angeles County Mus. Contr. Sci., no. 56, 24 p. 1966, The Miocene pinniped Allodesmus: California Univ. Pubs. Geol. Sci., v. 61, p. 1—46. 1968, The Mio-Pliocene pinniped Imagotaria: Fisheries Research Board Canada Jour., v. 25, no. 9, p. 1843—1900. Mitchell, E. D., Jr., and Repenning, C. A., 1963, The chronologic and geographic range of desmostylians: Los Angeles County Mus. Contr. Sci., no. 78, 20 p. Mitchell, E. D., Jr., and Tedford, R. H., 1973, The Enaliarctinae: a new group of extinct aquatic Carnivora, and a consideration of the origin of the Otariidae: Am. Mus. Nat. History Bull., v. 151, art. 3, p. 205—284. Morejohn, G. V., 1975, A phylogeny of otariid seals based on morpholoy of the baculum: Symposium on the Biology of Seals at the University of Guelph, Guelph, Canada, 1972, Proc., Rapports Proces Verbaux des Reun., v. 169, p. 49—56. Mourlon, M. M., 1877, Sur le classement stratigraphique des Phoques fossiles receuillis dans les terrains d’Anvers: Acad. Royal Sci., Lettres, et Beaux-Arts de Belgique Bull., v. 93, ser. 2, p. 603—609. Naeser, C. W., Izett, G. A., and Wilcox, R. E., 1971, Zircon fission- track ages of Pearlette-like volcanic ash beds in the Great Plains [abs]: Geol. Soc. America, Abstracts with Programs, 1971 Ann. Mtg., v. 3, no. 7, p. 657. REFERENCES CITED 85 Nagao, T., 1941, An occurrence of a fossil sea lion in the Miocene deposits of Sinano, Japan: Hokkaido Imp. Univ. Fac. Sci. Jour., ser. 4, v. 6, no. 2, p. 75—84. Nirei, H., 1969, On the Utatsuyama formation around Kanazawa City, Central Japan: Geol. Soc. Japan Jour., v. 75, p. 471—484. Ognev, S. I., 1935, [Mammals of the U.S.S.R. and adjacent countries, Vol. III Camivora (Fissipedia and Pinnipedia) : Moscow and Leningrad, State Publishing House, p. 316-600 (in Russian). __ Orlov, J. A., 1931, Uber die Reste eines primitiven Pinnipedier aus den neogenen Ablagerungen Westsibiriens: Acad. Sci. U.S.S.R. Comptes Rendus, ser. A, v. 3, p. 67—700. Packard, E. L., 1947a,. A fossil sea lion from Cape Blanco, Oregon: Oregon State Mons. Studies Geology, no. 6, p. 15-21. 1947b, A pinniped humerus from the Astoria Miocene of Oregon: Oregon State Mons. Studies Geology, no. 7, p. 25-31. Palmer, T. S., 1904, Index generum mammalium: a list of the genera and families of mammals: US. Dept. Agriculture, North American Fauna, no. 23, p. 1—984. Peters, W. C. H., 1877, Uber die Ohrenrobben, Otariae, als Nachtrag zu seiner im vorigen Jahre iiber diese Thiere glesenen Abhandlung: Kiiniglich Preussichen Akad. Wissen, Berlin Monats., p. 505—507. Pierce, R. L., 1970, Preliminary revaluation of late Miocene bio- stratigraphy of California [abs]: Am. Assoc. Petroleum Geologists Bull., v. 54, no. 3, p. 559. Ray, C. E., 1960, Trichecodon huxleyi (Mammalia: Odobenidae) in the Pleistocene of southeastern United States: Harvard Univ. Mus. Comp. Zoology Bull., v. 122, no. 3, p. 129—142. Reinhart, P. W., 1943, Mesozoic and Cenozoic Arcidae from the Pacific slope of North America: Geol. Soc. America Spec. Paper 47, ix + 117 p. Repenning, C. A., 1972, Underwater hearing in seals: functional morphology, in Harrison, R. J ., ed., Functional anatomy of marine mammals: London, Academic Press, Chap. 6, p. 307- 331. 1976, Adaptive evolution of sea lions and walruses, Syst. Zool., v. 25, p. 375—390. Repenning, C. A., Peterson, R. S., and Hubbs, C. L., 1971, Contri- butions to the systematics of the southern fur seals, with particular reference to the Juan Fernandez and Guadalupe species, in Burt, W. H., ed., Antarctic pinnipedia: Antarctic Research Ser., v. 18, p. 1—34. Repenning, C. A., and Ray, C. E., 1977, The origin ofthe Hawaiian monk seal: Proc. Biological Soc. Washington, v. 89, no. 58, p. 667—688. 7 Repenning, C- A., Ray, C. E., and Grigorescu, D., 1977, Pinniped Biography, in A. J. Boucot and J. Gray (eds.), Historical biogeography, plate tectonics, and the changing environ- ment: 37th Biological Colloquium of the Oregon State University, Corvallis, University of Oregon Press (in press). Robinette, H. R., and Stains, H. J ., 1970, Comparative study of the calcanea of the Pinnipedia: Jour. Mammalogy, v. 51, p. 527—541. Rutten, L. M. R., 1907, On fossil trichechids from Zealand and Belgium: Koninkl. Akad. van Wetenschappen, Amsterdam, v. 10, p. 2—14. Scheffer, V. B., 1958, Seals, sea lions, and walruses: Stanford, Calif, Stanford Univ. Press, 179 p. Schreber, J. C. D. von, 1776, Die saugethiere in Abbildungen nach der Natur, mit Beschreibungen: Leipzig, pt. 3, sec 12. (Date according to Allen, 1880, p. 194.) Shikama, Tokio, 1953, On a fossil remain of a sea-lion, Eumetopias (?) kishidai n. sp. from Japan: Yokohama Natl. Univ. Sci. Repts., sec. II, no. 2, p. 10—14. Shikama, Tokio, and Takayasu, Taisuke, 1971, Fossil mammals from the Shibikawa Formation in Oga Peninsula, Akita Prefecture: Yokohama Natl. Univ. Sci. Repts., sec. II, no. 18, p. 43—47. Shotwell, J. A., 1951, A fossil sea-lion from Fossil Point, Oregon [abs.]: Oregon Acad. Sci. Proc., v. 2, p. 97. Simpson, G. G., 1945, The principles of classification and a classification of mammals: Am. Mus. Nat. History Bull., v. 85, 16 + 350 p. 1961, Principles of animal taxonomy: New York, Columbia Univ. Press, xii + 247 p. Sivertsen, E., 1954, A survey of the eared seals (family Otariidae) with remarks on the antarctic seals collected by M/ K Norvegia in 1928—1929: Det Norske Videnskaps-Akademi i Oslo. Scientific results of the Norwegian Antarctic Expedition 1927—1928 et seq., instituted and financed by Consul Lars Christensen, v. 36, p. 1—76. Smirnov, N. A., 1908, [Review of Russian pinnipeds]: Acad. Sci. St. Petersburg Mem., Ser. 8 (Phys.-Mat.), v. 23, no. 4, p. 1—75 (in Russian). 1929, [Key to pinnipeds of Europe and North Asia]: Izves- tiya otdela prikladnoi ikhtiologii i nauchnykh oromyshovykh issledovanii, v. 9, no. 3, 231 p. (in Russian). Snavely, P. D., Jr., Rau, W. W., and Wagner, H. C., 1964, Miocene stratigraphy of the Yaquina Bay area, Newport, Oregon: Ore Bin, v. 26, No. 8, p. 133—151. Stirling, Ian, and Warneke, R. M., 1971, Implications of a com- parison of the airborne vocalizations and some aspects of the behavior of the two Australian fur seals, Arctocephalus spp., on the evolution and present taxonomy of the genus: Australian Jour. Zoology, v. 19, p. 227—241. Teilhard de Chardin, P., and Leroy, P., 1942, Chinese fossil mammals: Inst. de Ge/o-Biologie, Pekin, no. 8, 142 p. Thunberg, C. P., 1811, Mammalia capensia recensita et illustrata: Acad. Imp. Sci. St. Petersbourg Mem., v. 3, p. 299—323. True, F. W., 1905, Diagnosis of a new genus and species of fossil sea lion from the Miocene of Oregon: Smithsonian Misc. Colln., v. 48, no. 1577, p. 47—49. 1909, A further account of the fossil sea lion, Pontolis magnus, from the Miocene of Oregon, in Dahl, W. H., The Miocene of Astoria and Coos Bay, Oregon: US. Geol. Survey Prof. Paper 59, p. 143—148. Turner, D. L., 1970, Potassium-argon dating of Pacific Coast Miocene foraminiferal stages, in Bandy, O. L., ed., Radio- metric dating and paleontologic zonation: Geol. Soc. America Spec. Paper 124, p. 91—129. Van Beneden, P. J., 1877, Description des ossements fossils des environs d’Anvers. Pt. I. Pinnipedes ou Amphithériens: Muse/e Royal Hist. Nat. Belgique Ann., Tome I, 4 + 88 p. Van Couvering, J. A., 1972, Radiometric calibration of the European Neogene, in Bishop, W. W., and Miller, J., eds., Calibration of hominoid evolution: Edinburgh, Scottish Academic Press, p. 247—271. Van Couvering, J. A., and Miller, J. A., 1971, Late Miocene marine and nonmarine time scale in Europe: Nature, v. 230, no. 5296, p. 559—563. Van Deinse, A. B., 1964, De fossiele en recente walrussen van Nederland: Leiden, Zool. Mededelingen, v. 39, p. 187—205. von Boetticher, H., 1934, Die geographische Verbreitung der Robben (Pinnipedia): Z. Séugetierk., v. 9, p. 359—368. Weaver, C. E., (Chm.), and Western Cenozoic Subcommittee, 1944, Correlation of the marine Cenozoic formations of western North America: Geol. Soc. America Bull., v. 55, p. 569—598. Weston, R. J., Repenning, C. A., and Fleming, C. A., 1973, Modern age of supposed Pliocene seal, Arctocephalus caninus Berry 86 OTARIOID SEALS OF THE NEOGENE (=Phocarctos hookeri Gray), from New Zealand: New Zealand Jour. Sci., v. 16, no. 3, p. 591—598. Winterer, E. L., and Durham, D. L., 1962, Geology of the south- eastern Ventura basin, Los Angeles County, California: U.S. Geol. Survey Prof. Paper 334—H, p. 275—366. Wood, H. E., and others, 1941, Nomenclature and correlation ofthe North American continental Tertiary: Geol. Soc. America Bull., v. 52, p. 1—48. Woodring, W. P., and Bramlette, M. N., 1950, Geology and paleon- tology of the Santa Maria district, California: U.S. Geol. Survey Prof. Paper 222, 185 p. Woodring, W. P., Stewart, R., and Richards, R. W., 1940, Geology of the Kettleman Hills oil field, California: U.S. Geol. Survey Prof. Paper 195, 170 p. Wornardt, W. W., Jr., 1967, Miocene and Pliocene marine diatoms from California: California Acad. Sci. 000. Paper 63, 108 p. INDEX [Italic page numbers indicate both major references and descriptions] Page A Abbreviations ...... 7 Acknowledgments .. 7 Acronyms .............. 7 Addicott, W. 0., cited ................................................... 67 Aivuk Aiuukus ................. Cedros Island dental formula distribution North Pacific Aiuukus cedrosensis . . 5, 6, 14, 31, 43, 44, 46, 47 48, 55, 67, 81; pls. 1, 2, 3, 4, 14 age ,,,,,,,,,,,,,,,,,,,,,,,,,,, 14—16, 69,81 Baja California Cedros 4 .......... Cedros Island description. discussion .. distribution ear ,,,,,,,,,,, humerus .. mandible metacarpals .. North Pacific radius ...................... Rancho e1 Refugio . ,,,,,, 21 scapholunar .......... 36, 46; pl. 2 skeleton, postcranial ............................................ 20 skull .................................. 19, 16, 56; fig. 1, pls. 1, 2 type .. 19 teeth 111111 18, 19, 45, 56; fig. 1, pls. 1, 2 trapezium 21, 37; pl. 4 trapezoid ........ 21, 51; pl. 4 type specimen 16, 20‘ unciform.. .. 37; p]. 4 Alachtherium 12, 20, 55. antuerpiensis . .. Europe antwerpiensis cretsii .............. Europe Allen, J. A., cited quoted . . alleni, Prorosmarus .. 5, 6, 13, 16, 19, 20, 21,22, 55,56, 57, 81; pls. 2, 5 70 Allison, E. C., quoted Allodesmidae 10 Allodesminae .. 74 Allodesmus . 5, 10, 11, 74, 76, 78, 79 ancestor 76 dental formula. 75 desmatophocid .. 7’7 diagnosis ............ .. 75 Japan 79, 81 range. 75 scapula. 5 species, included .. type species courseru' Cali . 6, 75, 76, 79; pl. 9 78 Japan .. 78 Luisian deposits, Los Angeles County, Calif. . 75, 76 . 75, 76 5 75 40 skull, fossil, Japanese teeth, cheek ,,,,,,, type specimen kelloggi. Page Allodesmus—Continued kernensis 6, 54, 55, 74, 75, 76, 79 Calif,.... 76 femur... 75 0s penis . 75 San Joaquin Valley.. 78 teeth, cheek .. Temblor Sea packardi ,,,,, Calif femur.. . type, female 75 type specimen, Menlo Park, Calif. 75 sinanoensis... 76 Japan ,,,,, . 76, 78 6, 71, 75, 79, 81 . 6, 44, 56, 64 Almejas Formation age .................. 14, 16, 61, 69, 70 avifauna.... .. ..... 15, 70 Cedros Island, age. . 14, 16, 69 Baja California, Mexico. . 14, 44, 46, 47, 56, 61, 69, 7O cetacean fauna ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 15, 69 fur seals .. . 15, 16 fur seal, primitive marine invertebrate fauna . .. molluscan fauna. odobenid fossil. pinniped fossils stratigraphy ,,,,, Ameghino, F, cited Anadara obispoana ,,,,,,, Antarctic Convergence. antarcticus, Arctocephalus . antverpiensis, Alachtherium Trichecus ............... (Odobenus) antwerpiensis, Alachtherium. . A,‘ ’ ,, 79 Archaeohippus ..... Arctocephalinae . Arctocephaline aff. Callorhinus Arctocephalini .. Arctocephalus .. carpals dental formula. diagnosis ........ living, radius . species. louse fauna os penis skulls... species, included . living... type species antarcticus austmlis ,,,,,, Atlantic coast Pacific coast .. skull .............. South America. caninus.... doriferus.. fischeri 57 forsteri 57, 59,62, 72 Australia 57, 72 Arctocephalus forsteri—Continued endocranial casts ................ humerus, measurements ...... limbs, proportions New Zealand ,. skuH ............. measurements galapagoensis ............ . 57, 59, 72; pl. 19 Galapagos Islands . 57, 72 gazelle .......................... 57, 72 Antarctic Convergence... .. 57, 72 philippii ...................................... . 57, 73 J‘uan Fernandez Islands 57, 73 living, teeth, cheek .............. 64 pusillus .......................... 20, 57, 62, 72, pl. 21 Africa, southern 57 Australia, southeastern ..... 57, 61, 72 humerus, measurements limbs, proportions ........... living, mandible os penis ........ South Africa skull, female male ............ measurements South Africa. . teeth, cheek ...... doriferus..... Australia humerus. living... townsendi ..... California . mandible Mexico ....... 66 64; pl. 23 64 Pacific coast, North America 73 Point Conception, Calif 73 San Miguel Island, Calif.. . 57 teeth, cheek 64 tropicalis ............... 57, 59, 72 Antarctic Convergence . 57, 72 williamsi .. . 72, 73 Arctoideaw .. 1 Army Doblegado, tuff be 49 Astoria Formation. ........... . 6 Oregon 74 atwoodi, Ostrea 49 australis, Arctocephalus . 57, 66, 72 Avifauna 70 B Bacula ............................................................ 10, 65, 71, 78 Baja California ........ Baja California Sur Bakersfield, Calif., Miocene deposits. Balaenoptera sp .. 14, 43, 48,55, 60 Balaenopterid.... .. 15 Barnes, L. G., cited 7, 11, 22, 34, 69, 74, 76 quoted .............. 75 Barron, John A., cited .. Bean Creek, Scotts Valley, Calif .. Bears ................ Bering Island Berkeley Hills, Calif .. Birds, fossil .................... byronia, Otariu 88 Page C Calabrian deposits, Italy ................................................ 5 California, central, odobenids inland sea ............................. 49, 78, 81 Miocene ........... 49 californianus, Zalophus japonicus, Zalophus Callorhinus AAAAAA astragalus braincase ...... brain casts .. characteristics dental formula diagnosis ............. endocranial casts . 72 10,57, 67, 71, 73, 77, 81 66 living, genus male ........ mandible . louse fauna ,. origin. as penis species, included skulls ,,,,,,,,,, type species. ursinus California Japan ,,,,,,,,,,,,,,,,,,,,,, limbs, proportions .. male, adult ,,,,,,,,,,,, mandible ,,,,,,,,,,,,,,,,,,,, Pacific coast, North Seward Peninsula, Alaska ........ Canada del Rincon en el Rio land grant .. caninus, Arctocephalus Canoidea ................. Cape Blanco, Oreg 57 20, 57, 68, 73; pl. 21 Capitola State Beach, Calif 67, 68, 70 Careaga Formation, Calif... 15 invertebrate fauna .. 16 Carnivores, fissiped ..... 1, 19 land, patella 39 phalanges 38 terrestrial .............................................. 32 Castlecliffian deposits, New Zealand 72 Cedros Island, Baja, Calif . 7, 14, 21, 44, 46, 47, 56, 64, 81 cedrosensis, Aivukus ........................ 5, 6, I4, 31, 43, 44, 46, 47, 48, 55. 67, 81; pls. 1, 2, 3, 4, 14 Celite Co. quarries, diatomite unit ............................ 58 Celite Co. Quarry No. 9, Lompoc, Calif 58 Celite Co. Quarry No. 15, Lompoc, Calif.. 58 Celite Co. Quarry No. 38, Calif. 24 Central American Seaway .. closure . Central Valley, Cali inland sea ,,,,,,,,,,, . ,. . Cetacean fauna, Almejas Formation Drakes Bay Formation ................... Purisima Formation .. Cetaceans .......................... fossils ,,,,,,,,,,,,,,,,,,,,, .. 47 Santa Cruz, seacliff 44 cinerea, Neophoca .. . ,,,,,,,,,,,,,,,,,,,, 20, 72; pls. 1, 21, 23 Clarendonian land mammal fauna 24 Clinocardium .. 71 colmaensis, Neptunea 15, 68, 69 Comanche Creek, Kern County, Calif ,,,,, 24, 26 Comanche Point local fauna ,,,,,,,,,,,,,,, . ,,,,,,,,,,,,,,,,,, 24 Coos Conglomerate Member, Empire Formation .. 42 courseni, Allodesmus ...... 6, 75, 76, 79; pl. 9 Cowell Beach ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, . 44 Coyote Mountains, Imperial County, Calif 53 12, 21, 55 45 cretsii, Alachtherium.. Cystophora ,,,,,,,,,,,,,,,,, INDEX Page D Delphinapterine .............................................................. 15 Desmatophoca .. 10, 11, 74, 75, 76, 79 age ............ 78 dental formula. 74 description, original. 74 diagnosis .................. .. 74 humerus ..................................................................... 74 til) 74 species, included . 74 type species ....... 74 oregonensis 6, 74 Alaska 76, 79 fauna] record Oregon type locality . sp .................................... Desmatophocid, ancestral .. 76 incertae sedis 76 Desmatophocidae 2, 74, 76, 77 Desmatophocids . 19, 40, 50, 74, 78 California .. 78, 80 characteristics . . 11, 74 unique ..... 74 classification 76 discussion ...... 76 ecologic restriction. 78 evolutionary stage ,,,,,,,, 78 geographic distribution .. . 78 Pacific Ocean, North ........................................ 76, 77 seals 74 teeth 71 Desmatophocinae ................ . . . 10, 74 Desmatophocine A of Barnes 6, 27, 74, 75, 76, 78, 80 California ................................................................. 76 Desmatophocine B of Barnes 6, 75, 78, 79 California ............................ 76 San Joaquin Valley.. 78 Temblor Sea ............. . 78 Desmatophocine C of Barnes 6, 75, 76, 79 California ,,,,,,,,, 76 Desmostylian fossils . . 25 7‘ ‘ylus 75 Diatom flora ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 24 Diatomite... 24, 26, 41, 58 quarry. 41 Dicalite quarry 24 Dog, anatomical terminology .. 5 doriferus, Arctocephalus . 61 Arctocephalus pusillus . 64; pl. 23 downsi, Imagotaria ,,,,,,,,,,,,,, . 6, 7, 15, 19,22,46, 60, 74, 75, 79; pls. 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 Drakes Bay Formation ,,,,,,,,,,,,,,,,,,,,,,,,,,, 6, 42, 44, 67 age ..................................... 69, 78, 81 basal glauconite bed 15, 47, 67, 69 cetacean fauna fur seals ........ fur seal, primitive mollusk ........................ Monterey Pine cones pinnipeds ........... Point Reyes, Cali Drakes Beach, Point Reyes, Calif ................. Point Reyes, National Seashore, Calif . 15, 47, 81 44 Dusignathinae ..... . 9, 22, 38, 42, 43, 55, 78 characteristics. .. 48 evolution, parallel. . 22, 48 distribution ............................. 78, 81 Dusignathus. 8, 9, 10, 15, 43, 55, 56; pls. 5, 15, 16, 18 age ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 53, '78 dental formula. 43 diagnosis ....... .. 43 distribution 43, 81 North Pacific 78, 81 parallel specialization. 56 sexual dimorphism .................................................. 47 Page Dusignathus—Continued teeth, cheek. 45; pl. 18 type species. ..... 43, 78 santacruzensis ............ 6, 15, 43, 50, 51, 55, 67, 81 pls. 5, 15, 16, 18 associations . 48, 61 Baja California . California ...................... discussion, concluding . 48 fibula. 46 forelimb holotype.... . 43, 48 humerus.... 44, 48, 50, 52, 53; p]. 16 magnum... .......... 44, 46, 47; pl. 15 material, referred ........... referred, questionable metacarpals, measurements odobenid features radius ramus, mandibular scapholunar .......... skeleton, postcranial .. 43, 48, 52 35, 44, 45, 51, 52 specializations specimens, referred, discussion referred, dimensions ................ 47 referred, questionable 47 stratigraphic association. 46 trapezoid ..... 51 type locality type specimen discussion ulna ............... . 45, 50 unique features ................................................ 48 E Fchinnids Elk River Formation, Cape Blanco, Oreg Empire, Oreg ...... Empire Formation Coos Conglomerate Member Enaliarctidae .. Enaliarctids . characteristics evolution .......... fissipeds Neotherium.. Oligocene. otarioid pinnipeds Pacific Ocean, North undescribed.. Enaliarctinae ...... Enaliarctos ...... association diagnosis .. limbs ....... species, included type species mealsi ............ .. fauna] record .............. 2, 11, 76, 79 ........ 74,76 6, 12, 76, 77, 79 Miocene, early early .. 79 Pyramid Hill, Calif .................... 77 Epoch terminology, Atlantic Coast usage .. 5 European usage ..... West Coast usage. Erignathus 19 Eskimos, Bering Strait region .................................... 14 Etchegoin Formation .................. . 6, 46, 50, 51, 52, 53 age ........................... 54 fur seals, specimen . . 70, 81 Kettleman Hills. .. 49, 50, 54 Macoma zone ..... . 49, 52 Pseudocardium bed San Benito River area . . teeth 49 Upper Pseudocardium zone 49 Etchegoin — San Joaquin sea ..... 54 Etchegoin sea .................................................................. 54 Page 20, 28, 33, 41, 42, 43, 45, 65, 68, 73, 81 66 Eumetopias astragalus ,,,,,,,,,,,,,,,, characteristics, primitive endocranial casts living, radius teeth..... metacarpals . metatarsals ,,,,,, species, extinct. limbs, proportions living............... 71 teeth, canine. 33 kishidae ................. sinanoensis watasei 72 sp .. 71, 73, 77 California . Japan . Oregon Europe ,,,,,,,,,,,,,,, Evaporitic deposits, Italy 3 Evolution, parallel ............................................ 22, 32, 48 F Faunas ............................................................................... 77 fischeri, Arctocephalus. Foraminifera ,,,,,,,,,,,,,,, Delmontian stage Mohnian ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, Foraminifers, planktonic, zone N9.. . 5 farcei, Hipparion.. 25 Formation, unnamed. . 6 farsteri, Arctocephalus. . 57, 59, 62, 72 9, 10,20, 47, 57, 81 Fur seals Alaskan .. classification dispersal ,,,,,,,,,, distribution, circumpolar . endemic ...... Galapagos .. female .. history, summary ................................................... 73 living ,,,,,,,,,,,, 10, 59, 61, 62, 64, 66, 68, 70 Africa .. . Australia . diversification . Etchegoin Formation, specimen humerus .. radius . South America North Pacific ........ origin other .. polyphyletic .. 10' primitive ,,,,,,,,,,,,,,,,,,,, 15, 66, 81 Purisima Formation . .......... 67 San Diego Formation, specimens. South Africa ,,,,, South Atlantic .. 'scuulpaafic species, insular G,H 57, 59, 72; pl. 19 57, 72 galupagoensis, Arctocephalus .. .. gazella, Arctocephalus ...... Geologic time, terminology .. . Glaciation, ranges of seals .. ,,,,,, 82 Glauconite, age . 15,26, 44,47, 56 Glauconite bed... 15,26, 44,46, 67, 69 Glen Canyon Road, Santa Cruz, Calif Glen Canyon Road quarry, Calif ......... Granite Rock Co. quarry, Olympia, Calif . Great Lakes Carbon Co quarry Gressigrada INDEX Page Grizzly Peak Basalt.............. Gulf of California, ancient inland sea ............................ Hachimanyama area, Japan 75 Hagerman fauna, Idaho 54 Hallianasa . 73 uanderhoof 25 Halichoerus ..................... 45 healeyi, Patinapecten . 70 Hemicyonine ursids” 79 Hemphillian land mammals 53 Hendey, Q. B., cited .................. 7, 80 Hipparion . forcei mohauense Europe .......................... Hoffstetter, Robert, cited hookeri, Phacarctos .. Horse, tooth ......... tooth, Hemphillian Howard, H., quoted ............... . Howenegg fauna, Germany .. Human, anatomical terminology . Humboldt County, Calif, deposits ............................ 7] Humboldt State College, Geology Department 71 huxleyi, Odobenus . 12, 13, 55 Trichecus Hydrurga ....... I Illiger, J. C. W., cited ....................................................... 1 Imagotaria ........ 5, 9, 22, 43, 48, 55, 56, 60, 76, 78, 81 association astragalus ..... California ......... dental formula. diagnosis .......... distribution North Pacific range, temporal. Tejon Hills type species... downsi .............. 6, 7 15, 19, 22, 46, 60, 74, 75, 79; pls.4,,5,67,,,,89101112,,,131415 age... 24, 69 astragalus-m . 39; pl. 14 atlas ............. 41; pl. 7 < braincase... . ........ 31; pl. 10 calcaneum. 39, 42; pl. 15 California ..... 55 cuboid ..... 40, 42; pl. 9 cuneiform 36; pl. 13 cuneiform bones 40 diagnosis 24 discussion. .. 42 ear .. .......... 34; pl. 10 femur 38, 42; pl. 15 23, 24 ..................... 23; pl. 11 22, 24, 27, 32, 41, 43 35, 42, 50; pl. 12 fossil localities. front limb, male . holotype .. humerus .. dimensions 35 joint occurrence 42 magnum.. . 37, 42, 46, 47; pls. 14, 15 mandible .. 45; pls. 5, 7 female .. 23, 27, 28 juvenile ...................... 28 material, other localities 24 Santa Cruz .. 27 metacarpals . 37, 42,49, 51, 55; pls. 4, 11, 14 dimensions . 38, 52 metatarsals .. 40 navicular ..... 40 odobenid features 42 patella..................... 89 Page Imagotaria downsi—Continued phalanges .. .. 38; pl. 11 pisiform .. .. 36; pl. 11 radius .. 35, 42, 51; pls. 11, 12, 13 dimensions 36 male ......... 55 referred material. 22 age 24 males .. 52 Santa Cruz area . 24, 27 referred specimen, Sole a Canyon, Calif .................................................................. 27 referred specimens. scapholunar .. 36, 42, 46; pl. 13 scapula ........ sexual dimorphism. ............ 34,- pl. 8 .. 34 skull, dimensions .. 30 female .................... 23, 28, 33, 34, 41, 43,45; pls. 6, 7 male, juvenile .................... 23, 33, 41, 43, 58; pls. 8, 9 Santa Margarita Formation ............ 22, 40, 41 teeth .................. 22, 23, 28, 41; pls. 5, 6, 7, 8, 9 canine .. 33 cheek.... .. 32, 33, 34 upper, dimensions . 31 tibia .......................... 39 40, 41 Towsley Formation trapezium.... ......... 37; pl. 11 trapezoid 37,51; pl. 13 type locality .. ulna .................. 35, 42, 46, 50; pls. 11,12 dimensions ...... 36 USNM 13487 unciform.. vertebrae sp .............................................. 6, 22, 39, 40, 54; pl. 9 ImagoZaria-Pithanotaria pinniped fauna .............. 27, 49, 79 Imagotariinae ...... 9 Imperial Formation 6, 53 age ......................... 53, 79 Imperial Valley, Calif .. .. 79 imperialensis, Valenictus 6, 49, 52, 53, 55, 81; pl. 16 Individual II, Kellogg. 58 58 Individual III, Kellogg. . Instituto de Geologia, Univers1dad Nacwnal Autonoma de Mexico ......................................... 7, 61 International Geological Congress . Introduction ........... Invertebrate fauna Invertebrate fossils .................................. Invertebrate paleontologists, Miocene J Jacalitos Formation .................. Jacalitos invertebrate fauna .. Jacalitos Stage, megainvertebrate japonicus, Zalophus californianus Jewett Sand .......................................... Pyramid Hill Sand Member Johns Manville Corp... jubata, Eumetopias ..... K Kanomata-zawa formation 75 Kaseno, Y., cited ................... 71, 81 Kellogg, R., cited... . 39, 45, 50, 51, 58, 59 sketch map 44 specimen, type, description .. 44 kelloggi, Allodesmus . 40 kernensis, Allodesmus 6, 54, 55, 74, 75, 76, 79 Kettleman Hills, Calif. 49, 54, 70 Middle Dome . North Dome 90 Page Kilmer, F. H., quoted,,................... 14 kimitsensis, Zalophus 72,73 kishidae, Eumetopias 72 Knapp, Jane, cited ..... 44 Kokolik River, Alaska 82 koninckii, Odobenus . , 55 Trichecodon ............................................... 12, 20, 21 L Land bridge ............................................................ 57 Land mammal age boundary, Blancan-Hemphillian 3 North American ,,,,, 3 Langhian Stage, Europe , 5 Law of Priority ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 12 Leon 60 Linnaeus ................ 12 lobatus, Zalophus 72 Lomita, Calif 43 Lompoc, Santa Barbara County, Calif 24 San Bernardino County, Calif ,,,,, 58 Louse fauna .. 10 endemic .. 58, 71, 73 history, fossil . 58 Luisian Stage ........... 54 microinvertebrate 79 M macnallyae, Thalassoleon .................. 6, 60, 61, 66, 78, 81; pls. 23, 24 magnus, Pontolis ,,,,,,,,,,,,,, 6, 24, 39, 42, 43, 54, 55, 81; pls. 10, l4, 15, 18 Malaspina District, Alaska," 81 Mammalian age, Barstovian Clarendonian" Hemphillian, North America Mammalian fauna, Blancan Mammals, land, Alaska ..... land, Arikareean age Barstovian age Blancan ,,,,,,,,,,,,,, Clarendonian age ,. fauna ,,,,,,,,,,,,,,,,,,,,,,, Hemphillian ,,,,,, , Irvingtonian age Pleistocene marine Mammoth. Mammuthus .. Man, ranges of seals. Margaritan invertebrate fauna Margaritan Stage, Corey ,,,,,,,,,, megainvertebrate ,,,,,,,,,,,,,,,,,,, Marine faunas, Soledad Mountain Marine fossils ...... .. Marine invertebrate section, West Coast 3 Marine invertebrates, West Coast, provincial epoch terminology, marinas, Ursus" . , Mascall- equivalent beds, Oregon mealsi, Enuliarctos ,,,,,,,,,,,,,,,,,,,,,,, 6, 12, 76, 77, 79 Mediterranean Neogene Committee ,, ,, 3, 5 Megainvertebrate fauna, Mascall-equivalent beds ....................................... 79 Megainvertebrate stage, Jacalitos , , 81 Margaritanwm", ,. ,, 81 Temblor ,,,,,,,, Melbourne harbor, Menlo Park, Calif Merced Formation, fossils, mollusks, marine San Francisco area . Messinian Stage ,,,,,,,,,,,,,,, , , 3 mexicanus, Thalassoleon .......................... 6, 60, 61, 68, 69, 72, 81; pls. 20, 21,22, 23 Microinvertebrate stage, Luisian ,,,,,,,,,,,,,,,,,,,,, , 79 Microinvertebrate fauna, Mascall-equivalent beds, 79 INDEX Middle Dome, Kettleman Hills, Calif ,,,,,,, Miocene, informal subdivisions, redefining Langhian Stage Mediterranean .. Messinian Stage ,. Serravallian Stage. Tortonian Stage ,,,,, Miocene Epoch, duration, Pacific Coast Miocene inland sea, California Miocene-Pliocene Boundary ...... 49 typology Mirounga,,.,, mirum, Neotherium ................ 6, 40, 54, 55, 56, 77, 79; pls. 9, 11 Mitchell, E. D., Jr., cited ........ 2, 7, 24, 34, 41, 43, 45, 53, 68, 71, 74, 75, 76, 82 Mitchell, seals, classification ....................................... 2 mohavense, Hipparion ,,,,,,,,,,,,,, , 25 Mohnian- Delmontian boundary 58 Mollusks ....... 14, 15, 16, 25, 67, 68, 69 marine . ............................... 49, 67 Tertiary .............. 75 Vaqueros Stage 79 Monterey Formation , 25 diatomite facies .. 24 Monterey Pine, cones 15 Monterey Shale ................................. 6 Valmonte Diatomite Member 43 Monterey-Sisquoc formational boundary,.,,.........,,. 24 Moonstone Beach locality, Humboldt County, Calif .......... 71 Moraga Formation, , 25 Moss Beach, San Mateo County, Calif , 44, 48, 50 Moss Landing Marine Laboratories ........................ 23 N Nagano Prefecture 777777777 75 Narrow Cape Formation .. 6 Kodiak Island ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 74 National Museum of Natural History, 7, 75, 77 Marine Mammalogy ........ 44, 74, 79 Neogene epochal terminology ,, .......... 5 Neophoca ,,,,,,,,,,,, 20, 62, 66, 73 metatarsals 66 Ohope Skull ,, 72 species, extinct. 72 cinerea 20, 72; pls. 1, 21, 23 Australia ,,,,,,,,,,,,,, 73 limbs, proportions , 65 sp 73 New Zealand 73 Neotherium 54, 55, 76, 79 calcaneum ., 54 cuboid 54 discussion .. 54 enaliarctid ,_ 79 limbs" 55 referred material 54 type species . 54 mirum ...... 6, 40, 54, 55, 56, 77, 79; pls. 9, 11 California 55 lectotype..,,, , ., 54 Neptunea colmaensts 15,68, 69 Neuerita zone, San Joaquin Formation ,,,,,,,,,,,,,,,,, 49, 53, 54 North Island, New Zealand ,, 72 O Odobaenidae,,,,.,,.,........,,.,....., ,, ,,,, 8 Odobaeninae 8 Odobenidae.... 2, 8, 11, 12, 32, 43, 55, 74, 77 Odobenids 32, 33, 35, 36, 38, 40, 54, 76, 78 astragalus 54 distribution . 6 78 diversified , ..... 78 22, 45, 47, 48, 51, 52, 55, 56, 61 dusignathine Odobenids, dusignathine—Continued North Pacific ,,,,, enaliarctid ancestor evolutionary stage extinct ,,,,,,,,,,,,,,,,,,, fossil, geographic range stratigraphic range humerus ,,,,,,,,,,,,,,,,,,, North Pacific region odobenine .................. Tertiary, Pacific origin ......................... parallel specialization primitive ,,,,,,,,,,,,,,,,,,,,,,,, . 19, 80 radius .................................................................... 47, 49 teeth 71 Odobeninae ,,,,,,,, 2, 8, 12, 16, 19, 22, 38, 42, 43, 55, 78 Odobenine, distribution ,,,,,,,, 78 evolution, North Atlantic 56, 57, 78 North Pacific .................................................... 81 Odobenus ,,,,,,,,,,,,,,,, 8, 12, 13, 18, 20, 21, 30, 31, 32, 35, 42, 43, 49, 50, 52, 55, 56 Alaskan ...................................................................... 57 astragalus 40 braincase ,. 52 calcaneum 39 cuboid ........ 40 cuneiform 36 femur ., 38 fibula. 46 humerus ,,,,,,,,,,,,,,,,,,,,,,,, 53, 82 living ,,,,,, 18, 19, 20, 21, 46, 56 fibula, , 46,52 magnum,,, , 37, 46 metacarpals 37, 38; pl. 4 metatarsals, 41 navicular ,,,,, 40 os penis 75 patella... 39 radius ,,,,,,,,,, . 35,36 scapholunar 36, 46 scapula ................ 34 teeth 45 tibia .............................................................................. 39 trapezium. 37 trapezoid , 51 ulna 12, 13, 55 55 55 honinckii , 55 Europe . .............................. 55 rosmarus ,,,,, 9, 12, 14, 22, 55, 82; pls. 1, 2, 4, 10, 14, 15 living ................................................................... 55 North Atlantic, 55 North Pacific Pleistocene (Odobenus), Trichecus antuerpiensis, Trichecus . Oligocene-Miocene boundary . Omma Formation, age ,,,,,,,,, Eumetopias sp . ,,,,,,,,,,,,,,,,,,,,, Ishikawa Prefecture, Japan ,, Oregon coast ........................... oregonensis, Desmatophoca Oriensarctos watasei, Japan Orionina vaughani ......................................... 14 uaughani ostracode assemblage zone,. 14 Ostracode, Orionina vaughani assemblage zone 14 Ostrea atwoodi ,, atwoodi, stratigraphic range Otaria ,,,,,,,,,,,,,, Argentina byronia". Atlantic shores, South America , 73 Argentina ................................... 73 Pacific shores, South America. 73 Page Otariid, incertae sedis VVVVVVVVVVVVVV 73 Otariid seals, fossils, history 73 Otariidae ...................... 2, 7, .9, 50, 57, 7 2, 74, 77, 79 Otariids . 19, 20, 31, 32, 36, 40, 43, 5 ancestor, enaliarctid bacula. California . dispersal ..... diversification ................ evolution, sea lion stage. 81 evolutionary stage fossil iiiiiiiiiii Japan . limb proportions new North America, west coast Pacific, South, western fossil record . 0. 6 7 , 60, 72, 78, 81 lice, endemic . 78 living ............... 33, 35, 40, 54, 62, 65, 66, 67, 68 astragalus . 66, 70 fibula... 65 genera. 77 limb proportions 65 louse fauna, endemic 71 metacarpals .. 60 metatarsals .. 66 scapula 34 tibia ,,,,, 65 magnum ................... mandibular ramus metatarsals .............. North Atlantic Pleistocene ...... postcranial elements primitive .................. Purisima Formation San Diego Formation .. scapholunar ................ South Pacific Ocean species, extinct teeth...........,.,... . teeth, Elk River Formation Otariina Otariinae Otariini Otarioid evolution, North America, Pacific Coast .................................................. 5, 77, 82 stages ........................ Otarioid families, living.. Otarioid fauna, age ...... Otarioid pinnipeds, hemicyonine ursid 76 Neogene record 77 Otarioid seals, desmatophocids... 74 history, late Pliocene 81 ranges ........................... 82 Otarioidea 1, 2, 7, 74 North Pacific .. . 8 origin 8 Otarioids, ancestor... 76 diversification 78 epochal assignments .. 5 humerus ................ living, phalanges .. North Pacific .. pinniped fauna .. rib Ouliphocinae .. P Pacific Basin, North 48 North, fauna] record ....... 77 Pacific Coast, Baja California. 60, 61 California ....... 60, 61 North America... 3, 71 Pacific Ocean, North, latitudinal ranges 77 INDEX Page pacifica, Pliopedia ................................ 6, 4.9, 55, 56, 81; pls. 4, 14, 16, 17, 24 packardi, Allodesmus llllllllllll 6, 75, 79 Paleoparadoxia ........................ .. 75 Palos Verdes Hills, Los Angeles County, Calif... 43 Paralephas protomammonteus zone ....................... 72 Paso Robles Formation . 6, 54 age ................................ 49 conglomerate member 49 invertebrate fauna . Patinopecten healeyi .. Pelukian (Sangamon) transgression . 82 philippii, Arctocephulus 57, 73 Phoca rosmams .............. 12 ursina .. Phocarcto haokeri limbs, proportlons . New Zealand ........... Phocid seals, monachine . Phocids.... Phocoenid .. Phocoidea .............................................................................. 1 origin 8 Phocoids, Atlantic area .................. 1 North Pacific 81 Pinnipedia .............. 1 Pinniped fauna, Woody ............................................... 77 Pinnipeds.... 10, 32, 46, 55, 73, 79, 80, 81, 82 fauna." . .. Alaska California . 81, 82 early Pleistocene. 81 Japan ..................... . 81, 82 late late Miocene and early Pliocene ........ 81 late Pleistocene ....... 82 Miocene, early early 79 early middle ............ 79 late early .............................. 79 late middle and early late .. 79 Neogene ..... 79 Oligocene .. 7.9 Oregon . 81, 82 Pacific Coast, North America .. Pliocene, late .. Point Reyes, Cali .. skull, North America fossils 2, 7, 69, 81 skuH ................. geographic range living ....................... evolution North Pacific postcranial elements 81 ranges ..................... 27 rates of evolution temporal correlation 2 Pithanotaria 2, 26, 27, 42,58, 60, 72, 73, 76, 78, 81 association . 81 dental formula. 58 diagnosis . 58 distribution 58 Individual II. 58 Individual III... 58 Miocene, late middle 79 species, included ........ 58 teeth 71 type species. 58 starri 6, 15, 24, 27, 41, 58, 70, 72, 75, 79; pl. 19 age .. 58, 69 braincase 59 California 72 characteristics 60 diagnosis 58 discussion .. 59 humerus 58, 59, 60 holotype . . . ...... 58, 60 limbs .......... 58 mandible .. 69 91 Page Pithanotaria starri—Continued mandibular ramus .......................................... 58 material, referred . metacarpal . scapula ......... skull ........... referred. type specimen .................. stratigraphic association .. teeth, cheek ............................ type description. type locality type specimen ............................. .. 58, 59, 60 sp 69 Plesippus-bearing rocks ............................................... 70 Pliocene Epoch, duration . Pacific Coast . ....... Pliocene faunas, marine Pliocene-Pleistocene Boundary Pliocene-Pleistocene Commission .. Pliohippus sp . Pliopedia . age .. . distribution .................................... inland sea, California, central diagnosis North Pacific . range......‘. ...... type species ................................................................ 49 pact/tea...“ 6, 49, 55, 56, 81; pls. 4, 14, 16, 17, 24 age ......... braincase. California ,. discussion .................. Etchegoin Formation fibula ................................ forelimbs holotype .. humerus ............................. Etchegoin Formation limb, rear, possible ......... material, referred metacarpal ........ measurements .. proportions right .............. phalanx radius .. skull ...... synonymy, possible trapezoid, right ....... type locality ...... type specimen .. ulna .................. Poederlian deposits Point Lay, Alaska .. Point Reyes, Calif‘ 4.9, 51 fauna, age Point Santa Cruz lighthouse .. Polyphyly Pontolis age discussion North Pacific Oregon .. . radius . type species... 42 magnus ......... 6, 24, 39, 42, 43, 54, 55, 81; pls. 10, 14, 15, 18 braincase 42 Oregon 55 ramus .......... 42 skull, type . 42 type locality .. . . 42 Port Orford Formation 71 Clinocardium-bearing pebble and cobble gravel . 71 Potassium argon, age .. 47 92 Page Prorasmarus ..................................... I3, 14, 21, 53, 55 alleni. 5, 6, 13, 16, 19, 20, 21, 55, 56,57, 81; pls. 2, 5 Atlantic coast, North America 111111111111111111 81 Virginia .. 55 Protophocoid 2 Protopinniped, North Pacific . 2 Protopinnipeds ............................................... 1, 2 Pseudocardium bed, Etchegoin Formation . 70 Purisima Formation .. age ............................ Capitola State Beach, Calif cetacean fauna . 6, 42, 43, 44, 48, 53 . 54, 67, 69, 70 ............... 67, 70 . 69 Dusignathus ,,,,,, 53 fur seal, primitive .. 15 fur seals ..................... 15, 16 glauconite, age. bed .................... molluscan fauna .. Moss Beach, San Mateo County, Cali odobenid ..................................................... Point Santa Cruz, Santa Cruz, Calif 67, 68 Santa Cruz, Calif 15, 42, 47, 54, 56, 67, 68, 69, 70 fibula ....................................... 52 Soquel Point, Santa Cruz, Calif . 15, 44, 56, 67 44, 46, 67 Pusa sibirica ....................................... 81; pl. 16 pusillus, Arctocephalus 20, 57, 62, 72; pl. 21 doriferus, Arctocephalus .. ,,,,,,,,,,,,,,,,, 64; pl. 23 Pyramid Hill, Kern County, Calif ............................ 77 Pyramid Hill deposits, Zemorrian-Saucesian Stages .. Pyramid Hill local fauna .......... Pyramid Hill Sand Member, Jewett Sand. Q,R Quarry 9, Santa Barbara County, Calif. 24 Radiometric dates, correlation ..... . 3 Rancho el Refugio locality, Cape region, Baja, Calif. Sur ...................................................... 16, 21 Ray, C. E., cited ,,,,, 7 References cited ,,,,, 82 Remington Kellogg Library, Marine Mammalogy, National Museum of Natural History. Reptile, anatomical terminology Rhinoceros, maxilla. Rock-stratigraphic units D 44 ms rosmarus, Odobenus ...................... 9, 12, 14, 22, 55, 82; pls. 1, 2, 4, 10, 14, 15 Phoca .......................................................................... 12 Round Mountain Silt Kelley Canyon... Kern County, Calif. Russell, D. E., cited ..... S Sabertooth breeding tooth ............................................ 42 Sacaco, Peru ....... Salada Formation Salas, G. P., cited . Salmon, giant San Andreas fault San Benito River ............................ San Diego, Calif, Reynard Way San Diego Formation age ..................... fossils, marine otariids ................ fur seals, specimens invertebrate fauna ...... San Francisco Peninsula ,,,,, San Joaquin-Etchegoin sea San Joaquin Formation age. .............................. Kettleman Hills, Calif Neverita zone Pecten zone.... . 57, 73, 81 INDEX Page San Joaquin Formation—Continued Pecten zone—Continued mammalian fauna . tuff bed... San Joaquin Valley, inland sea San Vicente Creek ......................... Santa Clara Formation, mammalian locality, Blancan ............................. Santa Clara County . Santa Cruz, Calif .................................................. 7, 48, 74 Santa Cruz, Seabright District, East Cliff Drive 44 West Cliff Drive ..... Santa Cruz Aggregate Co. quarry, Calif .. Santa Cruz City Museum .......................... Santa Cruz Mudstone .............................................. 24, 26 Santa Margarita, San Luis Obispo County, Calif Santa Margarita Formation conglomerate ................... gravel bed ......................... gray sandstone member horse teeth ................. lower part .......... Santa Cruz, Calif 22, 24, 25, 26, 27, 28, 32, 38, 39, 40, 41, 42, 58, 60, 74, 75, 81 S " Calif 58, 59 Tejon Hills, Kern County, Calif ........................ 41 upper part 6, 25, 37 Santa Maria area santacruzensis, Dusignathus. 6, 15, 43, 50, 51, 55, 67, 81; pls. 5, 15, 16, 18 Scaldisian deposits, Antwerp. 12, 13, 21 Scheffer, V. B., cited ............. .. 1 Schrader’s North Pacific Diatom Zone XI .. 24, 59 Sea lions .................... 2, 9, 10, 20, 35, 37, 42, 56, 57, 60 Ar ‘ ,_ "Lalus lineage 73 Argentina .. Australia California .. Capitola, Calif. 71 carpals .......................................................... 65 Castlecliffian deposits, New Zealand characteristics ............................................ .. 81 classification ............................................................ 72 Cuneiform 36 distribution, southern . .............. 82 early Pleistocene recor ....... 71 evolution ........................... 70, 72 evolutionary history . femur 71 fossil fossil record ..... fossil, Japan. fossil, Pacific, eastern Pacific coast, North America Pliocene record .......................... San Diego ..... history, summary humerus................. Japanese ....................... late Pleistocene record 72 living ............................ 22, 32, 62, 68, 71, 72, 73, 78 diversification . Pacific, North... South... radius . louse fauna modern ....... Ohope Skull . Omma Formation .................................................... 71 origin 70 Pleistocene .............................................................. 82 teeth 71 Seals, desmatophocids fossil, Miocene Epoch .. Pliocene Epoch fossils ......................... living .............................. monachine, Antarctic .. Page Seals, desmatophocids—Continued origin otarioid, epochal assignments. Neogene ................................ North Pacific Ocean phocids ....... walking Mitchell 8 classrfication .. wrigglers ................................. Seaside, Calif Sharktooth Hill Sharktooth Hill bone bed .. Sharktooth Hill locality, Kern County, Calif. 79 Shibikawa Formation, Oga Peninsula ......... 72 Shikama, Tokio, quoted .. . 75 sibirica, Pusa ........... 81; pl. 16 Simpson, G. G., cited ............... 1 sinanoensis, Allodesmus 76 Eumetopias .. 75 Siphonulia zone.. Sirenian genus... Sisquoc Formation diatomite facies.. Lompoc, Calif. Smilodonichthys.... Smirnov, N. A., cited... Smithsonian Institution. Soledad Mountain, marine faunas. Soquel Point, Santa Cruz, Calif ..... Soquel Point area, deposits Species, allopatric ................. Stanford University, Santa Clara County starri, Pithanotaria ................ 6, 15, 24, 27, 41, 58, 70, 72, 75, 79; pl. 19 Stenodelphine ........................... 15 Stratotypes, Miocene Series Pliocene Series ................... Suprageneric diagnoses .................................................. 7 S, ‘ Naturae 12 T Tapir, fossil ..... Taylor, R., cited Tejon Hills, Kern County, Calif . Temblor Stage, megainvertebrate . Terminology, anatomical .. other Thalassa Thalassoleon .. carpals .............. characteristics dental formula diagnosis ...... distribution .. etymology evolution .. Pacific Coast .. species, included type species .......... macnallyae.. 6,60, 61, 66, 78, 81; pls. 23, 24 age ...... 67, 68, 69, 81 California 72 Capitola State Beach . 68 diagnosis. discussion etymology glauconite holotype .......................................... 66, 67, 69, 70 material, referred, questionable 67, 68, 69 metatarsal........................................... .. Point Reyes specimen . Purisima Formation skeleton, male, adult skulls ...... juvenile. measurements teeth, cheek ......... Page Thalassoleon macnallyae—Continued tibia ............ type locality type specimen . 67, 68 ulna . mexicanus.... 6, 60, 61, 68, 69, 72, 81; pls. 20, 21, 22, 23 age ............................................................ 61, 69, 81 astragalus ........... 61, 65, 66, 69 auditory region . 62 Baja California ..... 72 Cedros Island ..... 64, 65, 66, 67, 69 characteristics 66 diagnosis ................. 61 discussion, summary 66 etymology evolution . fibula ......... holotype ............... humerus, female .. male... measurements .. limbs ........................ proportions males ..... weigh mandible ................ material, referred metatarsals. os penis patella... pelvis ............................ postcranial elements .. postcranial skeleton, discussion radius female male... scapula. skull, description female, adult male, adult .. 62 . 61, 62, 64, 70 61 juvenile ............ measurements .. stratigraphic association. 61 teeth, cheek . 62, 64, 66, 69 tibia ........... 65 type locality .. 61 type specimen .. 64 ulna ...................... . 61, 64, 68 vertebrae . , 61, 64, 67 sp ................... 68, 77 Tokyo, Pleistocene deposits .. .. 72 Tortuga Formation, Cedros Island, Baja Calif... 15 diatomaceous rocks ................................................ 15 townsendi, Arctocephalus ...................................... 20. 73 INDEX Page Towsley Formation ............................................ 39, 40, 54 lower part .. .............. 6, 24 Trichecodan ...... 12, 21, 47, 50 scapholunar .. ........ 36, 46 koninckii 12, 20, 21 Trichecus ........ antverpiensis huxleyi ........ (Odobenus) antuerplensts Trichophocinae ........ Tropic of Cancer ................. tropicalis, Arctocephalus. True, F. W., cited ............... quoted .......................................................................... 42 U,V University of Alaska, Department of Geology 7, 74 University of California, Museum of Paleontology ursina, Phoca,.,, ursinus, Callorhmus .. Ursus marinus .......... Utsunomiya shrine Valenictus,. age Central Valley, inland sea , discussion ................................. evolution ....................................... Gulf of California, inland sea humerus ......................................... type ......................... Imperial Valley, Calif North Pacific radius type species 53 imperialensis . 6, 49, 52, 53, 55, 81; pl. 16 California .. .. 55 humerus. 52, 53 Valmonte Diatomite Member, Monterey Shale... 43 vanderhoofi, Halianassa .............................................. 25 vaugham', Orionina ........................................................ 14 Vertebrate paleontologists, North American, chronology ........................................................................ 3 North American, Miocene ....... 3 Virginia ........................................ 55 W Waddell Creek, Santa Cruz County, Calif Walrus age calcaneum ..... classification fibula Walrus—Continued 93 Page groups. 55 history, evolutionary 2 fossil ................... 73 summary ............................................................ 55 living ,,,,,,,,,,,,,, 12, 19, 21, 27, 30, 31, 37, 42, 55, 56 phalanges .......................................................... 38 unciform magnum ......... modern tibia. North Atlantic basin. North Pacific basin phalanx ..................... radius ., tibia. watasei, Eumetapias. Oriensarctos ,,,,,,,,,,,,,,,, Whakatane, New Zealand.. Whale, sperm, pygmy" White—Seale locality... williamsi, Arctocephalus , Wilson, L. 13., collection, Yale University Woody, Kern River districts. Woody local fauna .................... X,Y Yakataga Formation, Malaspina District, Alaska ..................................................................... Yorktown, Va ............. Yorktown Formation Virginia . Zalophus ....... astragalus, brain casts carpals ....... endocranial casts living, teeth californianus Galapagos Islands Japan .................... limbs, proportions . North Pacific coast ., japonicus .................. kimitsensis Japan . lobatus Zanclian-Piacenzian contact Zanclian Stage, Italy Zayante Road locality . ......... 81 ........ 13 . 6, 14, 53 .21 20, 45, 62, 73 PLATES 1—24 [Contact photographs of the plates in this report are available, at cost, from US. Geological Survey Library, Federal Center, Denver, Colorado 80225] PLATE 1 [Abbreviafionsz OW—oval window. RW—round window. P—promontorium. A—apex. F—floccular fossa. V—vestibulocochlear canal. FN—facial canal. ICM—internal acoustic meatus] FIGURES 1,3,6,7,10.Aiuukus cedrosensis n. gen. and n. sp. Holotype, IGCU901. Late Miocene, lower part of the Almejas Formation, Cedros Island, Mexico. 1. Palatal view of skull, X 0.73. 3. Right metacarpal I, dorsal view, X0.50. 6. Right petrosum, ventral view, X1. 7. Right petrosum, medial view, X1. 10. Left third upper incisor, posterior view showing cementum coat and two wear facets, X1.0. 2,4,5. Odobenus rosmarus (Linnaeus). Recent. 2. Female right metacarpal I, dorsal view, X050; for comparison. 4. Male right petrosum, ventral View, X1; for comparison. 5. Same, medial view. 8,9. Neophoca cinerea (Peron). Recent. 8. Male right petrosum, ventral View, X1; for comparison. 9. Same, medial view. GEOLOGICAL SURVEY PROFESSIONAL PAPER 992 PLATEl AIVUKUS, ODOBENUS, NEOPHOCA domiwmfioo .8.“ ”5.0x £5: wEEmw Ema .w domEmQEoo you dex $393 255$ EME .w .Eoowm .Awswwnnfiv 3.5222 3:335 .w .w .omdx got» RES: 693 MO 350 £352 3:53:85 $3 .w> 5303.8? doflafihorm 55033? ”25035 .mvmm EZmD 65320: $in wax mnemonw .2353 magaEmEER .v .mvdx .30? 13mm. 5.823238,“ ”Ema .3 .mvdx .263 Hwfiioa .uwcflosnmom Eng .2 .mvdx $33 .85: .uwniogagm Ema . Amdx $59 EH35 £5: Em? ma muwfimauéoufi 13me . Amdx .253 2283 £in Emma we Ea: Hun—Em . 22> Hum—:08 .wEwm .wmdx $63 1335 £582 afldflwawfi $3 mo madofimwum . .mwdx 553 we oEm EME A .03me #523 mouvoO deans—“om 38E? 23 mo flan .833 .2583 33 .HomDOUH figs—om .am .: was dun .= 333238 35:33. .Sla .h .m .mIH mmmowfi maimtxm 53.5 1:3.— 23 we van 1812.— 05 no '38 Sauna.— .Su newton—4|: .3393 EEEIH .398 Euauln .503 oiaaolo .5355: .3.“ an ungoglm .Ecumonau you wanna:- 316Eu|< "Eouugvfinfi N gain MSZWMQQQ mob “YEMQMQ ME “MD beCV N whet—m mam Mmmm>mDm Aw.~v NvdX .mvaH MOD 65.88:: £2 wEEom vwflwmwfi .wlw .393 Sign/w .m 53> uoioamom .v is 382 .m .wvdx .Hom DOUH 69$ .2er5: Em? 29:8 35.8% .mlm .303 uaEEm .N .ng 1%qu A .mmdx 5me mob .Haeaomawfi 9mg $2 ENE wohwmmm .N ”H .03me £533 mahwmo .noflmfipom 3.382 2.: mo inn .833 65522 3nd .nm d 93 dam .: mwmfimmokwwo «.33:va .wIH $3505 :35 incuoonlom .3333 EBEUIEQ ”macaagmunnfi m Hakim PROFESSIONAL PAPER 992 PLATE 3 GEOLOGICAL SURVEY AIVUKUS PLATE 4 FIGURES 1—21. Aivukus cedrosensis n. gen. and n. sp. Referred male carpal elements. Late Miocene, lower part of the Almejas Formation, Cedros Island, Mexico. 1—4. Left scapholunar from UCR 15260 (reversed), X0.45. 1. Ulnar View; A-cuneiform articulation. 2. Proximal view. 3. Distal View; B-pocketed articulation for magnum. 4. Dorsal view. 5—7. Right trapezium from UCR 15241, X0.47. 5. Ulnar view. 6. Distal View. 7. Dorsal view. 8-11. Left trapezoid from UCR 15260 (reversed), X050. 8. Ulnar view. 9. Distal view. 10. Dorsal view. 11. Radial view. 12—15. Left unciform from UCR 15260 (reversed), X043. 12. Ulnar View. 13. Distal View. 14. Proximal view. 15. Radial view. 16-18. Left cuneiform from UCR 15260 (reversed), X050. 16. Ulnar View. 17. Proximal View. 18. Radial view; C-scapholunar articulation. 19—21. Left metacarpal II from UCR 15260 (reversed), X0.48. 19. Dorsal view. 20. Ulnar view. 21. Proximal view. 22—25. Right metacarpals III of Aivukus, Pliopedia, Imagotaria, and Odobenus, dorsal views, X 0.50. 22. Imagotaria downsi Mitchell. Referred male limb, USNM 23859. Early late Miocene, Santa Margarita Forma- tion, Santa Cruz, Calif. 23. Pliopedia pacifica Kellogg. Type, USNM 13627. Late Miocene, basal Paso Robles Formation, Santa Mar- garita, Calif. The platformed articulation for meta— carpal II has been roughly restored with clay. 24. Aiuukus cedrosensis n. gen. and n. sp. Referred speci- men, H.S.C. 309. Late Miocene, lower part of Almejas Formation, Cedros Island, Mexico. 25. Odobenus rosmarus (Linnaeus). Recent male. GEOLOGICAL SURVEY PROFESSIONAL PAPER 992 PLATE 4 AIVUKUS, PLIOPEDIA, IMAGOTARIA, ODOBENUS 398 me 3035 m> EBBxSV doswgom :BOH—iow 6:885 .mvmm EZmD 693303 .33m 38 303.5 .333 magufimcgogm .v «sz 550 35mm dofiafiuom «.5255 .2593: 82 3nd .HNKN AEOD dafiofiom .mwozmm mwmfiwmzxogzum mxfufimwmifl .m .330 550 gram doflmfiuoh Sing—«2 Sham 69532 32 3.3m .wmmmm EZmD .coEHowmm Ewfiwm Umfiwmwm .2vo:2 ESSSV ESEMEEN .m Gemiwafioo Mom ”unwoom dowmwd mszafitoxag SERRA“ A WEDGE .omdx ASS aflsnqumfi mo :63 it 3:58 «EN 3 1233 m Era/Vim Mb KVEMQ KO %& £st <2”:th .vfiwxv‘k 00 “xx: .MDEEOQVN w WSW—m Na¢ ”EASE Am>mDm A goiBmom .m .2me 22m: .3383 as mass Ebaom .Ew ..wv.ox .303 Rummy .minM .5 .wvdx $53 Han—mm. .MED .w .mvdx .263 133w .eEwm .m .wvdx h33> Sign?“ .mfimw .N .wvdx .269 Eufifl £5585: 35.8% A .mmwmm EZmD £5: Eat «38 80¢ EGQEBH .blmNJ 3:283 PE: SE35“ vohwmofi «:50 .550 Siam downs—pom gimmgz 32mm 65.532 83 3.8m dwaofiz Stacw ExEoMuEN NIH $5505 Haida» an: we mmoooa Ewuglmm ”noflmgwfihga NH MEL/wan VNMVLOU YEN 3 mh<1E Nam mun—.35 Am>MDm A 33% ”@8va 22m: .Eosnfifinmmm .: .263 3qu 68mm .m .33 563 383 «$va EZmD £33: a3 .v dune—:2.» 3mth 61mg .263 Ewaow 68mm .3 .253 3me 68mm .2 45¢?» .85: 685m .E .33 “32> RES “Eosgmé .2 45¢? Raga .ofiwm .NH .53 3% 58mm .: .mvdx ”33> .35: .Egofimnso .3 $33 Emuow 65mm .m .263 .85: .mem .w .26? 13me 68am .h .mvdx m33> 3:635 .umnflonmwom .m .mvdx $53 RED—E .miwam .m #53 1282 68mm .m .mvdx .263 aoiwanw 653 A .mmwmm EZmD £6: EMS ENE 89¢ £deme .wTwJTH .3558? AS: .8135». @25me @sz ~NED 35mm doflm—th SENMEE macaw 6:332 33 3.8m dwafinz 3:33» 3.293335 .5; mmmDQh T285 $5 .«o mmwuen ~m€§lm~m Ecfifi>op£n¢a mH Hakim VNKVKOUVSQ « 2 9:1: Naa mmmsm Am>¢Dm Am>m3w Am>MDm A Emuou rmzmo “n30 Scam dofimfiuom SENEME Scam 5:882 33 2.8» Amwovwfi EZmD 4.3.8338 umum 8.5.QO .m .3 .303 3:55 :Ewo .NEO Scam .qouafiuom waiwma—Z Spam 6:832 3S 3.38 ”@8on @203 dqmfiwafi 315%:«8 wwhomvm A .wmoHMoM than arguiufii .mnH 2:5”,va ma makfim m5 ‘Nvflrwmbbsbxv‘ .VNZVKQZVE&~K a 9:15 Nam yum—ASE Am>mDm A5 4 xi THALASSOLEON PLATE 23 FIGURES 1,3,5—9,15. Thalassoleon mexicanus n. gen. and n. sp. Late Miocene, lower part of the Almejas Formation, Cedros Island, Mexico. Holotype and referred specimens. . Astragalus, UCR 15249, tibial view, X050. . Same, calcaneal view. . Femur, UCR 15258, anterior View, X053. . Axis, type, IGCU 902, anterior view, X053. . Fibula, UCR 15258, anterior view, X050. . Tibia, UCR 15258, anterior View, X050. . Innominate, UCR 15258, lateral view, X050. 15. Metatarsal I, UCR 15246, dorsal view, X050. 2,4. Zalophus californianus (Lesson). 2. Astragalus, tibial view, X050, for comparison. 4. Astragalus, calcanear view, X050, for comparison. 10,18. Arctocephalus pusillus doriferus Wood Jones. Adult male. 10. Axis, anterior view, for comparison, X053. 18. Metatarsal I, dorsal View, for comparison, X050. 11—14,17. Thalassoleon macnallyae n. gen. and n. sp. Late Miocene, lower part of the Drakes Bay Formation of Galloway (1977), Point Reyes, Calif. Holotype specimen, UCMP 2535. 11. Mandibular fragment, medial view, X053. 12. Maxillary fragment, palatal view, X066. 13. Mandibular fragment, lateral view, X053. 14. Basicranium, ventral view, X 0.66. 17. Metatarsal I, dorsal view, X053. 16. Neophoca cinerea (Peron). Metatarsal I, dorsal view, for comparison, X050. ©0046:me GEOLOGICAL SURVEY PROFESSIONAL PAPER 992 PLATE 23 THALASSOLEON, ZALOPHUS, ARCTOCEPHAL US, NEOPHOCA .32.» Emhow .mmNodEE «53583 .w .95: EME 89¢ 3.8%? 380.3 559820 ”33> 3535 42:: $3 .m .253 12:35 £3ng :3 .v .ode dsz ”2:! ”88333“ :85 $8.8m Eowmfioem mo flan 33:8 65882 33 3w; .wmmbwfi EZmD .muofim 8%3 $32: .3 .30? 35:3» .m. is #832 25 .m 5.63 Egon A .omdx :Emo hflea—:50 doflmEgom MEEE—m .wnwoozm 3E 65va EZmD .33 593 wEEom BBQEEH .Qm d wmuANNBQSuE :owNommEuSm .mIH mamagm Totmamon 3.33830 magmlo .mfifloioamon m=£=mlx ”mcofiagefinfl vm HEm>MDm A Salinas . ‘— River '. Loose gray fine to _‘ medium silty_ san_d _ Very loose gra'y: ~ Very loose gray medium sand ‘ Very loose black to gray snlty . medium to coarse sand and ‘ ' l clayey sand- some gravel and organic matter: -_Slight|y compact gray IIghfly compact gray coarse-sand Ightly compact gra '- siltv veryfine sand Soft and loose gray Interbed; ed clay and silty very fine: 91' a... @699 /Soft gray-green organic cla dose giay silty very fine sand e: Stiffs and slightly compact gra alternate 1 ft: layers of silt clay, cla e silt,ands/iI}y/f)y sa d/ So/ft gray clay/ /. Soft ra cla fl 9 v v \_.. Stiff gray silty clay Soft gray clay ense to very dense_ gray medium d Slightly compact and stiff gray nterbedded sand and clay ' oft and loose to stiff and slightly - ompact clay.interbedded cla IIIintIy compact and stiff brow nterbedded silty clay and " Soft and loose to stIff and very‘ compact lnterbedded gray f’ d ' W .,\_/ery sIlty soft gray clay/ W very dense gray coars Very dense gray coarse ' sand and grav Very dense and dense Z-Vcoa'rse sand and gravel 7% ’ Sand Clay Interbedded sand and Clay 1 _ _ _ @ Water table in borings Approximate Water table Standard penetration- at time of drilling blows/ft (January 1963) upstream from the bridge damaged during the 1906 shock (fig. 10) (data from Division of Highways, 1964). 20 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES FIGURE 12.——Damage at Moss Landing caused by lateral spreading (100. 19). A. Damaged railroad bridge. Photograph previously published in Lawson and others (1908, pl. 135A) with caption "Lurching of ground toward Salinas River, to left, carried piles from beneath bridge timbers and caused bridge to collapse. Displacement 9 feet [2.7 m].” B Displaced building. Photograph previously published in Lawson and others (1908, pl. 134A) with caption "House, tree, and fence moved 12 feet [3.7 m] by lurching of ground toward Salinas River." (Photographs by A. C. Lawson, courtesy of The Bancroft Library, University of California, Berkeley). C. Ground ruptures in Moss Landing between Monterey Bay and old Salinas River. View eastward toward bluffs between Elkhorn and Moro Cojo sloughs. (Photo- graph courtesy of Monterey County Historical Society, Inc.. Salinas.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 21 FIGURE 13.—Ground failure scarp in a former community called Chinatown located near south bank of the Pajaro River across from Watsonville (loc. 25). (Photograph courtesy of Pajaro Valley Historical Association, Watsonville.) FIGURE l4.—Graben and damaged buildings caused by lateral spread near the Pajaro River at the foot of Marchant Street in Watsonville (loc. 25). (Photograph courtesy of Pajaro Valley Historical Association, Watsonville‘) 22 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES . Ii; ‘{ karma-”Wm ‘ v FIGURE 15.—Slump scarp and sand boil near Pajaro River (loc. 27). Similar photograph previously published in Lawson and others (1908, pl. 141B) with caption “Secondary crack, with drop of 7 feet, in alluvial flood-plain of Pajaro River.” (Photograph by Livennon, courtesy of The Bancroft Library, University of California, Berkeley.) FIGURE 17.—County road bridge over the Pajaro River near Chitten- den (loc. 30). Abutment displaced and fractured by lateral spread- ing of sediments toward the river channel. (Photograph by Liven- non, courtesy of The Bancroft Library, University of California,‘ Berkeley.) FIGURE 16.—Sand boil near Watsonville. Part of this photograph previously published in Lawson and others (1908, pl. 143B) with caption "Craterlets near Watsonville” (100. 27). (Photograph by J. C. Branner, courtesy of Stanford University Archives.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 23 SAN FRANCISCO BAY COUNTIES REGION SANTA CRUZ MOUNTAINS The Santa Cruz Mountains comprise the northern part of Santa Cruz County and the western parts of Santa Clara and San Mateo Counties (pl. 2 and table 6, locs. 45—135; figs. 18—25). Ground failures in the form of hillside landslides, flows, slumping of streambanks and ground cracks, both on the hills and in the creek bottoms, were numerous and widespread effects of the 1906 San Francisco earthquake in this area. The most common damage caused by these failures was ‘to high- way and railroad grades and to bridges. Rockfalls and dry sand flows along the coastal bluffs were particu- larly disruptive in this respect. For example, the ocean shore railway grade between Lake Merced and Mussel Rock (locs. 131, 132, 133, fig. 24), then under construc— tion, was almost entirely destroyed for a distance of 3 mi (5 km). Somewhat similar landslides were triggered in this same area during the 1957 Daly City earth- quake. Two catastrophic rockfalls occurred in the southern Santa Cruz Mountains. One on Hinkle Creek, a branch of Soquel Creek (10c. 50), buried the Loma Prieta Mill killing nine men, and the other on Deer Creek (loc. 74, fig. 19) became a rock avalanche that moved about 0.25 mi (0.4 km) down a narrow canyon destroying a shingle mill and killing one man. A number of flow failures were triggered in the Santa Cruz Mountains by the 1906 earthquake (locs. 110—114, 135; figs. 20—22, 25). These failures occurred in wet surficial materials, and many were associated with conspicuous amounts of surface water. In some instances these flows traveled considerable distances at relatively high velocities (for example loc. 135). In other instances the flows traveled rather slowly (for example loc. 113). Because the flows were generated in wet unconsoli- dated sediments, generally as a consequence of liquefaction, precipitation records for the Santa Cruz Mountains for the winter preceding the 1906 shock are relevant. Monthly and daily rainfall data for San Francisco, Santa Clara, and Santa Cruz (figs. 2—4) show that precipitation was unusually heavy during the winter, particularly in March, and that a 17-day rain-free period preceded the shock. Had the shock oc- curred when ground water conditions were signifi- cantly different, such as during the dry season of the year, the number of flow failures most likely would have been greatly reduced; on the other hand, had the shock occurred when the ground was even wetter, such as during or immediately after a major storm sequence, even more flows might have developed. SAN FRANCISCO BAY, SANTA CLARA VALLEY, AND EAST BAY HILLS AREAS Included in the San Francisco Bay counties region are Alameda County, the southern part of Contra Costa County, and the eastern parts of Santa Clara and San Mateo Counties (pl. 2 and table 6, locs. 136—182; figs. 26—29). (San Francisco City and County is consid- ered separately in the next segment of the report. Locs. 183—192 are discussed in the section on the “North Bay Counties Region”) The most common failures on the alluvial plain surrounding San Francisco Bay were lateral spreads, slumping of streambanks, ground set— tlement, and ground cracks (figs. 26—28). These failures were located chiefly along stream channels and margi— nal to San Francisco Bay. Lateral movements of 2—6 ft (0.6—1.8 m) were reported in the Colma Valley (loc. 136), on the San Bruno Marsh near Baden (loc. 137), along Coyote Creek north of San Jose (locs. 149, 150), at a point on the railway line between Niles and San Jose (loc. 162), and along Alameda Creek (loc. 170). These failures were mostly in sparsly populated areas; however, they caused considerable damage to railroad and highway grades and to bridges. Similar types of failures occurred along the lower reaches of Coyote‘ (locs. 149, 151, 153) and Alameda (loc. 170) Creeks dur— ing the earthquakes of 1868 and 1906. Figure 29 shows a geotechnical section across Coyote Creek at the State Highway 237 bridge. This bridge is very near the site of the old bridge (10c. 149) that was damaged by lateral spreading during the 1906 temblor. The section shows that a loose granular layer just below the water table extends at least 800 ft (240 m) westward from the creek. Cracks and sand boils were reported as far as 2,000 ft (600 m) west of the creek after the 1906 shock (loc. 149). Liquefaction in this or similar granular layers was most likely responsible for the sand boils and ground failures that occurred in that area. . Ground settlements of up to 2 ft (0.6 m) occurred around two well casings 4 mi (6.4 km) apart (locs. 147, 148) near the south end of San Francisco Bay. Flow from these and other wells in the vicinity (locs. 146, 149) was temporarily increased. Settlement of highway and railroad fills was also reported from several loca- tions near the margins of San Francisco Bay (locs. 138, 141; notes 11—18, table 2). No significant ground failures were reported on late Pleistocene and most Holocene alluvial fan deposits at points well removed from active stream channels. These older, more consolidated and denser materials (Youd and others, 1975) apparently resisted the shak- ing of the 1906 event without significant yielding. This inference is supported by several specific notations of lack of ground movement. For example, Derleth (in 24 Jordan, 1907, p. 188) stated that "San Jose’s water works, like that of Santa Rosa, was not injured; its sewers were left intact, showing that there was no un— equal displacement of the ground [loc. 157].” Other statements showing lack of ground failure are listed under locations 144, 156, and 164 in the descriptions of ground failures (table 6). The 1906 earthquake reports indicate that ground failures in the hills east of San Francisco Bay and the Santa Clara Valley were few in number and not very severe. Ground cracks were generated at several loca- tions north of Livermore (locs. 166, 167, 180). In addi- tion several shallow debris slides occurred along a road in the hills east of San Jose (loc. 160), and an acceler- ated movement of an already active slide occurred east of San Pablo (loc. 179). SAN FRANCISCO CITY AND COUNTY Because of the considerable amount of damage that occurred in San Francisco (pl. 3 and table 7, locs. 193— 248; figs. 30—54) and the greater development of that community compared with the surrounding region, more thorough 1906 postearthquake investigations and more quantitative reports were prepared than for most surrounding areas. This was generally true for earthquakes in 1865 and 1868 as well. Ground failures in San Francisco have been limited mainly to areas underlain by filled over marsh and bay mud deposits, filled-in ravines, loose sand deposits near Lake Merced, sand dunes, and steep slopes. Locations of failures are shown on plate 3 and for the downtown and waterfront sections on figure 30. General notes describing these failures are listed in table 4. Before examining the nature of seismically triggered ground failures in filled over marsh and bay mud de- posits, a brief review of the methods and materials used in constructing the fills is provided from Brown and others (1932, p. 29): Many difficulties were experienced in grading the streets in the swamp areas. Hittell [1878] states: “When streets were first made the weight of the sand pressed the peat down so that the water stood where the surface was dry before. Sometimes the sand broke through, carrying the peat down under it, leaving nothing but water or thin mud near the surface. More than once a contractor had put on enough sand to raise the street to official grade, and gave notice to the City Engineer to inspect the work, but in the lapse of a day between the notice and inspection, the sand had sunk down six or eight feet; and, when at last a permanent bottom had been reached the heavy sand had crowded over the light peat at the sides of the street and lifted it up eight or ten feet above its original level, in muddy ridges full of hideous cracks. Not only was the peat crowded up by the sand in this way, but it wasalso pushed sidewise, so that houses and fences built upon it were carried away from their original positions and tilted up at singular angles by the upheaval.” HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES The eventual construction of the streets in the lower section of the city caused the existence of many blocks surrounded on all four sides by high earth embankments, undrained, unsightly, with many stag- nant pools, and gradually filling with rubbish and trash of all kinds. Captain Folsom was the first to fill in one of the water lots, an expen— sive piece of work on California Street west of the present site of the Bank of California. Despite the cost, it was recognized as a valuable undertaking and others speedily followed. James Cunningham real- ized this was a job for a steam shovel (or “steam paddy”) and had one brought in by steamer. From 1852 to 1854 and from 1859 until 1873 a steam shovel was working constantly, loading into small railway cars and dumping in the bay, moving more than fifteen hundred cubic yards per day. As the shovel could work profitably only in sand and where there were comparatively long hauls, a large portion of the work was moved more economically by horse and cart even though a man with a horse and cart was paid $15.00 a day. There are no official figures on the amount of material moved. Hittell (1878) assumed a transfer of 21,000,000 cubic yards, while Bancroft (1882) estimated this to be 22,000,000 cubic yards. These figures are probably conservative. Three major ground failure zones have been iden— tified in the filled areas of San Francisco. These zones are designated here as (1) foot of Market zone, (2) south of Market zone, and (3) Mission Creek zone (fig. 30; pl. 3). These zones are generally congruent with three conspicuous zones mapped as intensity grade B after the 1906 earthquake (fig. 5). The most common ground failure types in these zones were lateral spreads and differential ground settlements. Because of the critical locations of these failures, they were responsible for considerable damage including the breaking of several major water pipelines that, in turn, left the city largely defenseless against the conflagration (fig. 49) that fol- lowed the 1906 earthquake. (See notes 14—19, table 4.) In the foot of Market zone, settlements as large as 4 ft (1.2 m) occurred during the 1906 earthquake (10c. 193). During the 1868 earthquake settlements of 1—2 ft (0.3 m—0.6 m) were reported (locs. 199—200), and some settlement also was reported during the 1865 earth- quake (loc. 202). Near the waterfront, permanent lat- eral movements as large as 2 ft (0.6 m) developed dur- ing the 1906 shock (loc. 193, fig. 31). The magnitude of these displacements decreased with distance from the waterfront. Some horizontal movements also appar- ently occurred in the foot of Market zone in 1868 (locs. 198, 200). Before a thorough investigation could be made, much evidence of 1906 ground failures and most of the consequent damage in the foot of Market zone were destroyed by the fires that devastated the area and by workers during cleanup operations. A geotechnical section beneath Market Street from the bay to Second Street (fig. 32) shows that a layer of fill composed mostly of loose fine sand with some rubble particles underlies the foot of Market zone. The fill is approximately 20 ft (6 m) thick near the waterfront and thins with distance inland, feathering out near DISTRIBUTION AND CHARACTER OF GROUND FAILURES 25 Sansome Street. In March-July 1964, the ground water level was at a depth of about 8 ft near the waterfront and increased in depth to about 18 ft at Sansome Street. The ground water level was probably at about the same depth in 1906. Beneath the fill are layers of soft silty clay (bay mud) with some sand lenses. These sediments are in turn underlain by more firm mate- rials. Liquefaction within the sand fill or an underlying sand layer is the only tenable explanation for the ground failures that developed in this area (Youd and Hoose, 1976). One evidence that liquefaction occurred in the foot of Market zone during the 1868 earthquake was the ejection of water from cracks and fissures (loc. 201). These ejections were probably a form of sand boils. No sand boils were reported in the foot of Market zone in 1906. However, some could have occurred only to be destroyed during the fire and ensuing cleanup operations or for some other reason were not reported. The south of Market zone lies in an area of filled-over marshland extending from Rincon Hill near Fourth and Brannan Streets westward approximately 1 mi (1.6 km) to near Market Street (locs. 205—212; figs. 33—36). Differential settlements in this zone were as large as 5 ft (1.5 m), and lateral displacements as large as 6 ft (1.8 m). The ground failures in that area were remarkable in two respects. First, the slope down the axis of the zone (from near Eighth and Mission Streets to Fourth and Brannan Streets) was only 0.8 percent (05°). Second, the lateral movement was not into a free face, but rather into Rincon Hill, a sandstone outcrop (Schlocker, 1974). At the head of this lateral spread, extensional features such as open cracks and pulled- apart rails, sidewalks, and curbs (loc. 205) were com- mon. Where the failure butted into Rincon Hill, com- pressional features such as buckled rails and curbs indicated that the lateral displacement was absorbed by compression. The US. Post Office Building at the corner of Seventh and Mission Streets (10c. 210, fig. 35) is astride the north margin of the south of Market zone. During the 1906 earthquake the ground in front of the Post Office settled 4 or 5 ft (1.2 or 1.5 m) and moved laterally about 2 ft (0.6 m), damaging the stone skirting around the base of the building (fig. 35A, C) and pulling the sidewalk apart at the construction joints (fig. 353). The building, which is apparently founded on a grillage foundation carried down to a firm granular layer (Gil- bert and others, 1907, p. 97—98), sustained little struc- tural damage and is still (1977) in use. A geotechnical profile across the south of Market zone at the latitude of the James Lick Skyway (fig. 36) shows a 5—8-ft (1.5—2.4-m)—thick layer of rubble fill in the ground failure area“. The rubble is underlain by a 7—11-ft (2.1—3.4-m)-thick layer of generally loose sand that is also probably artificial fill. Underlying the sand is a deep layer of soft peaty clay (bay mud). The water table in March 1952 was only a few feet deep across most of this section. Liquefaction within the loose sand layer is the only tenable explanation for the ground failures in this area (Youd and Hoose, 1976). The Mission Creek zone, which apparently contained several ground failure segments, extended up the sinu- ous former channel of Mission Creek from Old Mission Bay to near the intersection of 19th and Guerrero Streets (locs. 214—217; figs. 37—51). Settlements as large as 6 ft (1.8 m) and lateral movements at least as large as 6 ft (1.8 m) occurred at several locations within this zone. The average slope from 19th and Guerrero Streets to the freeway is 0.6 percent or 0.3°. Ground failures within this zone were responsible for some serious and spectacular damage, much of which is well recorded on photographs. Figure 37 is a schematic diagram of present streets in the lower part of the zone showing approximate loca- tions and directions from which the photographs in figures 38—41 were taken. These photographs show southwestward lateral spreading across both Ninth and Dore Streets between Bryant and Brannan Streets. The lateral spreading did not extend as far south as 10th Street, where there was but little evi- dence of ground failure (loc. 214). Figure 42 is a schematic diagram of several streets in the midsection of the Mission Creek zone showing approximate locations and directions in which photo- graphs in figures 43—47 were taken. These photographs show the magnitude, location, and character of the lateral-spreading ground failure in this area. Lateral movements of up to several feet developed down the axis of the zone and in some areas toward the axis, causing compressional deformation at some locations (figs. 43, 44, 47) and extension features at other loca- tions (figs. 45, 46). The damage at Valencia Street was the most catas- trophic of any that occurred in the Mission Creek zone (loc. 216; figs. 48A, 48B). Vertical and horizontal dis- placements here were both as large as 6 ft (1.8 In). Two main arterial water pipelines laid beneath the street were severed by the ground failure cutting off the water supply to a major part of the city which was soon to be in flames (fig. 49). Also, the catastrophic collapse of the four-story Valencia Street Hotel (loc. 216; fig. 48A) occurred at this location. This collapse may have been at least in part a consequence of the large ground movements that extended beneath the structure. Tens of people were killed in the collapse and the ensuing consumption of the building by fire. At the head of the Mission Creek zone, a substantial three-story brick building (fig. 50) was carried laterally 26 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES 6 ft (1.8 m) along with the curb, walk, and street in front of the building (loc. 217). There is no evidence indicating that lateral spreading continued up the old Mission Creek channel beyond this location. A geotechnical section across the Mission Creek zone at Mission Street (fig. 51) shows a layer of very loose fine sand fill underlain by layers of soft clayey sand and soft silty sand, which are in turn underlain by alternat- ing layers of firm clean clayey sand. In February 1964 the water table was as shallow as 5 ft (1.5 m) beneath the ground surface in that part of the section where the 1906 lateral-spreading failure took place. Liquefaction wtihin the loose sand fill or possibly the soft silty sand layer beneath the fill is the evident cause of ground failure in this zone (Youd and Hoose, 1976). Two ground failures in filled ravines, though small in size, produced rather spectacular results. One near the intersection of Vallejo Street and Van Ness Avenue (loc. 221, fig. 52) was characterized by differential set- tlements as great as 2 ft (0.6 m) and lateral displace- ments as great as 3 ft (0.9 In). The other, on Union Street between Pierce and Steiner Streets (10c. 222, fig. 53), had vertical and lateral displacements as great as 10 ft (3 m) and moved down a rather steep slope into an unfilled vacant lot. Several lateral spreads and rotational slumps, some of which were converted into flow failures, occurred around the shoreline of Lake Merced during both the 1906 and 1957 earthquakes (locs. 232, 242; fig. 54). These failures caused considerable damage to highway embankment and bridge structures around and across the lake. Liquefaction within natural and artifically placed sand surrounding and beneath the lake has been identified as the cause of many of the 1957 fail- ures (Bonilla, 1960) and is the probable cause of the 1906 failures. After an earthquake in 1852 fissures were found be- tween Lake Merced and the Pacific Ocean (loc. 244) through which the waters of Lake Merced were flowing into the sea. It is unclear whether these ruptures were generated by an earthquake or by other factors such as previous heavy rain storms that may have caused over- topping of or piping through the natural embankment between the lake and the ocean (Soulé and others, 1854). The sand dune section of western San Francisco was very sparsely developed in 1906. As a consequence, only fragmentary descriptions of ground failures in that section are available from the 1906 reports, and some of these descriptions are vaguely stated. The re- ports indicate that some areas of instability developed near the ocean (locs. 246, 247) and some cracking and lateral spreading occurred farther inland (locs. 230, 243). Liquefaction probably contributed to the de- velopment of these failures as evidenced by the erup- tion of sand boils at several locations (locs. 243, 247). Several hillside landslides occurred on the steeper slopes in San Francisco during the 1906 event (locs. 218, 226—229). These failures were responsible for structural damage to the Cyclorama, a building in Golden Gate Park (loc. 228), and a retaining wall at Laguna Honda reservoir (10c. 229). TABLE 4.—Excerpted notes containing general descriptions of ground failure and consequent pipeline breaks in San Francisco Note Reference Quotation 1 Lawson and others, 1908, It is evident that the intensity varies with the geology, or with the areal distribution of rocks and soils. p. 253. 2 Jordan, 1907, p. 238. 3 Hall, 1906, p. 32. 4 Carey, 1906, 297 5 Hyde, 1906b, p. 739. The areas that suffered most severely were those upon filled ground. Areas upon marshy ground showed destructive effects similar to artificial filled land. Next in intensity to areas of filled land are those upon incoherent sands. The damage in sandy areas was due partly to the shaking of sand like jelly and partly to settling and sliding. In San Francisco considerable tracts of “filled” land were shaken together and thus made to settle a few feet, and were at the same time slidden [sic] several feet toward the bay. Locating, by the street system, the points of surface movement*** the writer has demonstrated that every subsidence or movement of note caused by the recent earthquake is either just along the edge of or within an area which formerly was swamp, salt marsh, mud flat or bay estuary in character. It is interesting to see how the platting of the earthquake’s footmarks of this kind in our streets of today outlines the limits of the former soft spots of the formations on which San Francisco was built, and how the specially soft spots are again revealed by the greater movement now presented. [Continued under locs. 193, 205, and 216 in “Specific descriptions of ground failures in San Francisco City and County,” table 7.] Re-surveys in San Francisco show that property lines are considerably displaced where the land is loose or "made,” while on the hills little or no dislocation is observed. EARTHQUAKE EFFECTS—Within this filled district, the vibrations of the earthquake caused a general but irre ular settlement. The streets naturally followed the changes in elevation and a wave-like effect was profuced. Observations on Market, Mission, East and other streets frequently indicate an amplitude of wave height of two feet, while occasional places are found with greater differences in elevation. In the Mission district, and in certain other parts of San Francisco, as, for instance, along the courses of former tidal streams, large areas of filled land exist. In such cases, the effect produced by the earthquake was not generally in the form of waves, as already described, for the eastern portion of the city was basin-like, representing local settlement in the streets and adjacent areas, and in some cases, decided misalignment. The most important wave-like distortions were observed on lower Market and Mission Sts., and on East St. along the present water front.*** Where the ground was hilly and solid, it was not decidedly affected by the earthquake shock. DISTRIBUTION AND CHARACTER OF GROUND FAILURES 27 TABLE 4,—Excerpted notes containing general descriptions of ground failure and consequent pipeline breaks in San F rancisco—Continued Note Reference Quotation 6 10 11 12 13 14 15 16 17 18 19 Jordan, 1907, p. 123. Jordan, 1907, p. 98. Preliminary Report, 1906, p. 15. Leslie’s Weekly, 1 9066. Chicago Record-Herald, 1906b. Seattle Post-Intelligencer, 1906a. Derleth, 1906a, p. 503. Gilbert, Humphrey, Sewell, and Soulé, 1907, pl. 56. Duryea and others, 1907, p. 253. Schussler, 1906, p. 1. Schussler, 1906, p. 33. Gilbert, Humphrey, Sewell, and Soulé, 1907, p. 19. Derleth, 1906a, p. 503 Gilbert, Humphrey, Sewell, and Soulé, 1907, p. 56. STREET AND SURFACE DEFORMATIONS. Great distortion of the surface was best observed in the streets, and was found on the filled areas and in some places, on the sand dunes. The best localities for observation werez—(l), Market Street near the Ferry Building, Fig. 11 [loc. 195]; (2), the water front on both sides of the Ferry Buildin , Fig. 12 [loc. 193]; (3), the corner of Howard and Spear Streets, where the J. A. Folger Company’s bui ding was saved from fire [loc. 203]; (4), the corner of Mission and Seventh Streets, the location of the General Postoffice [loc. 210]; (5), Van Ness Avenue at Eddy Street [loc. 219]; (6), the north end of Van Ness Avenue and the streets on the hillside slope in that vicinity downward to the water front on the north [loc. 221]; (7), Howard Street, between 17th and 18th Streets, Fig. 13 [loc. 215]; (8), Valencia Street, between 18th and 19th Streets, where the Valencia Hotel was wrecked with great loss of life, and the main water pipes were sheared [loc. 216]; (9), Fourteenth Street, between Mission and Howard Streets [loc. 214]; and, (10), the water front near the Potrero District. I might enlarge this list, but these are the typical examples. Examples 5 and 6 represent surface distortions on sand dunes. The rest are examples of filled ground deformations. Upon the filled ground the surface was very generally thrown into billow-like waves, a type of disturbance which was best seen near the Ferry house. Upon the sand dunes the surface was shifted by sliding motion so that cracks and fissures appeared upon the streets at right angles to the direction of sliding. It was in these areas that the sewers and water pipes of the gridiron system were so generally crushed and broken. Even had the main conduits survived, water could not have reached the hydrants in the lower Mission District. The brick sewers were uniformly helpless to resist destruction in these regions and the cast iron water and gas pipes fared no better. The billow-like effects that appeared in the streets of San Francisco near the Ferry house are most excellent examples of deformations in soft, incoherent materials. The sliding and rolling effects observed on some of the sand dunes and especially along the hillside at the northern end of Van Ness Avenue may be cited as allied phenomena. [loc. 221] The most violent destruction of buildings, as everybody knows, was on the made ground. This ground seems to have behaved during the earthquake very much in the same way as jelly in a bowl, or as a semi-liquid material in a tank. The earth waves which pass through the highly elastic rocks swiftly with a small amplitude seem in this material to have been transformed into slow undulations of great amplitude which were excessively destructive. The filled in material and the swampy foundation upon which it rests behaved, in other words, as a mass superimposed upon the earth’s surface, rather than as a part of the elastic crust itself. In a less degree the same thing is true of the sand dune areas, where the ground was frequently deformed and fissured. . [Picture caption] ”Solid asphalt pavements were broken up like ice in a spring thaw.” . [Picture caption] “The Valencia Street Hotel, in which ruins 200 people were reported killed.” [loc. 216] . [Picture caption] "Curbstone lifted clear of the asphalt, which remained level.” . [Picture caption] "Porch of house demolished, the rest nearly uninjured.” . [Picture caption] "Toppling like houses of cards.” [Picture caption] SUNKEN FLATS SHOWING HUGE FISSURES MADE IN THE STREET*** Reproduced from the Los Angeles Times. [Picture caption] FISSURE IN SAN FRANCISCO STREET CAUSED BY EARTHQUAKE: “Flat Building on Left Sank One Story into the Ground.” In many parts of the city, side hills resting upon original sand dunes have been bodily moved and deformed by wave-like and sliding motions. Author’s summary of pertinent data on map. Principal earthquake breaks in streets are marked at the following places on this map. Valencia Street between 18th and 19th: Seventeenth Street between Mission and Capp: Howard street between 17th and 18th: Shotwell Street between 17th and 18th: Seventeenth Street between Folsom and Harrison: Harrison Street between 17th and 18th: Fifteenth Street between Harrison and York: Fourteenth Street between Mission and Harrison: Eleventh Street between Harrison and Bryant: and the corner of Seventh Street and Mission Street. [See also loc. 214] In San Francisco, all serious fractures of water mains, as a result of the earthquake, were due to lateral displacements, or subsidences of filled or soft ground across which, unfortunately, the main supply pipes *** passedl. ’fl‘he displacement laterally amounted in places to as much as 6 or 7 ft.; vertically, it amounted to severa eet. The most serious injury sustained by the [water] works were the ruptures caused by the earthquake, in hundreds of places, in our city main pipe distributing system, especially where the streets crossed filled ground and, particularly, where such filled ground covered former deep swamps, which swamps, during the earthquake, subsided, tearing off sewers as well as water and gas pipes. The city pipe distributing system was broken and in many instances torn and twisted off, especially in places where the ground, over which the streets had been constructed, had been poorly and loosely filled over old deep swamps and soft marshes. There were also a number of breaks in the streets that passed with deep loose fills over former ravines. The failure to control the fire by reason of the crippling of the water supply was not due to the failure of the system outside of the city, but to the breaks in the distributing mains within the city, which rendered unavailable about 80,000,000 gallons of water stored within the city limits. These breaks occurred (see the map, Pl. LVI) wherever the pipes passed through soft or made ground. No breaks occurred where the cast-iron pipe was laid in solid ground or rock. The damage by the earthquake which was the direct cause of the city’s great fire loss, occurred in the water system. The main conduits entering the city were greatly damaged, and the pipes running through soft materials were very generally destroyed. The importance of proper construction and distribution of the water mains in districts liable to earthquakes is demonstrated by the fact that the greatest damage in San Francisco, fully 85 per cent of the total, was by fire. The action of the earthquake in starting the fires which grew to a great conflagration seems insignificant compared to the breaking of the water mains, which left the city defenseless against the flames. 01$me 28 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 4.—Excerpted notes containing general descriptions of ground failure and consequent piepline breaks in San F rancisco—Continued Note Referen ce Quotation 20 21 Duryea and others, 1907, 22 23 24 25 26 27 28 Gilbert, Humphrey, Sewell, and Soulé, 1907, p. 66. p. 324 San Francisco County, 1908, p. 732. San Francisco Chronicle, 1906c. Huber, 1930, p. 270. Huber, 1930, p. 270. Holden, 1898, p. 78. Davidson, 1906, p. 180. Wood, Allen, and Heck, 1939, p. 6. From what was left of the ruins themselves, and from the testimony of competent observers, including engineer officers who were ordered into the business district of San Francisco immediately after the earthquake and before the fire had destroyed the evidences of earthquake damage, I think it is quite certain that the earthquake damage was extensive and severe. There were no available data on which to base an accurate estimate, but I formed a general impression that the damage done by the earthquake alone was at least as great as 10 per cent of the total damage by fire and earthquake combined. The damage from the earthquake, however, was localized in a remarkable degree. On the areas over the old tide flats, the walls of the old 3-st0ry and 4-story brick buildings were built on “rafts” made of layers of plank. These settled, and the space between them was arched up like the back of a turtle, sometimes as much as 6 ft. between walls 30 ft. apart. The Board of Public Works realized that a vast amount of emergency work confronted them, and for which work the funds available were wholly inadequate to defray the cost. They determined that conditions forced upon them the immediate performance of the following work regardless of the moneys set aside to their department:***. 10. The survey monuments that had been shifted by the movement of streets must be reset before lot surveys for building reconstruction can be furnished. The following sewers were found to be in bad condition, and upon which repairs were made: Fourteenth street between Valencia and Howard streets. Poe. 214 ***. Eighteenth street between Church and Valencia streets— loc. 216]***. Mission street between Eighteenth and Nineteenth streets. Hoe. 216]***. Powell street between North Point and Beach streets—***. Vallejo street adjacent to Van Ness avenue—175 feet of brick sewer and 125 feet of 18-inch pipe sewer were reconstructed. Hoe. 221]. Diamond street northerly from Chenery street—1,500 feet*** reconstructed. Seventh street between Folsom and Harrison streets—[loc. 212]***. Eighth street between Bryant and Brannan streets—[loo 214]***. Crossing of Bryant and Fourth streets—[loo 205]***. Crossing of Seventeenth and Howard streets—[loo 215]***. Hayes street between Broderick and Baker streets. Crossing of Seventeenth and Mission streets—[10c 214]***. Crossing of Central avenue and Turk street.*** Valencia street between Eighteenth and Nineteenth streets—130 feet of 5-foot sewer collapsed. Hoe. 216]. Eleventh street between Harrison and Bryant streets—70 feet broken. Union street between Steiner and Pierce streets—[loo 222]***. Four gangs of sewer cleaners worked on Fourteenth street between Folsom and Howard streets [loc. 214], and Valencia street between Eighteenth and Nineteenth streets. On Fourteenth street, between Valencia and Harrison streets; on Harrison street, between Twelfth and Thir- teenth streets; on Eleventh street, between Harrison and Bryant streets; on Ninth street, between Bryant and Brannan streets; on Dore street, between Bryant and Brannan streets; on Laguna street, between Greenwich and Lombard streets; on Shotwell street, between Seventeenth and Eighteenth streets; on Seven- teenth street, between Folsom and Harrison streets; on Howard street, between Seventeenth and Eighteenth streets. [1868 earthquake] Then as to the destruction of property. Not a single strongly and honestly built house on the solid land of the city has been materially injured, while very few houses of any kind on the solid land have been injured at all. The damage done was chiefly confined to the old and inferior structures, pretentious and imposing as some of them may have been, erected upon the flats formed by filling in about 200 acres of water lots along the northeast front of the city. This filling, owing to the deep substratum of mud, was essentially unsubstantial and unsafe, and even the iles driven through it have proved, as in the case of the Custom House, not to be a sufficient foundation or large brick or stone houses in a place liable to earthquakes. Beginning with the issue of October 23, 1868, the San Francisco Bulletin publishes a detailed itemized list of estimated damage to buildings, the individual buildings being listed by blocks. ***the principal damage occttilrred on land reclaimed from San Francisco Bay and which is known to afford unsatisfactory foundation con itions. 1868, October 21, San Francisco. The shock was longer and more severe than that of October 8, 1865. Several persons were killed by falling cornices. *** The surface of the earth visibly undulated. ***H0n. Horace Davis writes that the destruction in S. F. was greatest along the old beach-line of the city, beyond which the soil had been filled in. I was one of the committee of investigation of the 1868 earthquake, and it demonstrated that the course of greatest dislocation at the surface of the ground was on the line of contact between the "made” land or the alluvial soil with the rocky stratum. This is repeated in this [1906] earthquake***. 1865 October 8. *** IX at least. At San Francisco the greatest damage was to the less strongly constructed buildings on made land; structures on solid ground, or well constructed, suffered little damage; water mains and gas pipes broke in several places because of shifting ground, and a crevice opened in one street. Text continued on page 60 DISTRIBUTION AND CHARACTER OF GROUND FAILURES 29 FIGURE 18.—-Hillside landslide in redwood forest about 4 mi (6.4 km) above Alma (100. 55). The landslide has dammed Los Gatos Creek from the south. (Photograph by J. C. Branner, courtesy of Stanford University Archives.) 30 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES FIGURE ISL-Deer Creek landslide that advanced about 0.5 mi (0.8 km) downstream destroying the Hoffman Shingle Mill and taking one life (loci 74). Photograph previously published in Lawson and others (1906, pl. 124D) with caption "Deer Creek, Santa Cruz Mountains. Earth-avalanche from Grizzly Peak.” (Photograph by B. Bell, J. C. Branner collection, courtesy of Stanford University Archives.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 31 FIGURE 20.—Flow failure in hills southeast of Half Moon Bay (10c. 110). Note that the displaced sediments piled up in a ridge on the gently sloping terrain rather than spreading out. Similar photograph previously published in Lawson and others (1908, pl. 132A) with caption "Earth-flow in hills east of Half Moon Bay.” (J. C. Branner collection, courtesy of Stanford University Archives.) 32 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES FIGURE 21.——Flow failure east of Half Moon Bay (10c. 111). A. Cavity with 10-fi (3-m)—high walls from which sediments flowed. Note the man (circle) standing at left edge of cavity. Photograph previously published in Lawson and others (1908, pl. 132B) with caption ”Earth-flow in small valley near Half Moon Bay/’8. Panoramic continuation ofA showing ridge of deposited debris Note the pick leaning against debris in center of picture and man standing on rim of cavity behind debris at right. (Photograph by R. A. Anderson, J. C. Branner collection, courtesy of Stanford University Archives.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 33 FIGURE 22.—Flow failure down a hillside slope on the Nunez Ranch 4 mi (6.4 km) east of Half Moon Bay (loc. 114). Note the man (lower circle) standing at the base of the deposited debris and a second man (upper circle) standing on the lower edge of the cavity at the top of the slide. Additional incipient flow failures, with much smaller movements, also occurred on the convex hill to the left of the principal landslide. Photograph previously published in Lawson and others (1908, p1. 133B) with caption ”Earth-flow 4 miles east of Half Moon Bay.” (Photograph by R. A. Anderson, J. C. Branner collection, courtesy of Stanford Univeristy Archives.) 34 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES ‘ _ Q ,4 v,- figs FIGURE 23.—Rockfalls along coastal bluffs between Ocean Avenue, San Francisco, and Mussel Rock (10c. 132). (Photograph by H. 0. Wood, courtesy of The Bancroft Library, University of California, Berkeley.) FIGURE 24.—Landslide on steep hillside near Mussel Rock (10c. 132). These and similar landslides destroyed several miles of highway and railroad grades during the 1906 earthquake. (J. C. Branner collection, courtesy of Stanford University Archives.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 35 FIGURE 25.—Cavity of flow failure on the south side of San Bruno Mountain above Mt' Olivet Cemetery near Colma (1°C: 135)‘ (”19' FIGURE 26.——Rails of electric railway on marsh west of San Bruno tograph by A' 0' Lawson, courtesy Of The Bancroft Library, Un1- that were buckled by compression during 1906 earthquake (loc. versity Of California, Berkeley.) 141). Buckling was most likely a consequence of lateral spreading. Photograph previously published in Lawson and others (1908, pl. 97C) with caption “Roadbed and rails of electric railway on marsh west of San Bruno.” (Photograph by R. B. M., Branner Collection, courtesy of Stanford University Archives.) FIGURE 27,—Ground cracks in the vicinity of Coyote Creek west of Milpitas (10c. 149). A. Cracks caused by slumping and lateral spreading near the creek channel. B. Cracks in road several hundred feet west of Coyote Creek. (Photographs from J. C. Branner Collection, cour- tesy of Stanford University Archives.) 36 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES FIGURE 28.—Sand boils in field between Milpitas and Coyote Creek (loc. 149). Note bridge over Coyote Creek in center background. Photograph by J. C. Branner, previously published in Lawson and others (1908, pl. 143A) with caption "Craterlets in fields near Milpitas.” (Courtesy of The Bancroft Library, University of California, Berkeley.) West East 30 — Approximate 1906 ground failure zone — Ground line at Highway 237 Fine sand Sandy silt Sandy silt E Silty sand and Silty clay line gravel E Nov. 3, 1966 Silly sand with trace nl organic material E V 0:1. 7, 966 10 ELEVATION, IN FEET ABOVE MEAN SEA LEVEL Fina sand and gravel . Fm Sm EXPLANATION Silly clay Cnarsa sand and Navel Trace of organic material Dry unit weight-lbs/fl’ O . o 10° 200 :00 FEET Sand and gravel Standard penetration-blows/f! 1 _V_ Water table in borings at time of drilling a 25 so 75 mo METE as date' Show“ -10 | | I | I I I I I L I 0 100 200 300 400 500 600 700 00 900 1000 1100 1200 1300 FIGURE 29.—Geotechnical section across Coyote Creek at the Alviso-Milpitas road (State Highway 237) bridge, viewed north (loc. 149). A wooden bridge at this site was compressed about 3 ft (0.9 m) by spreading of the banks into the creek in 1906. The banks at this site were also "shaken together” during the 1868 earthquake (data from Division of Highways, 1967). DISTRIBUTION AND CHARACTER OF GROUND FAILURES 37 © .11 0 00 v @610 [FEET L_JL__J F—Uhfif'fi—T ' u©©© EMEWERS ‘ w @G Maw-12 atom: - 06»:me Widlfimy FIGURE 30.—Aerial photograph of commercial and shipping district of San Francisco showing locations of lateral-spreading and ground settlement failures that disrupted the city during the 1906 earthquake. Liquefaction within subsurface saturated granular layers is the evident cause of the failures. FIGURE 31.—Cracks and separations in roadway pavement near the San Francisco waterfront caused by lateral spreading in the footiof Market zone (10c. 193). Photograph previously published by Givens (1906) with caption “break and two—foot [0.6 m] sink in East Street near Ferry Building” and by Schussler (1906, p. 91) with caption "street on water front badly broken up.” 38 4O 20 ELEVATION, IN FEET ABOVE MEAN SEA LEVEL EMEAHCADERU HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES STEWAHT STHEET SPEAR STREET MA IN STREET Ferry Building ' 1906 ground failure zone Artificial fill Predominantly dune sand but includes silt, clay, rock waste, manmade debris, and organic waste Dry unit weight -1b/ft3 Approximate water table —- mesa has?) _ ‘11-:F'i‘rie'sllty‘lsanay‘fill‘ gangsta? ' "Z '5 : '{sniv'san'é m r’uhfilé tin. I 1 - 1'? :1" ' -~ 1' said-sigdflbblicfllfloL _' ' - ' '.(Iqosetolcqn1pact) -.- 7—3— . ‘\ . brick blocks,wood fragment i ' ' lli @‘ebmeis'ei-ri rams: lossxand . \q .. . ., . t @293. $.11 - -- H :®‘35::Finesand(luuse,cleanl‘, Organic silty clay, . . . _. .- CD some sheIIS. rock —n,;‘ t , . - 4 ,- Silty clay with decayed: ., . ' .' 2 Iragments,and // Organic s‘ilty clay - fl vegetation/ '- . i .-. ~. ' d d (1 ‘th k t f 75 - friicgamveent‘svmry soft) mu! sgflglgbsfg _ 70 3'] @ S'hv Clay (soft) . _ / // \ W E / ///////// Iayey silt /E I] , , fine sand is— Medium to fine sand E III] with silt,trace of 59 (Grading with very thin vertical silt' (Grading very thin horizontal Sgggtsaiyodntiicoarn/gd \ E lenses) / lenses of clean fine sand) / m Fine sand fl 2222///////////////m s'lgsvogf; a 44 Very fine sandy silty E clay, some sand pack. ////////// med- 70 Q ets (medium-stiff) ’ / Clayey siltv . gs- m /////////////////////////// ($23,322?) m N 12:," a Silty sand with zones / “z 111111111 “"9 75 of silty clay (compact) S'MV clavlsoft) ,sand 67 III //(////////////////_///////(/// (gradmg’/ “"an Sllty clav(medlum Sllff) sand) III é/(fifl fl // ////////////////////// 6:5,; 6.7 Irregular pockets // fl / of sand fl _' Clay with silt lenses ////////// I /_7; / /(medium stiff to stiff) . V V Lav / v ////////////////ll|l|l EXPLANATION Bay mud and clay Plastic gray silty clay; some lenses of sand, peat, and shell fragments, fluid to soft upper layers; moderately stiff clay at depth ® Penetration resistance-blow/ft 275 lb weight dropping 18 in. on type U Dames and Moore sampler FIGURE 32.——Geotechnical section beneath Market Street in the foot of Market zone DISTRIBUTION AND CHARACTER OF GROUND FAILURES BEALE STREET FREMDNY STREET FIRST STflEET MARKET STREET SECOND STflEEY Silty clay with de- 88 I caved veg etation/ ///////////////////// 101 Clya ey silty finesand ///////// Market Street ground line O——O 100 'l Silty clay with organic materia/l and lenses of fine sand stiff 200 l 50 ////////////////////////// 104 Silty cl ay irregular inclusions, medium l||| Clayey fine sand 107 Medium fine sand 11° (dense) l \ E GBQQ 636% "-2 © Siltv clay, some or- ganic material lllllllll 1 FInesand IIIIHII C ay,f ne sand <\ ®®®®®® Dune sand Clean well-sorted fine to medium sand; yellowish brown to lig/Itgray.Maximum thickness approx- imately 150 feet 1 Water table in borings at time of drilling (March-July 1964) Illlllllllllli Medium to fine 5 rd (compact) ill/ll“ 300 l l 100 400 FEET I l 150 l 200 METERS o\ Silly fine sandy clay /C//////////_/////////(/ Clayey fIne san \ E=§=l§E§ -E= 8 ® in W 113 m Silty clay with lenses 0 w clayey sand an m "I" 9 [(3-3 112 / Clayey,mediumtofine sand { no llll | Fine sand Medium fine sand (dense) I lllllll Quaternary deposits,undifferentiated Locally includes Colma Formationmnconsolidated, fine to medium sand and, in places, clay beds, 6 in.t0 5ft. thick 63 Location of borings from which cross section is constructed (10c. 193) (data from Dames and Moore, Inc., 1964; geology inferred from Schlocker, 1974). MONTGOMEH Y NEW STflEET 3: 39 40 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES FIGURE 33.—Disruption of block pavement and sidewalk on Columbia Street just south of Folsom Street (loc. 205). Photograph previously published in Lawson and others (1908, pl. 89A) with caption "Columbia Street, just south of Folsom Street, San Francisco. Slumping, depression, and furrowing of block pavement." (Photograph by H. 0. Wood, courtesy of The Bancroft Library, University of California, Berkeley.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 41 FIGURE 34.—Rails on Fifth Street near Harrison Street, San Francisco, pulled apart by extensional movements associated with lateral spreading of underlying sediments (10c. 205x (Photograph by G. K. Gilbert.) 42 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES FIGURE 35.—Damage to San Francisco Post Office, Seventh and Mission Streets, caused by ground failure (10C. 210). A. View of southeast entrance showing differential, vertical movement of lower, nonstructural facing around building. B. View northeastward in front of building showing differential, vertical, and lateral movement of sidewalk. (Photographs courtesy of H. J. Degenkolb and Associates, San Francisco). C. View northeastward in front of Post Office showing settlement around building and lateral displacement of sidewalk to the southeast. (Photograph courtesy of San Francisco Public Library.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 43 LEFTLANE TD OAKLAND HA Y 551065 o—‘—o ‘Silty'san'd Iiu $4M I I | | " I. t E E m Lu In In In In :5 EH E E .5 v: JAMES LICK m SKYWAV v: v: .\ o. ‘ m 7H . W W E E E w E E. 3" § E °< E ’5 Q ‘“ u: u : : 2000 40m) 5000 FEET I I 600 1000 1500 2060 METE as 1906 ground failure zone 's'andstuneiragments‘ ‘ . ‘ .50flrnlav,silt‘ . . . . - , ‘ ELEVATION, IN FEET ABOVE MEAN SEA LEVEL 'oqr‘ Artificial fill Predaminan 11y dime sand but include: silt. clay, rack waste, manmade debris, and organic waste " ’11 Bay mud and clay Plastic gmy silty clay; some lenses of sand, pear, and shell fragments, fluid to soft upper layers: modEVately stiff clay at depth .fiu-th-sfind yin - T...“— Colma, Formation Unconsolidaged fine to mediumrsand; in place: include: clay beds, 6 in. to 5 ft. thick; commonly light brown to orange @ Standard penetration-blows/ft surface _' _' in] anlruphlef ' Finesand.‘ '~ '~ . clay, 1 Water table in borings at time of drilling (September I951—March 1952) Approximate water table 6 Location of borings from which cross section is constructed FIGURE 36,—Geotechnical profile across south of Market zone at the James Lick Skyway between Third and Sixth Streets, San Francisco (10c. 205). (Data from Division of Highways, 1952; geology inferred from Schlocker, 1974.) HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES i)" LICK EXPLANATION \/ Approximate location and direction from which picture was taken JAMES FIGURE 37.—-—Schematic diagram of area between Ninth and Tenth Streets and Bryant and Brannan Streets in lower part of Mission Creek zone, San Francisco (loc. 214), showing approximate loca- tions and directions from which photographs in figures 38—41 were taken. DISTRIBUTION AND CHARACTER OF GROUND FAILURES 45 FIGURE 38.—View along Dore Street from Bryant Street toward Brannan Street, San Francisco (10c. 214). A. Photograph after the 1906 earthquake showing undulations as large as 6 ft (1.8 In) in street. As much as 6 ft (1.8 m) of lateral movement also occurred at this location. Similar photograph previously published in Lawson and others (1908, pl. 89D) with caption “Looking along Dore Street, from Bryant toward Brannan. Undulating and fractured condition of pavement due to earthquake. Houses thrown off their underpinning and pitched out of the vertical.” (Photograph from J. C. Branner collection, courtesy of Stanford University Archives). B. Dore Street today (September 1974) from approximately the same location as fig. 38A, showing ramps of the James Lick and Central Freeways and other structures constructed since 1906. C. Building at corner of Bryant and Dore Streets damaged by differential vertical and lateral movements. (Photograph by H. 0. Wood, courtesy of The Bancroft Library, University of California, Berkeley.) 46 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES FIGURE 38,—Continued DISTRIBUTION AND CHARACTER OF GROUND FAILURES FIGURE 39.—Scarps and right-lateral displacements caused by lat- eral spreading at two points on Bryant Street near the intersection of Ninth Street (loc. 214). A. Between Ninth and Tenth Streets (?). B. Between Eighth and Ninth Streets. (Photographs by G. K. Gilbert.) 47 48 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES FIGURE 40.—Views along Ninth Street between Bryant and Brannan Streets (10c. 214). A. View northwestward from near Brannan Street showing lateral displacement of street, rails, curb, walk, and buildings. Photograph previously published by Gilbert and others (1907, pl. 5) with caption “Results of earth flow, Ninth Street, San Francisco,” and in Lawson and others (1908, pl. 91B) with caption "Ninth Street, between Bryant and Brannan. Westward lurching of land toward former creek channel where Dore Street now is.” (Photograph by G. K. Gilbert.) B. View northwestward showing building damage, some of which is due to southwestward lateral displacement of the DISTRIBUTION AND CHARACTER OF GROUND FAILURES ground. (Photograph courtesy of San Francisco Public Library.) C. Close-up View of damage in midsection of block. (Photograph courtesy of Charles A. Smallwood.) D. Close-up of damage at northwest end of block. Photograph previously published in Lawson and others (1908, pl. 91A) with caption "Ninth Street, between Bryant and Brannan. Undulation and fissuring of pavement and sidewalks. Houses over trough have been dropt from their underpinning.” (Photograph by G. K. Gilbert.) 49 50 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES FIGURE 41.—View northeastward on Brannan Street, corner of Ninth Street (10c. 214). The scarp and settlement in the foreground marks the southern boundary of the ground failure on Ninth Street. (Photograph by G. K. Gilbert.) STREET H1008 SSEN NV/l FIGURE til-Schematic diagram of area between 17th and 18th Streets and Capp Street and South Van Ness Avenue in the central part of Mission Creek zone, San Francisco, showing locations and directions from which photographs in figures 43—47 were taken (10c. 215). DISTRIBUTION AND CHARACTER OF GROUND FAILURES 51 FIGURE 43.—Buckling of rails by compression on Howard Street (South Van Ness Avenue) near 17th Street (100. 215). Photograph previously published in Gilbert and others (1907, pl. 6B) with cap- tion “Buckling caused by earth flow, Howard Street, San Fran— cisco.” and in Lawson and others (1908, pl. 92B) with caption "Looking south on Howard Street from near Seventeenth Street. Compressional flexure of car rails.” (Photograph by G. K. Gilbert.) ‘1. FIGURE 44.—Looking north on Howard Street (South Van Ness Av- enue) from near 18th Street toward 17th Street, San Francisco (10c. 215). Rails offset laterally by lateral-spreading ground failure. (J . C. Branner Collection, courtesy of Stanford University Archives.) 52 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES W." dqm_w ‘ , “We. -, ,. FIGURE 45.———Scarps showing vertical movement and northward lateral movement (10c. 215). View eastward on 18th Street. Intersection of Howard Street (South Van Ness Avenue) is in the middleground. (Courtesy of San Francisco Public Library.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 53 FIGURE 46.—Scarps, lateral and vertical displacements in Capp Street between 17th and 18th Streets (loc. 215). Photograph previously published by Zeigler (1906) with caption ”Capp Street, near Seventeenth, damaged by earthquake” and in Schussler (1906, p. 95) with caption "Capp Street sunken.” FIGURE 47.—Buckled curbstone on Capp Street near 18th Street (loc. 215). Buckling was caused by sediments shifting toward old channel of Mission Creek. (Photograph by G. K. Gilbert.) HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES 54 FIGURE 48.—Valencia Street between 17th and 18th Streets (loc. 216). Al View northward shortly after earthquake showing collapsed Valencia Street Hotel in which tens of people were killed. Note lateral displacement of street in front of hotel. (Photograph courtesy of Berkey Photo Service, previously Bear Photo.)B. View southward after fire showing lateral and vertical displacements 0f6 ft (1.8 m) and two temporarily repaired arterial water pipelines that were ruptured by the ground movements, cutting off the water supply to a major part of the city. (Photograph by Moran, courtesy of The Bancroft Library, University of California, Berkeley.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 55 FIGURE 49.—-—San Francisco in flames after 1906 shock. Fire fighting efforts were hampered by the unavailability of water, a consequence of the many major pipeline breaks that were caused by ground failures. (Photograph from P. E. Hotz collection, US. Geological Survey.) FIGURE 50.—Lateral spread at the Youth’s Directory on 19th and Guerrero Streets (loc. 217). Photograph previously published in Lawson and others (1908, pl. 94A) with caption “View along Nineteenth Street, from Guerrero Street. Both ground and build— ings moved north about 6 feet [1.2 m] toward center of old marsh, with component of movement down the channel.” (Photograph by A. C. Lawson, courtesy of The Bancroft Library, University of California, Berkeley.) 56 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES L‘. I: '41 '44 E E "’ MISSION "’ 51-3557 A ea 5° — 63 69 Z \ 40 —— 0 200 400 600 FEET I | | | | I I I I O 50 100 150 200 METERS 30_ ‘. - - - IQBrownsandv grave Missmn Street ground IIne /. . , .31 . “ medium-moist . 6 $63 ELEVATION, IN FEET ABOVE MEAN SEA LEVEL 99 Broyvn'fine sand, Imedlum wet . —10 h—A ._. —2o JIM Artificial fiIJ Brawn sandy clay, moderately firm} Brown fine sand, ' medium I ‘ IIIIII ;® 1. ,0 @ o @Zofllexysand ‘lRiis't-brcnm clay-9y sand, ' , ' I I dense-moist] ‘ - I: medi_um_moistr .. ‘ Brown clayey sand, I Predominantly dune sand but includes silt, clay, rock waste, manmade debris, and organic waste @232 Bay mud and clay Plastic gray silty clay; some lenses of sand, peat, and shell fragments, fluid to soft upper layers; moderately stiff clay at depth medium EXPLANATION Colma formation Unconsolidated fine to medium sand; in places includes clay beds, 6 in. to 5ft thick; commonly light brown to orange Quaternary deposits, undifferentiated FIGURE 51.—Ge0technica1 section across Mission Creek zone at Mission Street between 15th and 20th DISTRIBUTION AND CHARACTER OF GROUND FAILURES STREET GB GD 20m 18m 19m 3; \ 17th STREET STflEET STREET | Approximate limits of 1906 9 . I ground failure zone a Red-brovvn fine sand‘ ‘ - umedium wet, - '32. ‘ : Water level 141,-“. e . '.'G'ray fine‘sandmery ' ‘: -|upse,saturated ', 9 V ND sample recovery ~ ' tpnloose. Gray-brown sandy clay, moderately firm ll 1 Nil Hed- brown silty clay, lllllinllHH Redvbrown clayey sand, llllllffliillllll Red -bruwn and gray silty clay, firm (some sand) ‘ ‘ A l l l l | | | | | ~— B:rown clayey sandstone -shale :and chart bedrock, very firm, :highly weathered ‘. No sample recovery, VT-Waterlevel'. " ' Ltoqlnnsefg- 102‘ '_ /Bruwn clayey sand, ///e//ry l//ou//s/e B/Iu/e//- gray Silt sand ‘ '- medium / //l u , Iu/B/e/- gray fine t ' 'Grading dark 'gray ® peat lense . 7%?» a®l Gray silty clay, sandy ‘ with abundant charcoal/ /f/r/agments (marsh deposi I) //////////////////d{ //k/-//9/r ay clayey san / m6l6\61 6| 6 )GH‘ 6 6 C; KJc 6 Brown~red silt sand, ® y very dense @ Gray clayey sand with decayed vegetation El® m @ Radiolarian chert and shale Standard penetration-blows/ft Alternate beds of ha rd brittle chert, 1—5 in. thick, and brittle, crumbly shale, 1/8 to ’éin. thick. 2 Locally includes bodies of massive chert Water table in borings at time of drilling 111 (January-February 1964) Dry unit weight-lb/ft3 63 Location of borings from which cross section is constructed Approximate water table Streets (10c. 216). (Data from Harding Associates, 1964; geology inferred from Schlocker, 1974.) 57 58 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES FIGURE 52.—Disruption of Van Ness Avenue over a filled in ravine (10c. 221). Lateral movements as great as 3 ft (0.9 m) and vertical movements as great as 2 ft (0.6 In) occurred at this location. Photo- graph previously published by Zeigler (1906) with caption “Break in asphalt paving on Van Ness Avenue near Vallejo Street.” FIGURE 53 .—Slump in Union Street between Pierce and Steiner Streets (10c. 222). Photograph previously published in Lawson and others (1908, pl. 88B) with caption "Slip of a fill on Union Street, just west of Steiner Street, San Francisco.” (Photograph by H. 0. Wood, courtesy of The Bancroft Library, University of California, Berkeley.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 59 FIGURE 54.——Flow failure along shoreline of Lake Merced triggered by the 1957 Daly City earthquake (10c. 236). Photograph previously published in Bonilla (1960, p. 22) with caption “Damage to roadway by landslides along south arm of Lake Merced. View looking northwest.” (Photograph by M. G. Bonilla.) 60 NORTH BAY COUNTIES REGION The north bay counties region includes Marin, Napa, Solano, Sonoma, and Y010 Counties, northern Contra Costa County, and the southern parts of Mendocino and Lake Counties (pl. 2 and table 6, locs. 183—192; pl. 4 and table 8, locs. 249—311; figs. 55-66). Several mountain ranges and intermontane valleys, a long segment of the Pacific Coast, and the northern segment of San Francisco, San Pablo, and Suisun Bays are in this region. The epicenter of the 1906 San Francisco earthquake (near Olema, Marin County) and several segments of San Andreas fault rupture between Bolinas Lagoon and Point Arena are also included. Only four events have been documented in addition to the 1906 shock that have triggered even as much as very slight ground failures in the north bay counties region. These shocks include an event near Mare Is- land (loc. 251) in 1898, a swarm of shocks along the west edge of the Sacramento Valley (locs. 255—259) in 1892, and the 1969 Santa Rosa earthquakes (loc. 295). Hence, the 1906 event is by far the most important event for earthquake-related ground failure studies in this region. Hillside landslides have been the most common form of seismically triggered ground failures reported in the north bay counties region. Furthermore, it is likely that only a fraction of the 1906 occurrences of these failures were documented. One of the landforms most vulnerable to seismically generated landslides is the coastal bluffs. In 1906 extensive landslides, mostly rockfalls, were observed along the bluffs near Bolinas Lagoon (ICC. 188) and between Bodega Head and Point Arena (loc. 300). Flows also occurred at at least two locations both near Tomales Bay (loc. 274, fig. 64.). A large landslide north of Santa Rosa, the Maacama landslide (10c. 298), is noteworthy because (1) it in- volved considerable displacement of a large mass of rock that in turn dammed up a creek, (2) it is particu— larly well described, and (3) it is readily relocatable in the field today. Differential settlement, lateral spreads, and ground cracks were common failure types in lowland alluvial areas. In many instances, sand boils were associated with these failures, indicating that liquefaction was a factor in their formation. For example, the flood plain of the Russian River inland for 25 miles (40 km) from the Pacific Ocean (loc. 299) was severely disturbed by these types of lowland failures. HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES Some of the largest lowland ground failures in 1906, both in areal extent and amount of movement, were generated beneath and marginal to Tomales Bay (locs. 255, 259, 266, 267, 269, 271, 278, 280). Settlements as large as 8 ft (2.4 In) were common beneath roadway fills (locs. 269, 279). Lateral displacements as large as 30 ft (9.1 m) or more caused considerable damage. For example lateral spreading toward the Lagunitas Creek channel (loc. 266) shortened a bridge by 6 ft (1.8 m), buckling the structure at its north end. A large failure occurred beneath Tomales Bay near Inverness (loc. 271, figs. 61—63) where westward lateral displacements as great as 25 ft (7.6 m) occurred beneath two boat piers 800 ft (240 m) apart. Gilbert (in Gilbert and others 1907, p. 8; and in Lawson and others, 1908, p. 79 (quoted under loc. 271, table 8)) concluded that these movements were caused by shifting of bay bottom sed- iments westward, upslope, at least 30 ft (9 m) as a consequence of earthquake vibrations. It seems un— likely to us that unconsolidated sediments at this local- ity would slide uphill during an earthquake. Hence, we offer an alternate possible explanation that the move- ments may have been caused by a combination of lat- eral spreading of shoreline sediments into the bay and down-channel movement of bay bottom sediments. This down-channel movement is consistent with that observed in other areas during the earthquake and is conformable with gravity rather than in opposition to it. Other areas where significant 1906 lowland ground failures occurred include the margins of Suisun Bay, where lateral movements of a few inches (loc. 252) and vertical movements of several feet (10c. 254) were found, and the west edge of the Cotati Valley near Sebastopol (loc. 293), where rows of trees in an orchard were shifted several feet laterally. Santa Rosa (10c. 295) sustained more damage than surrounding areas during earthquakes in 1868 (Law- son and others, 1908, p. 439), 1906 (Lawson and others, 1908, p. 199—203), and 1969 (Huffman and Youd, in Cloud and others, 1970). There is no evidence that ground failures were a factor in producing the greater damage in any of these events. In fact, several specific notes show an absence of ground failure in the main part of Santa Rosa in 1906 (Derleth, in Jordan 1907, p. 188 (10c. 295)) and in 1969 (Huffman and Youd, in Cloud and others, 1970, p. 54 (10c. 295)). Hence, earth- quake damage in Santa Rosa has been a consequence of ground shaking, rather than ground failure. Text continued on page 66 DISTRIBUTION AND CHARACTER OF GROUND FAILURES 61 FIGURE 55.—Cracks at Bolinas Lagoon (100. 187). Photograph previ- ously published in Gilbert and others (1907, pl. 4A) with caption ”Secondary cracks, shore of Bolinas Lagoon” and in Lawson and others (1908, pl. 49B) with caption “Cracks made by earthquake in tidal mud near head of Bolinas Lagoon.” (Photograph by G. K. Gilbert.) FIGURE 56.—Cracks along the edge ofa sag pond near Bolinas (10c. 187). Photograph previously published in Gilbert and others (1907, pl. 4B) with caption "Secondary cracks, with settling, Bolinas” and in Lawson and others (1908, pl. 52B) with the caption "Earthquake cracks in Bolinas at edge of an earthquake sag.” (Photograph by G. K, Gilbert.) 62 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES x -. ’ FIGURE 57.—Buildings tipped into the water at the Bolinas water- front (loc. 189). (Photograph by G. K. Gilbert.) FIGURE 58.—Cracks and settlement at the head of a lateral spread between Olema and Inverness (loc. 266). (Photograph by G. K. Gilbert.) FIGURE 59.—Cracks generated by lateral spreading southwest of Point Reyes Station (10c. 266). Photograph previously published in Lawson and others (1908, pl. 50B) with caption "Faults in road embankment, southwest of Point Reyes Station. Fault-trace is be- yond fence. Ground lurched toward marsh of Bear Valley Creek.” (Photograph by G. K. Gilbert.) FIGURE 60.—Cracking and subsidence of road grade across a marsh southwest of Point Reyes Station (100. 267). Photograph previously published in Lawson and others (1908, pl. 50A) with caption “Road embankment broken by shaking of soft ground beneath. Southwest of Point Reyes Station and 10 rods [50 m] from fault-trace.” (Photo- graph by G. K. Gilbert.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 63 FIGURE 61.—Ridged bottom of Tomales Bay after 1906 earthquake (loc. 270). Similar to photograph previously published in Gilbert and others (1907, pl. 8A) with caption "Earthquake ridges on tidal flat, Tomales Bay” and in Lawson and others (1908, pl. 54B) with caption ”Ridged mud plain 1 mile [1.6 km] from Inverness. Looking east-southeast. Mr. Hamilton’s barn at right. April 28, 1906. Tide is low. Pools occupy the deeper troughs.” (Photograph by G. K. Gilbert.) FIGURE 62.——Shifted sediments of Tomales Bay (10c. 270). Photograph previously published in Gilbert and others (1907, pl. 7A) with cap- tion “Shifted bottom of Tomales Bay” and in Lawson and others (1908, pl. 55B) with caption “South part of Inverness shoal, at low tide, April 28, 1906. Looking north-northwest. Lane of water separates firm, gravelly beach from mud shifted shoreward by earthquake.” (Photograph by G. K. Gilbert.) 64 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES FIGURE 63,—Boat piers at Inverness displaced during earthquake (10c. 271). A. Martinelli’s pier. Photograph previously published in Lawson and others (1908, p1. 57A) with caption "Martinelli’s pier at Inverness. Originally straight; shifted and broken by earth- quake. Repaired before photograph was taken.” B. Bailey’s pier. Photograph previously published in Lawson and others (1908, pl. 57B) with caption “Bailey’s pier at Inverness. Originally straight; shifted and much broken by earthquake. In subsequent repairs curvature caused by earthquake was retained.” (Photographs by G. K. Gilbert.) FIGURE 64.—Hillside flow failure 2 mi (3.2 km) west of Inverness and 1 ini (1.6 km) south of Sunshine Ranch (10c. 274). A hillside bog was set in motion by the shock and flowed down the slope as a stream of mud. (Photograph by G. K. Gilbert.) DISTRIBUTION AND CHARACTER OF GROUND FAILURES 65 FIGURE 65,—Hillside landslide in roadcut (10c. 276). Photograph pre- viously published in Lawson and others (1908, pl. 53B) with cap- tion "Landslide from road-cliff about two miles [3.2 km] west of Inverness. Slide occurred at time of earthquake.” (Photograph by G. K. Gilbert.) FIGURE 66.—Landslide on the Hutton Ranch, east side of Garcia River Valley (10c. 307). Toe of landslide with transported trees is at left. Toe of slide also caused “wrinkling” of barley field in foreground. (Photograph by F. E. Matthes, courtesy of The Bancroft Library, University of California, Berkeley.) 66 NORTH COAST COUNTIES REGION Included in the north coast region are all of Hum- boldt County and northern Mendocino County (pl. 5 and table 9, locs. 312—354; fig. 67). This region is gen- erally mountainous with narrow intermontane valleys. The San Andreas fault trends offshore near Point Arena at the south margin of the region. The 1906 rupture on that fault extended as far north as Point Arena and probably continued for some distance northwestward off the Mendocino Coast. Surface rup- tures as a result of tectonic faulting occurred at Shelter Cove in 1906 (Oakeshott, in Bailey, 1966, p. 361), indi- cating that seismic energy was released at least that far north along the coast. Major landslides were noted along the coastal bluffs from Point Arena to Cape Mendocino (locs. 312, 332, 337). These slides were generally of the rockfall type and were most notable on the higher bluffs between Point Delgada and Cape Mendocino (10c. 332). A large landslide called the Cape Fortunas landslide occurred in the coastal hills just north of False Cape, which is located about 30 mi (50 km) southeast of Eureka (10c. 339, fig. 67). The reports (Lawson and HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES others, 1908, p. 390) indicate that this slide, which was about 1 mi (1.6 km) in length and 0.25—0.5 mi (0.4—0.8 km) wide, moved as much as 1,300 ft (400 m) horizon- tally into the ocean. However, the total difference in elevation between head and toe of the landslide was only about 500 ft (150 m); hence the lateral compo- nent of movement was several times larger than the vertical component. The Cape Fortunas landslide thus may have had some similarities with lateral spreads that developed in the Anchorage, Alaska, area during the 1964 Alaska earthquake (for example, the Turna— gain Heights landslide (Hansen, 1966, p. A59—A66; Seed, 1968)). Very few hillside landslides were reported in 1906 at inland locations in the north coast area. This may have been a consequence of distance from the source of seis- mic energy release, sparse investigative coverage, or local geologic conditions. Locally, landslides were re— ported in areas such as Petrolia (10c. 338) both in 1906 and during earthquakes in 1923 and 1968. Moderate- sized landslides were also reported locally in the Eureka and Arcata areas during events in 1927 and 1954 (locs. 353, 354). FIGURE 67.—Cape Fortunas (False Cape) landslide, one of the largest landslides triggered by the 1906 shock (10c. 339). A. View toward by Lawson and others (1908, pls. 127A, B) with caption “Earth-slump at Cape Fortunas, Humboldt County.” SUMMARY AND CONCLUSIONS During the 1906 shock, considerable lateral spread- ing, differential settlement, and ground cracking de- veloped in flood plain sediments near the mouth of the Lel River (locs. 340, 344). These failures, which were similar to ground failures generated along other major rivers in northern California, were accompanied by the pervasive eruption of sand boils, indicating that liquefaction was a factor in their development. Differ- ential vertical displacements as large as 3 ft (0.9 In) were common in the Ferndale area where the effects were particularly severe. No measurements of differ- ential lateral movements were given, but the abun- dance of open fissures indicate that significant lateral displacements must have also occurred. During the 1954 earthquake, settlements as large as 0.5 ft (150 mm) occurred in the older part of Eureka, which is constructed on fill. Several waterlines were broken in that area, presumably as a result of ground displacements. scarp. B. View toward toe showing extension of toe into Pacific Ocean in background. Similar (Photographs by A. S. Eakle, courtesy of The Bancroft Library, University of California, Berkeley.) 67 SUMMARY AND CONCLUSIONS Historically, major ground failures triggered by earthquakes in northern California have been limited to the Coast Range province; however, minor failures have occurred at numerous localities throughout the region. The historical record shows that except for offshore shocks, the size of the geographic area affected and the number and the general severity of ground failures increase markedly with Richter magnitude. Hence, the largest historical event, the 1906 San Fran- cisco earthquake, has been the most important generator of ground failures. That shock triggered failures over a 370-mi (600-km)-long zone extending from southern Monterey County on the south to Eureka on the north and inland as far as 60 mi (100 km) from the Pacific Coast. Geologic, hydrologic, and topographic setting has great influence on ground failure development. Areas photographs previously published 68 especially vulnerable to ground failure have been over- steepened slopes, such as streambanks and coastal bluffs, and lowland deposits, principally Holocene flood plain deposits, deltaic deposits, and poorly compacted fills, where liquefaction has been the major cause of ground instability. Hillside landslides triggered by the 1906 earthquake were too numerous for the postearthquake inves- tigators to document each occurrence. Most of the land- slides occurred in the Coast Ranges within a few miles of the ruptured fault. Because there was little de- velopment in the mountainous areas and because most of the landslides were small or involved little dis- placement, damage from hillside landslides was small compared to other damage sources. Nevertheless, where hillside landslides impinged on the works of man, the results were generally disastrous. For exam- ple, 10 men were killed and 2 lumber mills destroyed by landslides in the Santa Cruz mountains, and 3 mi (4.8 km) of the Ocean Shore Railroad was practically obliterated by landslides along the coastal bluffs south of San Francisco. Liquefaction was a primary factor in the develop- ment of many ground failures including flows, lateral spreads, slumping of streambanks, and ground settle- ments. Lateral spreads were the most common and most damaging liquefaction-induced ground failure. Flood plain and loose sand fill deposits were the land- forms most vulnerable to this type of failure. Bridges, roadways, pipelines, and buildings suffered consider— able lateral-spreading damage. Pipeline breaks were particularly critical in San Francisco, cutting off the water supply to a city that was soon afterward in flames. Recent borehole data show loose saturated sand beneath five lateral-spreading sites. Flows developed on several sandy hillsides but did little damage be- cause of sparse development in the affected areas. Several instances have been reported of the same type of failure occurring repeatedly at the same loca- tion in more than one earthquake. For example, land- slides similar in nature and location occurred during both the 1906 San Francisco earthquake and the 1957 Daly City earthquake on steep slopes near Daly City (locs. 132, 133) and around the margins of Lake Merced (locs. 232—242). Ground cracks and slumping of streambanks occurred along Coyote Creek between San Jose and San Francisco Bay (loc. 149) during both the 1868 and 1906 events. Lateral-spreading failures of the types that developed in San Francisco during the 1906 shock (locs. 193—217), but with smaller displace- ments, also developed during shocks in 1868 (locs. 198—202, 205, 209) and 1865 (locs. 202, 209, 212). This evidence indicates that if geologic, hydrologic, and top- ographic conditions remain unchanged, similar types HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES of ground failures are probable at these locations dur- ing future strong earthquakes. Because of recent population growth and land de- velopment, the potential for landslide and other ground failure damage during a large earthquake is enormous now compared to that in 1906. REFERENCES CITED Anderson, Robert, 1907, Earth-flows at the time of the San Francisco earthquake [abs]: Geol. Soc. America Bull, v. 18, p. 643. Alameda Daily Argus, 1906, in Alameda: Alameda, Calif, April 18, 1906. The Argonaut, 1906, Photograph: San Francisco, Calif, April 28, 1906, v. 58, no. 1520, p. 2. Bailey, E. H., ed., 1966, Geology of northern California: California Div. Mines and Geol. Bull. 190, 507 p. Bancroft, H. H., 1882, History of California: San Francisco, A. L. Bancroft and Co., v. 6, p. 200. Blue Lake Advocate, 1954, Earthquake hits county—loss heavy: Blue Lake California, v. 66, Dec. 23, 1954, p. 1. Bonilla, M. G., 1959, Geologic observations in the epicentral area of the San Francisco earthquake of March 22, 1957, in San Fran- cisco Earthquake of March 1957: California Div. Mines and Geology Spec. Rept. no. 57, p. 25—37. 1960, Landslides in the San Francisco South quadrangle, California: US. Geol. Survey open-file report, 44p. 10 figs. Bronson, W., 1959, The earth shook, the sky burned: New York, Doubleday & Co., 192 p. Brown, A. A., and others, 1932, Subsidence and the foundation prob- lem in San Francisco; a report of the subsoil committee: Am. Soc. Civil Engineers, San Francisco sec., 1932, p. 29. Brown, R. D. and Wolfe, E. W., 1972, Map showing recently active breaks along the San Andreas fault between Point Delgada and Bolinas Bay, California: US. Geol. Survey Misc. Geol. lnv. Map I—692. The Bulletin, 1906, Earth cracks small river: San Francisco, Califor- nia, April 20, 1906. p. 3. California Department of Water Resources, 1971, Reconnaissance study of landslide conditions and related sediment production on a portion of the Eel River and selected tributaries: California Dept. Water Resources, Northern Dist. Middle Fork Eel River Devel., Memo. Rept., Sept. 1971. Carey, E. P., 1906, The great fault of California and the San Fran- cisco earthquake, April 18, 1906: Jour. Geography, v. 5, no. 7, p. 289—301. Chicago Evening American, 1906, J. R. Rainey visits cities also stricken: Chicago, 111., April 19, 1906, v. 6, no. 248, p. 3. The Chicago Record-Herald, 1906a, Flee to the parks to escape flames: Chicago, 111., April 23, 1906, v.25, no. 299, p. 5. 1906b, Sunken flats showing huge fissures made in the street: Chicago, 111., April 25, 1906, v.25, no. 301, p. 2. Cloud, W. K., and others, 1970, the Santa Rosa earthquakes of Oc- tober 1969: California Div. Mines and Geology, Mineral Inf. Service, v. 23, no. 3, p. 42—63. Coffman, J.L., and von Hake, C. A., 1973, Earthquake history of the United States (rev. ed.): US. Environmental Data Service, Pub. 41—1, 208 p. _ Dames and Moore, Inc., 1964, Soil investigation and analyses Em— barcadero regional station and cross-over structure, Bay Area Rapid Transit District, San Francisco, California: Unpub. rept., Dames and Moore, Inc., contract S—702, 21 p., numerous plates. Davidson, George, 1906, The San Francisco earthquake of 1906: Am. Philos. Soc. Proc., v. 45, no. 183, p. 164—182. REFERENCES CITED Davison, G., 1906, The San Francisco earthquake of April 18: Scien- tific American Supp. no. 1586, p. 25416, May 26, 1906. Derleth, Charles, Jr., 1906a, Report by Prof C. Derleth, Jr.: En- gineering News, v. 55, no. 18, p. 503—504. 1906b, Some effects of the San Francisco earthquake on water-works, streets, sewers, car tracks and buildings: En- gineering News, v. 55, no. 20, p. 548—554. 1906c, Additional examples of street subsidence in San Fran- cisco: Engineering News, v. 55, no. 21, p. 580—581. 1906d, The destructive extent of the San Francisco earth- quake: Engineering News, v. 55, no. 26, p. 707—713. Division of Highways, 1952, Bayshore freeway, 8th Street to 3rd Street foundation investigation: Unpub. rept., California Dept. Public Works, Bridge Dept, bridge no. 34—43 R/L, file no. T—7—H, drawing no. PR—1856—1. 1964, Route 68, bridge across Salinas River, log of test borings: Unpub. rept., California Dept. Public Works, contract no. 14— 020814, document no. 80001070. 1967, Boring logs at bridge no. 37—84 on State Highway 237, Mt. View-Alviso Road, at Coyote Creek: Unpub. rept., California Dept. Public Works. Duryea, E., Jr. and others, 1907, The effects of the San Francisco earthquake of April 18, 1906, on engineering constructions: Am. Soc. Civil Engineers Trans, v. 59, paper no. 1056, p. 208—329. Earthquake Engineering Research Institute Newsletter, 1975, Let- ter by H. H. Howard: Earthquake Engineering Research Inst. Newsletter v. 9, no. 1, p. 103. Engineering News, 1906, The San Francisco disaster, 1906; Earth- quake and fire ruin in the bay counties, California: Eng. News, v. 55, no. 17, p. 478—480. Environmental Data Service, 1971, California, Annual summary, Climatological data, total precipitation and departures from normal: U.S. Dept. Commerce, Natl. Oceanog. and Atmospheric Adm., v.75, no. 13, table 2, p. 475—481. The Evening Bee, 1906a, Sacramento suffers no real damage from the effect of the shock: Sacramento, Calif, April 18, 1906, v. 99, no. 16, 358, p. 1. 1906b, Trains operating over marshes: Sacramento, Calif, April 19, 1906, v.99, no. 16,359, p. 4. The Evening Post, 1906a, Circuit of shaken region: New York, N.Y., April 20, 1906, v. 105, p. 3. 1906b, Destruction at Alameda: New York, N.Y., April 19, 1906, v. 105, p. 2. 1906c, March of conflagration. Sinking of waterfront: New York, N.Y., April 19, 1906, v. 105, p. 6. 1906d, Shock felt at Stockton: New York, N.Y., April 18, 1906. 1906e, Railroad tracks sunk: New York, N.Y., April 18, 1906, v. 105, p. 1. Evening Sentinel, 1906a, Most terrible and destructive earthquake: Santa Cruz, California, April 18, 1906, v. 10, no. 269, p. 1. 1906b, Earthquake notes: Santa Cruz, California, April 19, 1906, v. 10, no. 270, p. 2. 1906c, Personals, Moss Landing Watsonville: Santa Cruz, California, April 19, 1906, v. 10, no. 270, p. 6. 1906d, Earthquake notes: Santa Cruz, California, April 19, 1906, v. 10, no. 270, p. 8. 1906e, Earthquake paragraphs: Santa Cruz, California, April 20, 1906, v. 10, no. 271, p. 3. The Evening Wisconsin, 1906, Big crevasse in street: Milwaukee, Wisc., April 18, 1906, v. 59, p. 1. Fallows, Samuel, 1906, Complete story of the San Francisco horror: San Francisco, Hubert D. Russell, 408 p. The Ferndale Enterprise, 1956, Enterprise tells of big quake: Ferndale, Calif, April 20, 1956, v. 78, no. 16, p. 1, 3. 69 Gilbert, G. K., 1906, Photographic collection; from United States Geological Survey Library Archives, Denver, Colorado. Gilbert, G. K., Humphrey, R. L., Sewell, J. S., and Soulé, F., 1907, The San Francisco earthquake and fire of April 18, 1906 and their effects on structures and structural materials: U.S. Geol. Survey Bull. 324, 170 p. Givens, J. D., 1906, San Francisco in ruins: San Francisco, Califor- nia, Leon C. Osteyee. Hall, W. H., 1906, Some lessons of the earthquake and fire; II, A record of earthquake disturbances: San Francisco Chronicle, v.88, no. 125, p. 32. Halley, William, 1876, The Centennial year book of Alameda County, California: Oakland, p. 257—269. Hansen, W. R., 1966, Effects of the earthquake of March 27 , 1964 at Anchorage, Alaska: U.S. Geol. Survey Prof. Paper 542—A, p. A59—A66. Harding and Associates, 1964, Soil investigation, Mission Street line, San Francisco Bay Area Rapid Transit District: Harding and Associates, San Rafael, Calif, unpub. rept., Soil Inv. M—701, 10 p. and numerous plates. Himmelwright, A. L. A., 1906, The San Francisco earthquake and fire: New York, The Roebling Construction Co., 270 p. Hittel, J. S., 1878, A history of the city of San Francisco and inciden- tally of the State of California: San Francisco, A. L. Bancroft and Co., 1878, 498 p. Holden, E. S., 1898, A catalogue of earthquakes on the Pacific Coast, 1769—1897: Smithsonian Inst. Misc. Colln., v. 37, no. 5, 253 p. Huber, W. L., 1930, San Francisco earthquakes of 1865 and 1868: Seismol. Soc. America Bull., v. 20, no. 4, p. 261—272. The Humboldt Times, 1906, Earthquake in Eureka— Some funny stunts; Damage in the north: Eureka, Calif, April 19, 1906, v. 63, no. 92, p. 5. 1954a, Earthquake damages mounting: Eureka, Calif, Dec. 22, 1954, v. 174, no. 304, p. 13. 1954b, Eel River Valley damage severe; Quake damages mount: Eureka, Calif, Dec. 22, 1954, v. 174, no. 304, p. 1, 22. 1955, Remember the earthquake?: Eureka, Calif, Dec. 20, 1955, v. 75, no. 302, p. 17. Hyde, C. G., 1906a, The structural, municipal and sanitary aspects of the central California castastrophe, pt. II: Eng. Rec., v. 53, no. 23, p. 700—705. 1906b, The structural, municipal and sanitary aspects of the central California catastrophe, pt. III: Eng. Rec., v. 53, no, 24, p. 737—740. ' 1906c, The structural, municipal and sanitary aspects of the central California catastrophe, pt. IV: Eng. Rec., v. 53, no. 25, p. 765—769. Jordan, D. S., ed., 1907, The California earthquake of 1906: San Francisco, A. M. Robertson, 360 p. Kachadoorian, Reuben, 1968, Effects of the earthquake of March 27, 1964, on the Alaska Highway system: U.S. Geol. Survey Prof. Paper 545—C, 65 p. Lawson, A. C., and others, 1908, The California earthquake of April 18, 1906; report of the California State Earthquake Investiga- tion Commission: Carnegie Inst, Washington, pub. 87, v. 1 and atlas, 451 p. Leslie’s Weekly, 1906a, The earthquake’s havoc in San Francisco’s suburbs; immense damage wrought in Oakland by the severe seismic convulsion which drove all the people of that city in terror from their homes: New York, N.Y., May 3, 1906, v. 102, no. 2643, p. 416. 1906b, San Francisco’s season of suffering and gloom: New York, N.Y., May 10, 1906, v. 102, no. 2644, p. 447. 1906c, First panorama photographs of fire-scarred ’Frisco: 70 New York, N.Y., May 24, 1906, v. 102, no. 2646, p. 505. 1906d, Curious capers of earthquake and fire; records of the terrible and unexplainable action of the elements in the Pacific Coast metropolis: New York, N.Y., May 17, 1906, v. 102, no. 2645, p. 473. 1906e, Spectacular ruins of San Francisco: New York, N.Y., May 17, 1906, v. 102, no. 2645, p. 483. Los Angeles Daily Times, 1906a, The procession to the San Francisco Ferry: Los Angeles, Calif, April 23, 1906, p. 5. 1906b, Photographs; Sunk into ground: Los Angeles, Calif. April 21, 1906, v.25, p. 1, 2. 1906c, Stirring incidents of the cataclysm—quake works havoc in adjacent towns: Los Angeles, Calif. April 19, 1906, V. 25, p. 14. Los Angeles Herald, 1906, Photograph: Los Angeles, Calif, April 26, 1906, v. 33, no. 208, p. 1. Louderback, G. D., 1947, Central California earthquakes of the 1830’s: Bull. Seismol. Soc. Am., v. 37, no. 1, p. 33—74. McAdie, A. G., 1906, California section, Climatological service, Weather Bureau, U.S. Dept. Agriculture, March 1906, April 1906: Portland, Oreg., Weather Bur. Off, p. 32—33, 35, 39, 44—45. McCulloch, D. S., and Bonilla, M. G., 1970, Effects of the earthquake of March 27, 1964, on the Alaska Railroad: U.S. Geol. Survey Prof Paper 545—D, 161 p. McLaughlin, R. J., 1974, The Sargent-Berrocal fault zone and its relation to the San Andreas fault system in the southern San Francisco bay region and Santa Clara valley, California: US. Geol. Survey Jour. Research, v. 2, no. 5, p. 593—598. Monterey County Democrat, 1906, Aromas railroad track displaced and other notes: Salinas, Calif, April 20, 1906, v. 9, no. 52, p. 2. Moore, GE, 1906, Earthquake effects at Santa Clara, Palo Alto and San Jose, California: Eng. News, v. 55, no. 19, May 10, 1906, p. 526—527. New York Tribune, 1906, Ruin and death widespread: New York, N.Y., April 19, 1906, v. 66, no. 21.704, p. 1. Nichols, D. R., and Wright, N. A., 1971, Preliminary map of historic margins of marshland, San Francisco Bay, California: U.S. Geol. Survey open-file map, scale 1:125,000. Oakland Enquirer, 1906, Site of magnesite works submerged as a result of quake: Oakland, Calif, April 23, 1906, p. 8. Oakland Tribune, 1906a, Big fissure caused at Newark: Oakland, Calif, April 18, 1906, v. 65, no. 49, p. 8. 1906b, Damage done in this city: Oakland, Calif, April 18, 1906, v. 65, no. 49, p. 4. 1906c, One-fourth of the city is saved: Oakland, Calif, April 20, 1906, v. 65, no. 51, p. 12. Preliminary Report, 1906, Preliminary report of the State Earth- quake Investigation Commission: Berkeley, May 31, 1906. Public Ledger, 1906, Over mile of track sunk: Philadelphia, Penn, v. 141, no. 26, p. 4. Ransome, F. L., 1906, The probable cause of the San Francisco earth- quake: Natl. Geog. Mag., v. 17, no. 5, p. 280—296. Rickard, T. A., 1906a, The San Francisco earthquake: Mining and Sci. Press, v. 92, no. 17, p. 270—272. 1906b, The San Francisco earthquake: Mining and Sci. Press, v. 92, no. 18, p. 287—288. Sacramento Age, 1857, Doings of the earthquake: Sacramento, Calif, January 10, 1857. Salinas Daily Index, 1906a, Terrible earthquake; Stopped the train; At Spreckles; Bridge condemned; Cracks in sand dunes: Salinas, Calif, April 18, 1906, v. 19, no. 122, p. 1. 1906b, Santa Rosa is wiped out; Earth cracked at Gonzales; Loma Prieta Co’s loss: Salinas, Calif, April 19, 1906, v. 19, no. 123, p. 3. 1906c, Latest earthquake news; The Santa Fe’s condition: HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES Salinas, Calif, April 20, 1906, v. 19, no. 124, p. 3. 1906d, Pajaro badly shaken; Freak of the temblor: Salinas, Calif, April 25, 1906, v. 19, no. 128, p. 3. 1906c, At Moss Landing: Salinas, Calif, April 26, 1906, v. 29, no. 129, p. 2. Salinas Weekly Index, 1906, Salinas visited this morning by a devas- tating temblor that wrecked some and damaged all brick build- ings in town: Salinas, Calif, April 19, 1906, v. 34, no. 40, p. 1. San Francisco Bay Conservation and Development Commission, 1975, Minutes, November 20, 1975, p. 17. San Francisco Chronicle, 1906a, Photograph: San Francisco, Calif, April 23, 1906, v. 88, no. 98, p. 1. 1906b, River changed by temblor: San Francisco, Calif, May 1, 1906, v. 88, no. 106, p. 1. 1906c, Tells of damage to sewers; City Enginer Woodward reports on the breaks and the estimated cost of repairs: San Francisco, Calif, June 17, 1906, v. 88, no. 153, p. 34. 1954a, $2,000,000 damage in Eureka quake: San Francisco, Calif, Dec. 22, 1954, no. 32,658, p. 2. 1954b, Eureka cleans up after quake: San Francisco, Calif, Dec. 23, 1954, no. 32,659, p. 12. San Francisco County, 1908, San Francisco municipal reports: San Francisco, Neal Pub. Co., p. 732-735 (repr. Sept. 1971 by George Lithograph Co., San Francisco). San Francisco Examiner, 1954a, 1 killed, 20 hurt as violent quake batters Eureka: San Francisco, Calif, Dec. 22, 1954, v. 201, no. 175, p. 1, 6. 1954b, Eureka cleans up wreckage after big quake: San Fran- cisco, Calif, Dec. 23, 1954, v. 201, no. 176, p. 1, 7. San Jose Herald, 1906a, Cities and the losses known: San Jose, Calif, April 20, 1906, v. 80, no. 95, p. 5. 1906b, Water pouring from immense fissures: San Jose, Calif, April 20, 1906, v. 80, no. 95, p. 8. San Jose Mercury, 1906a, Many buildings are in ruins at the surf city: San Jose, Calif, April 20, 1906, v. 70, no. 110, p. 8. 1906b, News from the coast towns: San Jose, Calif, April 21, 1906, v. 70, no. 111, p. 3.’ 1906c, Where scores met death: San Jose, Calif, April 21, 1906, v. 70, no. 111, p. 1. Santa Cruz Surf, 1906, Forty days after the earthquake: Santa Cruz, Calif, p. 5. Schlocker, Julius, 1974, Geology of the San Francisco North quad- rangle, California: U.S. Geol. Survey Prof. Paper 782, 109 p. 63 figs, 2 pl. Schlocker, Julius, and Bonilla, M. G., 1963, Engineering geology of the proposed nuclear power plant site on Bodega Head, Sonoma county, California: U.S. Geol. Survey TEI—844, prepared on be- half of the U.S. Atomic Energy Comm., 37 p. Schussler, Herman, 1906, The water supply of San Francisco, California before, during and after the earthquake of April 18, 1906, and the subsequent conflagration: San Francisco, Spring Valley Water Company, July 23, 1906, 48 p. Seattle Post-Intelligencer, 1906a, Fissure in San Francisco street caused by earthquake: Seattle, Wash., April 25, 1906, v. 49, no. 162, p. 1. 1906b, Steamer Senator brings refugees: Seattle, Wash., April 25, 1906, v. 49, no. 162, p. 1. 1906c, Alcatraz island felt no tremor: Seattle, Wash., April 25, 1906, v. 49, no. 162, p. 2. Seed, H. B., 1968, Landslides during earthquakes due to soil liquefaction: Am. Soc. Civil Engineers, Proc. Jour. Soil Mechanics and Found. Div., v. 94, no. SM5, paper 6110, p. 1053—1122. Soulé, F., Gihon, J. H., and Nisbet, J., 1854, The annals of San Fran- cisco: New York, D. Appleton and Co., 824 p. REFERENCES CITED 7 1 Sprague, Malcolm, 1941, Supplementary climatic notes for Califor- nia, in Climate and Man: U. S. Dept. Agriculture, p. 795—797. Steinbrugge, K. V., Bush, V. R., and Zacher, E. G., 1959, Damage to buildings and other structures during the earthquake of March 22, 1957, in San Francisco earthquakes of March 1957: Califor- nia Div. Mines and Geology Spec. Rept. 57, p. 73—106. Sunday Mercury and Herald, 1906, Seventeen men buried at Lorna Prieta: San Jose, Calif, April 22, 1906, no. 112, p. 3. Sunday Record-Herald, 1906, Twain tells humors of 1868 earth- quake: Chicago, 111., April 22, 1906, v. 25, no. 50, p. 7. Taber, S., 1906, Some local effects of the San Francisco earthquake: Jour. Geology, v. 14, no. 4, p. 303—315. Townley, S. D., and Allen, M. W., 1939, Descriptive catalog of earth- quakes of Pacific Coast of the United States 1769 to 1928: Seis- mol. Soc. America Bull., V. 29, no. 1, p. 21—252. US. Weather Bureau, 1912, Summaries of climatological data: US. Dept. Agriculture Bull. W., v. 1, sec. 1—57. Varnes, D. J ., 1958, Landslide types and processes, in Landslides and engineering practice: Natl. Acad. Sci., Natl. Research Board Spec. Rept. 29, NAS—NRC Pub. 544, p. 20—47. Varnes, D. J ., 1978, Slope Movement Types and Processes, in Land- slides: Analyses and Control, Transportation Research Board, Special Report 176, Chapter 2. Weatherbe, D’Arby, 1906, Effects of the earthquake: Mining and Sci. Press, v. 92, no. 24, p. 402. Weekly Humboldt Times, 1906a, Damage in the country: Eureka, Calif, April 26, 1906, v. 53, no. 17, p. 3. 1906b, Local phases of disaster; Earthquake sufferers; no title: Eureka, Calif, April 26, 1906, v. 53, no. 17, p. 5. 1906c, Petrolia badly hit: Eureka, Calif., April 26, 1906, v. 53, no. 17, p. 6. Wood, H. 0., Allen, M. W., and Heck, N. H., 1939, Earthquake his- tory of the United States; Pt. 11, California and western Nevada: US. Coast and Geodetic Survey, Serial No. 609, 25 p. Wood, M. W., 1883, History of Alameda County, California: Oakland, M. W. Wood, 1001 p. Youd, T. L., 1973, Liquefaction, flow, and associated ground failure, U.S. Geol. Survey Circ. 688, 12 p. 1975, Liquefaction, flow and associated ground failure: Natl. Conf. on Earthquake Eng., Ann Arbor, Mich., June 18—20, 1975, Proc., p. 146—155. Youd, T. L., and Hoose, S. N., 1976, Liquefaction during 1906 San Francisco earthquake: Am. Soc. Civil Eng., Jour. Geotechnical Div., v. 102, no. GT 5, May 1976, p. 425—439. Youd, T. L., Nichols, D. R., Helley, E. J., and Lajoie, K. R., 1975, Liquefaction potential, in Studies for seismic zonation of the San Francisco Bay region: U.S. Geol. Survey Prof. Paper 941—A, p. 68—74. Zeigler, W. G., 1906, San Francisco and vicinity, the story of the great disaster, April 18 to 21, 1906, told by pen and picture: San Francisco, California, Leon C. Osteyee, 22 p., 100 illustrations. TABLES 5—9; FIGURES CITED IN TABLES 74 ‘ HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 5.——Specific descriptions Location number is assigned to each reported ground-fai of ground failures in the Monterey Bay counties region lure site. Corresponding numbers are found on plate 1. Figure number refers to figure in this report showing damage described under "Quotation" column. Failure type is indicated by the following symbols. Corresponding symbols are found on plate 1. l2 Hillside landslides including rotational slumps, block glides, debris avalanches, and rockfalls River stretches with extensively fissured flood plains; pattern indicates stretches of river affected and not width of disturbed zone 0 Streambank landslides including rotational slumps and soil falls Lateral spread Ground settlement [:19 Ground cracks not clearly associated with land- slides, lateral spreads, settlement or primary fault movements Sand boils Disturbed wells Absence of ground failure noted E! [J C! C) Miscellaneous effects 4—{]—> Arrows showing extent of area affected. Symbol shows failure type Accuracy with which failure sites can be located is given as follows: A, a site that can be accurately relocated; B, a site that can be relocated to within a few kilometers and probably could be located more accurately with further inves— tigation; C, a site where the information is insufficient to allow precise location. Plate numbers in the "Reference” column refer to plates in the original source material. Loca— Fig- Fail— Accu- Year of Reference tion ure ure ra— earth- No. No. type cy quake Quotation 1 I B 1906 Lawson and others, 1908, p. 298. 2 ‘ C 1906 Lawson and others, 1908, p. 298. O C 1906 Lawson and others , 1908, p. 409. 3 B 1906 Lawson and others, 1908, *4E}+ p. 297. C 1906 Lawson and others, 1908, p. 293. I]! C 1906 Lawson and others, 1908, p. 297. C 1906 Lawson and others, 1908, p. 291. [18> = C 1906 Carey, 1906, p. 297. The railway station at Bradley, standing on made ground, settled 2 inches at one end. At San Ardo * * * the river bed is thought to have sunk about 2.5 feet, tho evidence of this was not obtained. At San Ardo, quicksand was thrown up in a well, seeming to lessen the flow considerably. (S. A. Guiberson, Jr )--As superintendent of the pipe line, I am in a position to say that we have no breaks whatever in any place between Coalinga and the Salinas river, and there were no fissures of any kind along the line between these points. This I know positively, as I have line riders who were instructed to look closely for any disturbance of this nature. The line of fissures seems to have ended north of Priest Valley. * * * the Salinas river bed sank nearly 6 feet at King City, and the wide sandy bottom at Three Mile Flat was much cracked * * * The river—bed sank nearly 6 feet in the vicinity of King City. Priest Valley. * * * There were slight landslides and cracks along the edge of the creek banks. * * *‘At Priest Valley * * * fissures in the ground [were] reported. TABLES 5-9 75 TABLE 5.—Specific descriptions of ground failures in the Monterey Bay counties region—Continued Loca- Fig— Fail- Accu— Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake 6 I] c 1926 7 B 1906 4—{i—e 8 553 c 1906 9 _ <) A 1906 Townley and Allen, 1939, p. 241. Lawson and others, 1908, p. 319, plate 125B. Lawson and others, 1908, p. 293. Lawson and others, 1908, p. 293. 1926 July 25. 9:58 a.m. VII. Near Idria, San Benito Co. This shock caused rocks to roll down hillsides in the ldria-Panoche region * * * Cantua Creek, Fresno County (S. C. Lillis). The shock was severely felt in this region, and its direction was southeast—northwest. A series of landslides caused by the earthquake were reported by Mr. Lillis, extending from the northwest corner of T. 18, R. 14 E., M.D M. to the middle part of T. 15, R. 11 E., a distance of about 23 miles. The features were not at first recognized by Mr. Lillis as landslides, and as they occurred on the east side of the Coast Ranges, on the border of a portion of the San Joaquin Valley, where the intensity was abnormally high, the hypothesis was entertained that there might have been a supplementary fault in that region along the edge of the mountain range. The remarkable alinement of the features lent support to this suggestion. The region was, how- ever, subsequently visited by Prof. G. D. Louderback, in company with Mr. Lillis, and the features reported by the latter were found to be landslides. Professor Louderback furnishes the following note regarding them: The phenomena reported by Mr. Lillis are several landslides. In each case the effect of the movement can be followed in detail and sharply delimited. The form of the moved body is typically that of the landslide in each case, with the cliff at the upper end curved and concave toward the lower part of the slope. -The mass has moved away and downward, leaving in some instances an open space or fissure, partially filled at the present time (May, 1907) by caving. The back cliffs, followed around, gradually pass into lateral planes of movement, which themselves are sometimes gaping on the more elevated side, showing a forward and slightly lateral movement of the mass. (See plate 125B.) No general fissure, fault, or rift was observed passing thru or near these landslides, altho a careful search was made for such features. I suspected at first that there might be such a rift— line, because the landslides are approximately along one line or belt. This appears, however, to be due to the fact that one particular formation is especially favorable to landsliding, all the slides that I saw along the lower part of the range being associated with a thick reddish—brown shale of a definite stratigraphic horizon (Tejon?). The general structure of the range causes the rocks of any given horizon to outcrop along a line roughly parallel to the range front (approximately north- west-southeast). The landslides all lookt fresh, and according to Mr. Lillis several of them (and probably all of those under consideration) were caused on April 18, 1906. I made a trip across the hills from the valley to New ldria and noted nothing that appeared to be a recent seismic line. * * * the southernmost extension of continudus cracks along the [Salinas River] bank was found to be about 2.5 miles south of Gonzales bridge. From here to the mouth of the river the cracks are parallel with the river banks. The movement at Gonzales bridge was mostly on the west bank of the stream. A wire fence trending north and south was torn 6 inches apart here, and wooden piles at the southwest end of the bridge, said to be driven down 75 feet, have been torn loose and moved from plumb, their original upright position. At the north- east end of the bridge the piles are undisturbed, but the surface soil and a wire fence have moved relatively 18 inches northward. (See fig. 59.) [That figure, not reproduced in this report, shows a plan view of the bridge and approximate relative movements of the piers.] 76 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 5.—Specific descriptions of ground failures in the Monterey Bay counties region—Continued Loca- Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake North of Gonzales bridge the fissures are mostly on the west side of the stream channel, and reach a maximum width of 18 inches. No evidence of shearing could be found. I: C 1906 Salinas Daily Index; EARTH CRACKED AT GONZALES. Albert Hansen came to the Index this 1906b. morning. He had come from Gonzales. The bridge there was broken at the further end, and large areas of land opposite Gonzales across the river were under water. The river bed has opened and for miles the earth is cracked open from inches to feet in irreg- ular cracks. 10 ‘-<)—* C 1906 Lawson and others, In the creek bottoms west of Chualar, sand craterlets begin to 1908, appear and become numerous along the stream northward. p. 293. II C 1906 Salinas Daily Index, STOPPED THE TRAIN. Passenger train no. 3 was near Chualar when 1906a. the shock occurred and the engineer says that the track moved back and forth with a worm-like movement. ll B 1906 Gilbert, Humphrey, The road leading to Spreckels's sugar mill, 4 miles south of <)’ Sewell, and Soulé, Salinas, was also greatly damaged by the slips. Spreckels's 1907, sugar mill (Pl. XII,A) is located on soft alluvium. p. 21, plate XII,A. B 1906 Duryea and others, From the Salinas highway bridge to Spreckels, a distance of about 0 1907, 1 1/2 miles, the surface of the ground was scarred so much that p. 311. the roadway became impassable for vehicles. B 1906 Lawson and others, South of the Spreckels factory, the low bottom-land between the 1908, banks of the river is considerably cracked, although there is no p. 295. prominent vertical dropping of the land along the cracks. This low land lies west of the present course of the stream, and is inter- sected by sloughs and former water courses. All of the ground is of a deep sandy nature, consequently it was much disturbed and <)* fissured by the quake, and the fissures became filled with water (D and sand, forming a quicksand, this wet sand frequently being spouted into the air. No one noticed gases coming up. The posi- tion of the cracks is now marked by patches of light, bluish—gray sand in the field, from the drying out of the quicksands. Houses on this low land were thrown out of plumb, and chimneys were destroyed. The cracks diminish in number as one goes southward, and practically end in the vicinity of the Gonzales bridge. A 1906 Lawson and others, The ground [at Spreckels sugar mill] to the south had been much 1908, ment of the piers at the southern end. (Plate 123A.) On the west p. 293, bank near the bridge a series of peculiar cracks have torn up the plate 123A, road and adjacent field, along what is probably the path of an old plate 136, water course. These are shown in plates 136, 137. plate 137. A 1906 Lawson and others, On the road eastward to Salinas from Del Monte, no visible signs 1908, of the earthquake were encountered until the Salinas River was 0 p. 292. reached. The Salinas bridge was moved southerly several feet, according to report, and the framework was broken so as to render the bridge unsafe. A 1906 Duryea and others, Fig. 1, Plate LV, shows the south abutment of the Salinas bridge. <)- 1907, The ground at this point moved about 6 ft. into the river, * * *. p. 311. A 3—in. oil pipe line which crossed the bridge was ruptured on the south approach, one length of pipe being bent in the form of the letter S. The northern approach to the bridge was not affected. A 1906 Gilbert, Humphrey, The alluvial or soft soil forming the banks of rivers generally <)- Sewell, and Soule, moved toward the river under the earthquake vibrations, the set- 1907, tling of the ground being most marked (P1. VIII, B). The country p. 21, in the vicinity of Salinas River presented interesting features plate VIII,B. of this character. The county road crosses the river near Salinas on a wooden bridge, and the slipping of the banks carried the south abutment 6 feet toward the river. The ground was badly cracked and 0 there were a number of slips in the neighborhood. A 1906 Derleth, The land on the south bank of the Salinas river for a considerable 1906d, area has moved into the river in a northerly direction, on an aver— p. 712. age, through a distance of about 6 ft. Figs. 15 and 16 show the south abutment of the Salinas highway bridge. The ground has moved under the super-structure about 6 ft., bending the pile bent foundations as shown in the pictures without seriously injuring the trusses. A 3-in. oil pipe-line crossing the bridge was ruptured on the south approach, one length of pipe being bent like the letter "S." The northern end and approach to the bridge were hardly affected. In fact, the northern bank of the river shows little disturbance at this point. From the Salinas bridge eastward a distance of about 1 1/2 miles to Spreckels, the south bank of the river is continuously scarred and rent so that the old road is impassable for vehicles. B 1906 Monterey County The Salinas County Bridge at Hilltown is reported unsafe. The I( Democrat, earth this side of the river sank about four feet and the approach 1906, on the other side moved some five feet from its former position. p. 2. * * * On the Buena Vista side a fissure four feet wide and of unknown depth extends for several hundred feet. Near Agenda the C) earth opened twenty five feet across from which gushes a stream of cold water completely saturating and covering the neighborhood. B 1906 Mbnterey County Near the county bridge across the river just opposite the factory ll Democrat, [Spreekels] the earth is depressed in some places twenty-five feet 1906, and extends back into the hills several hundred feet. * * * p. 2. The earth has also sunk in many places along the river bank on both sides of the Salinas river. B 1906 Salinas Weekly Hiram Corey, of Las Palmas ranch across the Salinas river, had to come to town today on horseback, because of the impassibility of the roads. At his place no damage was suffered except the toppling over of some chimneys. Mr. Corey reports that there is a depres- sion of ten feet for a quarter of a mile at the Jacks prune orchard. The road is impassable and he had to make a detour into the orchard to get into town. He reports that the sediment land at Spreckels and in front of Las Palmas, which was dry yesterday, is flooded today, and that the county's iron bridge at Hilltown is out of commission for teams. TABLES 5—9 79 TABLE 5.—-Specific descriptions of ground failures in the Monterey Bay counties region—Continued Reference Quotation Loca- Fig— Fail— Accu— Year of tion ure ure ra- earth- No. No. type cy quake A 1906 Salinas Daily Index, BRIDGE CONDEMNED. The county bridge across the river suffered 1906a, badly and had been closed to travel by order of supervisor J. A. American, feet and the rails have been twisted all shapes. [See also loc. 1906, 16 for adjoining text.] p. 3. 15 B 1906 Lawson and others, It may be said, regarding the soil movement along these streams, 1908, that along the Salinas River from Gonzales to near Blanco, every— <) p. 294. thing shows a movement down the river. From Blanco to Neponset )l the movement seems to have been a settling of the alluvial mate- rials, while from Neponset to the mouth of the Pajaro River [see 10c. 20] the ground (in several places, at least) moved eastward or inland. A 1906 Salinas Daily Index, The bridge at Neponset is wrecked out of line and although it is 1906d. in use repairs will have to be made on the same as soon as laborers and material can be obtained. A 1906 Lawson and others, At Neponset and Salinas the piling under the county bridge was <)> 1908, moved in some of the bents at least 10 feet toward the river. A p. 293. section man who stood in the midst of the cracks at the end of the () Neponset bridge was drenched with Spurting water. <3» A 1906 Lawson and others, Both the railway and county bridges at Neponset were moved, the 1908, northern concrete piers of the former 2 inches east and the cen- p. 293. tral wooden pier of the latter apparently 4 feet south. 16 C 1906 Lawson and others, From Morocoho [Nashua] to Moss Landing fissures rarely show in <)- 1906, the marshy land, but the narrow gage railway track has been shifted p. 293. a few inches in several places. C 1906 Ransome, Between Castroville and Monterey the ground is said to have open— I) 1906, ed and shut and mud to have spurted from the fissures. p. 295. 80 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 5.—Specific descriptions of ground failures in the Monterey Bay counties region—Continued Loca- Fig- Fail- Accu- Year of Reference tion ure ure ra— earth- No. No. type cy quake Quotation C 1906 Chicago Evening American, 1906, p. 3. II II (D Watsonville, Cal. April 19.——As we passed through the country between Monterey, Castroville and Tajaro [sic] we saw unmistakeable signs of the terrific wrenching given to that section by yesterday's tremendous uphevals of the earth. Great sinks, extending along the tracks as far as the eye could reach and ranging from four to six feet in depth, have been left in the surface of the earth, a mute testimony of the awful twisting and wrenching of the internal forces. For distances of from one- quarter to three-quarters of a mile the road bed has dropped from four to six feet. MUD GEYSERS EXCITED INTO SUDDEN ACTION. Between Castroville and Monterey, along the railroad tracks and into the fields, mud geysers have been excited into action spouting a boiling hot, bluish, shale-colored mud to a h eighth [sic] of from ten to twelve feet. In places these geysers are from four to ten feet apart and in other sections they are fifty feet or more apart. At Fairman Section Foreman H. J. Hall and Roadmaster Goldman saw these geysers in violent action. The mud was spouted through the sand and loam. I saw this mud along the tracks for several miles, with here and there places where the geysers had been recently at work. The railroad tracks for almost the entire distance are twisted out of all semblance of tracks. Between Seaside and Del Monte the tracks have settled fully four feet and the rails have been twisted all shapes. [See loc. 14.] Castroville, being on solid ground, was not seriously affected. CREVASSE OPENS AND CLOSES IN THE EARTH. Near Castroville, while the disturbance was at its height, Foreman H. J. Hall grabbed his two children and left the section house. As they passed through the door they saw the earth open and a crevasse, which Hall described as fully six feet wide, open and close several times. I visited the scene at midnight and found the section house stand— ing in a pool of geyser mud. This mud was like quicksand, and of unknown depth. * * * The LePoncet [sic] bridge [see 10c. 15] is in bad condition. The Jajaro [sic, see loc. 24, 25] bridge is gone. Castroville bridge is in a bad condition. There has been a heavy slide at Newria [sic, see loo. 6] besides several other trifling damages. Prunedale. (H. H. McIntyre).-— * * * Water started flowing in many places where there had been none, or but little before. There were 2 small landslides from springy places, the direction of the slip being from north to south. At Moss Landing, where the river runs parallel with the shore line [the Salinas River disharged into the ocean north of Moss Landing in 1906], the strip of land is seamed for miles. A crack, or rather a sink, about 20 feet wide and 4 or 5 feet deep ran under the buildings and rent them asunder. The office build- ing between this crack and the river has been moved bodily--land and all—rabout 12 feet toward the river. Some of the cracks run into the ocean. TABLES 5—9 81 TABLE 5.—Specific descriptions of ground failures in the Monterey Bay counties region—Continued Loca- Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake B 1906 Lawson and others, At Moss Landing many small cracks occur in the mud on the west side 1908, of the river, and the condition of the wharf indicates an eastward 0 p. 293, movement of the sand—spit. (See plates 134B, 135A, B.) It is plate 1343, reported that at places along the pier where the water was formerly 12A plate 135A, 6 feet deep, it now has a depth of 18 or 20 feet. plate 135B. B 1906 Lawson and others, Along the beach or sand-spit which separates the Salinas River from 1908, the Bay of Monterey at Moss Landing, there was a marked lurching of p. 401, the spit toward the trench of the river as illustrated in plates lZB plate 134A. 134A, B and 135A, B. A 1906 Salinas Daily Index, The fill leading up to the bridge crossing the mouth of the Salinas 1906d. river at Moss Landing has been shattered and will have to be rebuilt. This fill and the flood gate there was put in about twenty years ago ' and its reconstruction will entail a great expense. A 1906 Mr. Albert Authors summary of notes taken during an interview with Mr. Albert Vierra Sr., T. Vierra, Sr., of Moss Landing, Dec. 3, 1974. Mr. Vierra was 12 Personal commu— years old at the time of the earthquake. Two areas between Monterey A 1906 Mr. Hans Struve, 1974, personal communication. (3 21 C 1906 Lawson and others, 1908, <) p. 293. 22 A 1906 Mr. Jarvis McGowan, 1973, personal communication. <3> 1907, manner [to the Salinas bridge, 10c. no. 12] due to the shifting of p. 312. the bank deposits. A 1906 Salinas Daily Index, Supervisor J. L. Mann * * * reports that the damage done to the ][ 1906d. Monterey side of the Pajaro river bridge was * * * caused by a sink which extends along the bank of the river on this side and allows Chinatown to drop about four feet. This sink, or fissure, followed the bank of the river and enters under the approach from this side, throwing the whole bridge out of line. * * * another fissure fol- lowed the Watsonville side of the river, but was not so bad as on this side. C 1906 Ransome, The district drained by the Pajaro River, between the town of 1906, Hollister and Monterey Bay, was much disturbed. The railroad for ][ p. 295. distances up to a mile was depressed from 4 to 6 feet. 84 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 5.——Specific descriptions of ground failures in the Monterey Bay counties region—Continued Loca- Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake C 1865 Lawson and others, At Watsonville there was a heavy shock. The earth opened in several C) 1908, places (secondary cracks), throwing up water. p. 448. A 1906 Salinas Daily Index; The wagon bridge, the Southern Pacific and the Pajaro Valley rail- 1906a, road [loc. 24] bridges across the Pajaro river are rendered impassable. p. 1. 26 C 1906 San Jose Mercury, Between Endicotts [sic] and Watsonville the road bed is said to have ll 1906a, sunk from two to fifteen feet. p. 8. B 1906 Evening Sentinel, WATSONVILLE. One thousand feet of S.P. Co.‘s track between Laguna ll 1906c, and Ellicott sank from 5 to 15 feet. p. 6. A considerable area of ground of J. P. Thompson's property has been covered with water. <)» The tracks of the Southern Pacific Co. in all directions are widen- ing [sic] like snake tracks. )( C 1906 Evening Sentinel, Between Ellicotts and Watsonville the roadbed is said to have sunk 1906d, from two to fifteen feet. p. 8. 27 15 I: C 1906 Lawson and others, Between Pajaro and Vega the ground cracked along the 2 to 6 foot 16 1908, bluff, marking the old river bank on the south side of the present C? p. 294, channel, and the side toward the river has settled several feet. plate 1413, This is well shown in plate 1418. This displacement has caused () plate 143B. numerous sand craterlets and pits (plate 1433); the largest pit noted being oval in shape, 6 by 20 feet in diameter, and 4 feet deep. :3 C 1954 Coffman, 1954. April 25. East of Watsonville. * * * ground cracked, and O 1973, loose earth slid onto the road. p. 175. 28 C 1906 Lawson and others, Northeast of Vega the movement seems to have died out, the last 1908, evidence found being mud caps on some old piles in the channel of ]( p. 294. the stream, showing a settlement of the ground amounting to 8 inches. Between Vega and Chittenden no evidence of movement of the river-bed could be found. a C 1953 Coffman, 1953. December16. Watsonville. * * * Boulders were strewn on the 1973, road some 5 miles west of Chittenden Junction. p. 175. 29 )K B 1906 Lawson and others, The track at the southern end of the Pajaro bridge sank from 2 to 1908, 4 feet for a distance of 150 yards, and between Chittenden and the p. 279. bridge the track was bent in an S-shaped curve in several places. The concrete piers of the bridge were cracked, and the granite cappings shifted as before noticed. (See plate 65B and fig. 43.) [Damage to bridge piers most likely was caused by fault movement.] C 1890 Coffman, 1890. April 24. Monterey Bay region. * * * centered near )‘ 1973, Chittenden * * * . The railroad was damaged through settling of p. 159. ground and displacement of a bridge. C 1890 Bailey, On April 24, 1890, a strong earthquake damaged Watsonville, 1966, Hollister, and Gilroy. Mr. Joe Anzar, who was a young boy living p. 361 in the San Andreas rift valley in the nearby Chittenden Pass area at the time of that earthquake, was interviewed in 1963 by Olaf P. Jenkins and [Gordon 3.] Oakeshott. Anzar clearly remembered ground breakage, which caused Anzar Lake to drain, and landslides, which closed the railroad and highway where the fault trace crosses Chittenden Pass. He judged the motion to be stronger (at his home) than during the San Francisco earthquake of 1906. TABLES 5—9 85 TABLE 5.—Specific descriptions of ground failures in the Monterey Bay counties region—Continued Loca— Fig— Fail- Accu- Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake a C 1963 Coffman, 1.963. September 14. Vicinity of Chittenden and Soda Lake. * a: y. 1973, Landslides were reported in the Soda Lake and Pajaro Gap areas. p. 181. 30 17 B 1906 Lawson and others, The damage to the concrete abutments of the county bridge across the <3> 1908, Pajaro River is due to this crowding in of the alluvial banks of the p. 111. stream. B 1963 Coffman, 1963. September 14. Vicinity of Chittenden and Soda Lake. * * * 1973, Bridge footings of a highway bridge across the Pajaro River were p. 181. damaged slightly. 31 B 1906 Lawson and others, Several cracks a foot or less in width show on the ridge, but the 1908, fault seems to set off about 100 yards to the northeast and to consist p. 110. of east and west cracks, having loosened the whole slope for nearly [2 a mile northward of Chittenden, causing great landslides. l] B 1906 Lawson and others, Inland from the coast there were numerous earth—avalanches caused 1908, by the earthquake on the walls of steep canyons. One of the most p. 388, noteworthy of these was on the north side of a short but deep plate 126A. canyon west of Chittenden and close to the line of the fault. (Plate 126A.) The rocks composing the side of the canyon are the bituminous shales of the Monterey series. The slope rises very steeply for about 500 feet and was quite dry before the earthquake, altho it was covered with spring vegetation. Areas of bare rock appeared thru this vege— tation. At the time of the shock several earth-avalanches were started, and these slid suddenly down the slope, part of the material filling the bottom of the canyon and part remaining on the less steep lower portions of the slope. The larger masses were broken off up near the brink of the canyon. There was apparently little or no rotation of the sliding mass. The result was to gorge completely the lower part of the canyon with rock debris, to widen the upper part of the canyon, and to expose extensive surfaces of unweathered rock. I] C 1954 Coffman, 1954. August 12. East of Watsonville. Rockslides were reported 1973, on a road near Logan. p. 176. [2' C 1959 Coffman, 1959. March 2. Near Gilroy. .2 it it Minor earthslide occurred on 1973, Chittenden Pass Road east of Watsonville, and boulders fell on p. 179. Hecker Pass Highway between Gilroy and Watsonville [10c. 47]. 32 B 1906 Lawson and others, There is much sulfur, oil, gas, and water in the hills here [vicinity O 1908, of Chittenden]. A marked increase was noted in the flow of oil and p. 279. water, and more gas and sulfur became associated with them. 33 3:: C 1906 Lawson and others, The disturbance affected the banks of the Pajaro River from (—63—) 1908, Chittenden to Sargent, causing a cracking and sloughing of the banks p. 111. into the stream but not a settling of the stream bed. The San Benito River was similarly shaken for about 3 miles up from its junction with the Pajaro. Cracks are also noticeable all along the Riverside road wherever it runs close to the river bank. (2 C 1906 Lawson and others, Near Chittenden the banks are caved in. 1908, p. 294. C 1890 Holden, 1890 * * * April 24 * * * between Pajaro and Sargents. It is <)> 1898, reported that the [railroad] track was moved a foot out of line, )[ p. 150. and that the ground settled six inches in places. The bridge, fifty feet high, is impassable at both ends, the rails being pulled a foot apart. 86 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 5.—Specific descriptions of ground failures in the Monterey Bay counties region—Continued Loca- Fig— Fail— Accu- Year of Reference Quotation tion ure ure ra— earth- No . No . type cy quake = C 1885 Coffman, 1885. March 30. 1. * * Cracks occurred in soft banks of the Pajaro 1973, and Benito Rivers. p. 159. 34 C 1906 Lawson and others, In the hills between San Juan and Natividad the ground is not cracked, “W 1908, except for a few places on hillsides where there was some sloughing p. 295. off. 35 C 1906 Lawson and others, In the lowland to the southeast there is little evidence of the 1908, fault, but crossing at right angles the county road running north p. 111. and south about a mile east of San Juan, is a band of small. cracks ‘ 15 feet wide, causing the road to sink 8 inches and making a marsh of the field beyond. C 1800 Townley and Allen, 1800 October 11 to October 31. [VIII or IX. San Juan Bautista.] = 1939, * * * [All the buildings were damaged and cracks appeared in the p. 21. ground near the Pajaro River —- Hutchings’ California Magazine, 5, 310.] C 1800 Coffman, 1800. October Zl{?). San Juan Bautista. * * 4. Cracks were = 1973, observed in the ground of the rancheria; a deep fissure was reported p. 155. along the slope of the Pajaro River. 36 C 1906 Lawson and others, One old settler remembers when the business part of Hollister was m 1908, a slough. An artesian belt also passes thru the town, which may p. 288. have affected the intensity along its path. C 1906 Lawson and others, There were no changes in the ground at Hollister save some slight = 1908, cracks in the vicinity; * * * p. 289. 37 O C 1906 Lawson and others, Paicines, tho south of Tres Pinos, * e * . Water is said to have 1908, spouted up in the flat land along the [San Benito] river, 0.25 mile p. 289. from the stream. 38 E C 1906 Lawson and others, a: a. * a small peak near Santa Ana showed a landslide down its steep 1908, face, plainly visible at a distance of 6 miles. A huge rock, rolling p. 289. down a hill in Santa Ana Valley, crashed thru a house and killed a man. 39 = C 1906 Lawson and others, The road at Corralitos is said to have been slightly cracked, and 1908, in the low hills between Valencia and Corralitos a few cracks were p. 110. found; but the fault evidently runs fully 0.5 mile east of Corralitos. The mountain roads east and northeast of Corralitos were rendered I] impassable by landslides and by bridges being injured. 40 C 1906 Lawson and others, On the higher ground between Watsonville and Aptos, the shock was 1908, little felt. There was no movement along Aptos Creek, both wagon D p. 292. and railway bridges being unaffected. 41 E B 1906 Lawson and others, * * * Capitola * * * . Much earth fell from bluffs near the town, 1908, but there was no appreciable effect on the surf. At the country p. 292. bridge across Soquel Creek, the ground at the east abutment shoved o inward, cracking the concrete and buckling a water-pipe. 42 B 1906 Lawson and others, The east abutment of the concrete wagon bridge over Soquel Creek <) 1908, cracked vertically, showing that the soil movement extended this far p. 292. up the creek. TABLES 5—9 87 TABLE 5.—Specific descriptions of ground failures in the Monterey Bay counties region—Continued Loca- Fig— Fail- Accu- Year of Reference Quotation tion ure ure ra— earth- No. No. type cy quake 43 A 1906 Lawson and others, At the Southern Pacific bridge, crossing the San Lorenzo River, 1908, there is a network of fissures varying from 2 to 15 inches in width, p. 271. running thru the sandy soil. The direction of the main fissures is east and west, and they are on the south side of the river, which is )[ nearest the bay. The ground has settled about 10 inches from the abutments and piers of the bridge. The depth of the fissures was <) indeterminable, as they had filled with sand. B 1906 Evening Sentinel, There are many cracks in the earth in various parts of the city 1906a, [Santa Cruz], as for instance near the depot, and at the corner of Front St. and Soquel Av. ll Near the Riverside Hotel the street has dropped several inches. * * * The railroad embankment extending from Casino [on arcade] to river has sunk. B 1906 Evening sentinel, Along the water front there were quite a number of friekish [sic] :: 1906b, pranks of the temblor. The earth along the esplanade was opened in places and left gaping wide. B 1906 Evening Sentinel, At the Riverside Hotel, ground in the orchard opened and river () 1906e, bottom sand and water were thrown up upon the surface of the ground, the earth closing again. A similar occurrence happened at Watsonville [loc. 25], and well drillers say that the soft blue soil exuded is such as is found not less than 100 feet beneath the surface of the earth. A 1906 Lawson and others, At the north end of the bridge crossing the San Lorenzo River, at :3 1908, Third Street, there were 4 fissures running practically parallel and p. 270. almost due east and west. These fissures are about 700 yards in length, and vary in width from 2 to 8 inches. They run thru an apple orchard and are in sandy soil, the softness of the land near the river-bed being apparently responsible for their presence. The river at this place runs about east. 44 )[ B 1906 Lawson and others, Along the San Lorenzo River, at Santa Cruz, this settling action 1908, [as along the Pajaro River] also took place for a mile or more p. 294. upstream from its mouth. A 1906 Lawson and others, In going thru the town of Santa Cruz in the direction of Boulder :: 1908, Creek, a fissure at the intersection of Bulkhead and River Streets p. 270. was noticed. This fissure was about 1.5 inches wide and ran east and west. C 1906 Santa Cruz Surf; [Picture caption] Fissures in soft mud of river bank. 1906, p. 5. [See also locs. 45 and 46.] 88 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.———Specific descriptions of ground failures in the San Francisco Bay counties region Location number is assigned to each reported ground-failure site. Corresponding numbers are found on plate 2. Figure number refers to figure in this report showing damage described under "Quotation" column. Failure type is indicated by the following symbols. Corresponding symbols are found on plate 2. River stretches with extensively fissured flood plains; pattern indicates stretches of river affected and not width of disturbed zone a Hillside landslides including rotational slumps, block glides, debris avalanches, and rockfalls O Streambank landslides including rotational slumps and soil falls Sand boils 0 Lateral spread Disturbed wells 1 Ground settlement Absence of ground failure noted = Ground cracks not clearly associated with land- slides, lateral spreads, settlement or primary fault movements Accuracy with which failure sites can be located is given as follows: 0 O E] m Miscellaneous effects <-D-> Arrows showing extent of area affected. shows failure type Symbol A, a site that can be accurately relocated; B, a site that can be relocated to within a few kilometers and probably could be located more accurately with further inves- tigation; C, a site where the information is insufficient to allow precise location. Plate numbers in the "Reference" column refer to plates in the original source material. Santa Cruz Mountains Loca— Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake 45 a B 1906 Lawson and others, (R. Collom.)——Going north from Santa Cruz, a small fissure ran north— 1908, west and southeast on the Boulder Creek road, about 0.75 mile north— p. 271. west of the California Powder Works. Along the lower end of this a road were several small and unimportant landslides. In general, the shock in this region does not seem to have been as severe as it was farther north. 46 B 1906 Lawson and others, (G.A. Waring.)-—The city of Santa Cruz furnishes excellent evidence 1908, of the effect of soil formation on the intensity of the earthquake 0 p. 271. shock. * * a: The San Lorenzo River was churned into foam, the banks 0 cracking and settling several inches; and sand, said to have come from a depth of 100 feet, was forced up in several places. The bed of the river is also said to have sunk several inches, and the current to be slower than before. A 6-inch water-main, running east and west across the river at the covered bridge, was broken at each end of the 0 bridge and moved 5.5 inches eastward. O C 1865 Lawson and others, At Santa Cruz * a. * The lowlands along the river opened and spouted O 1908, water like geysers. Some wells went dry or were filled with sand. p. 448. [See also locs. 43 and 44.] 47 E C 1947 Coffman, 1947. June 22. * a: * Hecker Pass was reported closed by slides. 1973, p. 172. E C 1959 Coffman, 1.959. March 2. Near Gilroy. * * * Minor earthslide occurred on 1973, Chittenden Pass Road east of Watsonville [loc. 31], and boulders fell p. 179. on Hecker Pass Highway between Gilroy and Watsonville. 48 B 1906 Lawson and others, The Saunders ranch is 3.5 miles southwest of Madrone, on the 1908, Madrone road. * a: * portions of what appeared to be quite solid I] p. 283. and massive rock outcrops were thrown from the steep hills near the house. I] C 1967 Coffman, 1.967. September 28. Rockslides occurred in the Morgan Hill area. 1973, p. 183. TABLES 5—9 89 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San ta Cruz Mountains — - Cont inue d Loca- Fig— Fail- Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake 49 C 1906 Lawson and others, Two miles west of Uvas P.0., and half a mile east of the sumit * * * II 1908, in wet places there was a noticeable settling of the ground. p. 288. 50 I] B 1906 Lawson and others, Near Olive Springs, 12 miles north of Santa Cruz, an earth-avalanche 1908, demolished Loma Prieta Mill and killed several men. p. 389. B 1906 Lawson and others, At Santa Cruz the inhabitants reported that near Olive Springs, 12 fl 1908, miles north of Santa Cruz, a landslide demolished Loma Prieta Mill p. 271. and killed 9 men. B 1906 Lawson and others, * * * the [fault] crack goes into Hinkley's Gulch, in which the 1908, Loma Prieta Mills are situated, and which are buried under the a p. 278. slides. B 1906 Lawson and others, On the northern side of Bridge Creek Canyon there are typical cracks fl 1908, from 1 to 8 inches wide, and here also occurred a great landslide p. 110. which buried the Loma Prieta Mill. D B 1906 Jordan, [Picture caption] Wreck of Lorna Prieta Sawmill, Hinckley's Gulch, 1907, Santa Cruz County. p. 30. E B 1906 Jordan, [Picture caption] Site of Loma Prieta Sawmill, covered to a depth 1907, of 125 Feet. p. 31. B 1906 Salinas Daily Index, Loma Prieta Lumber Company's Mill. The mill, boarding house and 1906d. other buildings of the plant were situated in a gulch, and were I2 overwhelmed by a portion of the mountain-~1500 feet long, 400 feet wide and 100 feet deep which slid down upon them. The mill and every- thing in the gulch were forced up the opposite slope of the mountain and there buried to a depth of one hundred feet. Pine and redwood trees 100 feet high came down with the slide and are now standing over the mill site as though they had grown there. Nine men were killed * 9: * , B 1906 Salinas Daily Index, LOMA PRIETA CO'S LOSS. When the earthquake occurred yesterday a 1906b. morning it caused a large mountain of earth to slide into the Canyon and completely covering the new mill. Continuing its course up the mountain on the other side it covered what is known as the bunk house and buried ten men, who were asleep at the time. 51 a C 1906 Lawson and others, The slides which obliterated Fern Gulch at Skyland do not seem to 1908, have come from the [fault] crack, but seem to lie to the west of p. 278. the crack. 52 a C 1906 Lawson and others, Skyland, Santa Cruz County a: a: * Large landslides occurred in the 1908, neighborhood. p. 278. S3 B 1906 Lawson and others, Gulches appear to have been contracted, as the bridges crossing 1908, them show that they were squeezed. The banks of Burrell Creek p. 276. appear to have approached each other, so that the creek has become very much narrowed. Water-pipes were broken and twisted, and filled with dirt. 54 B C 1906 Lawson and others, Here [ridge just west of Skyland?] 4: a. * great landslides occurred, 1908, p. 110. and redwoods were snapt off or uprooted. 90 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6,—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains-—Continued Loca- Fig— Fail- Accu— Year of Reference Quotation tion ure ure ra— earth- No. No. type cy quake 55 18 a B 1906 Lawson and others, A large landslide also occurred close to Wright Station, partly 1908, damming up the stream. p. 110. B 1906 Lawson and others, Just north of Wright's Station, on the west bank of Los Gatos Creek, a 1908, there was a landslide 0.5 mile wide which had slid into the creek and p. 276. dammed it. The top of this slide was near the Summit school—house and was close to the main fault-line. a B 1906 Lawson and others, A large slide close to Wright Station partly dammed the stream. 1908, p. 389. 56 B 1906 Lawson and others, At Freely's place, 4 or 5 miles north of Morrell's, some 15 acres of B 1908, woodland have slid into Los Gatos Creek, making a large pond. There p. 278. are many other slides in the neighborhood and many broken trees. B 1906 Jordan, Into this [Los Gatos] creek, from the Feely ranch, some ten acres of a 1907, land was thrown in a great landslide. At the head of the creek is p. 27. the long tunnel which cuts under the saddle, from Wright's to Laurel. [2 C 1906 Carey, Landslides were abundant, especially in the Santa Cruz Mountains, 1906, where the topography is more rugged. One slide, a few miles from p. 297. Wright's Station, involved eight to ten acres of ground. 57 B 1906 Lawson and others, On the ranch of Dr. Tevis [presently the site of Alma College], about 1908, a mile from Alma Station, where the land is rolling and wooded, the p. 275, ground was fissured and the bottom of an artificial lake was upheaved. plate 139C, (Plate 139C, D.) The cracks and fissures, of which there are many, plate 139D. run mostly north and south, and vary in length up to 100 feet, and in width from 0.5 inch or less to 20 inches. While a good many of the openings were parallel to the slopes and were caused by the ground starting to slide, others crost the roads and could be traced some distance up the banks. A board fence was splintered where it <) crost a fissure. The upheaval of the lake was caused by a closing together of the sides, shown by the heaving up of parts of the retaining dam at the lower end of the lake. The rise of the bottom is roughly 10 feet. [Some cracks described above may be fault ruptures.] 58 [2 B 1906 Lawson and others, Mr. Carey also reports another earth-avalanche located on the Petty 1908, ranch, about 4 miles southeast of the one just described [Deer Creek p. 388. landslide, 10c. 74]. Here a huge rock mass, which embraces an area of about 12 acres at the headwaters of Cauley Gulch, broke away from a ledge and dropt, leaving a vertical scarp of 40 feet or more. The rock mass in this case was not shattered. It practically main- tained its integrity. The narrow gulch below was unfavorable for free downward movement. As the block readjusted itself, its upper surface became nearly level, but was lower at the foot of the scarp than at its outer edge, thus indicating that it had suffered rotation. 59 C 1906 Lawson and others, The whole ridge west of the reservoirs [about 2 miles south of I: 1908, Congress Springs] was severely shaken, however, for cracks 4 or 5 p. 109. inches wide opened near Grizzly Rock and several large slides occurred in its neighborhood [loc. 74]. One water-pipe running north and south on the Beatty place was broken, while one trending east and west was unhurt. No cracks were found crossing the ridge between Grizzly Rock and White Rock. The cracks were next found on the road about a mile east of B.M. 2135 of the U.S. Geological Survey, but they do not show in the vineyard to the southeast. TABLES 5—9 91 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains—-Continued Loca— Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra— earth- No. No. type cy quake 60 C 1865 Holden, 1865, October 8. At Mountain Charley's, on the Santa Cruz road, the C) 1898, earth opened in several places, and steam and water were thrown up p. 67. through the cracks. C 1865 Coffman, 1865. October 8. At Mountain Charley's, the earth opened and E] 1973, boulders obstructed the road. p. 157. 61 I] C 1906 Lawson and others, Great slides on both sides of Aptos Creek have almost made a valley 1908, of the canyon for fully 0.75 mile. Following across the ridges and p. 110. canyons, the discontinuous line of slides and sinks in upland marshy places marks the course of the fault-line down into the lowland. C 1906 Lawson and others, On the western slope of the ridge just west of Skyland, several earth- E 1908, avalanches were caused by the shock; and great slides of a similar p. 389. character occurred on both sides of Aptos Creek for 0.75 mile. Besides these, there were many smaller earth-avalanches in many parts of the Santa Cruz Mountains which can not be enumerated. 62 C 1906 Lawson and others, About four miles south of Wright Station * * * The ridge on which :: 1908, we camped was full of cracks, ranging up to 2 and 3 feet in width, p. 278. and in length from a few rods to 0.25 mile, all trending west of north [2 to northwest. * * * The canyon south of us was filled with landslides. In this canyon the stratification of the rocks is plainly shown. The strike is northwest-southeast and the dip is almost vertical. The cracks coincide in direction with the strike of the strata. Cold () water was flowing from some of the cracks. I obtained a small bottle of crude oil from Mr. Sutton, which he said was dipt up from the ground on his neighbor's ranch, several hundred gallons of oil having run out of the ground since the earthquake, where there had been no sign of oil before. 63 [2 C 1906 Lawson and others, Landslides and cracks are reported between Scott Valley and Felton, :: 1908, and the dam across a small lake was cracked. p. 272. 64 [j C 1906 Lawson and others, At Ben Lomond no fissures nor other such evidences of the earthquake 1908, were to be seen. Inquiry showed this condition to continue in the p. 268. country about the town. 65 B B 1906 Lawson and others, At the dam on Big Creek (at 48, map No. 22), no harm had been done 1908, * * * . A half mile from this point cracks caused by slides were p. 269, noticed on a very steep bank. map 22. 66 I] B 1906 Lawson and others, A long, narrow landslide above a house 0.75 mile northeast of the 1908, mouth of Waddell Creek had landed against the end of the house, p. 274. taking out a strip of earth below a spring and causing a good supply of water to issue forth. This slide appeared to be partly due to the large amount of water present. 67 B 1906 Lawson and others, Half a mile southeast of where the main road crosses Finney Creek, I] 1908, a ledge of shale had been knocked into the gulch. The largest piece p. 274. which fell had an unbroken surface of about 4 square feet. The almost horizontal edges of shale beds near a house at this point were knocked down. 92 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains- —Continued Loca- Fig- Fail— Accu- Year of Reference Quot ation tion ure ure ra— earth- No. No. type cy quake 68 E C 1906 Lawson and others, At the north end of Ben Lomond Mountain, a slide carried trees and 1908, brush down to the creek. p. 389. B 1906 Lawson and others, (H. W. Bell.)—-* * * Near a deserted mill at the north-end of Ben I] 1908, Lomond Mountain, a small landslide had carried trees and brush down p. 269. to the creek, and tall trees had fallen along the road. 69 C 1906 Lawson and others, [Near Bloom's Mill on Waddel Creek] A steep bank beside the road I: 1908, showed small cracks, which could apparently have been easily made in p. 269. the loose soil. 70 B A 1906 Lawson and others, A small earthslide had started (at 45, map No. 22), and a crack, C: 1908, perhaps due to the same slide, was noticed. p. 268, map 22. A 1906 Lawson and others, Mr. Bloom, owner of a sawmill at the edge of the Big Basin, reports 1908, that the shock was less severe in the Big Basin region than at +{:F> p. 268. Boulder Creek; that there were no landslides on the road between the two places; and that, the he had been nearly to the summit on the day of the earthquake, he had seen only one crack where the earth had started to slide. 71 C 1906 Lawson and others, Near the junction of the first rOad leading from Boulder Creek into a 1908, the Big Basin, an old landslide which covered about 2 or 3 acres, p. 268. dating back to the previous winter, had been widened by the shock and its direction had changed. 72 C 1906 Lawson and others, At Boulder Creek a large portion of the soil was shaken loose from I] 1908, an abrupt hill 150 feet high, and fell to the level of the creek, p. 389. carrying trees with it. C 1906 Lawson and others, (R. Collom.)——At Boulder Creek, on the east side of the stream, a 1908, small hill of about 150 feet elevation rises rather abruptly. E p. 268. * a: * Near the top [of the hill], a large portion of the surface soil had been shaken loose, and had slid to the level of the creek, carry— ing trees with it. 73 E: C 1906 Sunday Mercury The earth opened during the great shock in the Bear Creek road, and Herald, five miles above Boulder Creek to a depth of two and one—half feet. (D 1906. From the fissure immense quantities of inky black water is pouring. 74 A 1906 Lawson and others, On Deer Creek, in the Santa Cruz Mountains, an extensive earth- E 1908, avalanche started near Grizzly Rock and moved westward down a steep, p. 388, narrow canyon for about 0.25 mile. (Plates 124D and 125A.) It then 19 plate 1240, changed its course thru an angle of about 60° as it entered a wider plate 125A. canyon of lower grade, and following this for another 0.25 mile, finally stopt at the Hoffmann Shingle Mill, which was wrecked. A fine growth of redwood, some 200 feet in height, was mowed down, and covered to the extent of 10 acres or more with from 30 to 60 feet of debris. The trees were from 3 to 10 feet in diameter. The main canyon was filled with earth and rock for an average width of 80 yards and a length of 400 yards. The entire area of the slide was about 25 acres. The difference in altitude between the point where the slide started and the shingle mill, where it stopt, is 500 feet. According to Mr. G. A. Waring, the slide material has a depth of 300 feet and is composed of soil, clay, and shale. Mr. E. P. Carey, who examined and photographed this interesting earth—avalanche, states that it originated in rock that broke away in pieces from the steeply TABLES 5—9 93 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains--Continued Loca— Fig- Fail- Accu— Year of Reference tion ure ure ra- earth— No. No. type cy quake Quotation [2 A 1906 Lawson and others, 1908, p. 267. 75 a A 1906 Lawson and others, 1908, p. 268. 76 I] B 1906 Lawson and others, 1908, p. 389. 77 C 1906 Lawson and others, 0 1908, p. 262. 78 C 1906 Lawson and others, 0 1908, p. 109. 79 B 1906 Lawson and others, 1908, p. 262. I] C 1906 Jordan, 1907, p. 69. inclined slope at the head of the gulch, leaving a large theater-like space, the bare, light-colored rock walls of which were in sharp contrast with the surrounding green vegetation. The movement was faster in the center or deepest part of the gorge than on the margins. The rock was in general piled up higher along both sides than in the center, and many pieces became entangled in the standing or uprooted trees. A steep-walled tributary to the southeast of the main gulch supplied rock material to the main avlanche, and the 2 streams joined much as confluent glaciers do. The material involved in the avalanche showed every gradation from powder to angular pieces 30 feet or more in diameter. The surface was uneven throughout. Near the mill a man was killed by a tree that fell as the avalanche was advancing. Oh Deer Creek a large landslide started from near Grizzly Rock and slid westward, but changed its direction 60° or more farther down toward the creek. The mill in the creek bottom below the slide was partly buried, and one man was killed. It is 500 feet from the mill in the gulch to the top, at the point where the slide started. The slide covered about 25 acres of ground, and destroyed a lot of virgin timber from 3 to 10 feet in diameter. The slide material, which is 300 feet deep, is composed of soil, clay, and shale. A small landslide had moved across the road (at 44, map No. 22), which 20 men spent one and a half days clearing away. A similar [to loc. 85] earth—avalanche was caused by the earthquake on the ranch of Judge Welch, not far from Long Bridge and within 2 miles of Saratoga. Mr. Herre reports that here the soil on the north— west side of a small creek coming down from the Castle Rock Ridge, was shaken down for perhaps 0.5 mile, the not continuously. In places the slide material filled up the creek—bed and totally changed the contour. It destroyed the road to the ranches farther up the canyon, and wrecked some bridges. Along the upper part of the area affected, a vineyard was destroyed; while farther down the canyon a heavy forest growth, consisting mostly of redwood, oak, alder, and laurel, was obliterated. This slide lies in the path of the San Andreas fault. Congress Springs * * * The car tracks on the curve near the path to the spring had been thrown over toward the bank for about 20 feet of the curve, a 4-inch displacement resulting. Following the Stevens Creek road on down toward Congress Springs, several landslides were noted, mostly small ones due to caving in of the banks of the creek. Just west of the springs the road was badly broken, twisted, and shoved up in places, the downthrow being first on one side and then on the other. In some places along the bank the west side projected 2 inches farther than the other, while the fence showed an offset of 2 feet. The large stone bridge across the creek appeared intact, but west of it a large patch of ground had slipt down 2 feet. 0n the Azule Springs road * * * Near the place where five roads fork, one mile north of Azule Springs on the road running southeast from the forks, there was a 6-foot drop on the road caused by a section sinking in a solid piece on a long slope, without much disturbance in its vicinity. [Picture caption] Rift Across Road near Azul Springs, Santa Clara County. 94 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.——Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mbuntains——Continued Loca- Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra— earth- No. No. type cy quake 80 C 1906 Lawson and others, On the Stevens Creek road, just after leaving the Saratoga road * * * 1908, a crack 2 inches wide showed a downthrow of 2 inches on the west p. 262. side. * * * at the next turn, 0.5 mile southeast of Stevens Creek * * * a large area of ground, extending for 150 feet, had been torn R up in a direction of N. 3° W., and a slide formed which almost blocked the road. B 1906 Lawson and others, At the southeast corner of the same grant [San Antonio] * * * . 1908, The road in front of the house [Sellinger's] was cracked, but probably p. 261. on account of the steep slope below the road. South of the house, ‘2 across Stevens Creek, there was a landslide 100 feet in width on the steep face of a bluff. B 1906 Lawson and others, Only one more effect of the shock was noted in this vicinity; namely, 1908, the bridge over Stevens Creek, on the road running due east and west p. 262. from West Side, was rendered unsafe for horses by being shoved a foot out of place. 81 A 1906 Lawson and others, On the northeast side of the creek, 0.25 mile south of the place 1908, where a road turns northeast from the Stevens Creek road to go up I] p. 262. Monte Bello ridge, there was a large landslide about 0.5 mile long and terraced from the top of the mountain. 82 G) C 1906 Lawson and others, There were numerous slides along Stevens Creek, due chiefly to the 1908, caving of the creek banks. p. 389. C 1906 Lawson and others, The short road which runs northwest along Stevens Creek for a couple 1908, of miles beyond the junction with the cross-road which connects with C: p. 262. the Monte Bello ridge showed an exposure of serpentine with cracks running along it N. 3° W. The cracks at the widest point measured about one foot. In the serpentine area the ground was badly broken up, and in one place it was covered with 3 feet of water. (Observa- tion made April 22-23.) [Some of these cracks may have been fault ruptures.] Following the road northwest beyond the terminus shown [2 in the map, many cracks were seen, due to big landslides. Fallen trees have rendered the road impassable; boulders and dead trees still fell occasionally; even while the observer was there a large tree fell not 10 feet from him, loosening rocks and soil. Just south of the two houses near the southern end of the cross—road leading toward the Monte Bello road from the Stevens Creek road, a break ran due east and west; it was 2 inches wide with a downthrow of 0.25 inch on the west side. * * * Another crack 4 inches wide was found in the road above the house. 83 C 1906 Lawson and others, (F. Lane.)--Along the ridge road southwest of Stevens Creek, sepa- 1908, rating Santa Clara and Santa Cruz Counties, there were some cracks B p. 264. due to landslides. Sandstone blocks, some of them 6 feet in diame— ter, had rolled down the hills toward the creek. C 1906 Lawson and others, Along the ridge road southwest of Stevens Creek, sandstone blocks, B 1908, some of them 6 feet in diameter, rolled down the hills toward the p. 389. creek. 84 a C 1906 Lawson and others, Mr. Herre further reports a large slide on the Mindego Ranch, 20 1908, miles southwest of Palo Alto. Here, on the north side of Alpine p. 389. Creek, a tract of some 50 acres sank at the time of the earthquake, with little or no apparent forward movement. The tract sloped to the south and west, and formed part of a great, open hill pasture, with trees and umderbrush about the lower or creek side. The creek- bed itself is filled with a growth of Douglas spruces and other trees. The land, which before the earthquake was steeply inclined, TABLES 5—9 95 TABLE 6.—-Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains~—Continued Loca- Fig- Fail— Accu- Year of Reference Quotation tion ure ure ra— earth- No. No. type cy quake is now comparatively level, the eastern and northern part having sunk perhaps 100 feet, while that on the west has sunk but 10 or 15 feet. The surface of the sunken tract was greatly seamed and cracked, and part of it was flooded, owing to the springs uncovered; but other— wise it was unchanged in appearance. There was no piling up of earth, nor sliding of one portion over another. A fence crost the tract, and the posts on it sank so that but a few inches protruded above the surface; while some Douglas spruces also sank several feet into the earth. A number of cattle were on the land at the time of the earth- quake, but were uninjured. It was a work of great difficulty to remove them, block and tackle being necessary. The creek-bed was apparently not affected, nor were the trees in it disturbed. There was no apparent movement of the earth into the canyon, but the whole mass seems simply to have been dropt from a steep slope to a nearly uniform level, surrounded by the high, blank, almost perpendicular walls of earth and rock from which it had been sundered. 8S [2 C 1906 Lawson and others, * * * on the ranch of Andrew Stengel, an earth—avalanche is reported 1908, * * * on a small tributary of Alpine Creek, and about 4 miles south- p. 388. west of the San Andreas fault at the point where the latter crosses Black Mountain into the head of Stevens Creek Canyon. The creek here is in a narrow, steep-walled canyon in the bituminous shale of the Monterey series. The soil on the canyon side was very shallow, and at the time of the earthquake it was shaken down into the bottom of the canyon, leaving the walls absolutely bare in places for a hundred yards at a stretch. The slide extends for 0.25 mile on both sides of the canyon. 86 B 1906 Lawson and others, Four miles from the town of Pescadero, on the east side of a bridge 1908, over Pescadero Creek, the ground had sunk 2 inches and the aperture a p. 273. filled by the land sliding. A mile nearer the town, the road had dropt 5 feet, but had been filled by a big slide. A house at this point was quite intact, but the chicken—house near it was carried u down and partly buried by the landslide. On Eues Creek, near its junction with Pescadero Creek, a hillside had started to slide and apparently needed only to become rain—soaked to continue the slipping. 87 B B 1906 Lawson and others, A small landslip, 0.25 mile up the east side of the short creek 1908, which flows into Gazos, just west of the fork of the road which p. 273. continues northwestward to Little Butano Creek, showed a 2-foot vertical displacement at the top, and the land had shoved into the road below. This slide measured 150 feet from its top to the road, and its width at the road was 100 feet. 88 E: C 1906 Lawson and others, Near a house on the level creek bed of Little Butano Creek, 4 cracks 1908, averaging 3 inches in width and about 20 feet in length ran N. 33° p. 273. E. The only crack noticed along the trail toward the coast was 1 mile northwest of the place where Little Butano Creek turns from southwest to northwest, and was about the same length, but ran N. 3° W. 89 C 1906 Lawson and others, (G.A. Waring.)--0n Butano Creek there were slight cracks in the road, GD 1908, and the streams were muddy. * * * the banks beside the road showed p. 273. traces of caving, there were only slight cracks, the longest one being‘in the middle of the road above the creek, running N. 67° E. for a distance of about 50 feet. 96 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains — -Cont inued Loca— Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra— earth- No . No . type cy quake 90 a C 1906 Lawson and others, A l-inch crack at the first fork of the road a mile from the town 1908, of Pescadero extended north and south for about 50 feet, * * * p. 273. and water oozed out of level ground near by. 91 = C 1906 Lawson and others, In the town of Pescadero * 4. * . Cracks were visible in the streets. 1908, * * * Cracks in the road also appeared, and dust spurted up. * * * p. 272. Going eastward from Pescadero, a small crack 30 feet long, with an east and west strike, was observed. In an orchard near by there were several cracks, the widest one measuring 8 inches, with a vertical displacement of 1 foot. m C 1906 Carey, Near Pescadero natural illuminating gas is reported to have escaped 1906, from fissures made in the ground, but during subsequent shocks the p. 297. amount that escaped steadily decreased. C 1868 Holden, 1868, October 21, Santa Cruz Mountains, near Pescadero. * it * E 1898, large pieces of rock rolled down the mountains. It is said that the p. 79. waters of Pescadero Creek became muddy in a moment, and that the m surface was covered with bubbles, which burst with a slight report and a small flame when a match was applied to them. 92 C 1906 Lawson and others, About 2 miles east of the town [Pescadero], on the north bank of O 1908, Pescadero Creek, a landslide in the shape of a half—moon, its axis p. 272. lying N. 23° W., had slipt down toward the bed of the stream. The greatest vertical displacement at the top of the slide was 15 feet; the distance from its apex to the road about 85 feet; and the span from end to end along the road about 220 feet. No solid rock was exposed by the slide. The road had dropt 6 feet at the south end, and 8 feet at the north. Only a few cracks appeared on the surface of the part which had slipt. The creek lying directly below the road had apparently received very little soil from the landslide. 93 E C 1906 Lawson and others, * * * at San Gregorio * * * Cracks from 12 to 18 inches wide appeared 1908, in the cultivated bottom-land * * * p. 266. C 1906 San Jose Mercury, Chief Engineer Rogers and party came down from up the coast Wednesday. 1906a. He reported that at San Gregorio, San Mateo county, a few miles beyond = Pescadero, he saw fissures in the earth from a few inches to fifteen 0 feet in width from which a little sand and water was being forced out. 94 A 1906 Lawson and others, * * * a couple of miles farther east [of San Gregorio], the creek 0 1908, was dammed up to a depth of 6 feet by a slide from its southeast p. 266, bank (at 32, map No. 22) * * * map 22. 95 I] C 1906 Lawson and others, On the Pomponio Creek road * * a: . A big slide above the last 1908, house forced the observer to leave the road and take the trail, p. 273. which rejoins the road a half mile farther on. 96 m A 1906 Lawson and others, Miss L. E. Bell reports that near Bellville a small alkali flat was E 1908, raised about 3 feet. There was a landslide into the road for a p. 266, distance of 300 feet, the height of the slide being 100 feet (34, map 22. map No. 22). 97 B 1906 Lawson and others, Near the Weeks ranch house, between La Honda and the summit of the 1908, ridge on the road leading to Redwood, an inconspicuous crack was p. 266. noticed running east. It was about 2 inches wide, with no vertical movement evident. The north side of the crack, however, had moved TABLES 5—9 97 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains--Continued Loca- Fig- Fail- Accu- Year of Reference tion ure ure ra- earth- No. No. type cy quake Quotation 98 C 1906 Lawson and others, 1908, p. 264. El] 99 a C 1906 Lawson and others , 1908, p. 264. C 1906 Lawson and others, I] 1908, p. 389. C 1906 Lawson and others, a 1903, p. 106. 100 a C 1906 Lawson and others, 1908, p. 389. 101 C 1906 Lawson and others, 1908, I: p. 108. fully 3 feet eastward. The crack simply marks a big slide which has been slipping for years, and which descended 3 feet during the earthquake. Farther west up the road which loops toward Langley Hill, a big crack running east and west, caused by a slide, showed a drop of 8 inches on the north side; and from here on down to the Alpine road the road was badly cut up with slides, but was not impassable. 0n the steep grade of Langley Hill a-slide had moved 30 feet. Following the Alpine road up Corde Madera Creek, cracks were common on the outside or filled portion of the road, and these were generally parallel with the embankment. The steep southern slope of the ridge just north of the Alpine road, along its lower course, was favorable to landslips. At many places huge masses of rock had been thrown down from these steep bluffs into the road, completely blocking it up. On the south side of the creek the slopes were not favorable to land- slips, but there were several of them; and at one point, about a mile from the summit of the ridge where this road enters the Page Mill road, one slide carried away the entire roadbed for a distance of about 300 feet. At many places on the south side of Corte Madera Creek, huge masses of rock had been thrown down from the steep bluffs into the road, completely blocking it. About a mile from the summit of the ridge, where the Alpine road enters the Page Mill road, a slide carried away the entire roadbed for a distance of about 300 feet. Road from Judge Allen’s southward.--Between 3 and 4 miles southeast of Portola, many cracks were visible extending in all directions. Several showed an uplift on the east or northeast side, which is also the downhill side. Some cracks were from 4 to 5 inches wide, and had a vertical throw of nearly a foot. In other places the downhill side had been thrust upward, and pieces of the crust shoved as much as 4 inches over the uphill side. Near the top of the ridge, just before reaching the point where the trail branches off, a 4—inch crack running S. 63° E. showed a 4-inch upthrow on the northeast (downhill) side. Southwest of the ridge and about 100 feet below the trail, an old landslide dating back to some time within the past year, covers about 2 acres. Around this slide the ground appeared to have been much cracked recently. Many other earth—avalanches of minor importance were caused by the earthquake in various parts of the Santa Cruz Mountains. At Hidden Villa, 2 miles northwest of Black Mountain, large blocks of rock are reported to have rolled down the slopes. Page Mill road.--In following the Page Mill road up Corde Madera Creek [now named Matadero Creek] from Mayfield, the first noticeable trace of the earthquake was a crack crossing the road due east and west, its width varying from 0.5 to 1 inch. Wagon—tracks showed a lateral displacement of 1 inch, the north side of the crack having moved west, relatively to its south side. This crack was traced a short distance into the fields beside the road, where it disappeared. Several smaller cross—cracks intersected it at intervals. There was no apparent vertical displacement. About 100 yards farther south were 3 smaller cracks varying from 0.25 to 0.75 inch in width. One ran N. 53° W., and another N. 23° W. The latter, being only 8 feet from a culvert crossing under the road, appears to have been deflected by this from a course running more nearly east. Here again was no evidence of vertical throw. Going on up toward the Alpine road from this point, more and more cracks were found, running approximately east and west, with the exception of several 98 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6,—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains-—Continued Loca— Fig- Fail— Accu- Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake north and south ones where the road ran closely parallel to the stream. Less than a mile from the first crack, groups of cracks were accompanied by small slides of dirt from the hill to the west of the road, and farther on from the bluff to the east of it. The cracks ran nearly parallel with the axis of the branch valley lying northeast and southwest. Farther up the road, large cracks began to appear among smaller ones running parallel. The first of these was 2.5 inches across and ran S. 13° E., with a downthrow of 1 inch on the east side, and could be traced from 50 to 100 feet on either side of the road. For a mile farther up the road, the cracks became so numerous and complicated that it was impossible to map any indi- vidual ones. They intersected and ran in all directions, and were all of varying widths, the largest seen measuring 8 inches across. The size of this crack, however, was probably partly due to its position on the side of a hill. The larger cracks could be traced for several hundred feet. In some places crushing had taken place, and the layer of macadam on the road had been humped up and broken. '2 In this same area are many small landslides, some large enough to cover the road; one has occurred since the earthquake. [Some of these cracks may have been fault ruptures, see McLaughlin, 1974.] 102 C 1906 Lawson and others, On Purissima [sic] Creek [unclear whether in Santa Clara County or B 1908, on a different creek of the same name in San Mateo County] a slide p. 389. filled the road for a length of about 100 feet; another, between 0.25 mile and 0.5 mile long, dammed the creek to a depth of 25 or 30 feet. 103 A 1906 Lawson and others, On the road from Clarita Vineyard to the Allen place (at 18, map. :3 1908, No. 22), several small cracks 0.25 to 0.5 inch across ran east p. 264, and west; numerous cracks intersected (near 18, map No. 22) in map 22. various directions, while some large ones running parallel to the contour lines were probably due to earth slipping. [Some of these cracks may have been fault ruptures, see McLaughlin, 1974.] 104 C 1906 Lawson and others, On the Bear Creek road, southwest of Woodside, there were many a 1908, cracks caused by landslips down steep banks. p. 265. 105 E: C 1906 Lawson and others, Small cracks appeared in the ground at Lobitos, and a small slide El 1908, occurred in the road 0.25 mile up the stream. p. 266. 106 O C 1906 Lawson and others, * a: * a crack east of the road below Purisima, due to a landslip, 1908, extended for about 1,000 feet nearly north and south; and an earth- B p. 266. slide on the side of a hill a mile or more farther south was about 100 yards long and 80 feet across. 107 A 1906 Lawson and others, Following the trail from King's Mountain House down Purisima Creek, a 1908, a large slide on the northeast side of the creek had filled the road p. 265, to a width of about 100 feet (at 23, map No. 22). The buildings at map 22. Hatch's Mill, just below (24, map No. 22) were not damaged, but a :: little farther down several cracks were found, one 8 inches wide and running S. 23° E. 108 A 1906 Lawson and others, On the northeast side of the creek, just below Borden's Mill, a B 1908, big slide had dammed the creek to a depth of 25 or 30 feet (at 25, p. 265, map No. 22). The slide was between 0.25 and 0.5 mile long. The map 22. buildings at the mill showed no damage, but a bridge just above the mill was crusht by a slide from the south side of the creek. TABLES a9 99 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains——Continued Loca- Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake 109 A 1906 Lawson and others, * * * the bridge over Pilarcitos Creek, north of the town of 1908, Half Moon Bay * * * (at 30, map No. 22) * * * was badly cracked, as p. 266, were the approaches at both ends. Just south of the bridge, several :: map 22. small cracks in the low ground west of the road permitted water to () spout up, bringing sand with it. 110 20 [2 A 1906 Lawson and others, One * * * [flow landslide] was formed in the hills bordering the 1908, terrace at Half Moon Bay, immediately south of Frenchman Creek, 1.5 p. 395, miles north of the town, and a mile from the sea, at an elevation plate 132A. of 100 feet. It is pictured in plate 132A. At this place the earth caved away in a crescent-shaped area on a slope of only 18°, and flowed out in two long arms so as to leave a hole 4 feet deep, surrounded by vertical walls of unaffected soil. The flow occurred at a fairly high point on a gently undulating incline. The dis- charged earth was divided by a mound, at a point 150 feet below the summit of the arc, and followed two courses which were determined by gullies on both sides. Much of the debris over-flowed the central mound at the same time, and inundated the barley fields to a depth of 2 to 4 feet, for 100 feet farther. On both sides of the central mound the caving away continued to the same depth. In the left-hand fork it stopt within a few feet, and the flow did not extend very far beyond. In the right—hand fork a cut 100 feet long and 50 feet wide was made, the earth flowing down from it 250 feet farther over the grain field, as shown in plate 132A. Thus the whole length of the slide was 500 feet. The width of the main hole was on the average about 100 feet, and the length, as already mentioned, 150 feet not including the arms. In this hollow in the hillsides many dry blocks of sod carrying growing grain—-usually in an upright position--were left stranded 4 feet below the surface of the hill by the removal of the subsoil. The fence that crost this area was broken and carried away and partly buried. Where the caving ceased in the right fork, a ridge of débris was piled up across the mouth of the hole, much higher than the stream of loose material that flowed farther. Similar ridges were heapt up across the path of the flow, where the break- ing away of the hill stopt in the other arm and at the upper end of the central mound. The south or right arm of the flow extended down the hill at an angle gradually decreasing from 18° to less than 5°. Large parts of the fence were carried on its surface for 300 feet. Plate 132A gives a detailed view of the lower extremity of the right arm. The stream came to an abrupt stop, like a quickly cooled lava flow, and preserved a face 1 to 2 feet in height above the grain field. The surface of the flow consisted largely of blocks of sod, usually almost upright, which were carried down from the hole without much moistening, or transformation into material capable of flowing. The bulk of the flow was a moist aggregate of earth fragments possessing something of their previous form and grading into mud, which assumed a semi—fluid consistency underneath. The bottom of the hole, and the flow itself, remained too muddy to walk on for weeks after the earthquake, and the field below the lower end of the large arm was left marshy, tho it had not been so before. It is to be noted that several fairly heavy rains followed the earthquake after an interval of several days, and before these earth—flows were visited; but these were not sufficient to account for the amount of moisture observed. The chief effect of the water was in the ground at a depth of 3 or 4 feet below the surface. It rendered the soil sufficiently fluid to enable it to flow down the gentle slope, probably partly oozing from under the surface crust and partly transporting the sod with it. Most of the surface was carried down with the main flow, the stranded surface blocks that remained in the cavity being account— able§for as fragments from the broken edges subsequently giving way and being carried only a short distance as the upper end of the flow came to rest. In this way, probably, the walls were trimmed, for the cut in general was left remarkably clean. 100 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains--Continued Loca- Fig- Fail- Accu- Year of tion ure ure ra- earth- No. No. type cy quake Reference Quotation 111 B c 1906 21A 213 112 C 1906 Lawson and others, 1908, p. 396, plate 1328, plate 133A. Lawson and others, 1908, p. 252. Another flow of similar character [see locs. 110, 112, 113, 114 and 123] took place 3 miles north—northwest of the town of Half Moon Bay, on the creek next west of Frenchman Creek. It is shown in plates 1328 and 133A. On the morning of the earthquake an acre of the gently sloping alluvial floor of a broad, short valley tributary to the main creek on the east caved and flowed out, leaving an excavation 10 feet deep, where before it had been almost level and where there had been no stream channel. In this case, the water already gathered in this basin—like valley, which here had had no means of prompt escape, was an important aid in the formation of the flow * * * . The presence of a large amount of water and the forci- ble movement during the earthquake shock resulted in the loosening and undermining of the ground and its transportation as a fluent mass. The angle of slope was about 5°. The flow carried out thousands of tons of earth in this manner and spread it over about 2 acres of meadow land, to an average depth of 1.15 to 3 feet. Plate 1328 gives a view of this earth—flow, showing the pit from which it was derived. Covering much of the surface of the flow and the floor of the hole are to be seen blocks of sod which have been carried right side up as if the material had moved en masse. The amount of water in evidence shows clearly how the earth was softened and enabled to move. The picture was taken two weeks after the earthquake. At that time water was still seeping up from under- ground, and out of the lower portions of the broken walls, while the ground near the surface of the valley was quite dry. The water had formed two definite rivulets thru the debris, at an elevation above the surrounding meadow, and was running in continuous streams, fast cutting a channel for itself and removing the soft material. Consid— erable water was dammed back in the hole by a 4-foot ridge of debris piled across the mouth of the hole, as in the case of the previously described earth—flow [1°C. 110]. This mound of earth, along the line where the stream left the caved-in area and flowed over the preexisting slope, was probably piled up at the last by the remnants of the flow gliding down and heaping themselves up as a barrier at the mouth of the hole. The cavity, about an acre in extent, has 10-foot walls which gradu— ally decrease in height lower down the valley, the bottom of the hole being more nearly level than the valley—floor. Plate 133A shows part of this flow in detail. Some of the great blocks of sod around the edges have not been removed, altho the material from underneath has gone. Concentric cracks not visible in the pictures extend around the edge of the hole and for 50 feet above its upper end, showing that the area affected is broader than appears at first sight, and that the work is not yet all accomplished. The material of the valley-bottom is a coarse, arkose earth, derived from decomposing granite, and containing many rock fragments. A flood of earth covers about 2 acres of the meadow. Water was present in this earth-flow in greater amount than in any other that was examined. The nature of the material may be judged of by the abrupt face of the stream where it stopt. The edge makes a steep angle with the meadow and rises to an average height of 2 feet above it. Yet the fact that this mass of earth was able to move more than 300 feet after it left the lower end of the hole, and spread into an even and thin layer over a wide extent of nearly level meadow, shows that it was fairly soft. It was moved on a basal layer of semi-fluid mud and sand, with the aid of the weight of the over— lying and partly disintegrated earth. About 4 miles east of Half Moon Bay, just off the south edge of the San Mateo sheet, there was another large earth—slide similar to the two [earth—flows, locs. 110 and 111] already mentioned. TABLES 53 101 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains——Continued Loca- Fig— Fail- Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake 113 a B 1906 Lawson and others, The largest of the earth-flows seen occurred in the canyon south of 1908, the house of Mr. Nunez, 2.5 miles east-northeast of the town of p. 397. Half Moon Bay, at an elevation of about 500 feet. It originated in a manner similar to the others Llocs. 110, 111, 112, and 114], but in a canyon along which there is a distinct but ordinarily dry stream channel. A long, irregular hole from 4 to 7 feet deep was excavated near the head of the valley, and a great volume of earth flowed down its curving course for 0.25 mile, as far as the Nunez house, and there stopt, being in part diverted into the main creek to which the valley is there tributary. According to the testimony of witnesses the flow reached the end of the 0.25 mile in 0.5 hour after the earthquake shock. It was seen gliding slowly down and engulfing the orchard just back of the house. According to observers on the Nunez ranch, the earth-flow was not accompanied by any water; but two weeks later, when examined by the writer, it preserved every evidence of having been muddy. Especially was this true at the bottom, where great masses of mud still had the consistency of jelly. It is probable that there was no flowing water on the surface of this or other earth-flows at the time of their formation, and that the presence of water in the flow was not evident to the casual observer because of the comparative dryness of the material on its upper surface. The slope of the canyon down which the moving body of land crawled is about 25° near the head and decreases to 15° farther down. The flow filled this to a width of 100 feet on the average, and to a depth varying from 10 to 20 feet. The inertia of the mass is illus- trated by the fact that in the early stage of the flow the earth was piled 20 feet higher on the hill, on the inside of the big curve made by the canyon, not far below the pit, than it was when the flow came to rest. The marks at this elevation were probably made very soon after the main mass was discharged from the cavity, before it had spread very widely. The central portion of this earth—flow is plate lSlB pictured in plate 1313, where it appears as a ridge many feet high rising above the tall grass on the hillside, on the right of the picture. The pressure of the material at the head of the flow, as it started, was so great that the earth bulged up over the sides in places, in such a way as to force upward great blocks of sod and turn them on edge or completely over, away from the rim of the hole. The flow assumed the form of two lateral ridges and a central de- pression, or channel. The ridge on the west or inner side of the curve was considerably the higher. The form was due partly to the concavity of the valley; but chiefly, it is thought, to the tendency of the more fluid material to follow the deepest possible path along the gully under the center of the flow. Thus the drier mate— rial was retarded at the sides. Subsequent to the first starting of the flow, a stream of semi—fluid mud and sand continued to run down the central channel, covering its sides with a coating of mud and leaving flowage striations on it. This channel and its markings are exhibited in plate 1313. Two weeks after the earthquake, when the photograph was taken, water was running in this channel and had cut down into it several feet deeper. Its bottom, however, was still from S to 10 feet higher than the bottom of the underlying pre- existent water course, where water had not flowed before at this time of the year. The man in the picture is standing at the bottom of the gully. To the left of him, the hammer and note—book mark the top of one of the parts of the lateral ridge which is here di- vided into several hummocks. To the right is the other and higher lateral ridge. The foreground was formerly covered by a dense thicket of willow trees. These willows have been completely buried, except at the sides where some dead branches protrude. A fence that crost the canyon was torn away for 100 feet, and not a trace of it could be found. The fence shown in the picture is one newly built in its place. 102 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.-—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains—-Continued Quotation Loca- Fig— Fail— Accu— Year of Reference tion ure ure ra— earth- No. No. type cy quake 114 B B 1906 Lawson and others, 1908, p. 397, 22 plate 133B. 115 I] c 1906 Lawson and others, 1908, p. 252, plate 124C, plate 126B. [2 C 1906 Lawson and others, 1908, p. 389. 116 C 1906 Lawson and others, a 1908, p. 253. 117 B 1906 Lawson and others, 1908, p. 253. 118 B 1906 Taber, 1906, p. 308, fig. 7. Two other smaller earth-flows occurred just over the hill westward from the last one described [loc. 113]. They are shown in plate 133B, the canyon on the left being the one occupied by the Nunez flow. One of these 2 earth—flows, that at the right of the picture, started near the top of the ridge in a depression in the slope, formed a hole 75 feet long and 40 feet wide, and coursed down a narrow runnel having a gradient of 25° to the bottom of the hill, a distance of 600 feet. Enough earth issued to fill up the rather deep ditch in the gully clear to the bottom of the hill and to bury the grain field on both sides to a depth of l to 2 feet. In this case, as in the preceding one, there were formed lateral ridges higher than the center, so as to leave a groove between. Down this channel there flowed softer material, which lined the sides of the lateral ridges with a smooth coat of mud and left conspicuous flow— age marks. The flow thus raised a ditch for itself above the level of the slope. The earth-flow probably assumed this form by leaving behind, at the sides, the material least capable of flowing, and by concentrating its most liquid parts along the deep central line. The other earth-flow was near by, on the convex face of the knoll in the center of the picture. A similar cavity was produced, from which the contents were spread out broadly. It is a good example of the starting of a gully, as there was no depression before. One branch of this earth—flow came straight down the hill and slightly toward the canyon on the left; the other branch came down toward the gully in which the first-mentioned of these two earth-flows occurred. Thus drainage lines were started which ultimately may separate the central hill from the ridge on the right, of which it is now a continuation. The left arm of the flow on the hill may develop a channel, as explained below, which will cause the drain- age from this hill, which is now toward the foreground, to pass into the canyon on the left. From Half Moon Bay to San Mateo, there were several large slides of different character from those already mentioned [earth-flows]. These resulted from the slipping of large masses of rock, many of the fragments in one of the slides being over 20 feet in diameter. (See plates 124C and 1268.) Near Half Moon Bay considerable masses of granite were dislodged on a.steep slope. (Plate 124C.) On the road along Pilarcitos Creek, an earth-avalanche brought down big blocks of sandstone upon the road. (Plate 1263.) [See also 10c. 121.] Just southeast of the house [on Cahill's ridge] is a depression in the ridge, across which furrows and cracks formed similar to those along the main fault-line, but not extending more than several hun— dred feet. These cracks do not seem to have been landslide cracks, for they are on top of the ridge and on a flat piece of ground. [See also 10c. 120.] A house on the northwest side of Half Moon Bay road, 2,000 feet southwest of the dam [through Crystal Springs Lakes], was thrown from its foundations, while some 200 feet northwest of this house there was a slide in the canyon. A striking evidence of [fault] displacement is shown in the earth dam that divides the Crystal Springs Lake. This dam is about 500 feet in length, and the road from San Mateo to Half Moon Bay runs along its crest. The accompanying sketch (Fig. 7) [See fig. 7 at end of tables} shows the position and direction of the cracks that were formed in the dam. The larger cracks [clearly secondary] are about 6 inches wide and are parallel with the dam. Smaller inter- secting cracks were formed near the northeast end of the dam along TABLES 5—9 103 TABLE 6.——Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains——Continued Loca- Fig- Fail- Accu- Year of tion ure No. No. type ure ra- cy earth- quake Reference Quotation 119 120 123 C C C C C C 1906 1906 1906 1906 1906 1906 Lawson and others, 1908, p. 398. Lawson and others, 1908, p. 246. Lawson and others, 1908, p. 253. Lawson and others, 1908, p. 265. Lawson and others, 1908, p. 252. Lawson and others, 1908, p. 252. the probable line of the fault, and the road was offset about 6 feet at this point. The fences on both sides of the road were broken in a number of places, and the unbroken boards were bent and arched so as to give a serpentine appearance to the fences. The wires of a telephone line crossing the dam sag in great loops. It seems probable that the total [fault] displacement is greater than the amount that may be directly measured at any place along the line of the fracture, for there is evidence of drag in the soil for a considerable distance on both sides. Water—pipes at a distance of several hundred feet from the fault-line have been pulled apart, telescoped, or bent in the direction of the movement, and fences for- merly straight have been bent into a slight curve for a distance of 200 or 300 yards from the fracture. There were examples of such slips [landslides] along the coast hills north San Pedro Point, near the road halfway between San Bruno and San Andreas Lake, near the road from Belmont to Crystal Springs Lake, 0.5 mile southeast of the San Mateo Alms House, and in many other places on the San Francisco Peninsula. In some places bare ridges had their lines of symmetry broken into little knolls and irregularities by these slips, a common occurrence in the hills of soft sand formations in the northern part of the San Francisco Peninsula. Thru the hills west of Belmont no cracks nor big landslides were found, but there were small landslides along the road leading from Belmont to Crystal Springs Lake. Another peculiar phenomenon was observed upon Cahill's Ridge, less than 1 mile northwest of the cracks mentioned [1°C. 116]. In an area of limestone, a small patch some 30 feet in diameter was torn up as tho it had been plowed and harrowed, and no large pieces of sod were left intact. Around this in various places were cracks of a few inches in width, with one or two over a foot wide. There was a slight downthrow on the uphill side to be noticed in some of these cracks, which eliminated the possibility that they were cracks pre— paratory to landsliding. Following the road along Pilarcitos Creek toward Half Moon Bay, many cracks and slides were found on the ocean side of the ridge, but few on the east side. All of these seemed due to slipping of the earth. At one place there had been such a large slide [see also loc. 115] that big blocks of sandstone had fallen down into the road. Here and there along the road big cracks had opened, parallel with the road and the creek where the slope is very steep, and pro— mising to make the road impassable by landslides, should a heavy rain come. On the south face of Scarper Peak, and on the southwest face of OK Hill, there were several landslides both large and small. South of Montara Point, in the low foot-hills north of Half Moon Bay, there were two large low—angle landslides or earth—flows. One of these landslides was on the low foot-hills facing the ocean; the other on the northeast bank of Frenchman's Creek [loc. 110], several miles northeast of Half Moon Bay. 104 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountainanont inued Loca- Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake 124 C 1906 Lawson and others, On the southwest face of Montara Mountain, nearly all of which is [:I 1908, visible from the road, no landslides of any size were observed. p. 252. 125 C 1906 Lawson and others, On Sawyer's Ridge, about 9 miles north of the region described on = 1908, Cahill's Ridge [locs. 116, 120], there were cracks several hundred p. 253. feet long almost at the top of the ridge. These were parallel to the line of the main fault, which is a mile to the east, and there was a marked downthrow of from 2 to 3 inches on the southwest side, which in this case was the uphill side. 126 B C 1906 Lawson and others, There were also several such slides [small earth-avalanches] on the 1908, granite slopes of Montara Mountain, farther north in the San Francisco p. 390. Peninsula. 127 C 1906 Lawson and others, South from the Devil's Slide to the first small coast valley, there E 1908, were landslides along the cliffs. The rock in this vicinity is mas- p. 252. sive granite, but the landslides showed that the rock had disinte— grated for a considerable distance below the surface and the slides were in this decomposed rock. Wherever the railway bed [Ocean Shore Railway] was filled or built out with this material, there was more or less sliding and settling, caused by the earthquake. 128 a C 1906 Lawson and others, Just north of the point known as Devil's Slide, there was a land- 1908, slide of the whole face of the west end of Montara Mountain. It p. 252. started at about 800 feet above the sea, and swept down carrying many hundred feet of roadbed along with it. The material that slid was sandstone and granite, but it seemed to be much weathered and softened in places, so that it was loose ground. B C 1906 Lawson and others, One earth—avalanche to the north of the Devil's Slide started 1908, about 800 feet above the shore and swept the face of the cliff, p. 387. carrying away several hundred feet of roadbed. The slide occurred near the contact of sandstones reposing on granite, and both kinds of rock were involved. Smaller earth—avalanches occurred farther south on the sea—cliffs. 129 C 1906 Lawson and others, From San Pedro Point southward for about 1.5 miles, the cliffs 1908, rise to heights of from 400 to 800 feet. The railway company had p. 252. cut a bench for its roadbed several hundred feet above the ocean. This roadbed, being largely in solid rock, was for the most part not muCh injured; but in some places it was obliterated by rock [2 slides that came from above. 130 I] C 1906 Lawson and others, To the south of Mussel Rock there were several small earth- 1908, avalanches along the cliffs, and numerous cracks were formed near p. 387. the brink of the cliffs which will in future doubtless lead to fur- E ther falls from the cliff-face. Near San Pedro Point there was a large movement of the earth on the face of the high cliff. 131 C 1906 Lawson and others, In the valley of Laguna Salada, the Ocean Shore Railroad had a 1908, temporary trestle erected for making a fill in the valley up to p. 251. required grade. This trestle was twisted and thrown out of line, 1 and the earth sank along the newly filled roadbed. * a. * Along the base of the cliffs south of Laguna Salada, there were B several small slides, some from the face of the hills and others in the newly graded roadbed. There were many small cracks along TABLES 5—9 105 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains——Continued Loca— Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra— earth- No. No. type cy quake the tops of the cliff, parallel to its edge, showing that the face of the bluff was shattered, and that more earth might slide. One big rock pinnacle, which had been left above the roadbed as a land— mark, and which had seemed a little dangerous before, was shaken down. 132 23 I] C 1906 Lawson and others, South of the Golden Gate, the most notable earth-avalanches were 1908, along the sea—cliffs between the city and Mussel Rock. This cliff p. 387, has a length of about 6 miles and ranges in height from about 100 plate 129C, feet up to 700 feet, and is cut almost wholly in the strata of the plate 129D. Merced (Pliocene) series, which are inclined at angles varying from 15° to 75°. The rocks are for the most part rather soft and inco- herent, tho there are numerous well—cemented and indurated beds in the series. This cliff converges on the fault at a small angle, and intersects it at its south end near Mussel Rock. The cliff was severely shaken and great quantities of earth and rock were caused to fall or slip down. The great earth-slump at Mussel Rock (Plate 129C, D) was also notably accelerated. C 1906 Lawson and others, All along this line of cliffs [between Lake Merced and Mussel 1908, Rock], and for a short undetermined distance inland, the rock masses p. 241. were cracked, broken, and traversed by narrow fissures. These effects grow more and more numerous and of greater and greater mag— nitude until, a short distance north of Mussel Rock, the fault is a reached. * * * All along the faces of these cliffs, much material fell or slid down to the beach. 24 C 1906 Lawson and others, Along the coast from Mussel Rock to Lake Merced * *_* . Along the 1908, face of these cliffs the Ocean Shore Railway had started a grade at p. 250. an elevation of about 300 feet above the tide level. Along this B bluff a large amount of earth slid down the slopes at the time of the shock. This caving of the banks was due to the nature of the soil, the proximity to the fault-zone, and the disturbance of natu- ral slopes due to the railroad terrace near the top. In places this slope toward the ocean was brought about to the angle of the repose of this material and the roadbed was entirely destroyed for a distance of 3 miles. E C 1906 Lawson and others, Near Mussel Rock part of the roadbed slid for about 500 feet and 1908, on the hillside above the road there was a long crack which was the p. 250. beginning of a slide that might have taken a large part of the hill. B C 1958 Coffman, 1958. December 11. * * * A minor landslide occurred on State High- 1973, way 1, west of Daly City. Magnitude 4.7. p. 179. A 1957 Coffman, 1957. March 22 and 23. * * * State Highway 1, near Mussel Rock, n 1973, was blocked by landslides; highway pavement was cracked extensively. p. 178. * * * a large reinforced concrete reservoir cracked. 133 I] B 1906 Lawson and others, There were several large landslides on both the southwest and 1908, northeast sides of the [Wood's] gulch, and at the ocean the amount p. 250. of dirt that had fallen was very large. A 1906 Lawson and others, On April 25, the writer was on the edge of the cliffs near Wood's 1908, Gulch. About 3 P.M. of that day there was a shock with an inten- p. 250. sity estimated to be between VI and VII. At that time the cliffs shook like so much gelatine, and it was necessary to hold on to a prevent falling. On the north side of the canyon, hundreds of tons of earth fell even with this light shock. 106 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains——Continued Loca- Fig- Fail- Accu- Year of Reference tion ure ure ra— earth— No. No. type cy quake Quotation a C 1906 Townley and Allen, 1939, p. 135. I] B 1957 Bonilla, 1959 p. 34. X B 1957 Bonilla, 1959, p. 35. I] I: 134 C 1906 Lawson and others, 1908, p. 250. (2 Along the top of the cliffs large cracks were formed to a distance of several hundred feet from the edge. Many of these cracks were a foot or even as much as 3 feet in width, and small scarps were often present, 4 or 5 feet high and 20 or 30 yards long. The general ten- dency was for everything to slide into the ocean, but this was not always true. Miniature scarps of more than 6 feet were seen with a downthrow upon the northeast or inland side. The Merced beds, as a whole, were badly shaken, and broke up all along the coast sec— tion. 1906 April 25. 3:15 p.m. * * * Landslides along the cliffs on the coast. Landslides occurred principally in two places: along the sea cliff between Mussel Rock and Alemany Boulevard, and along the shores of Lake Merced. [See 10C. 232.] Cracks resulting from lurching or settlement of artificial fill occurred principally along the highways but some occurred in resi- dential streets and in stony artificial embankments on the west shore of Lake Merced. Most of the cracks along the highways were on the downhill side, were straight, and either paralleled the cen— ter line or trended diagonally toward it. Some, however, were arcuate in plan. In two places in Westlake Palisades cracks were associated with small settlements of the street, sidewalk, and lawns. Small vertical movements occurred along some of these cracks and only horizontal separation in others. Cracks due to lurching or settlement of fills were seen along the coast hi hway north of Mussel Rock; along Chinese Cemetery Road [loc. 134?fi; along Juni- pero Serra Boulevard at the intersection with the cemetery road; in Westlake; and southeast of the intersection of Skyline Boule- vard and Westmoor Avenue. Cracks from landsliding occurred near the tops of the scarps of old or new landslides. These cracks were typically arcuate in plan, commonly showed small vertical movements, and in most places were on the downhill side of highways. This type of crack was prominent along the coast highway and along the cemetery road. Along several of the northeast-trending streets and one northwest- trending street in Westlake Palisades the sidewalks and pavement were arched. The arching may have been caused by compression re- sulting directly from earthquake waves or by some secondary effect, such as downslope movement of the pavement and sidewalks. * * * Another unusual type of crack was observed in the shoulder of the coast highway just south of Woods Gulch. Here the ground was ir- regularly fragmented as though an explosion had been set off beneath it (photo 1). 0n the east edge of the hills west of the Chinese Cemetery and 9- mile house, a line of cracks extends for a distance of about 1,000 yards. These cracks are more than a foot wide in places, and there is an apparent downthrow on the northeast; in one place there is a long line of crusht earth, such as occurs along the main fault-line. Inspection showed that these cracks were caused by a slight land- slide. The line of crumbled earth was due to the earth above it on the hillside sliding slightly, and the crumbling represented a line of buckling of the crust. These cracks are upon the top of a hill, at an elevation of about 400 feet; their general direction is about N. 40° W., and parallel to the San Andreas fault, and the line of hills here has the same general trend. TABLES 5—9 107 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region-'—Continued Santa Cruz Mbuntains——Continued Loca- Fig- Fail- Accu— Year of Reference Quotation tion ure ure ra— earth- No. Noo type cy quake 135 25 B 1906 Lawson and others, Mount Olivet Cemetery (A. C. Lawson). -- Perhaps the best illustra— [2 1908, tion of an earth~flow caused by a sudden accession of water to the p. 392, 393 and 394. incoherent materials of a slope, in consequence of the earthquake shock, is that which occurred in the upper part of Mount Olivet Cemetery, near Colma, 9 miles south of San Francisco. The locality is at the base of the San Bruno scarp, and about 2.75 miles northeast of the San Andreas fault at Mussel Rock. The steep slope of the scarp is underlain by hard sandstone of the Franciscan series, with but a thin veneer of soil, or none at all. At the base of the scarp is the gentle slope of Merced Valley, underlain here by Pleis- tocene and recent sands. The sands, partly eolian, lap up on the lower flanks of the scarp, and mantle the trace of the auxiliary fault which follows its base. The sands thus vary in thickness from a feather edge to an unknown thickness, which it is believed may be as much as a few hundred feet at no great distance from the base of the scarp. * * * At the moment of the earthquake there was a sudden outgush of sand and water at a point at the upper end of the cemetery, close to the base of the scarp and quite near, if not immediately upon, the line of the buried fault-trace. This stream of sand and water, admixed with the loam of the slope, flowed rapidly down the course of a shallow arroyo on a grade of about 1:25 with a depth of from 13 feet in its upper part to about 3 feet in its lower. The front of the stream stopt abruptly at a point just beyond the roadway about half a mile from the origin. The flow was so rapid that it carried away many small trees; a wind-mill was wrecked and the heavy concrete blocks which served for its foun- dation were swept down, with other debris. One of the pumping sta- tions of the cemetery was demolished by it, and 2 horses were car— ried off their feet, and were extricated afterwards with difficulty. plate 130A (See plates 130A, B and 131A.) plate 1303 According to Mr. M. Jensen, the superintendent of the cemetery, the plate 131A entire flow had been accomplished within 3 minutes from the time of the shock, and he was at its source within 20 minutes after it occurred. The height of the flow within a few hundred feet of its source was attested by the mud upon the trunks of some eucalyptus trees near its margin. This mud extended up to 13 feet above the bottom of the arroyo. This, however, doubtless indicates the height of the front of the stream as it past this point. As the flow ad- vanced, its surface near its source rapidly dropt; and by the time the front had reached the roadway the stream was probably no deeper at its source than at its terminus. Indeed, it seems to have been somewhat less, as there was a marked tendency for the sand to pile up at the front by reason of the negative acceleration at the front due to loss of water. After the moving mass had come to rest and partially dried out, it was found that it had left a streak of muddy sand on the bottom of the arroyo averaging 100 feet wide and about 3 feet thick. Taking the length of the flow as 900 yards, this gives the total volume of the compacted wet sand as 89,100 cubic yards. The cavity in the slope caused by the evacua- tion of this sand and loam was not measured, but was estimated to have a width of 150 yards, a length of 300 yards in the direction of the flow, and an average depth of 2 yards, On this estimate, its volume would be about 90,000 cubic yards, which agrees quite closely with the estimated volume of the material ejected. * * * There was no disturbance of the soil on either side of the cavity, even in its immediate vicinity. On the shoulder to the southeast, where the trace of the auxiliary fault passes over practically bare rock, no evidence of movement was detected on critical examination. 108 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6,—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Santa Cruz Mountains--Cont inued Loca- Fig- Fail— Accu— Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake I] B 1906 Lawson and others, Northeast of Mount Olivet Cemetery there was an earth-flow in the 1908, sandy soil at the base of the San Bruno Mountains. The angle at p. 249 and 250. which the materials slid was hardly more than 10 degrees. The sand and water forming this slide came out of a hole several hundred feet long and 150 feet wide, flowed down the hill several hundred yards toward the cemetery, carried away a pile of lumber, and knocked the power-house from its foundations. The front of the mud- flow piled up in a bank when it reached the nearly level ground, and dammed up the mass behind it. The earth was harder several weeks later than it must have been at the time of the flow, but it was still slushy and there was still a little water flowing along the path of the earth-flow, coming from a small spring where the slide originated. TABLES 5—9 109 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San Francisco Bay, Santa Clara Valley, and east bay hills Loca- Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake 136 A 1906 A 1906 the base of the V pointing northeast, the lateral displacement being about 1.5 feet. These are about 60—lb. rails, and at the V-shaped bend mentioned the rails were broken in three places. C 1906 Duryea and others, Of the three, long, riveted-iron, pipe lines leading from storage 1907, reservoirs into San Franicsco, the middle one, 36 in. in diameter, p. 252. leading from San Andreas Reservoir, was fractured at one point only. This was at Baden, where the pipe crossed a piece of marsh land on a wooden trestle. B 1906 Lawson and others, Just east of the station at Baden, where a creek crosses the county 1908, road, there were cracks in the filled soil * * * . p. 248. m C 1906 Gilbert, Humphre , Near Baden the line had been telescoped 42 inches, shearing off Sewell, and Soule, an 8—inch gate valve. The [University Mound] reservoir itself was 1907, undamaged, yet its three days' supply was rendered useless by the p. 18. breaks in the cast-iron distributing mains. [See also loc. 140.] 138 B 1906 Lawson and others, North of San Bruno Point, at the Southern Pacific tunnel along the 1908, bay shore cut-off, no damage was done, except for the sliding and II p. 248. settling of the debris in the newly filled area. 139 C 1906 Lawson and others, The buckling of the tracks of the South San Francisco car line 1908, between the town [South San Francisco] and San Bruno Point * * * p. 248. is significant of the contrast in the intensity of the shock at the <) two places. The rails are bent and broken in a number of places, where the track crosses the marsh between the two places. The difference of intensity is striking when it is taken into consid— eration how close they are together. 140 m A 1906 Schussler, One feature of the destruction of the bridge and pipe across the 1906, San Bruno marsh was that some of the pipe was thrown to the west p. 31. and some to the east as much as four or five feet. * * * We re-established by careful survey, the straight line and grade of pipe, and found that neither the original straight line of the piles nor the grade of their tops had been disturbed by the earthquake. m A 1906 Duryea and others, * * * [The break in the 44-in. Crystal Springs pipeline] was the 1907, p. 252. most extensive which occurred at any point about the Bay, save those immediately on the fault line * * * . This was at the crossing of the San Bruno Marsh, where, for a distance of about 2 000 ft., the pipe was supported upon trestle bents resting on a pile foundation. The pipe was pulled apart at a number of places, and nearly all of it was thrown entirely off the supports on either side. TABLES 5—9 1 1 1 TABLE 6,—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San Francisco Bay, Santa Clara Valley, and east bay hiZZs--Continued Loca- Fig- Fail- Accu— Year of Reference Quotation tion ure ure ra— earth— No. No. type cy quake 1906 Gilbert, Humphrey, In company with Herman Schussler, chief engineer of the Spring Valley m Sewell, and Soule’, Water Company, the writer made a detailed examination of the principal 1907, conduits and reservoirs. 0n the San Bruno marsh the 44-inch line to p. 18. the University Mound reservoir had been thrown off the trestle for a distance of 1,300 feet; and, while the pipe was readily repaired, the trestle had to be rebuilt, as many of the timbers had rotted. m 1906 Derleth, The Crystal Springs conduit * * * is ruptured in a number of places, 1906b, but mainly where it crosses the marshes. The worst destruction has p. 551. occurred in a distance of about 1,600 ft., where the pipe crosses a marsh between San Bruno and South San Francisco. In this place the pipe rests upon a wooden floor, supported by pile bents. These piles on the average penetrate the mud to a depth of about 40 ft. * * * the pipe was alternately thrown from one side to the other of the trestle floor and its box covering was generally smashed. 1906 Lawson and others, From South San Francisco to San Bruno, there is a line of big steel m 1908, water-mains, supported on a trestle frame, where it crosses the p. 248. marsh. This line did not break, but was bent and twisted into S— shaped figures. 1906 San Jose Mercury, San Bruno. The approaches to the bridges between San Bruno and )( 1906b. South City were sunken, making the bridges difficult of passage. 1906 Lawson and others, Near San Bruno, where the county road crosses a small stream, there 1908, were numerous cracks in the ground from 3 to 10 inches wide, parallel p. 247. to the line of the road, which is N. 25° W. The road at this place was built 8 feet above the mud flats, so that these cracks are I[ plate 97C accounted for by the settling of the fill. * * * Plate 97C illus- reservoir toward Menlo Park had been pulled apart. C 1906 Hyde, PIPES DESTROYED BY UNEQUAL SETTLEMENT.—-Wherever filled ground 1906c, existed, settlement in greater or less degree took place as the p. 767. result of the temblor. * * * Southward, in the vicinity of Palo Alto, where the effects of the earthquake were very pronounced, it is interesting to observe that the long 12-in. cast-iron pipe, known as the Searsville line, supplying the Leland Stanford, Jr., University, was broken in three :: places and in addition received many smaller cracks, principally along the bottom. 144 <—[]-> B 1906 Schussler, Neither the Sunol filter beds, on the Alameda Creek System, nor 1906, the Sunol Aqueduct and 36-inch pipe line, on the east side of the p. 32. bay, nor the four submarine [under San Francisco Bay] pipe lines were injured; only a slip—joint, on one of the two 16-inch shore connections, was pulled apart several inches, on the east side of the bay * * * . E] A 1906 Gilbert, Humphrey, (See the maps, Pls. LVI and LVII.) Some subaqueous pipe lines Sewell, and Soule, crossing the bay seem not to have been injured. 1907, p. 116. 145 m C 1906 Lawson and others, People reported new holes formed in the slough near Cooley's 1908, Landing, but their statements were not verified. p. 259. 146 A 1906 Lawson and others, In one case a wind-mill (at 6, map No. 22), which had been in use O 1908, for years to pump water from the well, was no longer found necessary, p. 260. but the artesian water was muddy. 147 B 1906 Lawson and others, At the Ynigo Ranch, 3 miles northeast of Mountain View Station, @ 1908, there was an artesian well which had, before the shock, flowed p. 408 and slightly or not at all, and a wind-mill was used to raise the water. II p. 261. After the shock, it was found that the casing had been shoved up 2 feet, damaging the pump. The flow of water was increased, and black sand was brought up. Another well at this ranch was unaffected. Along the Jagel Landing road, 2 artesian wells had increased pressure after the shock. An old artesian well filled with stones had begun to flow for the first time in several years. 148 C 1906 Salinas Daily Index; On Alviso slough two rows of trees in an orchard had parted off 1906b. and slid into the slough. TABLES 5—9 113 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San Francisco Bay, Santa Clara Valley, and east bay hiZZS--Continued Loca— Fig— Fail- Accu- Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake O C 1906 Taber, A well near Alviso, at the head of the bay, formerly required a 1906, wind-mill to pump the water. At the time of the earthquake the II p. 315; casing was driven 2 feet out of the ground, wrecking the pump, and also in since that time the well has been flowing under a heavy pressure. () Jordan, In some of the lowlands small cracks were formed, out of which water 1907, issued, bringing up mud and sand. p. 279. II C 1906 Ransome, [Railroad grade subsidence of several feet] is reported from Alviso, 1906, a town at the south end of San Francisco Bay. p. 294. 149 A 1906 Weatherbe, * * * some rather remarkable demonstrations of the earthquake have 1906, occurred along the levee paths following the Coyote river north from p. 402. the bridge on the road from Milpitas to Alviso. * * * At the locality o mentioned large fissures, as much as eight feet wide and of nearly equal depth, have been opened and as partial filling ensued immedi— ately, they must have been of much greater depth when first formed. 0 In some places the road has been completely precipitated into the creek and at a point about half a mile below the bridge both the banks and the bed of the stream, including a heavy growth of willows, have been cut by a series of parallel cracks and the trees and banks thrown into the stream, thus forming a partial dam. All of these cracks are roughly parallel with the stream * * * . C) Simultaneously with the above phenomena, dozens of small geysers or spouting craters were formed along the creek and in the adjacent fields. The mouths of these varied in diameter from three inches to about 15 in., though the actual orifices probably do not exceed four or five inches in diameter. Mud and water were spouted to a height of over twenty feet, and continued to flow for several days. On some of the miniature craters incrustations of salt were deposit- <)> ed. The bridge above-mentioned was shifted on its concrete supports, the two ends moving in opposite directions, and throughout the same locality rows of trees in the orchards are said to have been twisted and staggered out of shape. 27A C 1906 Lawson and others, Cracks like those which were observed in the ground on the 1908, Milpitas-Alviso road reappeared on both sides of the Coyote River p. 282. at intervals all the way to San Jose. Altho they occur in a general north—south direction, it seems probable that their origin was due to the unstable condition of the alluvial deposits which underlie the valley. B 1906 Lawson and others, From 1,500 to 2,000 feet west of the bridge over Coyote Creek,

1908, cracks cross the road in front of the Boot ranch—house, and several p. 281, of them occur in the road leading to that house. (Plate 140B.) plate 1403. Some of these cracks are about 6 inches wide and have a general bearing of N. 43° W. Immediately after the earthquake, water flowed () from some of them and brought up sand, which was heapt up about 6 inches high. The water ceased to flow after the second day. 273 II Near the dwelling house on the Boot place, the ground settled 11 inches on the east side of the crack. The fissures past under the corner of the dwelling house and the building was partly thrown from its foundation. The cellar beneath it was filled with water to a depth of from 2 to 3 feet. There is a capped artesian well in the yard of this house, and about this well is a pool of water 12 feet across. The west side of the pool was lifted 1 foot higher than the east side, and fish were thrown out of the pool. A hundred feet east the fissures past under the barn, and the ground settled on the west side. Water flowed from cracks in the yard and piled up sand 6 inches high on both sides. 114 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6,—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San Francisco Bay, Santa Clara Valley, and east bay hiZZs-—Continued Loca- Fig- Fail— Accu— Year of ra— earth— cy quake tion No. ure No. ure type Reference Quotation 150 28 ‘9 O I C 1868 C 1868 B 1906 C 1906 B 1906 plate 140A plate 143A Lawson 1908, p. 446. Lawson 1908, p. 444. Lawson 1908, p. 280. Carey, 1906, p. 297. Lawson 1908, p. 281. and others, and others, and others, and others, People living near Coyote Creek state that the water rose between 2 and 3 feet immediately after the earthquake; and up to April 26 the water in this stream had not returned to its normal level. At the bridge over Coyote Creek, on the Alviso—Milpitas road, the concrete abutments were thrust inward toward each other about 3 feet. A pile driven in the middle of the stream, which had been cut off below the water—level, was lifted about 2 feet and now rises above the water. About 150 feet north of this bridge the banks of the stream cracked, the fissures running parallel with the channel and the land on the creek side sliding toward the stream. (Plate 140A.) West of the stream, in an adjoining field, water rising thru cracks built up many craterlets of sand. (Plate 143A.) Residents of the vicinity state that the water rose 3 or 4 inches above the tops of these craterlets while they were being formed, and that it ceased to flow toward the end of the second day after the earthquake. In the road running northward along the west side of Coyote Creek from the bridge, many large cracks opened. Five hundred feet north of the bridge the cracks were 2.5 feet wide and 3 feet deep when the place was visited April 26. Farther north the cracks were very abundant, nnstly parallel with the road, and some were 4 feet deep and 3 feet wide. A quarter of a mile north of the bridge, the whole road was shoved eastward into the channel of the creek, and with it a large number of willows and cottonwood trees that grew along the banks. Just south of this place the road was broken up badly for a distance of 300 feet. One of the largest cracks in the road was 5 feet wide, 6 feet deep, and more than 100 feet in length. The bearing of the fissures at this place was N. 23° W. For the most part the principal features were approximately parallel with Coyote Creek. Where the Milpitas road crosses Coyote River, the banks were shaken together and the river-bed filled up. Milpitas.—-Along Coyote Creek the ground was cracked from Boot's ranch to the San Francisco Bay, the cracks being on the bay side and following the winding of the creek. As in 1906 much water was ejected from the cracks, and Coyote Creek rose. (W. Bellou.) Of the two bridges over Coyote Creek, the northern one suffered some damage by displacement of end supports. It was unsafe to travel over at the time of the visit. The southern bridge [100. 154] was found intact, the end supports showing signs of but small movement. Near Milpitas along the Coyote Creek artesian water gushed out of cracks and holes in the ground, making temporary fountains several feet high, bringing up silt and forming cones about the outlet, as shown in photograph. Similar outpourings of water are reported from the Salinas valley. In some cases artesian wells stopped flowing, but in general the water level was raised, so that many ordinary wells overflowed. The water in most cases was more or less muddy. At Mrs. North Whitcomb's ranch, on the south side of the Alviso- Milpitas road, between Coyote Creek and Milpitas, the prune orchard was cracked and the ground shifted at several places. The ranch- house, of concrete with a wooden upper story, was cracked across the northwest corner and settled slightly on the northwest side. In the back yard were fissures 1 foot wide, running about N. 13° W., with a downthrow of 1 foot on the east side. Some of the prune trees in the orchard are 2 feet out of alignment, and some as much as 6 feet. The lateral displacement here shows a relative movement of the south side toward the east. Considerable sand was brought up by water flowing from the cracks in this orchard. TABLES 5—9 115 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San Francisco Bay, Santa Clara Valley, and east bay hillsuContinued Loca- Fig— Fail- Accu- Year of Reference Quotation tion ure ure ra— earth- No. No. type cy quake 151 (j C 1868 Lawson and others, Cracks in the vicinity of Milpitas flowed artesian water for 48 1908, hours after the shock. (Mr. Durkee.) p. 444. <)- C 1906 Jordan, [Picture caption] Slump in Soft Ground, Milpitas. 1907, p. 66. C 1906 San Jose Herald, The district between Milpitas and Alviso is seamed with immense C) 1906b. fissures from which water is pouring. The country is being rapidly inundated. The road between Milpitas and Alviso is beneath water <’> in many places, making passage impossible. The bridge over the creek is down. Sections of land varying from small plots to entire ]( fields have sunk to depths of six inches to five feet. 152 C 1906 Lawson and others, Milpitas—San Jose Road (G. F. Zoffman).——About 0.5 mile south of = 1908, Milpitas, on the Milpitas—San Jose road, cracks were formed across p. 282. the road. They did not, however, appear to have any definite direc- tion, and were so small that no lateral movement was discernible. C 1906 The Evening Post, The roads skirting the bay were slightly injured in places by 1906a. fissures, but no serious faulting of the underlying rock was ob— served. * * * From San Jose to Alviso the road is lowered in places by the shocks, and the front of the principal hotel at Alviso has sunk at least ten feet. 153 <> C 1868 Lawson and others, I was told at the time that the water spurted up in the streets of 1908, San Jose, and out in the road between Milpitas and San Jose, to the p. 444. height of several feet. 154 A 1906 Lawson and others, On the north side of the bridge which crosses Coyote River, on the 1908, San Jose-Milpitas road, some cracks were found but they were evi— 0 p. 282. dently caused by the sliding of the banks. The bridge was not damaged. 155 B 1906 Lawson and others, (M. Connell.)—-0n the farm of Mr. Fox, 3 miles north of San Jose, O 1908, the water pipe of an artesian well was broken off 60 feet below the 0 p. 286. surface and carried by the heave of the land in a northwesterly direction 4 feet from its original position. 156 «4C}+ C 1906 Lawson and others, Alum Rock Road (G. F. Zoffman).-- * * * No cracks were found 1908, between Coyote Creek and the mountains * * * . p. 282. 157 E] C 1906 Jordan, San Jose's water works, like that of Santa Rosa, was not injured; 1907, its sewers also were left intact, showing that there was no unequal p. 188. displacement of the ground. E] C 1906 Lawson and others, (E. C. Jones.)—-There was only one broken gas main in San Jose, 1908, caused by the high wall of the building falling over; the bricks p. 285. penetrated thru the soft earth to the main and broke it. 158 C 1906 Lawson and others, Water and mud in many instances are reported as having spurted O 1908, from the artesian wells, but in a few days they resumed their normal p. 284. condition. * * * m Data‘were obtained of the directions in which the chimneys fell thruout the town. After the data were collected and tabulated as shown below [2710 chimneys], it became evident that chimneys usually fell with the slant of the roofs. 116 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San francisco Bay, Santa Clara Valley, and east bay hills--Continued Loca- Fig- Fail— Accu— Year of Reference Quotation tion ure ure ra— earth- No. No. type cy quake C 1865 Holden, 1865. October 8. The streams at McCartysville and Los Gatos have 1898, risen greatly since the earthquake, tapping the sources of the G p. 67. artesian wells in the Santa Clara Valley, many of which have ceased to run since the earthquake. 159 B 1906 Lawson and others, Just northwest of the 12-mile house, where the county road crosses C: 1908, to the Fisher ranch, there were cracks from 2 to 6 inches wide in p. 286. the coarse gravelly bottom of the Coyote River. There was evidence of water having been ejected from these cracks, as there were heaps (3 of clean, fine material surrounding small orifices. It was said at the ranch-house that muddy water came out of these openings follow— ing the shock. 160 C 1906 Lawson and others, Along the road down Penetencia Creek, a considerable amount of a 1908, debris had slid into the road, in many places obstructing all travel p. 283. except for pedestrians; but no evidence of cracks could be found. 161 C 1906 Lawson and others, Between this place [Calaveras Valley] and the head of Alum Rock I: 1908, Canyon, the residents stated that cracks appeared across the road p. 282. in several places; but altho this was in the proximity of the Calaveras Valley fault-line, which passes thru this region, it was not possible to verify their statements. 162 C 1906 Duryea and others, Between Niles and San .1056, on the Southern Pacific, there was at <3} 1907, one point a displacement of 3 ft. horizontal, but the vertical dis— p. 258. placement was only 6 in. 163 C 1868 Lawson and others, On the mountain above the old Mission, just above a place called 1908, Peacock Springs, a great crack in the earth appeared, which lookt I] p. 444. as if the lower part of the mountain had parted and slipt down. Many times I have crost the bridge which was built over the crack, and stopt and thrown rocks down to see if I could tell how deep it was. (Mrs. N. Ainsworth.) 164 *4:F* B 1906 Lawson and others, According to the track-boss, the railroad track suffered no dis- 1908, placements anywhere between Niles and Irvington. p. 304. 165 A 1906 Lawson and others, While at Niles, a visit was made to one of the new tunnels of the 1908, Western Pacific Railway, which is about 1 mile east of Niles in the p. 306. Niles Canyon. The tunnel had penetrated about 130 feet into the hillside, but had not yet passed thru anything but a sandy clay During the previous winter the walls at the portal, and also on the inside, had stood without timbering. Since the earthquake it had been impossible to break out more than 4 feet of ground ahead IX] of the timber sets without caving taking place. There had been an apparent movement in the soil which had removed its consistency and made it incoherent. The amount of water present in the tunnel was perceptibly changed. The foreman said that there was more water since the shock than there had been even in the wettest part of the winter. I] C 1933 Coffman, 1935. May 16. Niles Canyon * a: * Landslide * a: * . 1973, p. 169. TABLES 5—9 117 TABLE 6.—Specific descriptions of ground failures in the ‘San Francisco Bay counties region—Continued San Francisco Bay, Santa Clara Valley, and east bay hills-~Continued Quotation Loca- Fig- Fail— Accu- Year of Reference tion ure ure ra- earth- No. No. type cy quake 166 B 1906 Lawson and others, 1908, l: p. 309, plate 141A. B 1906 Lawson and others, 1908, p. 308, plate 141A. :3 167 B 1906 Lawson and others, 1908, :3 p. 309. (3 168 C 1861 Holden, fl 1898, p. 58. C 1861 Coffman, D 1973, p. 157. 169 <) C 1906 Lawson and others, i 1908, p. 280. I: C 1906 Oakland Tribune, 1906a. (D At the Alviso ranch, a little over a_mile north of the town [Livermore], the top of a small hill the earthquake. The breaking of the was broken up at the time of ground did not consist of fissur- ing along a line, but was in the nature of an uplift of a limited area. There were 3 fairly well marked concentric rings where the ground had broken, the inside ring in each case outside ring. being forced higher than the The effect was similar to that obtained by placing three plates of different sizes within each other. * * * Mr. Still reports that where the ground was deformed in con— centric ridges, as described by Mr. Matthes and Mr. Crandall, there was an alkaline spring years ago. [A somewhat similar phenomenon was seen on Cahill's ridge (locs. 116, 120) in San Mateo County.] An interesting feature appears 0.25 mile north of Meyn's ranch, west of the road leading north from Livermore, about 2 miles north of that place. It is on the summit of a smoothly rounded hill, sloping gently down to an even, peaty meadow traversed by the arroyo of Cayetana Creek. * * * The summit of the hill a series of concentric deformations, A number of nearly concentric cracks into a sort of panhandle, along each the soil had apparently taken place. in question was found crowned by rising stepwise above one another. were found extending northward of which an upward movement of The uplift along the 2 princi- pal cracks was found to be 19 and 16 inches, respectively. Along the minor cracks the vertical displacement amounted to an inch or two only. The surface of each step or bench was found to slope inward, and in some places the edge even appeared to have curled inward. Santa Rita, 3 miles east of Dublin (F. E. Matthes).--A small, flat levee along the east bank of Tassajara Creek, immediately north of the main road, showed several somewhat crescentic cracks along which the ground had slipt down and toward the creek from 1 to 3 inches. These cracks extended farther south, according to local settlers, and crost the road; but this was no longer traceable at the time of the visit. 1861. July 4?; 16h. 11m. * * * in the San Ramon Valley * * * . It opened a large fissure in the earth, and a new spring of water. 186]. July 3. Contra Costa and Alameda Counties. Severe * * * . In San Ramon Valley a fissure opened, and a new spring of water appeared. The track suffered a slight shifting in several places north of the village [Newark]. Cracks opened in the ground in the vicinity of 2 small watercourses, but on a less extensive scale than that noted at Alvarado [1°C. 170]. Some of them crost the railroad track. In every case they emitted the same bluish sand (with the water) that had been found near the Alameda Sugar Mill. In one place, 1.5 miles northeast of the village, considerable water was still left standing in shallow ponds. According to neighboring ranchmen, these ponds had not existed prior to the earthquake. Newark, April 18.~-About a mile north of this town a fissure was opened by the earthquake. This fissure is about a mile and a half in length and from eight to twelve inches in width. From the fissure quantities of water are being emitted, although the land is in what might be termed a dry district. 118 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San Eraneisco Bay, Santa Clara Valley, and east bay hills-—Continued Loca- Fig- Fail— Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake C) C 1906 The Bulletin, Fissure created by earthquake near Newark becomes running water, and 1906. pipe wells become gushers spouting twelve feet. One of the most peculiar freaks of the great temblor of last Wednesday C: morning is a fissure in the earth a short distance from Newark, several feet wide and about a mile and a half long, running with an excellent quality of water. This river was first noted by the crew of one of the Southern Pacific Company's trains coming up from San Jose, and investigation discloses O that many of the old pipe wells in the vicinity have suddenly gushed forth, in some instances the water spouting from eight to twelve feet into the air. 170 B 1906 Lawson and others, The [Alameda Sugar] mill stands on flat, alluvial ground 100 feet 1908, north of Alameda Creek. Along the banks of the latter a large number p. 305. of cracks extend, roughly parallel with the stream. Considerable O masses next to the stream-bed slumped toward the same, leaving gaping cracks 1 to 2 feet wide, and carrying with them small outlying build- ings, notably the fire-engine house, which moved bodily, concrete foundation and all, 2 feet south toward the creek. A small railroad trestle southwest of the mill moved 4 inches south on both of its 0 abutments, probably owing to slumping of loose ground on the north side of the creek. A 2—inch water-pipe, laid under the ground some 60 feet north of the creek and almost parallel with the same, shows m indications of having been submitted first to tension, causing rup- ture at one of the joints, then to sudden compression, causing it to be jammed together with violence. Cracks in the ground may be found as far as 250 feet from the creek. They were nearly all closed at the time of the visit (May 7), but (3 were easily traced by the streaks of bluish-gray sand which has issued from them, together with considerable quantities of water. According to the Chinese cook of the superintendent, the cracks nearest to his dwelling opened and closed several times in succession during the quake; and large volumes of mud-laden water gushed from them, splashing up some 10 feet in the air at each closing. A large crack of this kind opened under the northwest corner of the dwelling and the superintendent estimates that fully 500 gallons of water gushed from it, the flow continuing with decreasing volume for about an hour. The fence in front of the house shows that the ground there has been raised into a low hump. The sewer pipe leading west to the creek was detached from the house by a space of 22 inches. * * * () In the roadway south of the mill, water oozed out in a number of places, without the production of visible cracks. The water pipes and hydrants in this vicinity were crusht in several places. * * * A few cracks opened across the streets [of Alvarado], but these had been filled on the date of the visit. () C 1868 Lawson and others, Alvaradb. * * * The ground opened in several places and water issued. 1908, p. 443. B 1906 Duryea and others, At Alvarado, near the pumping station, the 30-in. riveted—iron force 1907, main conducting the water to Oakland crosses the Alameda Creek on a p. 254. short bridge. * * * In this pipe was placed a 24-in. stop-gate rest- ing on a concrete foundation. The piles under the pipe were not deeply II placed; and the shake operated to settle them irregularly, some going down as much as 6 in. B 1906 Hyde, ALVARADO FORCE MAIN.—- * * * The earthquake threw this pipe out of 1906c, alignment into a decidedly sinuous location for a distance of about p. 766. one-half mile. The flanges of the gate valve were broken by the )( unequal settlement of the trestles carrying the pipe and the concrete pier on which the valve rests. TABLES 5—9 1 19 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San Francisco Bay, Santa Clara Valley, and east bay hills—-Continued Loca— Fig— Fail— Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake B 1906 Derleth, The Alvarado force main, a riveted pipe, was wrenched in two places. 1906a, Three hundred feet from the pumps the pipe snapped, and a quarter p. 503. of a mile further on toward the city, in which distance the pipe lies upon piling, the flange of a large gate valve was broken by the )[ settling of the piles. B 1906 Duryea and others, [At the pumping station] on the marsh land near the Bay shore, one 1907, mile west of Alvarado. * * * II p. 249 and * * * the foundation settled about 2 ft., breaking all the pipe m p. 250. connections. During the quake the channel of the creek disappeared, its bottom being raised to the general level of the adjoining land. 171 ‘D B 1868 Lawson and others, The ground opened from 6 inches to 2 feet, and water with sand was 1908, ejected to a height of from 1 to 3 feet. North of the village p. 443. [Hayward] a ridge of ground 3 feet wide was raised 2 feet. * * * 0n the hills there were several new springs. B 1868 Lawson and others, The crack below Haywards Hotel was 12 inches wide. It ejected C) 1908, water and white sand. A fence which traversed a hill from north to p. 442. south was crost by the crack, and had the ends of the boards loosened from the posts. Gradually these boards lapt over one another, until within a couple of weeks they overlapt several inches, the progress of the overlapping being noted from time to time by a pencil mark. The "cap" board of the fence was also archt up in consequence of this III movement. Large waves were set up in the soil. The house' was moved southward, while a neighbor's was tipt northward. (D. S. Malley.) c) C 1868 Lawson and others, On B Street the ground opened about 2 inches, and water and sand 1908, were forced from the opening. Some springs were closed, while others p. 442. were opened or made to flow more freely. C 1868 Lawson and others, In the vicinity of Haywards it is reported that there were two 1908, branch cracks from the main one [fault?], trending off into the hills. C) p. 435. Water and sand were ejected from the crack in one place. 172 I] C 1838 Wood, Allen, 1838, late in June. * * * VIII at least. * a: * Landslide at San and Heck, Leandro. 1939, p. 3. C 1868 Halley, The bed of the San Leandro creek, which had been dry for several 1876, months, is now coursed by a stream of water six feet wide and one p. 263. foot deep. Back of San Leandro, in the mountains, there are numerous I: fissures in the earth, from which came clouds of dust, and from some ‘3 have come great volumes of water which flows into the San Leandro Creek. B 1868 Halley, A Mr. Davis, who resides on a farm near San Leandro, informs the 1876, Bulletin that the workmen on his farm at the time the shock occurred, p. 267. observed that the ground was disturbed and thrown about with a rapid and violent rotary motion, which continued several seconds. A creek (D running through the farm, and which was nearly dry, rose instantly I: to the depth of about three feet, and several deep gulches were formed in the plain. 173 )l B 1906 Lawson and others, * * * at Mills College * * * In the made ground there was a drop 1908, p. 304. of from 1 foot to several feet. C 1906 Alameda Daily Argus, [At Alameda] * * * damaged tracks in the marsh. 1906. One of the new tracks of the Oakland Traction Company across Webster St. sunk about four feet and the rails are twisted, blocking traffic. 120 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.——Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San Francisco Bay, Santa Clara Valley, and east bay hills——Continued Loca- Fig- Fail— Accu- Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake 174 m C 1906 Derleth, The [water] distribution system of East Oakland seems to be prac— ‘ 1906a, tically unaffected, but in West Oakland, upon filled ground, some p. 503. of the smaller pipes and some of the service connections have been broken, but the damage is, relatively speaking, slight. C 1906 Alameda Daily Argus, Bay Farm Island shows many crevices and cracks on the surface. 1906. C 1906 The Evening Post, Railroads are inactive, and wires are useless. Railroad tracks 1906b. across the marsh are twisted. * * * More than 600 feet of the ][ track of the Oakland Transit Comapny sank nearly four feet. m B 1906 Duryea and others, One 12—in. cast-iron pipe near the water front in East Oakland was 1907, drawn apart at the joint, and one or two breakages occurred on a p. 254. long line of 8-in. pipe leading to the Southern Pacific Railroad Broad—Gauge Pier. B 1868 Holden, 1868. October 21; IX. Oakland. * * * The draw of the railroad 1898, bridge was thrown twelve inches out of line. p. 76 and 77. B 1906 Oakland Enquirer, The magnesite works at the foot of Ninth avenue in East Oakland have )[ 1906. sunk several feet and now the ground on which they stood is under water. The sinking is probably due to the Violent earthquake of Wednesday morning. The land on which the works stand is very unsub— stantial, being simply built up by the process of dredging. C 1906 Leslie’s Weekly, [Picture caption] Frame structure which was cracked and wrenched )[ 1906a. and which sank for several feet into the yawning ground. B 1868 Halley, The drawbridge on the line of the S. F. and O. R. R. was thrown out ‘1> 1876, of place about eight inches, and as the locomotive and nearly all the p. 266. cars were at San Antonio, no train left Oakland at 8 o'clock. 175 A 1906 Derleth, The earthquake did not produce, relatively speaking, much destruc- 1906a, tion to these [Oakland, Alameda, and Berkeley water systems] works. p. 503. Breaks in the pipe lines invariably were found upon soft ground, or where the pipe lines passed from soft and yielding to more rigid foundation. * * * The 24-in. steel pipe crossing the Twelfth St. ll dam at Lake Merritt was also snapped from the settling of the flood gates, but the 37 1/2—in. pipe running parallel and slightly to the east, across the same soft foundation bed was only slightly deformed. II A 1906 Lawson and others, On the Twelfth Street dam, a cast-iron pipe was broken and displaced 1908, over a foot; while the high pressure steel pipe paralleling it was p. 302. practically undisturbed. A 1906 Oakland Tribune, Menaced by Water. The tide at the time of the earthquake was at its 1906b. lowest, and at the time of writing it has turned and running into Lake Merritt. What will happen when the lake is filled and the back- ing of the water pressures with great force on this portion of the dam is something that can not be foretold. The rock foundation of the dam at this point is cracked and broken and gives indications of giving away at any time. REPAIRS DAMAGE Secretary Hanson of the Contra Costa Water Company was early on the scene and with a force of men began at once to repair the water main so that the people of the city could have water for the protection of their homes in case of fire. It is not expected that water will be turned on before this afternoon The great main has sunken with the street and broken in two and parted for the space of several inches and the pipe will have to be uncovered and a new length put in or the pipe drawn together. TABLES 5—9 1 2 1 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San Francisco Bay, Santa Clara Valley, and east bay hills-~Continued Loca- Fig— Fail- Accu- Year of tion ure ure ra- earth- Reference Quotation No. No. type cy quake 0 Along the west shore of Lake Merritt the bank has been cracked and broken and caved off into the lake, showing the force of the shake at the water level. 176 C 1906 Duryea and others, A 24-in. riveted pipe lying in a street leading across a tide marsh <)* 1907, in Oakland was pulled apart 5 in. and displaced 8 in. laterally by )K p. 254. the settling of the entire Street. N C 1906 Rickard, * * * the lower alluvial flats of Oakland and Berkeley were seriously 1906a, disturbed * * * . p. 271. m B 1868 Wood, M.W., October 21, 1868. [Oakland] Portions of the wharves were carried 1883, away in some instances, while walls were cracked in almost every P. 665. house * * * . 177 A 1906 Bronson, Southampton Shoals Lighthouse in northern San Francisco Bay sits m 1959, firmly on piles that were tilted eleven degrees off vertical during p. 185. the quake and never righted. 178 C 1906 Salinas Daily Index, The Santa Fe's Condition. Santa Fe officials report that their line 1906c. at Point Richmond had suffered greatly from the temblor. * * * A E landslide choked tunnel No. 2 near Point Richmond but this was cleared by 6 o'clock. * * * N All water mains in Richmond and Point Richmond were broken. 179 a B 1906 Lawson and others, San Pablo earth-slwrp.--At the time of the earthquake a landslide 1908, occurred on Mills' ranch, which is about 4 miles east of San Pablo. p. 391, The slide is interesting from the fact that a previous geological plate 128A, mapping of the region indicated that the point where it occurred was plate 128B. on the line of a fault extending in a northerly and southerly direc- tion through the Sobrante Hills. The slide was examined by Mr. E. S. Larsen, who describes it as follows: E There are many other landslides in this vicinity, showing that the country is subject to such slides. In this particular case, one of the Castro boys informed me that the main part of this slide began during the winter rains, and had fallen a foot or more during these rains. The balance of the fall occurred the morning of the earth— quake. The slide is on the east slope of a steep hillside and extends from the top of the hill nearly to the bottom, about 400 feet on the slope. The width is about 1,500 feet.r At the northeast corner the scarp is greatest, reaching perhaps 50 feet. It gradually decreases, and is very slight for the southwest 700 feet. On this southwest 700 feet the only evidence of a slide is the crack near the top of the hill. The north 800 feet of ground shows every evidence of sliding. The dry ground is much cracked, and these cracks extend up and down the hill near the scarp and along the hill where the ground has been piled up. In some places there is a net- work of cracks. 0n the south side of the main slide the ground has piled up about 10 feet. This extends along nearly all of the south side, and this tendency to pile up to the south is shown in other places. Moreover, the north side shows that the ground has pulled away toward the south. The above shows that the movement was not directly down the hill, but was more to the south. The formation is sandstones and shales, with considerable soft surface soil. The same slide was subsequently visited by Mr. F. E. Matthes, and the following descriptive note is by him. (See figs. 68 and 69.) [See figs. 68 and 69 at end of tables.] 122 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San Francisco Bay, Santa Clara Valley, and east bay hiZZS--Continued Loca— Fig- Fail- Accu— Year of Reference tion ure ure ra- earth- No. N0o type Cy quake Quotation 180 181 182 183 184 1906 1965 1906 1906 1906 1906 1906 Lawson and others, 1908, p. 310. Coffman, 1973, p. 182. Salinas Daily Index, 19060. New York Tribune, 1906. The Evening Post, 1906d. Lawson and others, 1908, p. 195. Lawson and others, 1908, p. 75. The slip occurred east of a high ridge at the southern end of the Sobrante Hills. It covers the northeast half of an area whose terraced nature is indicative of a former landslide of much larger dimensions. The accompanying sketches show the general outlines, and a cross—section of the slide. It will be noticed that the slide does not extend all the way down the slope, its lower edges being fully 100 feet or more above the bottom of the gulch. The lower slopes were not materially changed, and but little debris fell into the stream-bed. A steep scarp has been produced east of the crest of the ridge. The downslip along this scarp does not exceed 50 feet, and decreases both to north and south. Along the north edge there has been a marked movement down and southward, the scarp there averaging 10 feet. Along the south side, on the other hand, the loosened mass had advanced over the old surface, presenting a bulging and cracked frontal scarp some 6 feet high. It appears from this that the movement took place, not along the line of greatest declivity, but in a direction somewhat more southward, as indicated by the arrow. The 2 hummocks probably existed before the slip occurred, but judging by their greatly cracked and rent surfaces, it seems likely that their height has been slightly increased. The main crack, which extends southward from the upper scarp, continues along the hillside in irregular zig—zags for some 300 feet south of the slide. (See plate 128A, B.) On a hillside above Peach Tree Spring, on the west side of Mount Diablo and very near the contact of the Knoxville shales and the Franciscan, a crack opened in the ground about 30 feet long, in a north and south direction, gaping 4.5 inches. 1965. September 10. * * * Minor rockslides were observed at Mount Diablo * ~k * . The big bridge at Middle River between Point Richmond and Stockton sank three feet and was shoved out of line. Stocton, Cal., April 18, —- * * * The Santa Fe bridge, over the San Joaquin River, settled several inches. [See also loc. 254.] Santa Fe Bridge Over the San Joaquin Settles. Stocton, Ca1., April 18.-- A sharp earthquake shock was felt here at 5:15 o'clock this nnrning. The Santa Fe bridge over the San Joaquin River settled several inches. At Point Reyes Light-house * * * . One of the light-house keepers stated that after the shock he lookt from the window of his room, which commanded a portion of the sea near the beach, and saw the water "boiling," but there was no change of the nature of a wave. About 6 miles farther south, at the head of Pine Gulch Creek, another road crosses the range, and in following this a group of cracks was seen. A short distance west of the divide, and about a mile in a direct line from the fault—trace, is a fault—sag trending northwest- southeast. On each side of it a crack was seen, the eastern crack being the wider and showing a small throw to the southwest. This crack was traced for about 0.75 mile and found to curve thru an arc of nearly 90° from southeast to southwest. At its southwest end, or at TABLES 5—9 123 TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued San Francisco Bay, Santa Clara Valley, and east bay hiZZS-—Continued Quotation Loca- Fig- Fail— Accu- Year of Reference tion ure ure ra- earth- No. No. type cy quake plate 52A plate 53A B 1906 Schlocker and Bonilla, 1963, p. 31. l] 185 . a C 1906 Lawson and others, 1908, p. 195. 186 B 1906 Lawson and others, 1908, [2 p. 192. 187 B 1906 Lawson and others, I] 1908, p. 198. = C 1906 Lawson and others, 1908, 55 plate 49B. :3 C 1906 Lawson and others, 1908, 56 plate 523. least the southwestern limit of tracing, it is on a ridge, and it there expands into, or else is replaced by, a group of cracks di— verging fan-wise. On each member of the group faulting took place, the downthrow being toward the northwest except in the case of two apparently short cracks with downthrow to the southeast. 0n four of these cracks the throw was greater than 1 foot, and at one place it was about 5 feet. Each crack was associated with a preexistent bluff or scarp, indicating that earlier movements have occurred at the - same place. The field in which the principal phenomena occur is cultivated with the exception of the steeper scarps, whose faces retain a bushy growth. (See plates 52A and 53A.) A tract lying between this locality and the coast, and extending several miles in each direction, exhibits a peculiar topography inter- mediate in type between that of the Rift and that commonly associated with landslides. Near the coast are a number of basins with ponds or lakes of much larger size than those along the Rift, and in asso- ciation with these are seen a number of sags similar to the fault sags of the Rift. On several lines which were thought from the physiography to represent partings between dislocated blocks, earthquake cracks were seen, and on one of these near the coast there was a vertical displacement of 3 feet, the downthrow being to the southwest. Vertical movement of as much as 5 feet, which Gilbert (in Lawson and others, 1908, p. 75) ascribed to faulting, occurred at locality 3 south of Mud Lake. The authors accompanied by Alan Galloway and Don Tocher, were able to positively identify locality 3 as the one that Gilbert described (op. cit., 1908, p. 75). This movement was probably caused by landsliding rather than faulting. Very large landslides are promi- nent just west of the locality, and study of aerial photographs reveals that an irregular scarp formed by a landslide joins the fractures described by Gilbert (1908). Moreover, the stream valley down hill from the site has been blocked by the landslide movement, causing ponding of the valley and deposition of sediments. Seven Lakes.—— * * * There were a few landslides, and a number of cracks already mentioned (page 75) [loc. 262] testified to movements of large blocks of ground; but I think these were due to a peculiarly sensitive condition of the country rather than to the violence of the shock. Daniel Bondietti lives 3.5 miles north from the head of the [Bolinas] lagoon, and his buildings are about 20 rods east of the main crack. His house was shifted 3 feet toward the fault and his barn moved in the same direction. Back of the Steele place, near the north end of the [Bolinas] lagoon, the hillside started [sic] eastward toward the lagoon, bulged upward, and cracked into several fissures from 30 to 100 feet long and from 5 to 18 inches wide. * * * The two bluffs along the stage road from the head of the lagoon to the town also broke and fell from 40 to 60 feet, completely blocking the stage road along the lagoon beach. [Picture caption] Cracks made by earthquake in tidal mud near head of Bolinas Lagoon. G.K.G. [Picture caption] Earthquake cracks in Bolinas at edge of an earth- quake sag. G.K.G. 124 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 6.—Specific descriptions of ground failures in the San Francisco Bay counties region—Continued Sam Francisco Bay, Santa Clara Valley, and east bay hiZZS—-Continued Loca- Fig— Fail— Accu- Year of Reference Quotation tion ure ure ra— earth- No. No. type cy quake B 1906 Gilbert, G. K., [Picture caption] Gilbert collection #2895. California earthquake. (from USGS Library House one mile northwest of Bolinas. Stands on alluvial fan of 0 Archives). Bolinas Creek. Was moved southward on fan. In the foreground is a secondary crack. Marin County, California. n.d. 188 C 1906 Lawson and others, The great ocean bluffs along the south and west of the entrance to a 1908, Bolinas Lagoon, some 165 feet high, crumbled and fell, crashing down p. 198. upon the ocean beach and reducing the slope of the bluff to half its former angle. C 1906 Lawson and others, On the sea-cliffs on the north side of Bolinas Bay and west of the n 1908, town of Bolinas, there was a very general crumbling and fall of the p. 387. sea-cliff upon the beach. C 1906 Weatherbe, * * * the mouth of Bolinas bay * * * the high cliffs -— about I] 1906, 150 ft. -- at the end of the peninsula have crumbled and fallen down, p. 402. carrying small trees with them. 189 57 B 1906 Lawson and others, Along the main street of Bolinas stand most of the houses a. * * . Q 1908, Of these about two-thirds were heaved, slid, tipt, and shattered into p. 198. minhabitable condition. * * * Along the bay shore were 7 buildings. Of these 6 went over or down. At the Flagg Staff Inn the tipping of the house has thrown it so far east into the bay that one may sit along the upper edge of the parlor floor and fish in 4 feet of water along the opposite edge of the same room. 0 B 1906 Weatherbe, At the village of Bolinas, the soil has slipped down easterly toward 1906, the lagoon and on the east side of the road, which runs north and p. 402. south, the buildings are entirely demolished, while those farther up the hill on the west side are not so badly affected. 190 C 1906 Lawson and others, On the steep southern face of Mount Tamalpais a number of rocks were [2 1908, loosened and rolled down the slope, some of them being large enough p. 77. to cut swaths thru the thicket which were visible for months afterward. 191 A 1906 Lawson and others, At the building occupied by Mr. George D. Shearer, 306-310 Fourth fl 1908, Street, on level land near the depot, there is a crack running north p. 207. and south v: * * . a A 1906 Lawson and others, A crack one block long, north and south, in low land near the station 1908, is reported. p. 207. = A 1906 Lawson and others, A Crack was formed in the ground 100 feet long, running north and 1908, south. The greatest damage was half a block north of the depot. p. 208. 192 B 1906 Lawson and others, At Scheutzen Park, 1.5 miles east—southeast of San Rafael, on land 1908, 7 feet above sea—level 4: * * water—pipes were broken, and there were a p. 208. many small fissures in the neighboring ground, running north and south. B 1906 Letter by William [William Hancock was living in Sausalito and working in San Francisco Hancock, 1906 After the earthquake he went down to the ferry in Sausalito.] earthquake collec— The asphalt pavement in front of the ferry was all broken up and in X tion; California one place sunk down, and the sea frontage looked as though it was = Historical Society about to fall into the water; long lines of cracks parallel to the Library, San Francisco. sea edge. TABLES 5—9 125 TABLE 7,—Specific descriptions of ground failures in San Francisco City and County Location number is assigned to each reported ground—failure site. Corresponding numbers are found on plate 3. Figure number refers to figure in this report showing damage described under "Quotation" column. Failure type is indicated by the following symbols. B Hillside landslides including rotational slumps, block glides, debris avalanches, and rockfalls Corresponding symbols are found on plate 3. 0 Sand boi ls E] Absence of ground failure noted 0 Streambank landslides including rotational slumps and soil falls Lateral spread 0 I Ground settlement a Ground cracks not clearly associated with land— slides, lateral spreads, settlement or primary fault movements Accuracy with which failure sites can be located is given as follows: m Miscellaneous effects )( Cracks in streets and pipeline breaks <4:}+ Arrows showing extent of area affected. Symbol shows failure type A, a site that can be accurately relocated; B, a site that can be relocated to within a few kilometers and probably could be located more accurately with further inves- tigation; C, a site where the information is insufficient to allow precise location. Plate numbers in the "Reference” column refer to plates in the original source material. Loca- Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake 193 31 ‘ C 1906 Lawson and others, 1908, p. 233. <) A 1906 Hyde, 1906b, p. 739. 0 A 1906 Jordan, 1907, p. 125. 0 A 1906 Derleth, 1906b, p. 553. C 1906 Jordan, 1907, p. 98. <)* [foot of Market ground failure zone] In spots the streets sank bodily, certainly as much as 2 feet, probably more. Accompanying this depression, concrete basement floors were broken and arched, as if to compensate for it. The surface of the ground was deformed into waves and small open fissures were formed, especially close to the wharves. Buildings on the water side, along East Street, generally slumped seaward, in some cases as much as 2 feet. The damage was greatest close to the water's edge, growing less as the solid land was approached, gradually at first, then more rapidly. These phenomena seem to suggest that the materials used in filling were shaken together so as to occupy less space with the accompanying development of waves, fissures, and structural damage. [Picture caption] The first picture shows the distortion of car tracks and the line of the local fault in East St. at its junction with Pacific St., along the water front. This territory is within the area of made land on the marshes of the old harbor front. [Picture caption] Fig. 12—-Rupture of Car Tracks and Pavement on East Street, Corner of Pacific Street. Fig. 14 shows destruction of car rails and street surface, corner East and Pacific Sts., near the Ferry Building. It is a general observation that the earthquake waves transmitted by the softer and less coherent materials and formations appeared to be much more destructive than waves which traversed the hard and more elastic rocks and other sound deposits. The billow-like effects that appeared in the streets of San Francisco near the Ferry house are most excellent examples of deformations in soft, incoherent materials. The sliding and rolling effects observed on some of the sand dunes and especially along the hillside at the northern end of Van Ness Avenue may be cited as allied phenomena. [loc. 221] 126 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 7 .—Specific descriptions of ground failures in San Francisco City and County—Continued Loca— Fig- Fail- Accu— Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake A 1906 Lawson and others, Editor‘s note: A part of the following quoted table was deleted, 1908, as it was not pertinent to the discussion. p. 143. Table 6.——Elevations of bench—marks before and after the earthquake ELEVATIONS AFTER BEFORE NEW-OLD EARTHQUAKE EARTHQUAKE COAST CHARACTER OF B.M. COAST AND COAST AND AND LOCALITY BENCH-MARK GEODETIC GEODETIC GEODETIC SURVEY SURVEY SURVEY 1906—1907 1877—1905 (Meters) (Meters) (Mm.) Cal. and Mont- Water table 41 5.1488 5.0173 +131.S gomery Sts. of Parrott Bldg. )[ East and Iron pillar 43 2.4828 2.8523 -369.5 Mission Sts. of brick bldg. Folsom between Granite post 44 5.4835 5.5516 - 68.1 Main and set in Beale Sts. brick wall. B 1906 Himmelwright, [Picture caption] EARTHQUAKE EFFECT. A fissure on East Street near El 1906, the water front. Note the settlement of the street surface as shown p. 19. by the exposed curb at the right hand side. In this locality the ground was "made," or artifically filled in. C 1906 Seattle Post— Water Front Suffered Less. * * * North of the ferry things were not Intelligencer, so good, but still they were better than anyone expected. Broadway m 1906b. wharf No. 1 was badly listed, and will have to be rebuilt, while Broadway No. 2 collapsed completely. The Union street wharf was blown up by the military authorities to save other property. A 1906 Hall, Here are, in general terms, the leading facts of earthquake effect 1906, within the area which we may designate as the city's principal soft p. 32. spot, even though it is not the largest area of marsh and mud flat which has been filled over and built upon. Beginning on the east and west streets at the north limit of the area of disturbance, we find the uppermost evidence on Pacific street at about the corner of Front, then on Jackson street at about the inter- section of Battery; next on Washington street at the intersection of Battery; next on Clay near the intersection of Sansome; then on Sacramento, also at the intersection of Sansome; on California it is doubtful whether the disturbance at Sansome is due to general subsi- )[ dence or only to subsidence toward a foundation excavation; on Pine street the upper limit of pronounced street disturbance is about 100 feet above Battery, on Market the line is very distinct at about the foot of Bush street. 0n Mission a sharp line of subsidence across the street about 100 feet before First, on Howard and in the line of Fremont to the east thereof the line is again plainly marked by a crack and subsidence below it; and finally, on the line of Folsom, about seventy feet below Beale, a line of subsidence is very distinctly )[ marked diagonally across the street exactly at the location and in the direction of the foot of the old hard ground and edge of the former mud flat. Platting these points on the map it is found that they either lie within or outline very nearly the limit of the former soft spot in the lower portion of the principal business part of the city. Below this bounding line the streets are nearly all waved, there being depressions of from six inches to as much as four feet in one TABLES 5—9 127 TABLE 7.—Specific descriptions of ground failures in San Francisco City and County~Continued Loca— Fig— Fail— Accu— Year of tion ure No. No. ure type ra- CY earth— quake Reference Quotation B 1906 Derleth, 1906a, p. 503. or two places and two or three feet at quite a number of points. While it cannot be said that the whole street area within this zone has sunk, a considerable portion of it has, and near the water front most of it has sunk from six to twelve inches, with several areas of greater depression. Davis street from Vallejo to California street, presents perhaps the extreme case, there being distinct depressions of from one to three feet at every street crossing or within every block, but it is apparent that some of this is due to slip of the street filling into building foundation areas, consequent upon failure of retaining walls and poor foundations of the buildings themselves. ‘ The area of about eighty feet frontage which has sunk to a maximum of about eighteen inches in front of the Market—street Ferry building and the depression of somewhat greater area and to a maximum depth of about three feet, at the northwest corner of the same building, are the extreme cases of subsidence along the main waterfront. Contrasted with the very heavy masonry Ferry building founded on piles and con- crete, which has not sunk at all, these subsidences will illustrate the point that it is only the soft mud and loose filling thereon which has been disturbed by the earthquake. There are places on the north and south streets where the whole street appears to have been thrown a few inches toward the bay, and at East street, which is the water— front street, there is much evidence of similar movement to about six to eighteen inches where the pavement has been shoved against wharves, piers and other water—front structures and caused to buckle up. Street and water-front railway rails are in a number of places buckled up six inches to two feet or are thrown as much as six inches out of line. Throughout the filled area above street-corner silt basins have been tripped out of plumb and bulged into sidewalk areas, and sewer manholes in street intersections are in several places canted up, showing sewer disturbance beneath; while granite curbs for 100 feet or more in length were tripped up by unequal movement of street pavements and the underlying ground and thrown out bodily on their sides upon the pavement or sidewalk. The Market-street Railway track, carried on a prism of concrete founded on piles for its length within this area, did not sink with the street on each side of it and is yet nearly on its original grade except at one point, where it has sunk apparently about four inches for several hundred feet. It is noticeable that streets have sunk least or not at all in front of the newer deep—piled foundations for adjacent buildings-—Market street in front of the Hotel Terminus and in front of the Buckley building, for instance--and this indicates that a part of the street movements is due to settlements into cellar and foundation excavations on failure of their retaining walls. A Curious revelation is noticed on the west side of Davis street, between Broadway and Pacific street. Here in 1857 was the water—front wharf. When the street was filled in it is evident that all the piles were not removed. The street pavement, which is basalt blocks, has sunk six inches or more for the full length of the frontage, and the position of the pile heads for about half the length is marked by their punching the pavement up in little pyramids, and fer the other half length the position of the pile bents with caps on is shown by the pavement sinking on either side of the caps, leaving ridges of paving blocks over them. Evidences of old structures beneath the surface and filling are brought out in a similar way at a number of points. All of the made ground between the Market St. water front and the region of Montgomery St. has been decidedly moved and deformed. Wave— like effects are common along lower Market St. and the water front. Wave-like depressions and crests amounting to four and five feet are found throughout this region. The same observations can be made in many other localities of the city, where soft ground is met. 128 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 7.4pecific descriptions of ground failures in San Francisco City and County—Continued Loca- Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake 194 A 1906 Derleth, The earthquake destruction was most marked on soft and sandy soil 1906a, and upon made ground. The Ferry building, at the foot of Market St. p. 503. is decidedly damaged. Its tower must be taken down. Bad cracks are found running horizontally between the windows on the second story. This building rests upon excellent foundations, but it is supported m upon material which seems to have acted like a viscous fluid. The building has been unharmed by fire, and with the exception of the tower can be repaired. A 1906 Gilbert, Humphrey, The Union Ferry Building [on piles] (Pl. XLVI, A), with the exception Sewell, and Soulé, of its high tower, was little injured, and the level of its floors was 1907, not perceptibly changed. At the same time, the streets at its front, m p. 135. which rested simply on the made soil, were rolled into waves 3 or 4 feet in height. A 1906 Himmelwright, UNION FERRY BUILDING. * * * The street surface at the N. W. corner )( 1906, settled 2', cracking the asphalt pavement and carrying down a large p. 226. section of the sidewalk. 195 C 1906 Duryea and others, On Market Street, near the Ferry House, the cable tracks resting )[ 1907, upon piles moved very little in comparison with the collapse of the p. 319. street surface on both sides of the car tracks. A 1906 Jordan, [Picture caption] Fig. ll--Street Surface in Front of the Ferry I! 1907, Tower, Showing Undulations and Cracks in the Asphalt Pavement. p. 123. II A 1906 Derleth, [Picture caption] Fig. lS.--Crack in Pavement in Front of Ferry 1906b, Building. Figs. 14 and 15. Views of Street Deformations, San p. 552. Francisco Earthquake. B 1906 Lawson and others, * * * the cable-car system on lower Market Street. On account of 1908, the constant tendency of the whole district to subside from year to p. 236. year, as the filling material became more closely compacted, these conduits were constructed upon piling to secure permanence of grade. )I On both sides of them the street sank in places as much as 2 feet, and the pavement was broken, fissured, and thrown into waves. These tracks did not escape entirely, but for several days, before street repairs were made, they constituted a narrow raised path along the center of the street. I! C 1906 Hyde, The most important wave-like distortions were observed on lower 1906b, Market and Mission Sts., and on East St. [loc. 193] along the present p. 739. water front. C 1906 Rickard, In San Francisco the street-car tracks on Market street retained ll 1906a, their alignment fairly well, but the roadway was depressed fully four p. 271. feet. Market Street is paved with cobbles; where there was an asphalt pavement in the lower parts of the town below Montgomery street, the roadway was buckled so as to make tents, and in other spots there were depressions several feet below the normal level. <) B 1868 Gilbert, Humphrey, Sewell, and Soulé, 1907, p. 76, plate 25A. Gilbert, Humphrey, Sewell, and Soulé, 1907, p. 31 and 32. Himmelwright, 1906, p. 81 and 84. Leslie's Weekly, 1906c. Lawson and others, 1908, p. 244. Lawson and others, 1908, p. 236. Huber, 1930, p. 266. Huber, 1930, p. 268. AETNA (YOUNG, OR COMMISSARY) BUILDING. The steel—frame structure at the corner of Spear and Market streets, * * * rests upon piles, and suffered relatively small damage from the earthquake. Pl. XXV, A, shows the corner of the building and the subsidence of the street at this point. * * * There was a vault under the Market street sidewalk, immediately behind the wall at the curb line. The basement floor in this vault was of concrete and had a total thickness of 7 or 8 inches. The earthquake caused the earth to bulge up in the portion of the basement under the sidewalk, rupturing the concrete floor and turning it up on its edge, so that where there had previously been a clear headroom of 7 1/2 feet the highest point of the bulge was within 3 1/2 feet of the beams carrying the sidewalk. The five-story AEtna Building * * * was built on piles. * * * The basement floor, which was of concrete 7 or 8 inches thick, was pushed up under the sidewalk, reducing the headroom at this point from 8 feet to 3 1/2 feet, approximately. This bulging was probably caused by settling (Pl. XXV, A), as the foundation piling did not extend under the sidewalk. [Pl. 25A shows sidewalk did not subside; adjacent street did.] YOUNG or SELLER BUILDING. S.W. Cor. Spear and Market Streets. * * * Levels on the water table show that the N. E. and S. W. corners are 3” and 6" lower respectively than the N. W. corner. These facts would indicate that the foundations had moved sufficiently to tilt the entire building to the east. From marks on the curb of the side- walk, it is also apparent that the surface of the ground settled considerably around the N. E. corner. [Picture caption] Surface of Market Street sunk by the earthquake five feet below the curb level. Appraiser’s Building.—- * * * The northwest corner is 0.909 foot = 10.908 inches above the southwest corner. The northeast corner is ‘ 0.909 foot + 0.054 = 0.963 foot = 11.556 inches above the southwest corner.. The southeast corner is 0.080 foot = 0.96 inch above the southwest corner. The rod was held on top of the water—table at each of the four corners, and the sights were nearly equal in length. The south side of the building is about 11.23 inches lower than the north side. LBecause no immediately pre—earthquake measurements are avail— able, it is not certain that all settlement occurred during the earthquake.] Buildings erected upon good foundations withstood the ordeal well, even when the streets around them were deprest and fissured. The Appraisers' Building furnishes a good illustration of this; it is substantially built of brick upon a piling foundation, at the corner of Washington and Sansome Streets, and still stands without signifi- cant damage. The levels of its foundation walls were not disturbed. The shock was principally felt on "made ground” and the flats where the foundation is known to be unreliable at all times. Eastward of Montgomery Street, toward the Bay, there are a number of buildings injured, while some are utterly ruined. Along the old water line of the Bay, running just back of Macondray's old place on the corner of Pine and Sansome Streets, and thence diagonally north-eastward toward the corner of Front and Jackson Streets, something like a slide occurred and buildings suffered severely from the slipping of the made- ground foundations on the old mud bottom of the Bay. This effect is more marked at some points than others; at the old Railroad Hotel on Clay Street, below Sansome, it is more marked than elsewhere. * * * photographs of * * * the rear of the Railroad House, on Clay Street, between Battery and Front, as they appeared after the shock. 130 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 7.—Specific descriptions of ground failures in San Francisco City and County—Continued Loca— Fig— Fail- Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake 199 A 1868 Lawson and others, On Pine Street, near Battery, the cobbles on the south side of the IK 1908, street sank away from the curbstones to the depth of 1 foot in some p. 436. places; and the asphalt sidewalk on the north side was twisted and torn out of all shape, and its connection with the curb-stone severed. 200 II A 1868 Lawson and others, In many places the made land settled. At the junction of Market 1908, and Front Streets, the ground sank for a foot or two, and there was p. 436. evidence that the tide had risen in the adjoining lot at the same time, for a pond of water collected and remained until low tide. 201 A 1868 Lawson and others, At the corner of First and Market Streets, the ground opened in a C: 1908, fissure several inches wide. At other places the ground opened and C) p. 437. water was forced above the surface. A 1906 Himmelwright, CROCKER ESTATE BUILDING. S. E. Cor. Mission and First Streets. 1906, * * * The levels on the water table show that the N. W. corner and II p. 86 and 89. the N. E. corner are 1/2” and 1 1/2" lower, respectively, than the S. E. corner of the building. 202 I: A 1868 Lawson and others, At Fremont and Mission Streets the ground opened in many places. 1908, The general course of damage in the city was along the irregular p. 437. line of the "made land,” or low alluvial soil, where it met the hard or rocky base beneath it. Along the line of the old shore of Yerba Buena Cove, we found the damage to brick buildings much the largest. A 1906 Lawson and others, The floor of the Pacific foundry was raised about 2 feet in places. 1908, The center of Mission Street (opposite Fremont Street) exposed an I: p. 437, opening from 8 to 10 inches wide; and openings of the ground were plate 146. also plainly to be seen on Fremont Street, in the same vicinity. Outside of the immediate district described above, damage to the rest of the city was very meager. * * * the region of greatest agitation was confined to the low portions of the city, or the vici— nity of some old creek bed or swamp. B 1868 Huber, The shop of Mr. Garratt, brass founder, near the corner of Mission 1930, and Fremont Streets, has been raised to its original position. It was II p. 267. found to have been lowered about eight inches by the recent earthquake. A _l906 Himmelwright, SCOTT BUILDING. S. W. Cor. Fremont and Mission Streets. * * * The )l 1906, S. W. corner of the building is badly racked and cracked by the earth— p. 86. quake. B 1865 Huber, Stoddard's warehouse on Beale Street is said to have been thrown out 1930, of place several inches as though it had been lifted up and set down II p. 264. again, while the south side of the building appears to have settled considerably. After the shock, the water rushed into the cellar, or basement, but whether from a disarrangement of the water pipes, or from any fissure in the earth which might have opened, was not known. A 1838 Louderback, Mr. Spear informed me that during the earthquake of June, '38, before 1947, mentioned, a large sand—hill standing in the Vicinity of what is now p. 52. Fremont street, between Howard and Folsom, and between which and the bay at high tide there was a space of about twenty feet, permitting a free passage along the shore to Rincon Point (the coves of which were then much resorted to for picnics and mussel parties), was moved m bodily close to the water, so as to obstruct the passage along the shore. After that no one could pass there at high tide, and we were compelled to go around back of the sand—hill, and wade through loose sand to reach that point, a much more laborious walk. 203 A 1906 Duryea and others, The Foljer [sic] Building, on Howard Street, was of brick exterior 1907, p. 288. and wood interior, and was left standing after the earthquake, the fire not having reached it. As was generally the case in buildings of any TABLES 59 131 ‘ TABLE 7.—Specific descriptions of ground failures in San Francisco City and County—Continued Loca- Fig— Fail- Accu— Year of Reference tion ure ure ra— earth- No. No. type cy quake Quotation A 1906 Himmelwright, I 1906, p. 28. 204 A 1906 Hyde, 1906a, X p. 705. 205 C 1906 Lawson and others, 1908, p. 236, 237, and 238, map 17. plate 88C plate 880 33 plate 89A description, the chimneys suffered and a small portion of the parapet wall, but otherwise the building was not damaged; and yet, at this very site, the street in front of the building had settled about 2 ft. FOLGER BUILDING. S. W. Cor. Howard and Spear Streets. * * * The street level settled about 2 ft. at the northeast corner of the building, but there are no earthquake cracks visible in the walls at this point. The bascule bridge recently erected by the Southern Pacific Company across the San Francisco River, so—called, at the foot of Channel St., near the water front, was seriously crippled by the settlement, to the extent of 4 in., of the east corner of the south abutment. [South of Market zone.] High intensity was developed thruout a small elongate district having a width of about two blocks, which extends from near the corner of Eighth and Mission Streets to the vicinity of Fourth and Brannan Streets; from this point the boundaries are irregular and very sinuous, leading to the water-front at about the crossings of Third Street with Berry and Channel Streets. A glance at the geological Map, No. 17, shows that the regularly bounded portion of this district corresponds very closely with the area of a former tide—marsh, drained and flooded by one or two small tidal streams. The former shore line of Mission Bay was just north of Brannan Street, between Fourth and Fifth Streets, so that the irregu- lar seaward portion of the district lies outside the old shore. This is one of two localities in the city, the other being'a "made" land tract along the former course of Mission Creek [1°C. 214], in which destructive effects of great magnitude were conspicuously developed. Only in very close proximity to the fault was greater violence manifested. For blocks the land surface, paved streets, and building plots alike, were thrown into wave forms, trending east and west about parallel to the length of the area. The amplitude and wave-length of these earth billows, and the distances to which they extend, are indefinite and irregular. The fissuring and slumping, and the buckling of block and asphalt pavements into little anticlines and synclines (arches and hollows), accompanied by small open cracks in the earth, characterize the land surface. This slumping movement or flow took place in the direction of the length of the area, and its amount was greatest near the center, or channel, where the street lines were shifted eastward out of their former straight courses, by amounts varying from 3 to 6 feet. A satisfactory photograph of this phenomenon was not obtainable, owing to the quick convergence of parallel lines in perspective, but to the observer in the field it was a very striking result of the shock. The greater part of the district was occupied by wooden dwellings and shops, with a small percentage of mediocre brick buildings and a few of substantial construction. The fire swept the area clear. Not even heaps of debris remained to cover the ground, most of the destructive effects being obliterated, along with the structures in which they were developed. Enough remained, however. Foundation walls and sidewalk pavements were broken and flexed; sharp little anticlines were produced in the street by the arching of block paving, as on Russ Street between Folsom and Howard Streets (plate 88C); granite curbing was broken and thrust up into an inverted V, as on Moss Street, between Folsom and Howard Streets (plate 880); there were fissuring and slumping in the block pavement, as along Columbia Street between Folsom and Harrison Streets (plate 89A), and sharp flexures of the paved streets and car tracks, as on Sixth Street just south of Howard.Street [loc. 209]. These effects point simply and clearly to the great magnitude of the intensity thruout the greater part of this old swampy district. 132 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 7,—Specific descriptions of ground failures in San Francisco City and County—Continued Loca- Fig- Fail- Accu- Year of Reference tion ure ure ra- earth- No. No. type cy quake Quotation plate 893 plate 89C 34 Attention has already been directed to the slumping or flow movement to the east along the long axis of the area. The heavily ballasted car-tracks on Bryant Street, at the crossing with Fourth Street, were sharply flexed laterally, tho bounded by block paving. (Plate 89B.) This was at the eastern end of the district where the marsh formerly bent to the south around the flanks of Rincon Hill, a mass of firm sandstone rising from the floor of Mission Valley. No similar sharp flexures were encountered along east—west streets in the western or central portion of the district, tho lateral displacement and flat, sinuous curvings of the street lines were common enough; notably on Harrison Street between Fifth and Sixth Streets, and on Folsom Street between Fourth and Seventh Streets. Both these streets cut across the direction of the flow movement at a small angle. These phenomena are easy to understand if, as seems certain, Rincon Hill served as a solid buttress against which the flow to the east was arrested, causing sharp crumpling of the surface near the buttress, with less disturbance farther away. This was combined with a slight tendency to flow southward in the southeastern part of the district. The shaking caused the materials used in filling to settle together and occupy less space, so that the surface over the whole district was lowered by amounts varying from a few inches to 3 feet or more. This is clearly seen in the change of street levels along the margin of the solid ground, where the car rails are bent downward in little nonoclines. Occasionally a structure with a relatively good founda- tion remains at its former level, with the whole neighborhood deprest about it. Such a case is exemplified on Sixth Street, a little south of Howard Street [loc. 209], near the margin of the area. (Plate 89C.) The flow movement is thought to be due simply to the action of gravity, the loose, water—soaked material being compacted into less volume by the shaking. Besides this sinking of the district, and its flow movement, mention has been made of the deformation of its surface into irregular waves, trending approximately east and west parallel with the length of the district. Along the streets running approxi— mately north and south, at right angles to the elongation of the area, Car rails were bent abruptly to the side, or raised in arches, and sharp anticlines were formed in the block pavements. Large square concrete slabs, used for sidewalk paving, were thrust one over the other; and in one or two cases a slab entirely covered an adjoining one. These phenomena indicate shortening by compression in the north— south direction. On the other hand, however, a stretching of the surface is shown by fissures in the paving; by places where wedge- like blocks were deprest below the general level; and by the rails of car tracks which were pulled apart in amounts varying from 8 to 12 inches. Owing to the relatively great and very variable structural strength of paved streets and heavily ballasted car tracks, these phenomena are not developed regularly nor frequently enough to afford a satisfactory tezt of the hypothesis that they are directly associated with the wave forms into which the surface of this district was thrown. Besides owing perhaps to the varying rigidity of the materials which make up the surface of the streets and building plots, the wave forms themselves, tho generally prevalent, are not persistent in their ex— tension. The compression and distension effects, however, are believed to be due to the same cause as that which generated the wave forms; for there is no evidence of any true shortening, or lengthening, of the north—south dimension of this district, nor is there any proba- bility of this having occurred. In addition, then, to the flow movement and the settling together of the loose materials causing depression, there was some sort of rhyth- mic movement in this loose earth which produced wave forms in the surface, with places of compression and places of stretching. It probably was this movement which was most effective in producing structural damage. It is not believed that these surface waves were traveling waves "frozen" as the shock subsided. If they had been of that character, the ground surface should be more broken than it TABLE 7 .—Specific descriptions of ground failures in San Francisco City and County—Continued TABLES 59 133 ‘ Loca— Fig- Fail— Accu- Year of earth— quake tion ure No. No. ure type cy ra- Reference Quotation ‘9 B 1906 A 1868 A 1906 C 1906 San Francisco Chronicle, 1906C. Lawson and others, 1908, p. 437. Himmelwright, 1906, p. 175 and 179. Hall, 1906, p. 32. appeared to be; for in relatively rigid materials such waves must develop open fissures along the crests, which would close with crushing in the troughs. It must be noted, without any attempt at explanation, that the destructive effects of great magnitude which have been described above, are practically confined to the "made" land which occupies the old marsh site. Southeast of Brannan Street, where formerly lay Mission Bay, such effects are of less magnitude, in general; are less regular in their occurrence and are, on the whole, less prevalent. The Fourth and Sixth street sewers were also greatly damaged, some of them showing a vertical and horizontal movement of as much as five‘ or six feet, and portions will have to be reconstructed. At the corner of Fourth and Bryant Streets, walls were cracked and damaged; Fourth Street near Bryant opened in places and at the cross—‘ ing of Harrison and Fourth the railroad track settled about 8 inches, the planks between the rails rising about 10 inches. CALIFORNIA CASKET C0.'S BUILDING. and Sixth Streets. * * * The levels on the water table indicate that the northwest corner is about 1/2" lower than the northeast corner. Mission Street, between Fifth It would be tedious in these articles which the writer is endeavor—‘ ing to keep within readable limits to catalogue too many dry and hard facts made about hard and soft land, so he does not undertake to trace in detail the outline of the Mission bay and estuary, salt marsh and fresh swamp areas, as these have new again been made evident by the street subsidence and other movements caused by our king shake. The region is a large one. To go slowly step by step around it as we did in the case preceeding [foot of Market area, loc. 193], on foot, as it were, would take too long; so we move rapidly over most of it as in an auto, and, commencing on Townsend street, near the Southern Pacific Railroad yards, we notice a disturbance near Cook street, another in Brennan, near Ritch; another near Harrison and Fourth; another near Folsom and Fifth, and another near Howard and Sixth. We find that by these we may outline on the map the old salt marsh limit as far as the greater impress of the earthquake's heel, which is found in the neighborhood of the new Postoffice. ‘ MISSION AND SEVENTH STREET DISTURBANCE. It looks pretty bad on Mission street at and near Seventh, to see the whole street disturbed for about 700 feet in length, to see that this disturbance extended far down Seventh street, and that an area of the adjacent land had sunk. As an American one cannot but feel glad that the new Postoffice building escaped, though barely, being in this area of depression. As a San Franciscan who knew this spot fifty years ago, who saw it a marsh with a little stream running through it, who saw hunters wearing gum boots tramping about shooting jacksnipe in that very area, who later saw it drained for market gardening, and still later saw it filled to a depth of ten or fifteen feet with sand dumped off a bank from side dump cars, and then saw it occupied by light wooden houses for a score and a half of years, it seems entirely natural that a real earnest earthquake should make it settle and move just as it has ‘ settled and moved. It never had an inducement to get down to a good bearing before. Now it has been shaken to where it will probably stay, and San Francisco will be the better for it. In the block south and west of the Postoffice this old John Sullivan marsh formerly headed. Its course was toward the east, joining an" area of salt marsh which bordered Mision bay. * * * The facts now are that under the earthquake influence the filling over this marshy area has settled at a number of places and to depths of from a few inches to three or three and one-half feet. One of the most pronounced settlements is the one referred to on Mission and Seventh where the 134 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 7.—Specific descriptions of ground failures in San Francisco City and County—Continued Loca- Fig— Fail— Accu- Year of Reference Quotation tion ure ure ra— earth— No. No. type cy quake subsidence has reached a maximum of about three and one-half feet, 0 as judged by the floor of the Postoffice building and the movement toward the bay, as judged by the street railway track and alignment of the trolley line support poles, has extended for 300 to 600 feet in the length of Mission street, reached a maximum of about five feet at a point 175 feet south and west of Seventh. Platting the limits of this disturbance on the map, they are seen to coincide as near as can be measured with the outline of the old Sullivan marsh as shown by the Coast Survey map of 1857, and as the writer distinctly remembers it to have been. 206 A 1906 Duryea and others, * * * a number of brick warehouses on the north side of the Southern )( 1907, Pacific Railroad, between Fourth and Sixth Streets, * * * their settle- p. 265. ment during the earthquake, with respect to the railroad right of way, just south of them, being about 2 feet. 207 A 1906 Hyde, On Fifth St., between Folsom and Harrison Sts., the brick sewer )( 1906b, settled with the adjacent land, its crown was destroyed and the channel p. 740. had become filled with sand which blocked up the sewage to a dangerous extent. 208 II A 1906 Lawson and others, Mint.-— * * * The southwest corner is the lowest, being 0.498 foot = 1908, 5.976 inches (mean) below the northwest corner. * * * p. 245. The deformation indicated by the above measurements can not be wholly referred to the earthquake, since it is quite probable that the struc— tures had settled somewhat before that event. It appears, however, to be desirable to put the measurements on record for future reference. 209 ll B 1906 Hyde, [Picture caption] Settlement on Sixth Street. * * * just south of 1906b, Howard St. The amount of the drop at the lamp post * * * was fully p. 738. 2 ft. The sidewalk north of this point and the street at the junction of Howard appears to have settled but very little. II B 1906 Derleth, Fig. 13 shows a drop in Howard St. Where the men are standing a 1906b, sidewalk on foundations has remained in place. [Picture caption, p. 553. p. 551] STREET SUNKEN; SIDEWALK IN PLACE BECAUSE 0N FOUNDATIONS. C: B 1868 Lawson and others, A small crevice opened, as in 1865, on Howard Street beyond Sixth 1865 1908, Street. p. 438. I: c 1868 Holden, 1868. October 21; IX. * * * As in 1865, a small crevasse was opened 1865 1898, on Howard Street, beyond Sixth. p. 76. I: C 1868 Sunday Record— Twain Tells Humors of 1868 Earthquake. * * * A crack a hundred feet Herald, long gaped open six inches wide in the middle of one street and then 1906. shut together again with such force as to ridge up the meeting earth like a slender grave. C 1906 Lawson and others, [Picture caption] Sixth Street, near Howard. Once occupied by )l 1908, marsh. Street dropt nearly 3 feet. Sidewalk held up by piling founda— plate 89C. tion of a building. H. 0. W. 210 35A A 1906 Gilbert, Humphrey, The steel—frame and granite post-office building (Pls. XLII, B; XLIII; SSE Sewell, and Soule, XLIV) was carried on isolated grillage foundations, each column having 350 1907, its own footing. The diagonals of the building ran nearly north and p. 97 and 98, south and east and west, the south corner being at Seventh and Mission plate 62, streets. To the south and west of Mission street was an elongated, plate 63, narrow, curved area in which the earthquake damage was very severe plate 64. [see loc. 205]. It was commonly reported that this area, which was TABLES 5—9 135 TABLE 7.—Specific descriptions of ground failures in San Francisco City and County—Continued Loca— Fig— Fail— Accu— Year of Reference Quotation tion ure ure ra— earth— No. No. type cy quake not far from the south corner of the post—office building, was a stream bed or ravine that had been filled within the recollection of the older inhabitants of San Francisco. * * * Mr. Roberts, who is evidently a cool and accurate observer, seemed of the opinion that the material under the building was a natural deposit, and not an artificial fill. But toward the south it was not of a nature to inspire confidence in its carrying power at the depth shown on the foundation plans. He accordingly obtained authority to lower the footings wherever the material at the depth shown on the plans seemed unreliable, so that the footings of the south half of the building were lowered--some of them, as I remember his statements, to a depth of 20 feet or more below the basement—floor level. At any rate, he carried them to a point where the material, in his judgement, was sufficiently hard and compact. All this underlying material is very sandy; but at considerable depths, I understand, gravel appears, and the combination is almost as hard as hardpan. II A 1906 Gilbert, Humphrey, The ground at the corner of Seventh and Mission streets settled about Sewell, and Soulé, 5 feet (Pl. XLIII, B). The floor of the [Post Office] building was 1907, slightly cracked at that point, and Mr. Roberts stated that there was p. 44, a settling of about 1 3/4 inches. plate 43B. )( A 1906 Gilbert, Humphrey, The street went down about 4 or 5 feet at this point [Mission St. in Sewell, and Soulé, front of the Post Office] as a result of the earthquake (Pl. XLIII, B). 1907, p. 99, plate 43B. A 1906 Lawson and others, The new United States Post—office building, (plate 943), at the corner 1908, of Seventh and Mission Streets, was just on the margin of the district p. 238. [see ICC. 205]. It is a steel and granite structure, resting upon a foundation of piling driven to a considerable depth * * * At its and slid out from the building about one and one-half feet. A 1906 Duryea and others, The Post Office Building * * * was on piles * * * and the earthquake 1907, p. 288, plate 38. wave motion moved through this forest foundation, following and develop— ing lines of least resistance, with consequent promiscuous racking all over the building. At the four corners the cracks were most pronounced ***, 136 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 7 .—Specific descriptions of ground failures in San Francisco City and County—Continued Loca- Fig— Fail— Accu— Year of tion ure ure ra- earth- No. No. type cy quake Reference Quotation A 1906 I A 1906 0! A 1906 I A 1906 211 B 1906 212 C 1906 A 1865 '9 B 1865 Himmelwright, 1906, p. 192. Himmelwright, 1906, p. 193. Hyde, 1906a, p. 701. Derleth, 1906a, p. 503. Gilbert, Humphrey, Sewell, and Soulé, 1907, p. 99, plate 62B. Los Angeles Herald, 1906. Lawson and others, 1908, p. 449. Holden, 1898, p. 66. UNITED STATES POST OFFICE. N. W. Cor. Seventh and Mission Streets. * * * At the southwest corner, the ground settled about 2 ft. at the building line and about 5 ft. at the curb, the entire surface from the building line moving out about 5 ft. to the south. This distorted the sidewalk and steps of the two entrances, there being cracks in the joints of the cement sidewalk slabs 8" wide. It was necessary to place two temporary wooden steps of about 8" rise from the sidewalk in its settled position to that portion of the steps which remain approxi- mately at the original height. [Picture caption] UNITED STATES POST OFFICE. Northeast Corner of Seventh and Mission Streets. At the curb in front of the building, on the right-hand side, the ground settled 5 feet and moved to the east away from the building about 6 feet. At the building line, the ground settled about 2 feet, causing the displacement of the granite coping, steps, etc., at the sidewalk level, as shown. The sidewalk was originally a straight grade on the right—hand side where the sag is now shown. The northwest corner of the building was badly racked by the earthquake, and temporary shores were in position when the photograph was taken. This building was only slightly damaged by the fire. United States Post Office.—- * * * The effect of the earthquake throughout this vicinity has been most marked. Streets in this neigh— borhood have settled very considerably and the sidewalk has also seriously sunk. Mission St., at the corner of 7th, has been thrown bodily southward to the extent of at least two feet. It is evident that the building itself settled slightly, inasmuch as serious cracks were developed on the granite facing at the southwest, northwest, and the northeast corners, and these cracks presumably extend into the brickwork. The magnificent Post Office building, corner of Mission and Seventh Sts., rests upon sand, and under one end of it at one time ran an old creek bed. This building rests upon piles and heavy concrete walls, but has been badly cracked due to the severe convulsion of the ground. It has been unharmed by the fire, but will need exten— sive repairs. * * * there was a partially erected steel frame (Pl. XLII, B) on the southwest side of Seventh street, near the post-office. Before the earthquake all the columns were plumb and in true alignment. As a result of the shock there was a lateral shifting of the column bases—~the relative movement being almost 2 feet in some p1aces--at the cellar—floor level. The basement walls of the incomplete building were also shifted horizontally; at the east corner, where the walls had met at a right angle, they had been ruptured by a vertical crack and moved laterally in such a way that the angle between them was reduced to about 75° * * * . [Picture caption] This remarkable photograph shows how Mission street, San Francisco, sunk away from the curb to a depth of five feet. The postoffice is located one block north, where the ground was upheaved to a height of several feet. In this area of ”made” ground houses were completely overturned and in several instances sank for considerable distance below the level of the street. 0n the marshy lands in the vicinity of Howard and Seventh Streets the ground was heaved in some places and sank in others. Lamp-posts were thrown out of perpendicular, gas—pipes were broken, etc. 1865. October 8; IX; * * * 0n the marshy lands in the vicinity of Howard and Seventh streets, lamp posts, water pipes and gas pipes were broken and thrown out of position. The ground on Howard Street, from Seventh north to Ninth, cracked open, leaving a fissure nearly an inch wide. TABLES 5-9 137 TABLE 7.—Specific descriptions of ground failures in San Francisco City and County—Continued Loca— Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake 213 m C 1906 Lawson and others, The blocks between the old tide—marsh area, extending east from near 1908, the Post-office [loc. 205], and the former course of Mission Creek p. 229. [loc. 214], give evidence in the form of cracked foundation walls, broken concrete cellar floors, etc., of intensity values high in Grade C. 214 A 1906 Lawson and others, [Mission Creek zone] As stated briefly above, a similar district 1908, [loc. 205] of high intensity occurs in an area of made land along the p. 238 and 239. lower portion of the former course of Mission Creek. This district varies in width from 1 to 2 blocks, extending from near the corner of Ninth and Brannan Streets westward for about 3 blocks, then south- westward for about 2 blocks more; and finally, westward some 4 blocks more to a point on Nineteenth Street just east of Dolores Street. Mission Creek was formerly a sinuous tidal stream, with narrow fringes of salt marsh about its banks. Near its mouth the stream wound around a rocky point where the serpentine hills of the Potrero rose abruptly from its southern bank. Here along its margin, is found the most sudden transition from high to low intensity that is 38A anywhere encountered in the city. Along Dore Street, a narrow alley 383 running from Bryant Street to Brannan Street, between Ninth and Tenth 380 Streets, the street pavement was broken into a series of waves. The plate 89D photographs, plate 890, looking along Dore Street from Bryant toward plate 90A Brannan Street; plate 90A, looking from Brannan Street in the reverse plate 91A their underpinning, and a few collapsed. Plate 91A shows a wave trough near Bryant Street, with the resulting disturbance of the pavement. The dwellings immediately in the trough have dropt from 40A plate 91B their foundation posts. In Plate 91B, looking along Ninth Street from 403 near Brannan Street, is shown the depression and fissuring of the 400 street and its slumping or flow westward toward the former channel 41 of a short branch of Mission Creek, which occupied the present location was noted and the slumping of flow eastward (toward the channel of the little branch of Mission Creek) is scarcely noticeable. C 1906 Gilbert, Humphrey, As in districts outside of San Francisco, the greatest damage was Sewell, and Soulé, done to those structures having insufficient foundations built on II 1907, soft alluvium or filled ground. The settling of the ground in the p. 26. mud flats along San Francisco Bay and of the filled ground in old water courses was accompanied with great destruction. It was in such ground that the greatest number of breaks occurred in the cast-iron gas and water mains and the sewers. The breaks in the sewers were not so evident as those in the gas and water mains, for the reason that the latter were under pressure and breaks in them resulted in breaks in the streets themselves. The most noticeable destruction resulting from the settlin of soft or filled ground occurred in Howard [now South Van Ness% and Shotwell streets between Seventeenth and Eighteenth streets [loc. 215], Bryant street between Ninth and 138 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 7,—Specific descriptions of ground failures in San Francisco City and County—Continued As in Loca- Fig- Fail- Accu— Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake and Tenth streets, Dore street between Bryant and Brannan streets plate 6A (Pl. VI, A), and at the corner of Seventh and Mission streets [loc. 210]. The settling was greatest in Howard, Dore, and Bryant )[ streets, being in Dore street at least 5 feet. A 1906 Gilbert, Humphrey, The filled districts of San Francisco afford several examples, and Sewell, and Soulé, two of these are illustrated by Pls. V and VI, B. The view shown 1907, in P1. V is northwestward on Ninth street, near Brannan. Before the p. 8, earthquake the car tracks and curb line were straight and approximately 40A plate 5, level, and this condition was not disturbed on the relatively firm 43 plate GB. ground shown in the distance. In the nearer part of the view the street crosses a tract of made ground created by filling a valley tributary to a narrow tidal inlet called Mission Creek. The descent <) of this valley was southwestward, and the made ground flowed in that direction, carrying street and buildings with it. In taking the photograph reproduced in P1. VI, B, the camera stood on ground made by the filling of Mission Lagoon, an expansion of Mission Creek, and was pointed northward, commanding a portion of Howard street Lloc. 215]. The made ground here flowed northeastward and the buckling of street-car tracks was caused by its motion. Where the same earth A 1957 Bonilla, Table 1. Landslide number 8; slump in soil and sand flow; 100 feet 1960, long and 800 feet wide and consisted of several coalesced slides. p. 6. <)’ A 1957 Bonilla, Slide 8, shown on figure 4, probably started as a sand flow, but 1960, removal of support by the flow resulted in slump failure near the p. 21, head of the landslide. fig. 4. 237 A 1957 Bonilla, Table l. Landslide number 7; sand flow and slump in soil; 80 feet 1960, long and 225 feet wide. p. 6. 238 B 1906 Lawson and others, A * * * sudden movement of the ground occurred on the west side of 1908, Merced Lake, whereby a large section of the slope sank toward and p. 387. into the lake, and a portion of the bottom of the lake was lifted above the surface by the deformational rotation of the collapsed ground. B 1906 Lawson and others, Just south of the bridge across Lake Merced, a sand-bar was forced 1908, up out of the lake, from water that was previously 6 or 8 feet deep. p. 251. This bar is parallel to the west bank of the lake, and has a direc- tion almost due north and south. This was probably caused by the <) same thing that wrecked the bridge; that is, the displacement and settling of the west bank of the lake at the time of the earthquake. 0 A 1957 Bonilla, Table 1. Landslide number 4; slump in soil; 110 feet long and 200 1960, feet wide. p. 6. 239 A 1906 Lawson and others, Lake Merced.--About 6 miles north of Mussel Rock, where the Merced 1908, p. 251. beds disappear under aeolian sands, the disturbance seems to have been quite violent. An old railroad trestle, that crosses the northern end of Lake Merced in the narrowest place, was badly wrecked. This TABLES 5—9 145 TABLE 7.—Specific descriptions of ground failures in San Francisco City and County—Continued Loca- Fig- Fail— Accu- Year of Reference tion ure ure ra— earth- No. No. type cy quake Quotation O A 1957 240 [z A 1957 Bonilla, 1960, p. 6. 241 a A 1957 Bonilla, 1960, p. 6. 242 A 1957 Bonilla, 1959, p. 34. Bonilla, 1960, p. 6. <) A 1957 O A 1957 Bonilla, 1960, p. 21. 243 B 1906 Lawson and others, 1908, C) p. 241. bridge was broken in two places, and the intermediate piece was out of line with both ends. The direction of the offsets was very nearly due north and south. At one break the west piece was shoved 12 or 14 feet past the other section. The west end of the intermediate piece failed to join the section at the west bank by 6 or 7 feet. The west section that remained with the bank was from 4 to 5 feet lower verti- cally than the intermediate piece. The trestle was old, built of heavy timbers on a sharp curve, and not in use, which will in part account for its destruction. The swaying of this bridge destroyed a section of it 50 to 60 feet long. On the hillside where this trestle reaches the west bank of the lake, cracks parallel to the shore line suggest the cause of the destruction of the bridge. The displacements here are larger than any along the main fault line, and it is apparent- ly entirely local, due to the slipping and settling of the west bank of the lake. Twlel. feet wide. Landslide number 3; slump in soil; 100 feet long and 100 Table 1. feet wide. Landslide number 2; debris slide; 100 feet long and 150 Table 1. feet wide. Landslide number 1; debris slide; 75 feet long and 80 In addition to the failures along the roadway at the shore of Lake Merced, the failure of the artificial fill at the north end of the foot-bridge crossing the north arm of the lake also was probably caused by sudden liquefaction of sand. Some of the slides along Lake Merced were of the slump—earthflow type and displayed a backward rotation of the component blocks. These slides may have been caused by removal of support by sudden liquefaction of sand at the foot of the slope. Table l. Landslide number 5; sand flow; 70 feet long and 80 feet wide. Landslide number 6; slump in soil and earthflow; 125 feet long and 120 feet wide. Slide 5 is an example of this type of landslide. It severed the artificial embankment north of the footbridge that crosses the north arm of Lake Merced, as shown on figure 3. The exact dimensions of the slide deposit could not be determined because it was under water but about 80 feet of the embankment was destroyed. The vegatation displaced by the slide is visible on air photos taken four months after the slide. Measurements made on the photos show that the vegetation on the sides of the embankment moved at least 70 feet both eastward and westward from the original shore- line of the embankment. The material visible in the north end of the embankment was artificial fill composed of clean sand, and as a large deposit of dune sand is found a short distance north of the site, it is probable that all of the embankment was clean well-sorted sand obtained from the dunes. The slope of the embankment above water was on the order of 20°, and under water was presumably less. The earthquake vibrations probably liquefied the staurated sand at the base of the embankment and the unsupported embankment collapsed and spread over the lake bottom. On Ocean Avenue and X Street, near where the former outlet of Lake Merced flowed, fissures were developed in the street and in the sands on either.side, and water was squeezed out so as partly to flood the roadway. Drain pipes were unearthed and bent or twisted. 146 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 7,—Specific descriptions of ground failures in San Francisco City and County—Continued Our Loca— Fig— Fail- Accu— Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake 244 C 1852 Townley and Allen, 1852 November 22. 11 p.m. VIII. Near San Francisco. Severe shock a 1939, eight miles southeast (7) of San Francisco. Next morning a fissure p. 28. half a mile wide and three hundred yards long was discovered, through which the waters of Lake Merced were flowing to the sea. C 1852 Coffman, 1852. November 22 to 24. San Francisco Peninsula. This shock was 1973, severe 8 miles southwest of San Francisco. Apparently, considerable D p. 156. fissuring occurred from Lake Merced to the sea. C 1852 Holden, 1852. November 22; 11 p. m.; VIII. Severe shock eight miles south— : 1898, east of San Francisco. Next morning a fissure half a mile wide and p. 39. three hundred yards long was discovered, through which the waters of Lake Merced were flowing to the sea. C 1852 Soule/, Gihon, and November 23d.-—The waters of Lake La Mercede, in the vicinity of the Nisbet, city, and which cover several hundred acres, sank about thirty feet. 1854, Shortly before midnight of this day, a shock like that of an earth— p. 408. quake was felt by parties residing near the place; and the following morning it was discovered that a great channel between the lake and the sea had been opened, through a broad and high sand bank, during the night, by which the waters had found a way and been discharged. m * * * The most probable conjecture is, that the excessive rains of the season had simply forced open a passage through the broad and loose sand-bank from the lake to the ocean. Formerly the lake had no visible outlet whatever; and its waters had insensibly been kept about the same level * * * 245 A 1906 Schussler, * a: * west and southwest from Lake Merced * * * the outlet, or ocean 1906, end, of our brick Lake Merced drainage tunnel was completely covered a p. 9. up and closed by a large slide. 246 A 1906 Lawson and others, Forty-eight Avenue, between K and N Streets, is a district underlain 1908, by deep sand where extensive grading operations were undoubtedly p. 240. necessary to convert an area of sand—dunes into streets and building «0—» lots. Here small, substantial frame dwellings were shifted bodily from 1 to 2 feet out of position, and the streets were slightly dis- located. 247 A 1906 Lawson and others, * 9: * Forty—eight Avenue between K and L Streets, within a few hundred m 1908, feet of the ocean * a: * the sand in our basement raised from 1 foot p. 242. to 18 inches. A wide and long 3—foot depression was raised level. <) lot, which was 120 feet deep, was shorted at least a foot, which was shown by the folding of the fence. Electric—light poles in the street in front of us, which were in the sand, were thrown down north, east, south, and west. There was a fissure for about a block, between Forty— seventh and Forty-eighth Avenues, about 3 feet wide and 6 or 8 inches deep, which was of course in the sand. There were also other blow- 0 holes in the sand, which emitted water and sulfurous odors. 248 A 1906 Seattle Post- There is one place within pistol shot of ruined San Francisco that I] Intelligencer, the earthquake did not touch, that did not lose a chimney or feel a 1906c. tremor—-A1catraz island. Despite the fact that the island is covered with brick buildings, brick forts, and brick chimneys, not a brick was loosened, not crack made, not a quiver felt. TABLES 5—9 147 TABLE 8.—Specific descriptions of ground failures in the north bay counties region Location number is assigned to each reported ground—failure site. Corresponding numbers are found on plate 4. Figure number refers to figure in this report showing damage described under "Quotation" column. Failure type is indicated by the following symbols. Corresponding symbols are found on plate 4. [2 Hillside landslides including rotational slumps, 1:: River stretches with extensively fissured flood block glides, debris avalanches, and rockfalls "' plains; pattern indicates stretches of river affected and not width of disturbed zone 0 Streambank landslides including rotational slumps and soil falls <> Sand boils 0 Lateral spread E] Absence of ground failure noted X Ground settlement m Miscellaneous effects I: Ground cracks not clearly associated with land— <—{]—> Arrows showing extent of area affected. slides, lateral spreads, settlement or Symbol shows failure type primary fault movements Accuracy with which failure sites can be located is given as follows: A, a site that can be accurately relocated; B, a site that can be relocated to within a few kilometers and probably could be located more accurately with further inves— tigation; C, a site where the information is insufficient to allow precise location. Plate numbers in the "Reference” column refer to plates in the original source material. Loca- Fig— Fail— Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake 249 C 1906 Jordan, At Sobrante, in Contra Costa County, east of San Francisco Bay there 0 1907, are large slumps or cracks in the earth. p. 33. 250 C 1906 Salinas Daily Index, The Santa Fe's Condition. * * * The railway bridges and railyards )[ 1906c. at Pinole sank two feet. 251 B 1906 Lawson and others, More Island.--The earthquake was much less severe than that of 1898, 1908, which wrecked many of the Government buildings in the navy—yard. p. 212. None of the Government buildings was wrecked this time, nor was the damage at all serious except in the case of two or three new build— ings recently erected on the ”made" land near the water—front. Here EU the ground was thrown into violent undulations, and the buildings were so twisted that about $2,000 worth of repairs had to be made. On this soft ground the brick walls were cracked * * * . In the case of the older buildings resting on hard ground, no cracks were formed, nor any injury reported. I! B 1898 Townley and Allen, 1898 March 30. 11:43 .m. VIII. Mare Island. San Pablo Bay. P 1939, This earthquake wrought such damage at Mare Island Navy Yard that p. 105. it may properly be known as the Mare Island earthquake. * * * Admiral H. W. Lyon, U.S.N., has furnished the following information: " * * * The violence of the shock was greater than any shock pre: viously experienced on this island, as far as can be learned from the oldest inhabitants. "A detailed account of the damages done is set forth in a report to the commandant, dated April 5, 1898. 252 B 1906 Lawson and others, The railroad track east of Martinez, near Bull's Head Oil Works, <)> 1908, was thrown 3 inches out of alignment to the north. Many cracks p. 310. occurred in the embankment on both sides of the track. A series of 5 small transverse waves was found in the embankment about 0.5 mile west of Peyton Station. The distance between crests was about 10 to 15 feet; amplitude estimated at 3 inches. This embankment lies in flat marshy land. A small railroad bridge near Avon Station was <)’ thrown 4 inches toward the east abutment, but it had been repaired at the time of the visit. 148 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 8,—Specific descriptions of ground failures in the north bay counties region—Continued Loca- Fig— Fail— Accu- Year of Reference Quotation tion ure ure ra- earth- No. No. type cy quake 253 C 1906 San Francisco Stocton, April 30.--It was discovered to—day that in the vicinity Chronicle, of Woodbridge the Mokelumne River has fallen twelve feet, the bed of m 1906b. the river having dropped from the effects of the recent earthquake. The stream had been carrying a lot of water when it was noticed that the river was steadily falling, contrary to all precedent. The people could hardly believe their eyes. They watched the river recede for a day and made an investigation, with the result that the bed was found to be almost twelve feet lower than before. As the water way has been steadily filling up each year with silt from the upper portion of the river, farmers along that stream are highly pleased with the change, since it can carry far more water than heretofore, and not endanger their lands on either side of it. Another incident of the earthquake is the drying up of Tracy lake, in the northwestern part of San Joaquin county. Ever since the earthquake the water has been decreasing at a rapid rate, and at present it is almost on a level with the Mokelumne river. Whether or not there is a crack in its bed, or an underground passage connecting the lake with the river, is not known, but at the rate the lake has been falling it will soon be as low as the river. The benefit in both instances will be appreciated by farmers. C 1857 Sacramento Age, We have information of severe effects of the earthquake along the 1857. line of the lower Stockton Road. Below Benson's Ferry the waters of the Mokelumne river much swelled by recent rains, were thrown over m the banks, leaving the bed of the stream almost bare. Houses were shaken violently, destroying articles of glassware and over turning furniture. Limbs were broken off from trees and the trees in some instances settled down two or three feet into the ground. The inhabitants of that section were terror stricken, whilst dumb brutes appeared to be paralyzed. 254 C 1906 Duryea and others, At one point on the marsh between Benecia and Suisun, on the Southern )[ 1907, Pacific, the settlement was 11 ft.; at another point, 5 ft. These p. 258. were nearly vertical. C 1906 Ransome, On the north shore of Suisun Bay part of the track of the Southern l[ 1906, Pacific, laid on marsh, subsided several feet. p. 294. )I C 1906 Davison, Three miles of railway have sunk out of sight between Suisan and 1906, Benecia * * * . p. 25416. C 1906 Engineering News, Farther east [of Oakland and Berkeley] the Southern Pacific Co. 1906. suffered much disturbance of its railway lines * * * . A section of )( track between Oakland and Sacramento sank several feet; a railway bridge over the San Joaquin settled some inches [see 10c. 182]. C 1906 The Evening Post, Effect of Shock Between Susan [sic] City and Benecia. A telegram 1906e. from Sacramento to the Western Union Telegraph Company's office in II this city, reports that three miles of railroad sank out of sight as a result of the earthquake between Suisan [sic] City and Benecia, in Solano County, and all wires were taken with it. * * * reported sinking of a three-mile section of the railroad company's tracks between Suisun and Benecia, which are on the direct line between Sacramento and San Francisco. The road crosses some low land at the point where its tracks are reported sunk. The location of this sinking of the earth is about thirty miles from San Francisco. C 1906 The Evening Bee, Trains Brought Back. Trains which had been dispatched for San 1906a. Francisco early this morning had to be brought back, and they were sent to the Bay City by the Lathrop route. II It was at the spot where the track disappeared that the railroad company had so much trouble last Winter, when a loaded passenger train came near going out of sight. A great army of men was then set at work to fill up the sink. The task was a most difficult one, TABLES 5—9 149 ‘ TABLE 8.—Specific descriptions of ground failures in the north bay counties region—Continued Loca— Fig- Fail— Accu— Year of tion ure ra- earth- No. No. type ure cy quake Reference Quotation 255 256 OI] C 1906 B 1906 B 1906 C 1906 C 1892 C 1892 C 1892 Los Angeles Daily Times, 1906C. Public Ledger, 1906. New York Tribune, 1906. The Evening Bee, 1906b. Holden, 1898. p. 174 and 176. Holden, 1898, p. 187. Holden, 1898, p. 178 and 179. as enormous timbers which were thrown into the hole quickly sank from View, and the trainloads of earth dumped in disappeared like snow in a fierce sunshine. But the engineers finally mastered the situation, and after a week's time trains were sent over the route again. It is now feared that it will take several weeks' time to repair the present colla se, and in the meantime all trains will be sent by the Stocton [sicfi route. EARTH CRACKED OPEN. There are great crevasses on each side of the track through the Suisun marshes and it is reported that a great ocean of water has flowed over the lowlands between Suisun and Benecia. ENGINE SINKS IN EARTH. A short distance below Suisun, a Southern Pacific switch engine sank into the ground for a distance of three feet, not far from where the track disappeared. Sacramento, April 18.—-A short time after the big shock came a message from Suisun, Solano county, saying that a long section of track had disappeared from View. It was learned later that, in one place between Sprig and Teal stations, in the Suisun marshes, for a distance of one mile and a half, the track had sunk down three to six feet, and at another point nearly one thousand feet of track went out. * * * The track sunk by the earthquake is near the place where a loaded passenger train came near going out of Sight. There are great crevices on each side of the track through the Suisun marshes, and it is reported that a great ocean of water has flowed over the lowlands between Suisun and Benecia. A short distance below Suisun, a Southern Pacific switch engine sank into the ground for a distance of three feet, not far from where the tracks disappeared. *** Sacramento, April 18.-—A short time after the shock of the earth— quake a message came from Suisun, Solano County, saying that a long section of track had disappeared from view. One place between Sprig and Teal stations in the Suisun marshes for a distance of one mile and a half the track had sunk three to six feet, and at another point near 1000 feet of track went out. Sacramento, Cal., April 18—— * * * It was learned that between Sprig and Teal stations for a distance of one mile and a half the track had sunk three to six feet. At another point nearly a thousand feet of track sank from sight. The Southern Pacific Company repaired its tracks beyond Suisun yesterday afternoon and trains are now running direct to San Francisco. The local officials state that the rumor to the effect that the tracks had gone out of sight was not so, and it took but two or three carloads of dirt to level the tracks. This was done by 2:30 yesterday afternoon and last night trains were running through to Oakland on schedule time. 1892. places. April 19; Vacaville. * * * The ground was fissured in many 1892. April 20; Winters. At Winters there have been developed a number of fissures in the earth, water has been ejected, gas has escaped, and the bed of the creek has been filled up for a distance of over 70 yards. Many of the wells have been filled up by the collapse of the walls. . 1892. April 19; Winters; 2h. 50m. a. m. 0n Putah Creek, half a mile west of Winters, a phenomenon was witnessed by a young man named Fred Willis, who was riding past at the time of the big shake. There seemed to be an explosion, and the water was thrown from the creek to a distance of 20 feet on either bank. Then follow- ed a hissing sound as of gas escaping. At daylight several fissures 150 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 8.—Specific descriptions of ground failures in the north bay counties region—Continued Loca— Fig- Fail- Accu— Year of Reference Quotation tion ure ure ra— earth- No. No. type cy quake were found in the bed of the creek and in the roadway and fields adjoining. On each side of the creek where the explosion took place G) the banks caved in, the landslides being 75 feet in length and 12 feet deep. C 1892 Holden, 1892. April 19; Winters; * * * near the town the bank of Putah O 1898, Creek, ten feet wide, caved in, and along the bottom of the creek p. 178. for a great distance rents were made by the shocks. West of here about three miles, an acre of ground slid into the creek. I: C 1892 Coffman, 1892. April 19. * * * Fissures were found in the bed of Putah 1973, Creek, 1/2 mile west of Winters, and in the adjoining roadway and G) p. 160. fields; banks of the creek were caved in. C 1892 Holden, 1892. April 21; Winters. * * * The railroad track is all right, 1898, and telegraphic communication has been uninterrupted, but there are I: p. 195 and 196. many nasty cracks and fissures in the roadways, and driving is dangerous. C 1892 Holden, 1892. April 21; Winters. * * * The sand bars in Putah Creek near () 1898, Winters opened and from the fissures the water spurted high up on p. 195 and 197. the banks. In some places the creek became dry, in others it changed 0 to a torrent. The banks caved in some places and almost dammed the stream. 257 C 1892 Holden, 1892. April 19; * * * Up the Berryessa road the passage is blocked I] 1898, by immense bowlders [sic], some weighing several tons, which were p. 178. thrown down the hillsides into the road. It is near this point where the rents in the road were noticed. 258 m C 1892 Holden, 1892. April 21; * * a: It was reported that several boiling springs 1898, had burst from the foothills on the north and west and were flowing p. 195 and 197. steadily. 259 I: C 1892 Holden, 1892. April 19; Esparto. * * * The earth opened in several places 1898, between here and Capay. p. 184. 260 A 1906 Duryea and others, The draw—bridge at Black Point, over Petaluma Creek, on the Sonoma 1907, Branch of the California Northwestern, was open at the time of the p. 259. earthquake, and was thrown off its center 2 ft. to the east and 1 ft. to the north. This is a steel structure, 220 ft. long, on four iron caissons, filled with concrete, on pile foundations. 261 B 1906 Lawson and others, The buildings of the Dickson ranch, 2.5 miles south of Olema, are 1908, about 0.25 mile east of the fault-trace, standing on a hillside E p. 192. presumably on firm ground. They nearly all slid southwest--that is, downhill and toward the fault. 262 I: C 1906 Lawson and others, Bedrock cracks occurred at many points within the Rift, usually 1908, appearing as branches from the faults. They were seen also at a p. 75. number of points west of the Rift, their distribution reaching to the ocean in the vicinity of Point Reyes, ten miles from the fault— trace. At the more remote points they were quite small, often barely discernible, and no system of arrangement was discovered. They are peculiarly prominent along the summit of the ridge constituting the southwestern rim of the main Bolinas-Tomales trough. This summit was visited on four lines of road [locs. 265, 275, 276] and at each :: locality conspicuous cracks were found. 0n the road from Inverness to Point Reyes Post Office [loc. 276], about a mile in a direct line TABLES 5—9 1 5 1 TABLE 8.—Specific descriptions of ground failures in the north bay counties region—Continued Loca- Fig— Fail- Accu— Year of Reference tion ure ure ra— earth- No. No. type cy quake Quotation = C 1906 Lawson and others, 1908, p. 195. 263 B 1906 Lawson and others, 1908, p. 192. 264 B 1906 Lawson and others, 1908, p. 193. 265 = B 1906 Lawson and others, 1908, p. 75. 266 58 I: C 1906 Lawson and others, 1908, p. 74. plate 49B C 1906 Lawson and others, 0 1908, 59 plate SOB. from Tomales Bay, a crack was traced for more than 800 feet. Its general trend is east and west, but its course is not straight and it has a branch diverging at 45°. Along this crack there is a hori— zontal throw of from 2 to 6 inches, the south side having moved westward with reference to the north side. In this region [Bear Valley] I saw only a few cracks other than road-cracks, and the road-cracks were unimportant. Mr. Payne J. Shafter's place is near the village of Olema. The fault—trace is close to the house and other buildings. These stand on a bed of alluvium which is probably supported by bedrock at a short distance below the surface. In the barnyard men were milking, and were thrown violently to the ground, along with the cattle. The buildings were much damaged. During the earthquake a cow fell into the fault-crack and the earth closed in on her, so that only the tail remained visible. At the time of my visit the tail had disappeared, being eaten by dogs, but there was abundant testimony to substantiate the statement. As the fault-trace in that neighbor- hood showed no cracks large enough to receive a cow, it would appear that during the production of the fault there was a temporary parting of the walls. [Some have discredited this report. See Earthquake Engineering Research Institute Newsletter, v. 9, no. 1, p. 103. Olema.--The village of Olema is about 0.5 mile east of the fault—‘ trace and at the edge of the Rift belt, the greater part being included within the Rift. The residence of Mr. Pease, standing on alluvium, was shifted south about 2 feet, falling from its supports. * * * A neighboring piece of alluvial land bordering Olema Creek sank about 2 feet. On Mount Whittenberg there are two bedrock cracks. One of these crosses the northeastern spur of the peak near its junction with the main crest. Its trend is approximately northwest and southeast and at one point it margins a fault—sag. As it assumes in one place the ridge phase of the fault—crop, I infer that it has horizontal dis— placement. 0n the opposite side of the main crest is a crack which was traced for about 1,000 feet. Its general course is northwestr southeast, but it is not straight and exhibits a vertical throw of l or 2 feet to the southwest. At one point it touches a fault-sag. Between these two long cracks a group of short cracks occurred, with similar trend, on a knob constituting a portion of the main divide. [These cracks may have been caused by secondary faulting.] They [cracks in alluvium] were seen from the train in the bottom- land of Papermill [Lagunitas] Creek within a mile of Point Reyes Station. They were also seen in the delta of Papermill Creek, in the bottom-land of Olema Creek near Olema, and in the delta of Pine Gulch Creek. They were seen in the bottom-lands and deltas of a number of small creeks entering Tomales Bay from the west between Inverness and the head of the bay. Other localities were tidal marshes at the head of Bolinas Lagoon (plate 49B), at the head of Tamales Bay, and in small estuaries near Inverness. They were seen in the marsh of Bear Valley Creek near where the stream joins Papermill Creek; and a road embankment crossing that marsh was elaborately cracked and faulted thru much of its extent. [Picture caption] Faults in road embankment, southwest of Point Reyes Station. Fault-trace is beyond fence. Ground lurched toward marsh of Bear Valley Creek. G. K. G. 152 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 8.—Specific descriptions of ground failures in the north bay counties region—Continued Loca- Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake A 1906 Jordan, Paper Mill Creek [Lagunitas] runs past the same village [Point Reyes 1907, Station], * * * . The two banks of the stream were forced toward each <) p. 19. other so that the length of the bridge was shortened by about six feet and the bridge was correspondingly humped at its north end, an arch about six feet high being forced up. B 1906 Duryea and others, * * * a highway bridge across a creek tributary to Tomales Bay, near 1907, Point Reyes Station on the North Shore Railway, in Marin County, and p. 270, within less than 2 miles of the fault line. * * * plate 54. The bridge, originally, had eight panels, its total length being approximately 120 ft. It is located in a direction nearly north and south. The abutments are of piles and timber. The effect of the )[ earthquake was to settle the north abutment some 2 or 3 ft., and move it southward so much that, in patching up the structure temporarily, the north end panel was not utilized as part of the span * * * . B 1906 Duryea and others, Portions of the trestle over Lagunitas Creek, about a mile from 1907, Point Reyes, were thrown entirely off the piles, the piles themselves <) p. 258. being moved down stream. C 1906 Lawson and others, Beyond Garcia the creek has several reaches of alluvial bottom, and 1908, some of these were so badly shaken that the railway embankments and m p. 196. trestles had to be repaired. Railway traffic to Point Reyes [from Sausalito] was interrupted for about 10 days. 267 I: C 1906 Lawson and others, [Picture caption] Road embankment broken by shaking of soft ground 1908, beneath. Southwest of Point Reyes Station and 10 rods from fault— 60 plate 50A. trace. G. K. G. B 1906 Lawson and others, The road running southwest from Point Reyes Station and crossing 1908, the valley at the head of Papermill Creek [Lagunitas] delta was p. 71. offset 20 feet [by faulting]. * * * The embankment of the road rested on marshy ground so soft that a portion of the embankment )[ sank into it, and material of this character was in other localities <) demonstrably shifted. 268 C 1906 Lawson and others, I drove a few miles north and east from the [Point Reyes] station, 1908, over a high terrace separating the upland from the bay at the east. :: p. 196. * * * A few cracks were seen in the ground, but they were much smaller and less numerous than at a similar distance on the oppo— site side of the fault. 269 )K B 1906 Lawson and others, The ”fills” across the arms of Tomales Bay generally sank from 2 1908, to 8 feet. The 1,000-yard fill about 2 miles north of Point Reyes p. 197. Station sank from 6 to 8 feet; as did the next fill, which is some 500 feet long. In one or two instances the pile—supported bridge in the middle of the fill remained at grade. * * * The bottom of the bay in these arms is usually sand. B 1906 Duryea and others, On the North Shore, about 2 miles north of Point Reyes, the road, 1907, originally, had been constructed with pile trestles across several p. 258. arms of Tomales Bay; these trestles had been filled about 15 years ago, the road-bed being about 8 ft. above ordinary high tide. Two of these embankments, 2 200 ft. and 900 ft. long, respectively, )[ sank until the water at high tide washed over the rails. C 1906 Weatherbe, Along the coast, the railway was greatly disturbed, invariably )[ 1906, sinking in the low swampy land except where built on piling. p. 402. 270 C 1906 Lawson and others, A large portion of the delta was thrown by the earthquake into m 1908, gentle undulations, the difference in height between the swells p. 78 and 79. and hollows being usually less than a foot. The chief evidence , TABLES 5—9 153 ‘ TABLE 8,—Specific descriptions of ground failures in the north bay counties region—Continued Loca- Fig- Fail— Accu— Year of Reference Quotation tion ure ure ra- earth— No. No. type cy quake 61 plate 54B, of this is found in the distribution of pools at low tide, and where 62 plate SSE, vegetation is present the evidence from pools is supplemented by plate 56A. that from the condition of the plants. The undulations were not elongate and were not found to have a systematic relation to the fault. When the tidal mud was first seen after the earthquake, it was observed to be covered with ridges and troughs. (Plate 543.) This corrugation was gradually smoothed out by the action of the waves (plates 55B and 56A), so that at the expiration of a year its ex- ‘ pression was largely lost, tho a few of the larger ridges could still be traced, and much of the plain retained a pattern imprest on it by‘ the ridging. It is probable that the entire tract of tidal mud was thus affected, altho the ridges were not seen on the area lying nearest to the east shore. That area did not come under observation until after the spring floods of 1907, and it was then overspread by a fresh deposit brought by Papermill Creek [Lagunitas Creek]. The ridges varied somewhat in height, the amplitude from crest to trough ranging from 1 to 3 feet and possibly more. Their general trend was parallel to the fault-trace, but there were notable excep— tions, and over small tracts the direction was even at right angles to it. In some cases, where the minor ridges were parallel, there were larger ridges traversing them obliquely. Fig. 25 reproduces a sketch map of the locality showing the greatest complexity. [See fig. 25 at end of tables.] So far as the broad undulation of the tide lands were seen in conjunction with the ridging, the greater ridges were on the swells and not in the hollows. 271 A 1906 Lawson and others, [See text, ”North bay counties region," paragraph 5, for additional 1908, discussion of failure mechanism and direction of movement.] There Sewell, and Soulé, shifting of mud on the bottom of Tomales Bay. At the head of the 1908, bay and thence for a distance of several miles northwestward the p. 8 and 9, soft mud was moved bodily westward. It not only descended from the 62 plate 7A, northeast shore, so as to cause deeper water, but ascended toward 61 plate 8A. the southwest shore, creating a broad shoal (P1. VII). The hori- zontal change of position near the southwest shore was in places (p. more than 25 feet, and the vertical change as much as 2 feet. As the ascending movement can not be ascribed to gravity, it must be referred to the earthquake, even though the way in which the earth waves produced the effect is not evident. The locality is adjacent to the fault trace, the position of which is along the bottom of the bay, east of the shoaL The illustrations may require a few words of explanation. The upper view of P1. VII looks northward from the southwest shore of the bay. Tide being low, the newly formed shoal or mud bank is broadly exposed, but the receeding tide has left a lane of water to mark the separa- tion of the mud bank from the firmer ground that withstood the quak- ing. Immediately after the earthquake the mud was rigid, as in the tract shown in Pl. VII, A; but before the view of P1. VII, A, was taken (April 28, 1906) the surface had been largely smoothed by the action of wind waves. A single ridge which escaped that action appears at the left in the upper view of P1. VII and in the fore- ground of the lower view. B 1906 Lawson and others, Inverness is a village of summer residences on and near the south- 1908, west shore of Tomales Bay. The upland of the peninsula there closely p. 194, approaches the bay. The village occupies two narrow valleys normal plate 453. to the shore, and a mesa between them. Its site is within the Rift, :: and both valleys and mesa were traversed by many cracks, of which some had the character of branch faults. All the houses were of wood. About half of them were shifted on their foundations. To a certain TABLES 5—9 155 ‘ TABLE 8.—Specific descriptions of ground failures in the north bay counties region—Continued Loca— Fig- Fail— Accu- Year of Reference tion ure ure ra- earth- No. No. type cy quake Quotation 272 = B 1906 Lawson and others, 1908, plate 51B. 273 B 1906 Lawson and others, 1908, = p. 195. 274 64 C 1906 Lawson and others, [2 1908, p. 77. 275 I: C 1906 Lawson and others, 1908, p. 75. 276 C 1906 Lawson and others, 1908, p. 194. a C 1906 Lawson and others, 1908, plate 51A. E] C 1906 Lawson and others, 1908, 65 plate SSB. 277 B 1906 Lawson and others, I 1908, p. 195. 278 A 1906 Lawson and others, m 1908, p. 197. extent the direction of shifting was determined by the slopes of the ground, the houses moving downhill; but where that factor did not control, the movement was toward the west or southwest. In one instance I noted a southwestward movement of several feet uphill. [Picture caption] Roadside crack a mile southeast of lnverness. G. K. G. Sunshine Ranch and Vicinity.--I drove to the summit of the ridge southwest of the head of Tomales Bay, finding abundant and strong road-cracks all the way to the crest, which is about 1.5 miles from the fault—trace. There were also a number of landslides in this region, and a considerable number of trees were broken or uprooted. On the west side of the main ridge west of the head of Tomales Bay there occurred two wet slides. In one case a hillside bog was loosened from the slope on which it rested and descended as a flow of mud to a canyon bottom 100 or 200 feet below. In the other case, the earth beneath a wet meadow in a rather steep canyon flowed down the canyon for about 0.5 mile, overpowering trees on its way and leaving a deposit 15 or 20 feet deep in places. This was the largeSt individual slide observed. In all the cases mentioned the conditions were such that slides would have taken place at some time had the earthquake not occurred. On the next road to the southward [from loc. 276] a group of cracks was seen at a point a mile from the shore of Tomales Bay. These cracks occur on a crest trending northwest and southeast, and their trend makes a small angle with that of the crest. The arrangement of the cracks suggests horizontal shear, but no definite observation was made on this point. They extend for several hundred feet at least, but were not traced out. Inverness to Point Reyes Light—house.--For the first 2 miles of travel, covering a right—line distance of about 1.5 miles, road- cracks were numerous and often large. There were also numerous small falls of earth from the road cliffs. Beyond that point there was a rapid falling off of such evidence, and tho road—cracks were‘ frequently seen they were all small. [See loc. 262.] [Some of these cracks may have been caused by secondary faulting.] [Picture caption] Roadside crack 2 miles west of fault, between Inverness and Point Reyes P. 0. G. K. G. [Picture caption] Landslide from road-cliff about two miles west of Inverness. Slide occurred at time of earthquake. G. K. G. At the U. S. Life Saving Station, on the coast 3 or 4 miles from the light-house * * * the filled ground about the house settled several inches. , At Marshall's a hotel and a stable built on the west side of the track and on underpinning, resting in the tidal flat, went easily‘ and gently into the bay. The occupants of the hotel did not realize that the hotel had fallen, but at first thought the water had risen. [Photographic evidence in the Branner Collection at Stanford suggests this‘may have been structural failure, that the building simply slid off its piles.] 156 HISTORIC GROUND FAILURES IN NORTHERN CALIFORNIA TRIGGERED BY EARTHQUAKES TABLE 8.—Specific descriptions of ground failures in the north bay counties region—Continued Loca- Fig- Fail- Accu- Year of Reference Quotation tion ure ure ra— earth- No. No. type cy quake 279 )l B 1906 Lawson and others, Just above Hamlet a trestle—work which had been filled in settled, 1908, leaving the trestle—work some 2 feet above. The bottom of the bay p. 197. in these arms is usually sand. a At Hamlet quite an extensive landslide has started in the hillside above the track. The railroad cut is in old rock, and the arch of the head of the slide is some 70 feet above the track. The country wagon road has been carried away by the slide for possibly 100 yards. 280 C 1906 Lawson and others, At the mouth of Tomales Bay there are two points projecting west— 1908, ward from the east shore, and both of these, according to the obser— p. 65, vations of Prof .R. S. Holway, are crost by the fault-trace. * * * (D plate 1423. On each side of the crack [fault] are crater-like depressions, some of them being double or overlapping. Mr. Keegan, the owner of Dillon's Beach, reported that these craterlets were numerous and distinct. In some instances a great deal of sand and water had been ejected. Others are reported on the southwest side of the fault- trace, from which the belt containing them extends some 70 feet. The craterlets vary in size up to 6 feet in diameter and it is re— ported that on the day after the earthquake the water which stood in them could not be bottomed by a fishing pole. 281 B 1906 Lawson and others, [Near Salmon Creek] * * * a small mud-flat extends from the sea up C) 1908, to the road. Curious mounds of mud, shaped like truncated cones, were p. 191, thrown up by the earthquake. Subsequent examination showed that the plate 142A. line of the earthquake fissure must have past near this spot. 282 C 1906 Lawson and others, There are quite a number of cracks in the flat valley-bottom adja— 1908, cent [to Valley Ford, Sonoma County]. A landslide of several hundred p. 199. yards in length but of very slight movement is found on the side of the valley directly east of town. The slide has moved just enough to make a furrow—like ridge on the lower side and has developed cracks on the upper side. Other small slides occur in the vicinity. 283 B 1906 Duryea and others, Trestles.—-Trestles over marshes suffered more or less from the <)> 1907, movement of the soft material into which the piles were driven. The p. 258. most serious damage was at Fallons, on the North Shore, where a trestle, 600 ft. long and 70 ft. high, was thrown down. This trestle was constructed of framed bents on piles. B 1906 Duryea and others, Trestles.--The damage to trestles was small, except on the North 1907, Shore Railroad, where a trestle of framed bents on piles, 600 ft. p. 214 and 215. long and 70 ft. high, was thrown down, and portions of another trestle were thrown entirely off the piles, the piles themselves <3» being moved down stream. These trestles were across soft ground, and near the fault-line. 284 I: B 1906 Lawson and others, Tamales, Marin County. * * * Cracks were reported in the street 1908, and near the depot. Just north of the depot there was an extensive

1908, a layer of moist soil only a few feet in thickness moving down the p. 205. slope, introducing bends in various lines of cultivated plants. I saw another feature of this sort on an adjacent farm, and was told of others which I did not visit. B 1906 Lawson and others, [David Starr Jordan] At Burbank's farm, 0.5 mile west of Sebastopol, 1908, I noted these things: In the lot adjoining, to the south, the soil p. 204. being clayey, there is a large crack running northwest and southeast, or nearly so, and, according to Burbank, 0.25 mile long. It runs thru the fields and weeds, and was very distinct on August 6. The end of this crack comes up against the sandy hill occupied by Mr. Burbank‘s orchard. The crack does not show itself in the hill, but on the east side of the line of the crack the rows of trees and \ ' I‘\I'li(l5h‘«h ‘ , , ,. r, fkr’edra,‘ 'Azul /‘ , x‘MH H?“ H Mos-:91 La :19 g _ . " , R . , _, 1 I, M \ *_ _ ‘ ,, ,1 _ __ 3 / __ _. ‘ _ g . I Amway mum ammv m. u H 45’ Bolsa 45 malls J. ‘ ta Réta Vlllage Heights?! ‘32“. J . ‘2 l . , f Mama Vista J’Parky” <3 i T. 14 S. . , e9 Stordge \ MST __ T _._ *r‘v’wws ,. _ "r Pines . ‘WHGNTEREV/'ii ‘ _.\Ai’I-:l§$~’i3)5n mr . ‘ , (mercury ‘ I ‘I -. ’"Tf’ ~~fr~r~ .4 — “ ITIS 3W] 1 L _+|_ J Ptifqlpingsgfl‘grfl ,3 .— ,—— +— -- —— -— —— —— —— 4— * I . ,. Efngglfi... ‘Chiiney I mama ' I \I ~ T. 15 S. i Rt", 0;: ‘ I , ' ' i L \ i ‘ ' l ‘ ' / l 'i I: i’ .. . ._ y ,_ \ » r. . \ ~ _ 7,. / I Klan \ oz? : 30' 30' 'j\\ \/ {Semaphore , _ f , /_ * . 1 _ . \ \PH 1 1 5* W( ((17111 e: sVIi'I'm'? \\ Lax, (\ \vrnj‘UA ", . _,\. .Wate‘r\ 15' ' ‘ » ., - - , . - ‘ _ , , \ -- - , > » , £3le ., T. 19 S. 15' f’felffer Poser ‘ I‘\-H . , 1.. _ r; _ w Canyoq r, , _ ~ / (p > /',/" ' i Grimes Point <72; . I9 \(g .7 "5'2 “ T. 20 S. Parlingtoh Point . LOPE?‘PDIV‘I’ \ '\ \§. 14mi.S ' / < ., , . r , ‘ _ _ , __ T 225. "7's ,7 '< - ’ 1 7 < Bradley 121° R.9E. R.10 E. r ,/ _ 3 0 : R.14 E. R-15E- FI-I‘3'5-12o"156'00 {:7 Interior~Geological Survey, Reston, Va. —l978—G76254 36°00' 122° 12' 122 45' R.2E. R.3E. R. .5E. R. Base from US. Geological Survey Santa Cruz 1:250,000 series 5 0 5 10 15 20 MILES l 5 0 5 1 H H H H H '0 1.5 2° 25 3O KILOMETERS . 1 CONTOUR INTERVAL 200 FEET DOTTED LINES REPRESENT 100 FOOT CONTOURS DATUM IS MEAN SEA LEVEL LOCATION MAP OF GROUND FAILURES IN THE MONTEREY BAY COUNTIES REGION EXPLANATION Failure type is indicated by the following Quotations describing these ground fa found in table 5 block glides, debris avalanches, and I <3 <>, II I: Lateral spread Ground settlement Streambank landslides including rotational s and soil falls I n Hillside landslides including rotationa‘E‘slumps , PROFESSIONAL PAPER 993 PLATE 1 s", . «gimme avg.-. By; alls . F Ground cracks not clearly associated with landslides, lateral spreads, settlement, or primary movements not width of disturbed zone C) <8 C] RI Sand boils Disturbed wells Absence of ground failure noted Miscellaneous effects River stretches with extensively f issured flood plains; pattern indicates stretches of river affected and <—E}—> Arrows showing extent of area affected. Symbol shows failure type 44 Location number assigned to each reported ground failure site. Corresponding number found in table 5 fault, April 18, 19.06 122” Approximate trace of fracture on San Andreas 120° CALIFORNIA 37° — Plate 3 Plate 2 35 o 40 I—rJ—l—l—r o 40 8O 80 MILES _‘_l 120 KILOMETERS PROFESSIONAL PAPER 993 UNITED STATES DEPARTMENT OF THE INTERIOR PLATE 2 GEOLOGICAL SURVEY 121°15' o . 0 . R. 6 E. o , 38°63,3 15 R 7w R.6W 30' R AW. 15' -2W- 122 00 -1W' '1E' ‘ " 38 00 l Point ey T 2 N. \p < LI ' ‘\ R) O R /> K) / // 190 l2 (‘1 \ ‘ / C \_ _J ”’01) (:9 /r\\ ‘y, 300/ C?\\ ,’/’ 7 ‘\\ /r’“ / /\ \\ 2 ~ / \~/1//‘<,:/ i N \ 0 to 06 j I’rombrtea area ‘, :th 530 «c» p l D L) / UM \ o \\ «£3 a,“ M , \- \. I \- ” North Brailon 2) -\ FARALLON ISLANDS s » 6’ 45' _ 44% \\ % \ (f5 (o \ . (do \. \~\ ° Mr‘adre ‘\. Faralhq r9) \ . '\ <6 3/ ’50 '\, , § ‘ Iheast Farallon .J- ‘b 554% t , QNO " \ \ O \\ Q 30’ — T. 5 S. G! O O \ LL. I, o o o ’\ . 6 Sr .124 122 120 — \J I | EXPLANATION Failure type is indicated by the following symbols. Quotations describing these ground failures are 410 L _ found in table 6 a Hillside landslides including rotational slumps. block glides. debris avalanches, and rockfalls R o Streainbank landslides including rotational slumps 7 s ,, and soil falls ‘ ' Plate 5 _ o g 0 Lateral spread I RNIA , CAL F0 3\ ~ Ground settlement a 95 8? \\\ . . . ‘ = Ground cracks not Clearly assocrated With landslides, o lateral spreads , settlement, or primary fault a 39 —‘ . movements 91: 92 a 86 15, 15' — \\ River stretches with extensively fissured flood plains; plate 4 pattern indicates stretches of river affected and a 90 / not width of disturbed zone _ A MR1 _ 0 Sand boils 8988 . 8 3. {Plate 3 @ Disturbed wells = D Absence ofground failure noted a Plate 2 87 m Miscellaneous effects 37° — —‘ 59 éD—> Arrows showing extent of area affected. Symbol shows failure type a Plate 1 ~ 68 110 Location number assigned to each reported ground T 9 8 failure site. Corresponding number found in ' ' _ _. table 6 - [2 66 Approximate trace of fracture on San Andreas 67 * B 65 fault,April 18, 1906 ‘ a 350 l | r 0 40 80 MILES l_'_i_|__l_‘_l_l ‘ T. 10 s. 0 40 80 120 KILOMETERS T. 11 S o . _ " 37° 00' 3700 o , 0 . y r, R.5E. R.6E. 121°15 123 15 123 00 45 30 1" 7% Interior—Geological Survey, Reston, Va. —1978 —G76254 Base from US. Geological Survey San Francisco and San Jose l:250,000 series SCALE 1:250 000 0 5 1O 15 20 I—II—II—tl—lI—tI———-——I CONTOUR INTERVAL 200 FEET WITH SUPPLEMENTARY CONTOURS AT 100 FOOT INTERVALS TRANSVERSE MERCATOR PROJECTION 20 MILES 30 KILOMETERS l LOCATION MAP OF GROUND FAILURES IN THE SAN FRANCISCO BAY COUNTIES REGION UNITED STATES DEPARTMENT OF THE INTERIOR GEOLOGICAL SURVEY 1 39° 00' 45’ 30' 15' 38° 00' 124° 00' Base from US. Geological Survey Santa Rosa and Sacramento 1:250,000 Seres 24°00' 41 39 37 o M A N C H ’y / Arenamock‘ / Poinl Ar ‘ _ 0 C ESTHER B/EAQH , STA\ E 553x ‘3 , 120° CALIFORNIA Plate 1 80 80 MILES Hrfl—H 120 KILOMETERS 45’ Haveris Fisgkgi’ks: \ AN .IJ. Robinson Regfja-j‘: \ I”?! \ \ \K’ L x Gua‘Ialg Pm'fwe ‘ x ‘“ _. I2 0 <> I D mad Tut'szr EXPLANATION Failure type is indicated by the following symbols. Quotations describing these ground failures are found in table 8 Hillside landslides including rotational slumps, block glides, debris avalanches, and rockfalls Streambank landslides including rotational slumps and soil falls Lateral spread Ground settlement Ground cracks not clearly associated with landslides, lateral spreads, settlement, or primary movements River stretches with extensively fissured flood plains; pattern indicates stretches of river affected and not width of disturbed zone Sand boils Absence of ground failure noted Miscellaneous effects 6E]? Arrows showing extent of area affected. Symbol 30' 280 shows failure type failure site. Corresponding number found in table 8 LOCATION MAP OF GROUND FAILURES IN THE NORTH BAY COUNTIES REGION Location number assigned to each reported ground Approximate trace of fracture on San Andreas fault, April 18, 1906 (from Brown and Wolfe, 1972) 15' Badefig H sweat , \ may“! ietgps- 11*, _\ ~ 123°00' a. 9w. 45' n. 8W. R. Point l324i Monitor CL EA LAKE Luebow Point u'l/t'r’(' , \lx’tr fi/zlgzlfl l : RAN "59/ i M ILITA , ijESERVAI , , ‘65/ //'4:u Mun A v: 7m DRARES B 123°00' 45' SCALE 1:250 000 5 5 10 15 20 MILES I—I I—-l I—I I—I I—I ' | . J 5 5 10 15 20 25 30 Kl LOMETERS I CONTOUR INTERVAL 200 FEET WITH SUPPLEMENTARY CONTOURS AT 100 FOOT INTERVALS TRANSVERSE MERCATOR PROJECTION z? s x Hamilton AFB . V 1 4 30' Rise. ' Wm mill: Page“ *A x ,, . Midshipman Point Pomtz M Point : In! I "um:- ; ,B-Emcm , ARQEN a“ ‘5}, a /, Ill-l Y c)” 122° 00' (l NIZZL Y RlW ‘allorm “Mu! \ yéws TB . Water I LO COU NTY _ Judge ,C 253 m l ,t d sposal ma ISQAIQEL F / Y MY U NI J . Coflnty I a: TracyviIIe 7 m NW of Lodi Woodbridge 2.8 m NW of Lodi \ (Dorm i {MARY ATION S AT'I ( 7f R15 PROFESSIONAL PAPER 993 PLATE 4 121°45' 39°00' ___S_.U IT— (kw! China 8 nd Hiutl fluke, 45' Pia: ire}? N {my} , BO’Vis . Radio stataon t . V/ :13 fo ii}K I f,\‘ [gar-1)) 09 Island/\ rs'firlrmc 9 17 ‘ I75 NAM AI’ONS , 8 o 5,860 R I) 6267 ‘ 122°00’ (iv If, I ), Island a Chipps I 121°45' fimterior — Geological Survey, Reston, Va. — 1978— G762 54 15' UNITED STATES DEPARTMENT OF THE INTERIOR GEOLOGICAL SURVEY 122°30' 37° 50' 2 T.28. V: 1x3 9‘ Mile ROCK ‘- L1g‘rthoqse O“ 47’30” {a BAKERS BI; sway ' ' Huger) PHELAN BE d , (heficeg A Reserve} > eel} cm? L C P (Iiann , {I 1‘ HS , 45' . .; I» . '1 ."L , . “ 1‘ V.‘ SR OI 181*! ,3 , fathom ,1 . 3' UFISCI rm Flu T. 2 S. 37°42'30" Base from US. Geologizal Survey 1%: San Francisco North and San Francisco South TRUE NORTH GNEr /C NORTH MA APPROXIMATE MEAN DECLINATION. 1978 LOCATION MAP OF k0 C) Fort Point Rocks I“) > SAN FRANCISCO/ /,/'M/GOL/F/ CLU B I R. 5 W. 27'30" FORT M A SON MILITARY «Anita Rock ,3. , ' , gfigngfi)“ a (01)“ VI YE” Y4 grown 7" Sr'EfiiyvlimR- - \ R. 6 w. 2730” R. 5 w. SCALE 1:24 000 . ‘ a o 1 1 5 O . 1KILOMETER l—-l l-—- A (3‘ 9V 47130” R W incon Roirrt I‘sztt'dl Bash? . ‘I ?\ / ~ ' 1 _ xx , . Sham 8mm. /' " ID . {gouble Rock). - \l" ' VAL, Ilmanp Playgrour‘irx ‘ O _ andlegtlckw Point .‘I ) IADIUMII’ II I \\ \x ,_ \‘1> T.28. WM. . — 1978 —G76254 122022730" 37°42'30" 41 39° 37 35 PROFESSIONAL PAPER 993 PLATE 3 EXPLANATION Failure type is indicated by the following symbols. Quotations describing these ground failures are found in table 7 Hillside landslides including rotational slumps, block glides, debris avalanches, and rockfalls Streambank landslides including rotational slumps and soil falls Lateral spread Ground settlement Ground cracks not clearly associated with landslides lateral spreads, settlement, or primary fault movements Sand boils Absence of ground failure noted Miscellaneous effects Cracks in streets and pipeline breaks §xaao I119 on Arrows showing extent of area affected. Synbol shows failure type 195 Location number assigned to each reported ground failure site. Corresponding number found in table 7 Old shoreline before development from Nichols and Wright. 1971 124° 122° 120° Plate 3 Plate 2 Plate 5 _ CALIFORNIA Plate 4 Plate 1 O 40 80 MILES Infi—LI‘H 0 40 80 120 KILOMETERS ii \McKinleyvte , 354:: f *j V”, -_ \ / T '7 VINE? Shower/s L ‘ % Hydgs'yum ;_ “$7 /, tCarlotta’a -‘ _ “ ‘ .\ 1w, Mendoti: ‘ l _ 1/31 lmnttmfl (2 3 - In Walla Mug/«3 0 C a n W) n Swarm '- Punt: 6‘ Westport . . \ Bruhel Point VtZCBIHO Dome Paint Arm} 39° I ‘ [a , , Rowland ‘,V ’r Mffln ‘1 Mk” 1 Base from US. Geological Survey California State Map, 1971 SCALE 1:500 000 10 1o l—ll—ll—ll—ll—ll—————————l 10 0 10 20 30 CONTOUR INTERVAL 500 FEET DATUM IS MEAN SEA LEVEL DEPTH CURVES AT 100 FATHOM INTERVAL fiInterior—Geological Survey, Reston, Va.—1978—G76254 1230 40 MILES *I 60 Kl LOMETERS ‘I PROFESSIONAL PAPER 993 PLATE 5 EXPLANATION Failure type is indicated by the following symbols. Quotations describing these ground failures are found in table 9 E Hillside landslides including rotational slumps,block glides, debris avalanches, and rockfalls Streambank landslides including rotational slumps and soil falls Ground settlement Ground cracks not clearly associated with landslides, lateral spreads, settlement, or primary fault movements 0 0 Lateral spread l E River stretches with extensively fissured flood plains; pattern indicates stretches of river affected and not width of disturbed zone Sand boils Absence of ground failure noted Miscellaneous effects Arrows showing extent of area affected. Symbol shows failure type Location number assigned to each reported ground failure site. Corresponding number found in table 9 Approximate trace of fracture on San Andreas fault, April 18, 1906 CALIFORNIA Plate 3 Plate 2 Plate 1 0 80 MILES i‘T‘T‘l—l—FJ 0 40 80 120 KILOMETERS LOCATION MAP OF GROUND FAILURES IN THE NORTH COAST COUNTIES REGION