7 DAYS Geology and Resources of Fluorine in the United States GEOLOGICAL SURVEY PROFESSIONAL PAPER 933 b term i mat Mihi eu anus - I voor ore * COVER PHOTOGRAPHS 1 2 4 5 7 8 9 10 11 12 13 14 0 pw N - ~ o Asbestos ore . Lead ore, Balmat mine, N. Y. . Chromite-chromium ore, Washington . Zinc ore, Friedensville, Pa. Banded iron-formation, Palmer, Mich. . Ribbon asbestos ore, Quebec, Canada . Manganese ore, banded rhodochrosite 10. . Zinc ore, Edwards, N. Y. 12 13. 14. . Aluminum ore, bauxite, Georgia . Native copper ore, Keweenawan Peninsula, Mich. Porphyry molybdenum ore, Colorado Manganese nodules, ocean floor Botryoidal fluorite ore, Poncha Springs, Colo. Tungsten ore, North Carolina (Geology and Resources of Fluorine in the United States Edited by DANIEL R. SHAWE With sections by D. R. SHAWE, R. E. VAN ALSTINE, R. G. WORL, A. V. HEYL, R. D. TRACE, R. L. PARKER, W. R. GRIFFITTS, C. L. SAINSBURY, and J. B. CATHCART G EO LOGICAL SURVEY PROFESSIONAL PAPER 953 An evaluation of the geochemistry, geographic distribution, and geologic environments of fluorine, and descriptions of major United States fluorine mineral deposits UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON: 1976 UNITED STATES DEPARTMENT OF THE INTERIOR THOMAS S. KLEPPE, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Director Library of Congress catalog-card No. 76-600061 For sale by the Superintendent of Documents, U.S. Government Printing Office _ Washington, D.C. 20402 Stock Number 024-001-02901-4 APPRAISAL OF MINERAL RESOURCES Continuing appraisal of the mineral resources of the United States is conducted by the U.S. Geological Survey in accordance with the provisions of the Mining and Minerals Policy Act of 1970 (Public Law 91-631, Dec. 31, 1970). Total resources for purposes of these appraisal estimates include currently minable resources (reserves) as well as those resources not yet discovered or not currently profitable to mine. The mining of mineral deposits, once they are discovered, depends on geologic, economic, and technologic factors; however, identification of many deposits yet to be discovered, owing to incomplete knowledge of their distribution in the Earth's crust, depends greatly on geologic availability and man's ingenuity. Con- sequently, appraisal of mineral resources results in approximations, subject to constant change as known deposits are depleted, new deposits are found, new extractive technology and uses are developed, and new geologic knowledge and theories indicate new areas favorable for exploration. This Professional Paper discusses aspects of the geology of fluorine as a frame- work for appraising resources of this commodity in the light of today's technology, economics, and geologic knowledge. Other Geological Survey publications relating to the appraisal of resources of specific mineral commodities include the following: Professional Paper 820-'"United States Mineral Resources" Professional Paper 907-'"Geology and Resources of Copper Deposits" Professional Paper 926-'"Geology and Resources of Vanadium Deposits" Professional Paper 959-'"Geology and Resources of Titanium in the United States" f ; CONTENTS ADSHTACE RNs Li ri rs ress iA 1 Introduction, by R. E. Van Alstine............. fall up Geochemistry of fluorine, by D. R. Sh@We 4 * Abundance Of fIUOTINE T1] FOCKS ...?.1.}:1..» srs sa ners 5 Abundance of fluorine in gases and waters 9 Geochemical Cycle Of censor 12 Element associations, by R. E. Van Alstine and D. R. Shawe........ 14 Geographic distribution of fluorine in igneous rocks, by D "R- SHAWE -an iva nla cars ans inp nanan tha nous 15 Geologic environments and distribution of fluorine, by & D. R. Shawe and R. E. Van AlStINE 17 Fluorine in fluorite deposits and occurrences .......................... 17 Fluorine in topaz deposits and occurrences.......... 19 Fluorine in phosphate and fluorapatite deposits ..................... 19 Fluorine it- OUCY AEDOSILS 19 Tectonic setting of fluorine deposits, by D. R. Shawe - 99 Relation of fluorine deposits to faults......................... (98 f Relation of fluorine deposits to intrusive igneous rocks ......... 95 | Relation of fluorine deposits to geophysical properties........... 27 Descriptions of major fluorine deposits and districts and important types of fluorine OCCUIrences................................ 30 AMaskay by C. L. 3393s 18s esse ere ss 30 Descriptions of individual districts and deposits .............. 30 TL-OSUIRIVET ::: 32002709: ein s i830 Other deposits .. sige Washington-Idaho-Montana-Wyoming-South Dakota, By BAC PATKET .. ccr niet bes nt 34 | Descriptions of individual districts and deposits ............... 34 rss sen eine Hive s eni neces era's i114 es Aisi ns aas cnd eN aie ona er sean ce 35 Meyers (LOVE .. :?:: rs me ista cedes neste rbh nan read es hii 35 Bayhorse ........ Stanley area ......... «£86 Big Squaw Creek. A87 | Other Zee. 2s: dun 37 ssa rite -s in abate shee 38 Crystal Mountain cC 88 }: ::::+»: nese Per deena dr xxs Aba sa rn neh r Acuna sacs 38 BSDALs an rine res Ne NLC snr ec snd 39 Other deposits .. ~ 30 WYORIING .e: assad cns recs re nee 40 1 South cece so inta non 41 Nevada-western Utah-southeastern Oregon-southeastern Cali- fornia-western Arizona, by R. G. Worl and W- #2: io eat teens 41 Descriptions of individual districts and deposits............... 43 raise cl ANAL Teen asia rs 43 Broken Hills:......../...: im: 48 Quinn Canyon Range...... 44 UNAN 2011, Ire Rex raver rete res cabe ria 47 Spor Mountain. 47 aften. 47 Castle Dome .. i> 47 rie. 0.020. 48 Descriptions of major fluorine deposits and districts and important types of fluorine occurrences-Continued Colorado-New Mexico-western Texas-southeastern Arizona, by R. G. Worl, W. R. Griffitts, and R. E. Van AMISUNC 233 - iMe inane or Descriptions of individual districts and deposits .............. (COIOTACO , s.: Hr rre eines een ra navi eng :i + +1223 +1 x rx hap we le eee NOMHGALE 2: is reff d s i Other deposits .. New Burro Mountains.... OHA .-- sien ferret Cooks Peak-Fluorite Ridge. Sierra Cuchillo.:....;.......:..... ZUM MOUNIAIS ..:: 2m ce ieee recess evi gen SIEM AICADAILIOS ;::: :» itn ae cca evita erase cantons SOUtECASEETN UATIZONA -:. eds 2 iate DUNCAN . 1402 13 rede nie Pa HRH daven ries : : senaste Central and Eastern United States, by A. V. Heyl and KE - Van :..: :s ererrr! Feige ien ede er Nbr birr snene en nn roues HHNOIS-ICNEUCKY ::.: 114 ve cevir rif rede Illinois-Kentucky fluorspar district, by BR:: D. EIACE esr renet SELLING ns FJuQrSPAT QTE SUMMATY Central Kentucky......;.............. Cumberland River, Kentucky . Tennessee-North Carolina- Virginia.... Central: rete cans nith SWEELWALETY, ALENNESSEE . in cs devils Del Rio, Tennessee-Hot Springs, North Carolina ...... Hamme, North Carolina-Virginia........................ Faber, VITGIMIA .::: tree». Other deposits, seni Missouri-Arkansas-central Texas Southeast MISSOUFT + Other deposits, Missouri and Arkansas... Llano: uplift, Central NEW NOTES:: yi ern hns eerie Adirondack Mountains .. Other- CEDOSIES ?...... -/-. .at: ..... - NEW -EN@IANG:..:: ez i:: Hei q iyi naal in vive an seve Cheshire County-Westmoreland, New Hampshire ..... North Chatham, New Hampshire.................. Long Hill (Trumbull)-Monroe, Connecticut. € Long Island, Blue Hill Bay, Maine............................ Other districts and deposits of the Central and Fluorine in phosphate deposits, by J. B. Cathcart................... v Page 48 53 53 53 56 58 59 59 59 59 60 61 61 61 62 62 62 62 63 VI CONTENTS Page Present and future resources of fluorine, by D. R. Shawe, Present and future resources of fluorine-Continued R. G. Worl, R. E. Van Alstine, and A. V. Heyl.................... 83 Suggestions for likely future fluorine provinces: Summary of fluorine resources, by R. E. Van Alstine Speculative fluorine resources, by D. R. Shawe, and R. G.: WOT IRI: rier caters insite bevor rs 83 B: Worl, and A. V - iivenss Known favorable areas for fluorine exploration, by Geochemical and geophysical prospecting methods for fluorine, D. R. Shawe, R. E. Van Alstine, R. G. Worl, $ by W. R. Griffitts and R. E. Van and A. V. anv i 85 |. .. irie rnsn in rive rabi ness cso oe nen Ficures - 1-5. 8-17. 18. 19-21. 22. 23. 24. 25. 26. 27. 28. ILLUSTRATIONS Graphs: 1. Fluorine and silica contents of olivine basalt frOM Kil@Ue@, HAWAI ...... rere rrr enne rere rere rere ees 2. Fluorine and silica CONtENtS Of five I@gNCOUS SUItES reer $. Fluorine and silica contents of nine igneous suites from the United States...... 4. Fluorine and silica contents of basalts and andesites from Western United States and Antarctica 5. Fluorine and silica contents of hybrid granitoid rocks of Jurassic age and aplite of Oligocene age from the SOUtherN SN@ke NEV@GA rr reer renee rr rrr enn n rrr neer errr nnn . Map showing distribution of fluoride content of ground Water in the UNIited enne enne nene rre es . Scatter diagram showing relation Of flUOTiNE t0 SiliC@ iN VOIC@NIC TOCKS neer nene rer renee nere errr renee enne Maps: f 8. Distribution of fluorine-rich igneous rocks in the United States. 9. Distribution of fluorite deposits iM the UNited renee errr rrr nere rrr rrr neve rene rere nen renee ene 10. Locations of major topaz deposits, topaz occurrences, and topaz-bearing rhyolites in the United States................ 11. Locations of phosphate and fluorapatite deposits in the UNited SH@tES ...... seee reer nner ere ees 12. Major faults, distribution of fluorine minerals, and major belts of fluorite deposits in the United States... 13. Normal faults, and distribution of fluorine mineral deposits in the Western United States ..................... 14. Major Phanerozoic intrusive bodies and distribution of fluorine minerals in the United States .. 15. Bouguer gravity map and distribution of fluorine mineral deposits in the United recracesn finn 16. Fluorite mineral belts and high heat-flow regions in the Western United States and positive aeromagnetic anOMALES 111 iv..) {rare panne tha 6+ sere ceevet rene arbre bee «a bet nib eee 17. Crustal thickness, crustal seismic velocity, upper-mantle seismic velocity, and distribution of fluorine mineral depOSits in the UNit@d SUES renee enne enne rere rrr nene ne nene rre er nene ree nnn nn nnn ner Longitudinal projection along "main vein" of the Baxter (Kaiser) fluorspar mine, Broken Hills district , Nevada............ Maps: 19. Locations of fluorspar deposits in the Quinn CaNYON RANGE, errr renner enne 20. JAMEStOWN GiStTiCt, COIOFAGO esen neer rn ener rere nn nner ner er nn nner renner 21. Fluorspar deposit at the Burlington mine, Jamestown district, Colorado.. Sketch showing slab of fIUOTSPar frOM NOTthGAtE, COIO rere rere rr nner en rr rere nene renner rere renner nere rene en nen Map showing major structural features, igneous rocks, and distribution of known fluorspar deposits, Illinois- KENtUCKY fIUOTSDAT cc rere revere rere nene eer ern rrr nen renner rrr nene nere reer ern nner neenee neer renee nene neenee bebe nene neenee en nnn nnn Stratigraphic column, IIlinois-KeAtUucky fIUOTSDAT rere rere rr errr nene ene rrr neenee nene eevee nene nene nene een Generalized stratigraphic column, western Kentucky fluorspar district....... Map showing major structural features in the region surrounding the Illinois-Kentucky fluorspar district .. Cross section of Davenport mine, MOOre Hill @reA, enne neenee} Map of Cave in Rock fluorspar district, Hardin County, IIL................ Page . 86 +89 .\ 91 Page co ~1 & & 11 13 16 18 20 21 23 24 26 28 29 GEOLOGY AND RESOURCES OF FLUORINE IN THE UNITED STATES Edited by D. R. Snawr ABSTRACT Fluorine mineral resources are evaluated in this report in the per- spective of the geochemistry of fluorine and the geologic environment of fluorine mineral deposits. Rapidly increasing demand and the possi- bility of shortages in supply of fluorspar and other fluorine products that are vital to the economy of the United States prompted this review of fluorine resources. Continental crustal abundance of fluorine averages about 650 ppm (parts per million); basalts and gabbros average about 400-500 ppm, andesites and granodiorites 600 ppm, and rhyolites and granites 800 ppm; alkalic extrusive and intrusive rocks average much higher, 900-2,600 ppm. Despite the higher average fluorine content of silicic rocks, many silicic rocks contain as little fluorine as do low-fluorine mafic rocks. Intrusive rocks tend to contain higher amounts of fluorine than do extrusive rocks of similar composition. Metamorphic and sedi- mentary rocks have average fluorine contents similar to those of igneous rocks. Sandstones, carbonate rocks, and evaporites average about 200-250 ppm F, and oceanic sediments and shales 700-900 ppm. Marine phos- phorites contain remarkably more fluorine, averaging more than 30,000 ppm. Soils average less than $00 ppm F. In some comagmatic series of igneous rocks, fluorine increases with silica, and has been concentrated by magmatic differentiation. In other series fluorine decreases with silica, and the fluorine distribution in the rocks appears to have resulted from wallrock contamination. A further control of the abundance of fluorine in igneous rocks is the presence or absence of minerals that can readily accommodate fluorine in their crystal structures, so that hornblende- and biotite-rich rocks may contain more fluorine than do more silicic members of the same igneous series. Fluorine content is lowest in igneous rocks that have been irrupted through oceanic crust and highest in igneous rocks irrupted through continental crust. Fluorine appears to be highest in rocks of alkalic affinity and these appear to have originated mainly as magmas derived from the mantle beneath the continental crust. Some volcanic gases and sublimates contain notable amounts of fluorine; in general high-fluorine silicic and alkalic magmas exhale higher amounts of fluorine than do low-fluorine mafic magmas. Expulsion of fluorine from high-fluorine magmas may have occurred principally because of pressure drop due to magma rise. Surface waters average 0.25 ppm F, and higher contents are found mostly in arid regions. Ground waters contain an average of 0.4 ppm F. Pacific and Atlantic sea waters contain about 1.2 ppm F. Connate and related waters average 2.7 ppm F, and thermal waters associated with volcanoes and epithermal mineral deposits average 5.4 ppm F. Most of the fluorine transported to the sea is carried in minerals held in suspension and deposited in detrital sedimentary rocks. Only rarely do conditions develop where fluorine is deposited chemically in marine phosphorite deposits. Fluorine shows a strong geochemical affinity for the assemblage of elements _ beryllium-lithium-manganese-niobium-tin-tungsten-urani- um-yttrium-lead-zinc and is commonly found with these elements in a wide variety of geologic environments. High-fluorine igneous rocks are geographically restricted to two major provinces in the United States, one a broad belt extending from southern Maine southwestward into Alabama and the other a U-shaped region extending from southwestern Montana southward into southern California, eastward to southwestern New Mexico, and northward through western Colorado. Fluorine-rich igneous rocks in the Eastern United States are dominantly Paleozoic in age and those in the Western United States are dominantly Tertiary in age. Fluorine mineral deposits occur in a wide variety of geologic environ- ments and in a multitude of forms. Veins, mantos, pipelike bodies, breccia stockworks, contact zones, pegmatites, greisens, disseminated deposits, and bedded layers occur variously associated with or enclosed within volcanic rocks, intrusive igneous rocks, metamorphic rocks, or sedimentary rocks. Fluorite is the most common and abundant fluorine mineral, and fluorspar constitutes the ore of fluorite. Other significant fluorine minerals that likely constitute future resources are topaz, fluora- patite, and the carbonate-apatite of sedimentary phosphate rock. Additional minerals that may become economically significant sources of fluorine are cryolite, sellaite, villiaumite, bastnaesite, and the humite group of silicates. Fluorine occurs in epigenetic mineral deposits principally in broad provinces in the Eastern and Western United States that coincide closely with the broad provinces of fluorine-rich igneous rocks. On the basis of available age data the deposits also are similar in age to associated igneous rocks. Fluorine is found in marine phosphate rocks chiefly in the Southeastern and the Western United States. Distribution of epigenetic fluorine mineral deposits in the United States has been strongly controlled by tectonic environment. There is a close spatial association of fluorine mineral deposits with major fault zones. In the Western United States the distribution of Tertiary fluorine mineral deposits closely follows the distribution of Basin-Range structure; some deposits occur in tensional faults. The distribution of fluorine mineral deposits in northwesterly, northeasterly, and easterly belts suggests an association with deep-seated strike-slip fault zones of these orientations that seem to constitute a fundamental structural framework of the continental crust. Regions of low gravity in the United States that broadly represent crustal zones intruded by significant volumes of silicic igneous rocks correlate in a rough way with regions of fluorine mineral deposits. Regions of high heat flow in the Western United States that reflect the middle Tertiary to present episode of magmatic and hydrothermal activity of this part of the country coincide remarkably with the regions of rather young fluorine mineral deposits. Available regional aecromag- netic data reveal that zones of numerous sharp positive anomalies (that represent in part intrusive igneous centers) coincide with the positions of fluorine mineral belts. A broad region of low seismic velocity in the upper mantle in the Western United States closely encompasses the region of middle to late Tertiary magmatism and fluorine mineralization. Inferences on the composition of the low velocity upper-mantle material suggest that 1 5 GEOLOGY AND RESOURCES OF FLUORINE volatile components composing fluorine-enriched magmas were fractionated out of the upper mantle and injected into the crust which accounts for the higher heat flow, the fluorine-rich igneous rocks, and the fluorine mineral deposits of the province. Plate tectonics is invoked to account for extension of the continental crustal plate beginning in middle Tertiary time; this extension was accompanied by conjugate strike-slip faulting in the lower part of the crust and tensional faulting in the upper part of the crust. Crustal extension resulted in a phase trans- formation in the upper mantle, which accounts for the development of fluorine-rich magmas that were injected upward into the crust and were accompanied by fluorine mineralization. Fluorine mineral deposits that have had significant production or constitute a significant type are described in the text on the basis of geographic distribution and geologic environment. In the Lost River tin district, Seward Peninsula, Alaska, large carbonate-replacement and pipelike fluorspar deposits are associated with granitic intrusives. Else- where on the Seward Peninsula, fluorite occurs in other tin districts, in mineralized carbonate rocks north of Nome, in breccia pipes in the Kigluaik Mountains, and in carbonate rocks surrounding alkalic intrusives in the Darby Mountains. In other parts of Alaska, fluorite and topaz are associated with peralkalic granites and occur in some sulfide deposits. Large fluorine resources are known but currently inaccessible in marine phosphate rock in northern Alaska. In the northwestern part of the United States, important fluorspar- producing areas are the Meyers Cove and Challis deposits, Idaho, where fluorite veins occur in volcanics and porphyries and in dolomite, respectively; and the Snowbird, Spar, and Crystal Mountain deposits, Montana, which are pegmatitelike bodies. Numerous other fluorspar deposits associated with volcanic rocks, granitic intrusive rocks, or alkalic rocks, or occurring as pegmatites or disseminated in granite are widespread in north-central Washington, central and southwestern Idaho, western to central Montana, several areas in Wyoming, and the Black Hills of South Dakota. In the southwestern part of the United States manto replacement and pipelike bodies of fluorspar occur in limestone in the Fluorine district, Nevada; vein deposits are found in andesite and rhyolite at Broken Hills, Nev.; pipe deposits are known in carbonate rocks at Spor Mountain, Utah; and veins occur in volcanic rocks in the Indian Peak Range, Utah. Throughout much of Nevada, southeastern California, western Utah, and western Arizona numerous occurrences of fluorspar are known in veins, mantos, or pipelike bodies in and associated with volcanic rocks, in contact zones or stockworks of hypabyssal or plutonic igneous rocks, in volcaniclastic sediments, and in pegmatites. The Quinn Canyon Range district, the Wells Cargo deposit, and the Iowa Canyon deposits, Nevada, and the Castle Dome district, Arizona, offer promise of substantial fluorspar production. The southern Rocky Mountain region of the Western United States in Colorado, New Mexico, western Texas, and southeastern Arizona con- tains numerous fluorspar deposits. Many deposits are characterized as layered or crustified veins in young steep tensional faults or breccia zones occurring in Tertiary volcanic rocks, Paleozoic or Mesozoic sedimentary rocks, or Precambrian silicic igneous or metamorphic rocks. Such veins are in the Northgate and Browns Canyon districts, and near Crystal, Dillon, Poncha Springs, and Wagon Wheel Gap, Colo., and the Gila, Burro Mountains, Anderson, Gold Hills, Steeple Rock, Sierra Caballos, Zuni Mountains, Fluorite Ridge, and Cooks Peak districts, and deposits at Tonuco, Tortugas Mountain, Bishop Cap, and Sierra Cuchillo, N. Mex. Manto deposits occur in carbonate rocks in the Sierra Caballos and Sierra Cuchillo, and at Tortugas and Bishop Cap deposits, New Mexico, and in the Eagle Mountains and Christmas Mountains-Corazones Peak districts in Texas. At the Jamestown district, Colorado, significant stock- works, pipelike shoots, and mineralized breccia zones, containing fluo- rite and lead, silver, gold, and uranium minerals, occur in Precambrian granite near a middle Tertiary alkalic-silicic stock. In the Illinois-Kentucky district, large and numerous veins and re- placement bodies of fluorspar, barite, and sphalerite occur in faulted carbonate rocks in the vicinity of small mafic-alkalic dikes. These deposits have yielded more than 75 percent of the fluorspar mined in the United States. Large low-grade tonnages of fluorspar-mineralized breccia are known in the vicinity of Hicks dome, Illinois, but have not yet been developed. In the Central and Eastern United States outside of the Illinois- Kentucky district, numerous fluorine mineral deposits are known and some offer promise of substantial fluorine resources. High-temperature veins, lodes, greisens, tactites, and skarns commonly containing fluorite and topaz and associated minerals of tungsten, tin, silver, gold, lead, zinc, molybdenum,iron, rare earths, and thorium, occur in the Appalachian structural belt, the Ozark dome, Missouri, and the Llano uplift, Texas. Intermediate-temperature deposits are chiefly veins such as fluorite- quartz veins in the Cheshire County-Westmoreland district, New Hamp- shire, and complex fluorite-bearing veins near Faber, Va.; in the Connecticut Valley, Mass.; near Phoenixville, Pa.; on Deer Isle, Maine; and at Litchfield, Conn. Most of the fluorspar deposits in the Central and Eastern United States are low-temperature veins, mantos, breccia bodies, stockworks, pipelike bodies, and disseminated deposits. The largest and most numerous deposits outside of the Illinois-Kentucky field, are in the Mississippi Valley and in the Appalachian Valley and Ridge province in the vicinity of related zinc deposits. Other Mississippi Valley-type districts are the Central Tennessee district, the Cumberland River district, the Central Kentucky district, and the Rossie area, New York. Deposits of the Valley and Ridge type are numerous along the Appalachian belt from Alabama into Pennsylvania in Cambrian-Ordovician carbonate rocks. The largest deposit is in the Sweetwater district in eastern Tennessee. Others are the Del Rio, Tenn., and Hot Springs, N. C., districts and the Gilley deposit in northeastern Alabama. Disseminated deposits are found in central Pennsylvania, northeastern New York, northeastern Iowa, north- western Ohio, and southeastern Ohio. Residual deposits of fluorspar are known in western Kentucky, central Kentucky, and central Tennessee, and placer deposits of topaz are known in South Carolina and in the Llano uplift of central Texas. Economic fluorine-bearing phosphate deposits in the United States are confined to marine sedimentary phosphorites of Ordovician age in Tennessee, of Permian age in Idaho, Montana, Wyoming, and Utah, of Miocene age in North Carolina, Georgia, and South Carolina, and of Miocene and Pliocene ages in Florida. Fluorine content of the rock is about 3-4 percent, and billions of tons of phosphate resources are known. Present and future resources of fluorine in the United States are evaluated in the light of past production. In the past most fluorine production has come from fluorspar, and fluorspar will continue to be the major source in the near future. Phosphate rock and other byproduct and coproduct sources will become increasingly important as fluorspar resources diminish. From 1900 to about 1950 the United States changed from a minor producer of fluorspar to the major world producer, but from 1950 to 1970 has reverted to the seventh or eighth leading producer. Production generally has risen in recent years, however, until it now averages about 200,000 short tons CaF, annually. Until about 1950 United States production of fluorspar nearly equaled its consumption, but since 1950 consumption has increased dramatically to nearly 1,500,000 short tons of CaF, annually. In addition, since 1950 worldwide fluorspar consumption has increased spectacularly to more than 4,500,000 short tons of CaF, annually, emphasizing the fact that the United States is faced with strong competition for foreign supplies of fluorspar. United States reserves of fluorspar are about 25,000,000 short tons of crude ore ( > $5 percent CaF). Hypothetical resources of fluorspar in the United States are about 45,000,000 short tons. The marine phosphate rock deposits of Florida, North Carolina, Tennessee, Utah, Wyoming, Idaho, and Montana contain about one- third billion tons of fluorine in known reserves and about 2 billion tons of fluorine in identified resources. Marine phosphate rock thus consti- tutes by far the United States' and the world's largest fluorine resource. INTRODUCTION 3 Topaz reserves at the Brewer mine, South Carolina, are about 100,000 tons of schist averaging about 15 percent topaz; hypothetical resources are estimated at about 800,000 tons of rock. In the Front Range, Colo., a lens in Precambrian gneiss contains about 15 percent topaz and a reserve of about 600,000 tons of topaz-bearing rock. Byproduct resources of topaz and fluorite in molybdenum porphyry deposits at Climax and Hender- son, Colo., and Questa, N. Mex., may constitute several million tons of fluorine. Bastnaesite in carbonatite at Mountain Pass, Calif., contains an esti- mated 1,000,000 tons of fluorine and is a potential byproduct of rare-earth - extraction. Fluorine resources in the United States appear to be adequate for pro- jected needs for many years into the future. Adequate sources of fluor- spar are available for the present and near future; vast amounts of fluorine in marine sedimentary phosphate deposits are available as a more distant future resource, dependent on technological improvements in recovery. Most fluorine mineral districts and localities in the United States remain as areas favorable for further exploration for fluorine resources. In productive districts extensions of known fluorspar bodies should be sought along strike and in depth, and additional ore bodies may be found along nearby faults. Carbonate rocks susceptible to replacement by fluorite should be explored, as deposits in such rocks tend to be large and amenable to large-scale low-cost mining. An increasingly important source of fluorspar will be as a coproduct from the mining of iron, base- metal, barite, and rare-earth ores. More analyses for fluorine in earth materials should be made routinely to provide background and geochemical information as well as possibly to disclose new environments of fluorine mineral deposits. Exploration for and evaluation of speculative fluorine resources-those occurring in districts and environments yet to be identified-must be based on knowledge of the geologic environment and the geochemical cycle of fluorine. Geologic mapping in areas not now adequately covered by geologic maps will be essential to evaluation of geologic environment and to successful exploration in such areas. Major geologic parameters that partly define the geologic environment of fluorine mineral deposits in western North America are: (1) extensional tectonism related to deep-seated, mainly northwesterly, northeasterly, and easterly strike-slip structures; (2) high heat flow from the mantle; and (3) magmatism of alkalic affinity characterized by fluorine abundance. These parameters define environments and processes in which fluorine was intrinsically available in abnormal amounts in some part of the system, and could be mobilized into economic concentrations. On the other hand, geologic processes in other environments have concentrated minor (crustal abundance) amounts of fluorine into deposits of economic interest, such as marine phosphate rocks. Earlier concentrations of fluorine may have been dispersed, or re- mobilized and reconcentrated in new environments by later processes. In the Central and Eastern United States new districts of fluorspar in carbonate rocks may be found near Toledo, Ohio, near Serpent Mound, Ohio, in southern Indiana, in and near the Ste. Genevieve fault zone in southeastern Missouri and southern Illinois, along the Cumberland River, Ky., in central Tennessee and central Kentucky, and near Magnet Cove, Ark. In the Western United States fluorspar deposits undoubtedly underlie parts of the widespread cover of Tertiary and Quaternary sediments and volcanics. Fluorite-bearing Cenozoic tuffaceous lake beds and other similar deposits in the Western United States may offer potential for large fluorine resources. Fluorspar deposits related to silicic-alkalic volcanic rocks in the Western United States may be found in the igneous areas of the Big Bend region of Texas, Nacimiento region of Arizona, New Mexico, and Colorado, and the Shoshone province of Nevada, Utah, Idaho, Wyoming, and Montana, particularly southeastern Idaho, north- western Wyoming, and northwestern Utah. Caldera ring-fracture zones in addition to major deep-seated strike-slip fracture zones and major rift structures may be favorable for undiscovered fluorspar districts in these regions. Deposits of syngenetic fluorite associated with marine gypsum and limestone of Permian age, like the deposits in north-central Wyoming, may be found elsewhere in the United States. Large deposits of carbonate-fluorapatite in marine phosphate rock probably exist beneath the sea in Cenozoic coastal plain sediments extending east from the deposits in the Carolinas and Florida. Fluorine may be recoverable from deposits of fluorapatite in alkalic rocks and from deposits of fluorapatite and magnetite forming contact- metamorphic zones around intrusives. Probably millions of tons of topaz-bearing rock are yet to be discovered in widespread metamorphic terranes throughout the United States, as well as in presently unknown molybdenum porphyry deposits. Inasmuch as cryolite for many years constituted a significant fluorine resource at Ivigtut, Greenland, it conceivably may occur in similar deposits elsewhere. Sodium fluoride in saline lake beds may some day be a source of fluorine. Many other very low-grade geologic sources of fluorine offer potential only for the distant future. Some very large low-grade multicommodity ores, in which fluorine minerals are associated with metallic minerals, barite, rare-earth minerals, apatite, zeolites, feldspars and other minerals of commercial value, may become resources of fluorine. Geochemical and geophysical prospecting methods may be valuable additional aids to apply to exploration for fluorine mineral resources. Geochemical techniques as well as panning are useful for detecting detrital fluorite that has been dispersed, in some cases as much as 35 miles, from its source. Some pathfinder elements such as zinc, silver, antimony, cadmium, gallium, germanium, vanadium, and molybde- num may be useful in geochemical prospecting for fluorine mineral deposits. Uranium identified by radiation detectors might also be a useful guide to fluorspar-uranium deposits. Fluorine anomalies may be guides to deposits of associated valuable minerals. An induced activity method of measuring the short-lived isotope F* after activation by a polonium-beryllium source offers promise of accurate in-hole measurement of fluorine content of rocks. Earth-resistivity, refraction seismic, and self-potential geophysical methods have been used to detect structures that may contain fluorine mineral deposits. Induced potential methods may successfully outline alteration zones that are favorable for fluorine mineral deposits. INTRODUCTION By R. E. Van Austing® Fluorspar, the primary source of fluorine, is a vital and integral part of our economy, for it has great commercial and strategic importance because of its wide use in the metallurgical, chemical, and ceramic industries. Reviews (Industrial Minerals, June and July 1970) of the world's supplies of fluorspar predict shortages in the near future. The increased demand is chiefly the result of the change- over by the steel industry to the basic oxygen process requiring about three times as much fluorspar as conven- tional processes; continued expansion of the aluminum industry; and the growing market for fluorine chemicals, especially fluorocarbons (MacMillan, 1970). The expand- ing needs of these industries stimulated review of our fluorine resources. This report is part of an effort by the U.S. Geological Survey to present comprehensive and updated informa- tion on the significant mineral resources of the United States, with appropriate reference to world resources. The expanding industrial needs for fluorine stimulated the present review of our resources. Particular attention is 4 GEOLOGY AND RESOURCES OF FLUORINE given to the geochemistry of fluorine, for it defines and limits fluorine's geologic environment, and thus the origin, habit, and distribution of fluorine mineral deposits. Representative examples of the major types of deposits (fluorspar, calcium fluophosphate, topaz) and of other fluorine-bearing minerals are described in detail, by geographic area; many other deposits are described briefly, and references to sources of detailed information are given. In addition to this report, fluorine resources are also reported on briefly by Worl, Van Alstine, and Shawe (1973), and a map by Worl, Van Alstine, and Heyl (1974) shows distribution of fluorite in the United States. The map was intended as a supplement to the present report, and reference to it will be made in the following pages. Continuing supplies of domestic and imported fluor- spar are required to produce the variety of metals and materials essential to our industries (Bradbury and others, 1968, p. 35-60; Williamson, 1961). The United States con- sumption of fluorspar has doubled in the past decade and is expected to grow about 5 percent per year. About 80 percent of the fluorspar that we consume is imported, chiefly from Mexico, Spain, and Italy; fluorspar imports exceeded 1 million short tons for the first time in 1968. About 80 percent of our domestic fluorspar, totaling about 237,000 short tons in 1972, comes from the Illinois- Kentucky district, which has been the world's most productive district. Small quantities of fluorine com- pounds are produced as byproducts from the phosphate industry, and some fluorine is recovered by the aluminum industry from the waste streams of the Hall cell and by the chemical industry in plants producing hydrofluoric acid. These secondary sources of fluorine should provide for an increasing part of the fluorine supply as we become more concerned with reducing environmental pollution. Although fluorspar is still the primary source of fluorine, much more fluorine is lost in phosphate rock that has been mined than is recovered from fluorspar. The National Materials Advisory Board (NMAB-269, 1970) estimated that annual consumption of fluorspar in the United States would increase by about one-fourth between 1970 and 1975, but by only one-fifth between 1975 and 1980. As shown in the accompanying table, production of steel and hydrogen fluoride accounts for about 93 percent of the total domestic fluorspar consumption. This per- centage is not expected to change greatly in the near Consumption of fluorspar in the United States, in short tons Use 1970 1975 1980 (estimated) (estimated) SteE-pFOUUCHION :: 2.112; innit 548,700 723,000 895,000 Hydrogen fluoride production ...... 726,500 920,000 1,160,000 Glass and enamels 35,000 35,000 35,000 Foundry ...a}. esti 21,400 24,100 27,400 ATEOLRCT .::.. 1s» as ska 35,000 45,000 56,000 Total. levis aire deve 1,366,600 - 1,747,100 2,173,400 future, even though the requirements for the various uses are growing. In 1952 the President's Materials Policy Commission (Paley, 1952) predicted that the United States fluorspar consumption by 1975 would be 1.15 million short tons, a quantity that we have exceeded since 1968; this prediction is an example of the tendency to under- estimate future resource needs. In the United States about 3 tons of crude fluorspar is mined and beneficiated for each ton of finished product marketed. Fluorspar generally is sold in three grades -acid, ceramic, and metallurgical. Acid-grade fluorspar should contain at least 97 percent CaF». Limitations are sometimes put on S10;, CaCO;, and sulfide sulfur. Chemical and physical requirements are given in National Stockpile Material Purchase Specifica- tions P-69a-R1, June 1, 1967, General Services Admin- istration. Ceramic-grade fluorspar varies substantially with individual buyers but is usually finely ground and has a CaF,-content ranging from 85 to 97 percent. The FesOys- content usually is limited. Metallurgical-grade fluorspar in the form of gravel, lump, artificial pellets, or fine flotation concentrates contains at least 60 percent effective CaF, units, deter- mined by subtracting from the contained CaF, 2.5 percent for every percent of SiO; present in the complete analysis. National Stockpile Purchase Specification P-69b-R1, December 2, 1963, contains - requirements - for metallurgical-grade fluorspar. Published prices (1973) per short ton were $78.50-$87.00 for acid-grade fluorspar, $77-$87 for ceramic-grade, and $68.50 for metallurgical-grade pellets (70 percent effective CaF,). The (1973) tariff on fluorspar containing more than 97 percent CaF, was $1.875 per short ton and was $7.50 per short ton on fluorspar containing less than 97 percent CaF,. CaF, percentages in this report are weight percentages. In 1971 about 52 percent of the fluorspar consumed in the United States was acid grade, 43 percent was metallur- gical grade, and 5 percent was ceramic grade (U.S. Bur. Mines, 1973). About 97 percent of the quantity consumed in 1968 was used in the following 15 States: California, Colorado, Delaware, Illinois, Indiana, Kentucky, Louis- iana, Maryland, Michigan, New Jersey, New York, Ohio, Pennsylvania, Texas, and West Virginia (U.S. Bur. Mines, 1969). GEOCHEMISTRY OF FLUORINE By D. R. Snaws Fluorine has the atomic number 9, atomic weight 19.00 (lightest element of Group VII of the periodic table), and a valence of minus one in all naturally occurring com- pounds. It has no isotopes. Having an ionic radius of 1.36A, the fluoride ion is "readily isomorphous with the hydroxyl ion, (OH)~-! GEOCHEMISTRY OF FLUORINE 5 (radius 1.40), a fact of great importance in its geo- chemical occurrence and behavior. To a lesser degree, fluoride ion may isomorphously replace chloride ion, CI- (radius 1.81), and oxygen ion, 0-2 (radius 1.40A)." (Fleischer and Robinson, 1963, p. 58). Fluorine commonly is considered to be concentrated as a result of magmatic differentiation into silicic igneous fractions, residual solutions, and vapors (for example, Fleischer and Robinson, 1963); however, data compiled here suggest ambiguity in this conclusion. Although certain rhyolitic and granitic rocks, as well as most alkalic rocks, contain unusually high amounts of fluorine, some consanguineous suites of igneous rocks show higher fluorine contents in less silicic members, and alkalic suites do not necessarily contain greatest fluorine concentra- tions in their most differentiated rocks. Further, some hybrid granitoid rocks whose chemical character was strongly controlled by assimilation of wallrocks show a clear decrease in fluorine content with increasing silica content. ABUNDANCE OF FLUORINE IN ROCKS Averages of the fluorine contents of various igneous rocks are as follows: Fluorine, in parts per million (number of samples in parentheses) Rock type Fleischer and Robinson This report! (1963, tables I, II) Extrusive: Basalt his rf ie : 380 (268) 510 (187) Andesite..... 220 (83) 630 (85) Rhyolite .... 700 (145) 780 (261) Phonolite .::.: ris: 950. (14) y= .* =~ ..... Intrusive: 420 (47) :: _ ..... Granite," granodiorite... S10 (185) ": - ...... Alkalic rocks 960 (71) 32,640 (100) 'Average fluorine contents of igneous rocks compiled for this report include published data and many unpublished data in the U.S. Geological Survey files that are not specifically credited here. *Fuge and Power (1969) reported a mean fluorine content of 1,395 ppm for 90 unaltered granites from southwest England. *Gerasimovskiy and Savinova (I969) confirm the higher values for alkalic rocks given here. Fleischer and Robinson (1963, p. 61), in noting that fluorine appeared to increase with increasing silica content of igneous rocks, commented on the unexpectedly low fluorine content of andesites and suggested a possible regional effect owing to the preponderance of Japanese andesites in their summary. A more detailed view of the data in the foregoing table, as presented in the next table, may clarify this and other discrepancies in the averages for certain rock types. In all instances the discrepancy between data compiled by Fleischer and Robinson and data compiled in this report can be attributed to the inclusion of rocks of alkalic affinities in the groups with higher fluorine averages. For example, rocks such as alkali basalts at Craters of the Moon, Idaho, and basalts termed absarokites from the Mogollon Rim, Ariz., raise the average fluorine content of Fluorine, in parts per million (number of samples in parentheses) Rock type Fleischer and Robinson This report! (1963, tables I, II) Basalt: Alcutfan sis 91410... 1.0 (+) seis. 290 (13) Hawail 340 (80) 320 (71) Snake River Plain.......:.l....... C @ 0 400 (26) Columbia River Plateau........... | (==> ...... 440 (20) PucItO e (PR A earth 460 (5 ) (OLHEYT :s: Aerie dsr asa soars 500 (58) 920 (52) Andesite: Japan (dominantly) ................. 210 (77) ; . \: as ees; Aleutian Islands ....... ety. - ane. 250 (12) Other. ::: . AIHATRR AEs ieee PEL T Bhare 2700 (73) Rhyolite: XIKali ) 00 Oo sll. $2,180 (25) OTREL .se: 3 Hsien ihren dia cin dole 700 (145) 630 (236) Alkalic rock (mostly intrusive): Syenite, nepheline syenite (dominantly 960 (71). s 9 can..... Rapakivi granite, mafic alkalic . ~ * ...s. 2,640 (100) 'Compilation includes most of the rocks analyzed by the U.S. Geological Survey that were averaged by Fleischer and Robinson, plus additional ones, and also some data ott ic oa Idaho, Wyoming, Nevada, Utah, Arizona, New Mexico. "other' basalts shown under "This report" in the preceding table. In contrast, basalts from Iceland have an extremely low fluorine content (180 ppm, average of 5; Barth, 1947). Although andesites from the Aleutian Islands and Japan have similar fluorine contents of about 200-250 ppm, the average content for the group of "other" andesites compiled for this report is appreciably higher because more alkalic types (latites and trachytes) that have the same silica range as andesite were included. In Fleischer and Robinson's compilation of data on alkalic rocks, syenites and nepheline syenites dominate, whereas in the compilation for this report the alkalic rocks are chiefly rapakivi-type granites and mafic alkalic rocks that appear to contain significantly more fluorine than do syenites. Virtually no duplication is involved in the two compilations in this case. Fluorine tends to increase with silica in glassy volcanic rocks (Shawe and Bernold, 1966, pl. 5-C). Those data are summarized in the following table: Fluorine content of glassy volcanic rocks Fluorine, in parts per million Silica rahes (number of samples in parentheses) 70-80 DETCENL ic errors rans 1,040 (88) 10-75. 670 (49) 65-70.. 580 (13) AB :s» is irr nen h ssl t 340 (7) Some workers (for example, Coats and others, 1963, p. 963) have considered that volcanic glasses, having been cooled quickly, closely reflect original magma composition. If so, the previous table strongly indicates that fluorine is concentrated in later fractions by mag- matic differentiation. Virtual proof of the concentration of fluorine in igneous rocks by magmatic differentiation is afforded by the data 6 GEOLOGY AND RESOURCES OF FLUORINE of Wright (1971) and Wright and Fiske (1971). Twenty- nine olivine basalts from the summit of Kilauea, Hawaii, that are considered to be undifferentiated, range from 47.9 to 51.6 percent silica and contain an average of 410 ppm F (Wright, 1971, table 4). Nine other basalts from Kilauea classified as having formed by differentiation from magma chemically similar to basalts of the summit of Kilauea, on the basis of their general chemistry and geologic occurrence, range from 51.5 to 55.4 percent silica and contain an average of 660 ppm F (Wright and Fiske, 1971, table 13). Also, nine samples of fracture fillings in natural flows, or material that flowed into open drill holes in flow crusts ("oozes"")-that range from 49.6 to 56.2 percent silica contain an average of 820 ppm F (Wright and Fiske, 1971, app. 3, table 1). These data are presented graphically in figure 1. Certain other suites of igneous rocks from various parts of the world also exhibit a tendency for fluorine to increase along with silica, suggesting concentration by magmatic differentiation (fig. 2). Of the examples shown, however, rocks of the Hallett volcanic province, Antarc- tica, were considered by Hamilton (1972, p. 59) to represent two divergent petrologic series, and rocks from the Egan Range, Nev., constitute two groups (one high in fluorine, one low), believed by Shawe (1961, p. B181) not to be closely related genetically. 60 - ¥ c f #37" 355— /<:/’ x fir yA 2 a. T> *, o > C a - y" 5 #4 x i 50 |- ca 45 | | 1 ] 0 500 1000 1500 2000 FLUORINE, IN PARTS PER MILLION EXPLANATION +- Olivine basalt, Kilauea Summit (Wright, 1971, table 4) + Differentiated lavas, Kilauea (Wright and Fiske, 1971, table 13) x Kilauea "differentiates" (Wright and Fiske, 1971, appendix 3, table 1) FicureE 1.-Positive correlation between florine and silica contents of olivine basalt from Kilauea Summit, differentiated lavas of Kilauea, and Kilauea "differentiates," Hawaii. 80 ,- 75 70 o & er o O o SiO,, IN PERCENT oa O 45 40 35 1 | 1 1 1 0 500 1000 1500 2000 2500 FLUORINE, IN PARTS PER MILLION 3000 EXPLANATION Egan Range, Nevada (Shawe, 1961, table 1) * Mountain City-Owyhee, Nevada (R. R. Coats, unpub. data) x Schell Creek Range, Nevada (Drewes, 1967, tables 10, 11, 12) © Hallett volcanic province, Antarctica (Hamilton, 1972, tables 2, 3) &a Black Hills, South Dakota (Norton, 1970, table 2) Ficure 2.-Positive correlation between fluorine and silica contents of five igneous suites. In comparisons of fluorine content with beryllium con- tent (and indirectly with silica content) of volcanic rocks, Griffitts and Powers (1963, p. B18-B19) and Shawe and Bernold (1966, p. C8) suggested that positive correlations were poor within a group of genetically related rocks, but relatively good overall among samples from many populations (that is, genetically unrelated rocks). Other data, presented by Shawe and Bernold (1966, pl. 5-A), surprisingly show a negative correlation between GEOCHEMISTRY OF FLUORINE 7 fluorine and silica contents of volcanic tuffs, as summa- rized in the following table: Fluorine content of volcanic tuffs Fluorine, average in parts per million (number of samples in parentheses) 420 (20) 700 (24) 700 (22) 900 (5) 60-65. d r t.. e] Most of the tuffs summarized in the above table are from the Nevada Test Site, southern Nevada, and the table thus generally represents a single igneous petrographic province. Many of the tuffs are zeolitized, but it is not known if zeolitization had any effect on fluorine content of the rocks. Other examples of inverse silica-fluorine relations among rocks in consanguineous series are illustrated in figure 3. Not every example here necessarily represents a comagmatic series, but each example does represent a suite of igneous rocks generated in a unique segment of the crust or upper mantle, a fact reflected by the coherency of the field for each example shown. A special case is presented by the field for rocks from the Valley of Ten Thousand Smokes, Alaska (Lovering, 1957, table 1), in fugure 3; most samples are hydrothermally altered ash and thus the negative fluorine-silica correlation here is not directly attributable to magmatic processes. An inverse silica-fluorine relation is shown in yet another assemblage of volcanic rocks. Basalts and ande- sites-excluding basalts from the Aleutian Islands, Hawaii, the Snake River Plain, Idaho, the Columbia River Plateau, Wash., and Puerto Rico, which plot in restricted fields lying mostly between 45-55 percent silica and 100-600 ppm F-are shown in a much broader field in figure 4. The strong tendency for higher fluorine to cor- relate with lower silica is summarized as follows: Fluorine content of basalts from the Western United States and Antarctica Fluorine, average in parts per million silica sange (number of samples in parentheses) 60-65 .DENCENL - 1:2 100 (1) $5-00: :...... 510 (13) 50-55..... 860 (8) 45-50..... 1,110 (17) 40-45.. 1,040 (13) 2.1 s nicl 1,900 (1) Lee and Van Loenen (1971) have shown convincingly that the composition of hybrid Jurassic granitoid rocks in the southern Snake Range, Nev., was closely controlled by the composition of assimilated wallrocks. This suite of related igneous rocks whose compositions vary serially and hence suggest origin through magmatic differentia- tion are thus shown to have developed probably simul- 80- 75 70 65 & O SiO,, IN PERCENT 45 40 35 x \\ 30 mix K A X agl \\\\ \o 20 | | | | | | | 0 500 1000 1500 2000 2500 3000 3500 FLUORINE, IN PARTS PER MILLION EXPLANATION *- Railroad district, Nevada (K. B. Ketner and J. F. Smith, Jr., unpub. data) 0 Valley of Ten Thousand Smokes, Alaska (hydrothermally altered ash; Lovering, 1957, table 1) x Creede-Summitville, Colorado (Steven and Ratté, 1960 tables 2, 3, 7; Ratte and Steven, 1967, tables 2, 4, 6, 14, 18, 22) © Funeral Peak, California (Drewes, 1963, table 4) a Bullfrog, Nevada (Cornwall and Kleinnhampl, 1964, table 2) o Craters of the Moon, Idaho (H. A. Powers, J. G. Murtaugh, unpub. data) + Mogollon Rim, Arizona (T. L. Finnell, unpub. data) ® San Gabriel Mountains, California (G. J. Neuerburg, unpub. data) v Mount of the Holy Cross Colorado (Ogden Tweto, unpub. data) Ficure 3.-Negative correlation between fluorine and silica contents of nine igneous suites from the United States. 8 GEOLOGY AND RESOURCES OF FLUORINE 65)- 60 o o SiO,, IN PERCENT (eal («] B o 40 35 l | I 0 500 1000 1500 FLUORINE, IN PARTS PER MILLION I 2000 EXPLANATION A Basalt, Bullfrog, Nevada (Cornwall and Kleinnhampl, 1964, table 2) o Basalt, Cortez, Nevada (Gilluly and Masursky, 1965, p. 73: table 9) © Andesite, Klondyke, Arizona (Simons, 1964, p. 71, 72, 90) x Basalt, southeastern New Mexico (C. L. Jones, unpub. data) ® Basaltic andesite, Steamboat Springs, Nevada (White and others, 1964, table 1, analysis 14) +- Basalt and andesite, Blue Range, Arizona and New Mexico (Ratté and others, 1969, table 6) 0 Basaltic andesite, Twin Buttes, Arizona (J. R. Cooper, unpub. data) v Basalt, Horse Creek Valley, Nevada (Harold Masursky. unpub. data) & Olivine basalt and andesite, Funeral Peak, California (Drewes, 1963, table 4) -+ Basalt, Lassen County, California R. L. Smith, unpub. data) + Basalt, South Victoria Land, Antarctica (W. B. Hamilton, unpub. data) © Olivine basalt and trachybasalt, Hallett volcanic province, Antarctica (Hamilton, 1972, tables 2, 3) FicURE 4.-Negative correlation between fluorine and silica contents of basalts and andesites from the Western United States and Antarctica. taneously through assimilation of chemically contrasting wallrocks. In the granitoid rocks, fluorine varies inversely with silica (fig. 5). Lee and Van Loenen (1971, table 9) showed that the quartzite wallrocks adjacent to the high- silica granitoid rocks contain an average 200 ppm F whereas the shale and limestone wallrocks adjacent to the low-silica granitoid rocks contain an average 1,300 and 600 ppm F respectively. According to Lee and Van Loenen (1970, p. D199), most of the fluorine in the granitoid rocks is in biotite. Biotite content varies from less than 5 weight percent in rocks of about 75 percent silica to more than 25 percent in rocks of about 65 percent silica. Interestingly, fluorine content of the biotite is highest (about 10,000 ppm) in rocks of about 75 percent silica and lowest (2,500-5,000 ppm) in rocks of about 65 percent silica. These relations suggest that the biotite took up what fluorine was available in the magma as it crystallized. Moreover, it appears that biotite in the more mafic rocks could have taken up more fluorine had it been available. Total fluorine content of these rocks thus depended on total fluorine available in magma and assimilated wall- rock, and amount of biotite formed from the magma. Despite the likelihood of some initial fluorine in the magma, the data indicate that fluorine distribution in comagmatic igneous rocks of different composition need not be wholly dependent upon magmatic differentiation. Aplites in the southern Snake Range that are younger than the granitoid rocks just discussed, and that were intruded probably without appreciable assimilation, show a different fluorine-silica relation and therefore a different origin from that of the granitoid rocks (fig. 5). Fleischer and Robinson (1963, p. 63) indicated an average of 380 ppm F for 69 metamorphic rocks of a variety of types. Sedimentary rocks, except for shales, oceanic sediments, and phosphorites, contain generally lower amounts of fluorine than do igneous rocks, as shown in the following table based mostly on Fleischer and Robinson (1963, table IV). Fluorine, average in parts per million Rock type (number of samples in parentheses) sis 5 220 (98) DOIOMItE rense 260 (14) Sandstone and graywacke............... 200 (50) SHAC: sais: rennes deer 940 (82) Oceanic sediments........................... 730 (79) Volcanic ash and bentonite ............ 750 (270) ENADONILESH: SF- re eres en 200 (30) COAL ::: t:) errr reverse n 150-190 (10) er 231,000 (60) ' Range for coals of Southwestern United States used in 10 powerplants, according to Swanson and Vine (1972). *Average for Phosphoria Formation, Western United States, (Gulbrandsen, 1966, table1). GEOCHEMISTRY OF FLUORINE 9 80 SiO,, IN PERCENT 60 | ] 1000 1500 -l 0 500 FLUORINE, IN PARTS PER MILLION EXPLANATION * Aplite (Oligocene) of Pole Canyon-Can Young Canyon area (Lee and Van Loenen, 1971 table 10) © Quartz monzonite (Jurassic) of Pole Canyon- Can Young Canyon area, (Lee and Van Loenen, 1971, table 6) © Granodiorite and quartz monzonite (Jurassic) of Snake Creek- Williams Canyon area (Lee and Van Loenen, 1971, table 5) Ficure 5.-Correlation between fluorine and silica contents of hybrid granitoid rocks of Jurassic age and of aplite of Oligocene age from the southern Snake Range, Nevada. Carpenter (1969) indicated that the average fluorine contents of recent oceanic sediments range between 450 and 1,100 ppm. Sediments from freshwater lakes have similar fluorine contents. Analyses of some mineral com- ponents of sediments indicate averages ranging from 50 ppm F (opal) to 1,560 ppm F (glauconite); carbonate minerals ranged from 280 to 1,075 ppm F (Carpenter, 1969, table 2). According to J. B. Cathcart (oral commun., 1973), the high amount of fluorine in phosphorite may be due in part to the ""stripping'' of traces of F-! from ground water by calcium phosphate minerals after sedimentation and burial. Soils, which average 285 ppm F in 327 samples, are richest in fluorine where they are high in clays, phosphate, or micas (Fleischer and Robinson, 1963, p. 66). Analyses by the U.S. Geological Survey's Washington laboratory of 959 soil samples collected by H. T. Shacklette in a study of the geochemistry of soils throughout the United States indicate an average content of 291 ppm F (H. T. Shacklette, written commun., 1973). These values are sub- stantially below the averages for most crustal rocks. On the basis of average fluorine contents of igneous rocks, Fleischer and Robinson (1963, p. 67) estimated the average content of the continental crust to be 650 ppm. The average fluorine content of 533 volcanic rocks (rhyo- lites, andesites, and basalts) compiled for this report is 660 ppm, which is perhaps fortuitously close to the crustal estimate of Fleischer and Robinson. ABUNDANCE OF FLUORINE IN GASSES AND WATERS Some gases and waters of volcanic origin long have been acknowledged as containing large amounts of fluorine (Zies, 1929; Fenner, 1983). Zies estimated large tonnages of fluorine being exhaled from cooling volcanic tuff in the Valley of Ten Thousand Smokes, Alaska, and Fenner, quoting Perrét (1924) and other early observers, described vast volumes of fluorine-bearing gas and extensive fluorine-bearing sublimates given off in the eruption of Vesuvius and other Italian volcanoes. The hydro- thermally altered volcanic rocks of the Valley of Ten Thousand Smokes contain fairly abundant fluorine (500-2,100 ppm, Lovering, 1957). The hydrofluoric acid content of "active" gases from a number of volcanoes, given by White and Waring (1963, table 1) are as follows: HF, in weight percent Volcano Rock type (number of samples in parentheses) Hekla; Iceland.;................... Basalt trace (2) Kliuchevskii, Kamchatka.... Basalt 1.6 (7) Aso caldera, Kyushu ........... Basaltic andesite A (1) Vesuvius, Tephritic leucitite A (1) Showa-Shinzan, Hokkaido - Hypersthene dacite 2.5 (11) Katmai, Rhyolite 21 (9) In a general way these data show an increase in fluorine content of gases evolved from volcanoes with successively higher silica contents, from basalt through dacite to rhyolite. The anomalously low hydrofluoric acid amount for the one analysis of gas from leucitite lava is not in keeping with the reports of large volumes of fluorine emanating from Vesuvius. White and Waring (1963, table 2) also presented data on the hydrofluoric acid content of fumarole condensates from several volcanoes, as summarized here: HF, in milligrams per litre of water yglano Rock type (number of samples in parentheses) Hekla, Iceland..............;...... Basalt 15 A1) Santa Maria, Guatemala..... Andesite 20.4 (1) Sheveluch, Kamchatka........ Andesite 38 (4) White Island, New Zealand Hypersthene andesite 69 (7) 10 GEOLOGY AND RESOURCES OF FLUORINE Additional data on fumarole condensates (White and Waring, 1963, table 3), expressed in a different way, are: Fluorine, in parts Volcano Rock type per million (number of samples in parentheses) Kilauea Iki, Hawaii Basalt 20 (1) Ebeco, Kurile 8: Augite andesite 2.4 (1) Showa-Shinzan, Hypersthene dacite 195 (1) Again, fluorine content seems to increase with in- creasing silica content of the source lava, from basalt through andesite to dacite. Sublimates of volcanic fumaroles and eruption clouds were also reported by White and Waring (1963, table 5): Fluorine, in weight percent Volcano Rock type umber of samples in parentheses) Paricutin, Mexico ................ Basalt-andesite 0.55. (2) Bezymyannyi, Kamchatka.... Andesite 0067 (1) Valley of Ten Thousand Smokes; Rhyolite 14.9 (1) \ The general fluorine-silica correlation is evident here, too. Pertinent to the probable occurrence of fluorine- bearing condensates associated with volcanic materials are the data on fluorine content of tephra erupted from the volcano Hekla, Iceland, in 1970. Thorarinsson (1970, p. 47) reported as much as 2,000 ppm F in silicic tephra erupted in early phases of the 1970 eruption. This is a notably high amount compared to the 180 ppm average for Iceland basalts (Barth, 1947). Much of the fluorine in the silicic tephra was probably in the form of a readily soluble condensate, inasmuch as surface waters have rapidly leached fluorine from tephra shortly after Hekla eruptions (Stefansson and Sigurjonsson, 1957). The fluorine contents of gases extracted from igneous rocks, presented by White and Waring (1963, tables 6, 7), are summarized below: F,, in volume percent Sock type (number of samples in parentheses) DASAIE .l cit oy asan nere 6.16 (8) Andesite, dacite .. 2.03 (5) Obsidian.......... 3.19 (6) aa 2.27 (2) Here there appears to be no correlation of fluorine with silica, and basalts appear to contain gases carrying rather high amounts of fluorine. White and Waring (1963, table 4) gave the following data on the variation in fluorine content with tempera- ture of fumarolic gases of Showa-Shinzan: Temperature, °C Fluorine, in parts per million 760° 238 525° 169 220° 35 A clear decrease in fluorine content with falling temper- ature is evident. White, Hem, and Waring (1963, tables 1-27) compiled data on the fluorine content of subsurface waters. The data for ground waters collected in specific rock types, or as specific water or environment types are summarized in the following list: Fluorine, average in parts per million Rock a Gc. of wale DPC (number of samples in parentheses) Granite, THYOIE, EIC 0.9 (14) Gabbro, basalt, ultramafic... 2 (12) Andesite, diorite, syenite........ 1 (4) Sandstone, arkose, graywacke... A (16) Siltstone, clay, shale............ .6 (18) Limestone.. .3 (14) Dolomite .6 (5) Other sedimentary rocks. A (4) Quartzite, marble............ 2 (7) .6 1 (5) 0 Other metamorphic rocks............ 3 (13) Unconsolidated sand and gravel .... 2.6 (16) Sodium chloride connate waters ....................... 3 Sodium and calcium chloride connate waters .. 2.0 (3) Sulfate-bicarbonate connate waters ................... 4.5 (7) Spring waters similar to sodium chloride 2.1 (5) Spring waters similar to sodium and calcium chloride connate waters.................... 1.3 (6) Thermal waters, volcano-associated geysers...... 36.1 (12) Thermal sodium chloride-bicarbonate volcano- associated waters *4.1 (6) Volcano-associated acid sulfate- ChHOIOFiGE /:: ers iii are 528.4 (3) Volcano-associated acid sulfate springs ...... .6 (6) Sodium bicarbonate-boron spring waters ......... 5 (5) Thermal waters associated with epithermal imineral dEpOSILS ...:::. !:: rre: inne 5.4 (11) Non-thermal saline acid waters from mines, etc. 2.3 (3) Travertine-depositing spring waters ................. 4.2 (5) Thermal waters probably meteoric in origin.... .3 (3) Saline waters associated with salt (EpOSLES, ELC -s. :.1,,: -so 7.7 (8) 'Excludes one water sample containing 4.0 ppm F collected from hornblende gneiss, Transvaal, Republic of South Africa. "Excludes one water sample containing 24 ppm F collected from lake beds, Bruneau, Idaho. 'Five waters associated with rhyolite, dacite, and other silicic volcanic rocks average 10.0 ppm F; 7 waters associated with andesite and basalt average 3.4 ppm F. 'Four waters associated with rhyolite average 5.8 ppm F; 2 waters associated with andesite and basalt average 0.8 ppm F. *Excludes one strongly acid hot-spring water sample containing 806 ppm F collected from White Island (andesitic volcano), New Zealand. As seen from the preceding list, the fluorine content of waters collected from specific rock types in general reflects the relative abundance of fluorine in those types of rocks. The average fluorine content of 123 water samples from all rock types is 0.4 ppm. Connate waters of all types contain much higher amounts of fluorine; 26 samples average 2.7 ppm. Still higher are thermal waters associated with vol- canoes and epithermal mineral deposits, and related waters; 51 samples average 5.4 ppm F. Notably nine thermal waters associated with silicic volcanoes average 8.1 ppm F whereas nine thermal waters associated with mafic volcanoes average 2.8 ppm F. Fleischer and Robinson (1963, fig. 3) showed the distri- bution of fluoride content of ground waters, by counties, in the United States. A generalization of their map is shown as figure 6. Significance of the data in figure 6 is discussed in the following section on the geochemical cycle of fluorine. Livingstone (1963, tables 6, 7, 9-16, 18, 23, 25, 26) presented data from which it is computed that river and 11 GEOCHEMISTRY OF FLUORINE '12rem puno83 ut 124814 £961) uosurgoy pue 12y>staI| wor; 'sae§ portup ay} Ur puno13 Jo juaju0) apuon|; ;o uonngqLnst(-'9 © 10 4 vudd gf areotput seare popey§ '(g '34 «p£4 94 84 .08 28 98 88 .06 .26 «6 96 86 001 201 «FOT 901 801. O11 211 «pI L911 811 T T I I I I I I I I I I I I I I I I I I I T 2217 2C sy ! "% 005 0 s3uiw 00s «pe I~ ~ ve 92 [~ ~.9¢ «82 [~ ~.ee o€ [ ~.2g «v€ [C ~.ve 9€ [ ~.9€ »8€ [~ ~.se 0p [ ~.or & & Feesuth ¥ a N 4 au 0+ $s fi nis «2p |" ifiL % N.er 1,9: \ \ fair \ -- & «bp [7 1° ® } N.pP C3 a F A ~ p 1 > \ / N.or 9v \A * ev c_ \ \ \ \ \ \ \ \ \ 1 1 1 L 1 1 1 1 I I | 1 I 1 [ L-S.gp +99 99 89 07 24 «PL 94 +84 08 28 «v8 - .98 88 06 «26 «v6 - .9%6 .86 .O0L .2O1 .POL .901 .80T .OIL .ZII .911 .8H1 .¢ZL .9¢1 .82l 12 GEOLOGY AND RESOURCES OF FLUORINE lake waters of North America have an average fluorine content of 0.25 ppm (101 samples). (Compare Correns' (1956) estimate of 0.26 ppm F as the mean content of river waters.) Livingstone (1963, p. G41) also computed from the data of Konovalov (1959) that the mean fluorine content of rivers in most of the U.S.S.R. is about 0.09 ppm. Only the Rio Grande system seems abnormally high in fluorine among the systems averaged in the North American estimate given in the preceeding paragraph. Nine samples average 0.6 ppm F (Livingstone, 1963, table 16). In addition, 7 samples of surface waters from Devils Lake basin, North Dakota, have an average fluorine content of 1.2 ppm, and 11 samples of surface waters from the Basin-Range province (as defined by Gilbert, 1928, p. 1) and adjacent closed basins have an average fluorine content of 5.3 ppm (Livingstone, 1963, tables 20, 19). Fleischer and Robinson (1963, p. 72) on the basis of newer data endorsed the value of 1.2 mg/l (milligrams per litre) (1.2 ppm essentially) for the fluorine content of Pacific and Atlantic sea waters given by Thompson and Taylor (1933). GEOCHEMICAL CYCLE OF FLUORINE A review of the foregoing data allows some generaliza- tions on the geochemical cycle of fluorine, but because data are incomplete or lacking on many aspects of the problem, much uncertainty remains. Fluorine contents seem to be lowest in igneous rocks that have been irrupted through oceanic crust, such as andesites and basalts in the Aleutian and Hawaiian Islands and Iceland. The undifferentiated Hawaiian rocks clearly are derived from melting of the mantle, and the implication is that the fluorine content of the mantle is no higher than that of the basalts derived from it. Additional detailed chemical data on oceanic basalts other than Hawaiian are needed to indicate if mantle elsewhere in the world has similar fluorine compositions. Some Hawaiian basalts that have migrated to shallow chambers have become differentiated before eruption, and this differentiation caused not only an increase in silica content, but also a substantial increase in fluorine content. It is evident that such magmatic differentiation does con- centrate fluorine, as many earth scientists long have conjectured. Basalts irrupted through continental crust tend to contain appreciably higher fluorine than do oceanic basalts. These continental basalts tend to have alkalic affinities; whether their higher fluorine content resulted from differentiation toward alkalic composition or from contamination by fluorine-enriched continental crust is a moot point. More data are needed on the fluorine content of alkalic rocks differentiated from mantle-derived basalts in the oceanic environment to test if differentiation or con- tamination might be dominant in the continental environment. Silicic rocks that contain high amounts of fluorine also tend to be alkalic. Like the mafic rocks, those that show strong alkalic affinities tend to occur in continental environments. Even though differentiation toward silicic and alkalic types clearly does occur in the oceanic environment the process may be enhanced where it occurs in continental crust. Certainly, rocks-either silicic or mafic-that contain unusually high amounts of fluorine-5,000 ppm (0.5 percent) and higher-have formed in continental crust. Although silicic igneous rocks as a whole average higher in fluorine than do mafic rocks, many silicic igneous rocks have fluorine contents as low as those that characterize most mafic rocks (fig. 7). The siliceous character of such rocks likely was not acquired through magmatic differentiation. The data on distribution of fluorine in granitoid rocks of the southern Snake Range, Nev., indicate that fluorine content of the rocks was dependent on initial availability of fluorine and on the abundance of minerals that were capable of taking up fluorine. The negative correlation of fluorine and silica in numerous suites of igneous rocks thus might best be explained if their fluorine content represents that of the deep crust in which they were generated, and if they were not subject to strong mag- matic differentiation after initial melting and before extrusion or emplacement in the upper crust. Such rocks likely changed composition substantially throughout their fluid history, but probably by assimilation of (reaction with) wallrocks rather than by fractional crystallization. The evidence that crustal composition controlled fluorine content of some igneous rocks, and the strong association of fluorine with alkalic rocks, raise the question of the origin of alkalic rocks and the high concentrations of fluorine accompanying them. Without discounting the possibility of fractionating alkalic rocks from the mantle to account for their high alkali and fluorine contents, it seems quite possible that some alkalic rocks originated because alkalis became segregated and concentrated by the fluxing action of fluorine that was intrinsically high at a site of magma generation in the crust. Phosphatic rocks, for example, if melted, would develop a phosphate- and fluorine-enriched magma chemically similar to some known alkalic rock types. Some alkalic rocks therefore may reflect origin in fluorine- enriched segments of the crust, and their fluorine content may have been initial rather than a result of magmatic differentiation. The fluorine content of metamorphic and sedimentary rocks is similar to that of most igneous rocks. If some ig- neous rocks are generated by melting of crustal rocks, including metamorphic, sedimentary, and preexisting ig- neous rocks, it is expectable that they contain generally similar amounts of fluorine. Conceivably some igneous rocks would contain less fluorine if it were driven off during melting that formed the magmas, and conceivably GEOCHEMISTRY OF FLUORINE 13 80- £3.‘ es .‘ ® L Ci lit.: 7 "ams * ° 75_.5~~: " H *% A 0 ...:.~$.-,‘“. .... ® ® ag? 911. /: £ ‘éizziwf 'or > $ e e ..° # : * 70 |- 2.33: 3..::. R n: :$" x I t r*: G'. .°J:\ x 65 |- *> 2. all, t A. o.~.?§oc’.. # s ‘ e s e a 5 F i & s 860_ 'n * z Q. § I # ;* f a. fi'. ag bd Z F4. g ® . + .O-N55__' : $0 lut t* e ‘0 0 £ rj: b * ® e 0 50|- Ik + "*%. -§lj. U (le oP *% e 8 e ® -$g§i! vs * l: % s o.. .. 0 45—‘0 he : F 40|- * * 35 1 1 1 1 | 1 1 | 1 ) 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 FLUORINE, IN PARTS PER MILLION Ficure 7.-Scatter diagram showing relation of fluorine to silica in 582 volcanic rocks. Data from published sources, and unpublished U.S. Geological Survey sources. some igneous rocks would contain more fluorine if the parent rock contained unusual concentrations of fluorine. The common occurrence of a bimodal assemblage of young alkali rhyolites and basalts, for example in the Western United States, both enriched in fluorine, provides insight into the cause of some fluorine enrichment. The basalts undoubtedly are mantle-derived; fluorine enrichment of a type such as that at Craters of the Moon, southern Idaho, could have occurred by differentiation following irruption of magma into a high level of the crust. A strong negative correlation of fluorine and silica in these rocks (fig. 3), however, suggests concentration of fluorine by a mechanism other than magmatic differen- tiation. The related alkali rhyolites in the same region, compositionally akin to the young basalts and of similar age and geologic setting (erupted through tensional struc- tures), likewise may have been mantle-derived. For such rhyolites to have been derived from the mantle, extreme fractionation of the more mobile and volatile compon- ents of mantle material must have occurred. Such frac- tionation by partial melting of mantle material could account for the unusually high fluorine contents of the rhyolites. The fact that intrusive rocks tend to contain more fluorine than do extrusive rocks indicates that fluorine may be lost as a volatile component in a low-pressure environment. The data showing abundant fluorine in some volcanic exhalations corroborate this conclusion. 14 GEOLOGY AND RESOURCES OF FLUORINE Some intrusive rocks are surrounded by halos of altered rocks in which fluorine is substantially enriched, sugges- ting that even magmas under relatively high pressures lose fluorine. Analytical data on the fluorine content of volcanic gases, fumarole condensates and sublimates, and thermal waters associated with volcanoes, indicate a tendency for more fluorine to be exhaled by silicic volcanoes than by mafic ones. This relation is in keeping with the fact that silicic igneous rocks generally contain more fluorine than mafic igneous rocks, but more data are needed to correlate fluorine contents of exhalations directly with fluorine abundance in magmas. Data on the fluorine content of alkalic lavas and their exhalations particularly are needed. Nevertheless, it seems a fair conclusion that the higher the fluorine content of the magma, the more fluorine is given off in volatile phases, whether the magma is extruded or emplaced below the surface. Further, the temperature data from Showa-Shinzan, Hokkaido, suggest that the higher the temperature of the magma, the more readily fluorine is driven from the magma. Whether a magma is intruded and cooled below the surface, or extruded, it seems clear that release of volatile components-including fluorine-from magma results primarily from pressure decrease, due to magma rise or rock fracturing. Crystallization of magma by itself would not be expected to drive off much fluorine, although the fluorine content of the residual magma would increase unless the minerals that crystallized included some capable of taking up fluorine (micas or amphiboles). Noble and others (1967) showed that devitrified silicic volcanic rocks have about half of the fluorine content of glassy equivalents (0.09 percent F, average of 89 samples, compared to 0.21 percent F, average of 72 samples), indica- ting that appreciable fluorine is lost upon crystallization (largely devitrification). This difference may be the result of lack of minerals in the crystallized silicic rocks capable of taking up fluorine. The fact that soils average substantially lower in fluorine content than do rocks of the crust shows that fluorine tends to move into ground and surface waters during weathering. Apparently ground waters generally dissolve fluorine from the rocks with which they are in contact in propor- tion to the abundance of fluorine in the rocks. For example, ground water in North Carolina and Virginia shows high fluorine contents only in areas where granites are known to contain fluorite. An arc extending from northwestern Ohio westward through Iowa and north- westward through the Dakotas has ground waters with high-fluorine content. The arc is underlain by glacial materiais of the Mankato, Cary, and Tazewell Stades of the Wisconsin Glaciation, but not by those of the Iowan, Kansan, or Nebraskan Glaciations. Possibly the fluorine content of the waters can be related to the composition of glacial material, or in a complex way to bedrock geol- ogy. In southern Arizona some areas where ground waters have high-fluorine values coincide with areas of volcanic _ rocks with high-fluorine contents. Undoubtedly meteoric waters transfer some fluorine from the atmosphere (derived mostly from the oceans and in part from volcanic exhalations and industrial sources) to ground water near the Earth's surface, but data are needed to quantify the amount, seemingly less than that supplied to ground water from rocks. Surface waters of the United States, composed of meteoric water mixed with a certain amount of fluorine- bearing ground waters, appear to stabilize with a fluorine content of about 0.25 ppm. These waters flow into ocean waters containing 4-5 times as high a concentration of fluorine, and thus tend to dilute the ocean waters. Barth (1947, p. 422) argued that only a fraction of 1 per- cent of the fluorine derived from weathered and eroded rocks remains in solution after it reaches the sea; the rest is precipitated. According to Carpenter (1969) calcium carbonate precipitation dominates the removal of dis- solved fluoride from sea water; precipitation of calcium phosphate undoubtedly removes substantial amounts. Of course it is apparent that much fluorine is transport- ed by rivers to the sea in suspension, contained in clay minerals that formed by weathering and in micas that are residual from weathering. Carpenter (1969) concluded that most of the fluorine which moves from the land to the sea does so in this fashion. A comprehensive report on the geochemistry of fluorine (Koritnig, 1972) was published after this report was written. ELEMENT ASSOCIATIONS By R. E. Van Arstin® and D. R. SHaws Spectrographic analyses (Allen, 1952; Heyl, 1967, p. 25-26; Hall and Heyl, 1968, p. 658-661; Trace, 1960) reveal that minor elements associated with fluorite include aluminum, antimony, barium, beryllium, boron, carbon, cerium, copper, dysprosium, erbium, europium, gadoli- nium, iron, lanthanum, lead, lithium, lutetium, mag- nesium, manganese, neodymium, niobium, potassium, praseodymium, samarium, scandium, silicon, sodium, strontium, sulfur, terbium, thorium, thulium, tin, uranium, ytterbium, yttrium, zinc, and zirconium. In addition, galena associated with fluorite commonly contains silver and antimony, and the associated sphalerite generally contains cadmium, gallium, and germanium. The geochemical association of fluorine with beryllium in certain silicic volcanic rocks of the Western United States has been cited by Staatz and Griffitts (1961), Griffitts and Powers (1963), and Shawe and Bernold (1966). Coats, Goss, and Rader (1963) showed regional variations in the fluorine content of glassy rhyolitic vol- canic rocks of the Western United States, and Coats (1956) GEOGRAPHIC DISTRIBUTION OF FLUORINE IN IGNEOUS ROCKS 15 cited the correlation of fluorine and uranium contents in these Cenozoic volcanic rocks. Worl (U.S. Geol. Survey, 1970) found anomalous amounts of molybdenum in Pre- cambrian wallrock next to the fluorspar deposits of the Northgate district, Jackson County, Colo. Altered wall- rock contains as much as 800 ppm Mo, and the amount decreases away from the vein. For fluorite from Illinois, Virginia, and various districts of Colorado, New Mexico, and Utah, Blanchard (1966) gave the contents of strontium and yttrium, the two most abundant and widespread minor elements of the fluorite. Fluorite from the Illinois district (Ruch, 1968, p. 5) showed, in parts per million: < 0.04-0.60 scandium; 0.17-1.5 europium; 0.2-9.4 dysprosium; 0.2-4.1 samarium; 1-113 copper; 5-188 sodium; <0.1-1.2 manganese; <0.2-0.8 cobalt; yo ho f y: ® D. Sedimentary rock associations-Continued 3. Disseminations and vug fillings in carbonate rock (Permian of north-central Wyoming; Lockport, N. Y.) 4. Disseminated deposits in volcaniclastic rock (Thomas Range, Utah) 5. Tuffaceous lake beds (Rome, Oreg.) 6. Saline beds and basin brines (Searles Lake, Calif.) E. Miscellaneous 1. Major fault zones with fluorite (Illinois-Ken- tucky district) 2. Cryptovolcanic structures (Hicks Dome, IIl.) Some inevitable overlaps occur in this, or any other, classification. Therefore, we discuss the occurrence and distribution of deposits according to their primary min- eral character (fluorite, topaz, or phosphate) and we dis- cuss the geology of deposits and districts geographically, State by State. FLUORINE IN FLUORITE DEPOSITS AND OCCURRENCES Fluorspar is a mineral aggregate or mass containing enough fluorite (CaF,) to be of commercial interest; pure fluorite is 51.1 percent calcium and 48.9 percent fluorine. Fluorite is harder than calcite (CaCO,) and softer than quartz (S10,), two common mineral associates with which it may be confused, and is appreciably heavier than either one; it is not as heavy as barite, another common associate. Fluorite has a vitreous luster and a wide variety of colors, commonly shades of purple or green, but also white to yellow, green, blue, rose, purple, red, brown and purplish black. Fluorite generally crystallizes as cubes or octahedrons but may occur in massive form, crusts, globular aggregates showing radial fibrous textures, and earthy and banded cryptocrystalline forms resembling chalcedony. It has perfect octahedral cleavage, which with its characteristic crystal forms, color, heaviness, and hardness generally is sufficient to differentiate it from other minerals. Minerals other than calcite and quartz commonly found in fluorspar deposits are barite (BaSO,4), galena (PbS), sphalerite (ZnS), pyrite (FeS,), chalcopyrite (CuFeS,), hematite (Fe,0,), limonite (hydrous ferric oxide), manganese oxides, clay, and various oxidation products from breakdown of the sulfide minerals. In a few deposits the fluorite is accompanied by uranium, rare- earth, or beryllium minerals, where they locally might be coproducts; major fluorspar districts with associated uranium minerals include Jamestown, Colo., Thomas Range, Utah, and Bare Mountain, Nev. (Wilmarth and others, 1952). Fluorine is widely distributed throughout the United States in fluorite deposits (fig. 9). The patterned areas of figure 9 include reported occurrences of fluorite as well as productive fluorspar deposits. Generally speaking, most «be [~ ~ '(¥L61) 49H pue 'aunsry uea 'rom uo paseq 'sareig portug ay ut sitsodap ;o uonngqLost(-'6 140914 2 »94 08 ce «ve 98 -se .06 .c6 «v6 96 86 1061 !i .for .- pol-! _;! .gof s0ff _| [ffs Ari «911 «811 T T T T I I I I I I I I I I I I I I I I I I I N 00s o a por- Li- -H- s saum oos o \ x m R. \ 4 @ k» f 0\ /d C 3 "92 [~ g_ ® X = yxi108¥9N wanos } GEOLOGY AND RESOURCES OF FLUORINF 18 JIPJQ IIIWI'II'IAIIIIh ~ se 2 k 1 \ \ \ \ \ \ \ \ | 1 1 1 I 1 1 I I 1 | I 1 | I | 1 1 1 LS «P9 - .99 0.89 = .0L _ .24 .b4 _ .8L _ .08 - .28 = .98 - .88 .06 .26 .P6 .96 .86 .00L .ZOT .FOI .901 .801 .OIL «III .¥Il .911 .811 .0ZL .921 .821 GEOLOGIC ENVIRONMENTS AND DISTRIBUTION OF FLUORINE 19 of the fluorspar deposits and fluorite occurrences are found in the mountainous regions of the Western United States excluding the Colorado Plateau, and in the Appalachian Mountains in the Eastern United States. The principal producing area of the United States-the Illinois-Kentucky district-is outside of mountainous regions. FLUORINE IN TOPAZ DEPOSITS AND OCCURRENCES Topaz (Al, [SiO,] (OH, F),) generally has not been produced as a commercial source of fluorine although the fluorine content of topaz is high enough to suggest a potential source. Topaz ideally contains 32.6 percent silica, 55.4 percent alumina, and 20.7 percent fluorine (total 108.7 percent, less 8.7 percent O for F). Topaz is hard (harder than quartz and softer than corundum) and heavy (density 3.49-3.57). Color is variable, commonly colorless to yellow, pink, light blue, and light green. Its hardness and vitreous luster have made colored topaz a valuable gem mineral. Topaz usually forms orthorhombic prisms with perfect basal cleavage. Common mineral associates are quartz, fluorite, tourmaline, beryl, cassiterite, wolfra- mite, rutile, and muscovite. Topaz is found chiefly in silicic igneous rocks such as granites, granite pegmatites, and rhyolites and in veins, disseminated deposits, and greisens associated with these. It is also associated with other heavy minerals in placers near certain silicic intrusive rocks. Topaz deposits are widely distributed throughout the United States, as shown in figure 10. FLUORINE IN PHOSPHATE AND FLUORAPATITE DEPOSITS Fluorine-bearing phosphate deposits occur in rocks of all geologic eras from Precambrian to Cenozoic as igneous apatite, sedimentary phosphorite, and guano deposits, and as secondary accumulations formed by weathering, diagenesis, and reworking of any of the primary materials. Fluorine is present in economically interesting amounts in phosphate deposits in many parts of the United States, and in particularly large amounts in sedimentary phos- phorite deposits. The phosphate mineral of the sedi- mentary phosphorite deposits is carbonate fluorapatite-Ca,,(PO,,CO,),F;_;. The ratio of F to P;,0; in this mineral is about 1 to 10; P;0; content in minable deposits ranges from 28 to 38 percent, and F content of mined phosphate of the Phosphoria Formation in the Western United States is about 3.1 percent, that of Eastern phosphates is 3.6-3.8 percent. Presently commercial phos- phate deposits in which fluorine occurs as a potential by- product are confined to marine sedimentary phosphorites of Ordovician age in Tennessee, of Permian age in the Western United States (southwestern Montana, south- eastern Idaho, western Wyoming, and northern Utah), of Miocene age in North Carolina, Georgia, and South Caro- lina, and of Miocene and Pliocene ages in Florida, and to residual deposits derived from some of them. Locations of phosphate deposits in the United States are shown in figure 11. The phosphate deposits of the Atlantic Coastal Plain (Florida, Georgia, South Carolina, and North Carolina) are in unconsolidated rocks of Tertiary age. The deposits consist of quartz sand, clay, and irregularly rounded, highly polished, generally dark colored phosphate grains. The phosphate grains range in size from about 0.1 mm to several centimetres in diameter, but most are sand sized. The phosphate deposits of Tennessee are derived from calcareous rocks of Ordovician age. The deposits are residual and consist of rounded grains of sand- to silt-sized brown phosphate, quartz, and minor clay. Irregular masses of quartz and phosphate sand cemented by phos- phate are common. The phosphate deposits of the Western United States (Idaho, Montana, Wyoming, and Utah) are in rocks of the Phosphoria Formation of Permian age. The deposits con- sist of consolidated beds of carbonaceous, pelletal, sandy phosphorite associated with black shale, chert, and lime- stone. Nonsedimentary fluorapatite deposits of a variety of types may have potential for byproduct fluorine produc- tion (fig. 11). Fluorapatite occurs with magnetite at Ausable, Mineville, and Trembley Mountain, N.Y., Can- field mine, Dover, N.J., and Iron Mountain, Mo.; with rare-earth fluocarbonates and magnetite at Scrub Oaks mine, Dover, N.J.; in skarn in Hurdtown, N.J.; and in dikes with ilmenite near Roseland and Piney River, Va. Fluorapatite is found in magnetite-rich contact meta- morphic deposits at Iron Springs, Utah, and at the Barth mine, the Modarelli mine, the McCoy district, the Black Joe prospect, the Emery-Fisk prospect, and in the Buena Vista Hills, Nev. FLUORINE IN OTHER DEPOSITS Fluorite, topaz, and fluorapatite occur also in types of deposits other than those just described. In addition, some other fluorine-bearing minerals of possible economic significance are the fluorides, cryolite, Na;AIF,, sellaite, MgF;, and villiaumite, NaF; the humite group of fluorine hydroxysilicates; and bastnaesite, (Ce,La) (COs;)F, whereas other still less common fluorine-bearing minerals will probably be of little economic significance in the future. Fluorine is also present in many other minerals including silicates like tourmaline and the micas (phlogopite, zinnwaldite, lepidolite, muscovite, and biotite), and phosphates (monazite, xenotime, and amblygonite). Some types of deposits are not known domestically. A cryolite deposit at Ivigtut, Greenland, is virtually a monomineralic pegmatite although it contains also some complex fluorides, fluorite, topaz, microcline, columbite, GEOLOGY AND RESOURCES OF FLUORINE 20 'sareg partup atp ut Sunreaq-zedo; pue zedo1 'sitsodap zedo; 10few Jo «PL 94 84 08 28 Bus .98 88 .06 26 Bus 96 86 OOT 201 «POT 901 «SOT O1 »I «PI «911 811 T T I I I I I I I I I I I I I I I I I T go? M T T EE |T. mz ao squi3w071X 008 o 98 $ y _|_|r.L|1|.JL|_ <. ; S31iW OOS »¥2 [~ J, T.v2 t f; 92 [~ ® / ~.92 \ 3 a11j0Ay1 -x % fap 1oAy1 Buneag-zedo1 srl zedo| _. z. 7 9 W Tol" wijd « & sodap sofepy - o Kn fz tldn oue] '1a9e]d 12100119 x A 1 More e FET -s sv x a u aas ] TAs ol- ___ w. ~a jofinsip uebug 0 \_ s Sg: £ sissIN f~ 1 U O PN « & I ¢ 4_/:.,\\(<€/14>) | XX \ //\ul vs a [* 1 .\ | i 1 201%4 @ , j b ze Leow aX va n v % 4 Y1 y _ [x> "y q by M «M f vow 0 Hov Tx o | x* I Cid1 y" *,. x41. £3 § * _| auru Saw—40 a s 1 l el- \|L\||W_ ._||||||.||l|lll| Mu. /o # ‘ 7/ Tore 5~} ,., "00 uosipep 'sauip 1aajig H Inrllflrfl/f P x/ mL 1 I | ee o { mumsfilm—s; 8 v 8 K v 3 _ 0 a m 1 I‘- \ Tock & m o \ x] S ~.eg /_ ~ - Pune messes 5292an4 FE m, { uos1 4 c A opyay-pein" . durnu v3.5.5 E I df , se |- r’l4’um’ XXI 3 4 M.sf vous v rg a K | 1/~:/~ [ 3 I i | =% 1 F. & . Til/F 198 ( oye: "me *~ -~ <1 1 1 pomss # a n 0g _ | | I 0 X 1 1 HY q j» 1 Fes. ar mento f vasa N.er filllllllllllh =J I A Nop i §. 27. A \ | +B: 3 «pp xhxcj [Enzulff no 1 i fl/ eyseiy 'abuey sy001g | .s; sams &: ~ seare Bujuiu annoy \\© y pjord eibi0a9 Ewfisomnmntmt chafnmoz $ovoXx a a ¥ uis -ei610a9 U19y1ON \; Ask R y=a_ sisodap apeydsouy 8 jouisip yin0os-2! [Ow 5] | T 3 in § ints 4 1 to q ~ .0€ a av ---- Px NOILVN¥14X3 § isse prose t y. 1 | §: C ( uo1s9] 18 var vias 1 ~ QWMMLAM—OO‘. UCN—fl- n 1&4 m JPI/i _— 1 \ “VI i108 Y9 2 I X g P 1 1 «ce [~ sok - %. /k L y £ Kv 3 # v/ ,__ N . A4 6 / ¥ No ( % i \v\\m_\|l.| a“ x vow 0 Hov 7% o d | | 21 i y ”x Nase e Py g 15909 1 O rox, "atas * U Sisodap ,, 1201 uMmo1G,, |aassouus | h,- I 1 f wy 8 a a= L I u [ ""*" mat guy0129 ( o (L pue pusissot anuofien f a r sg - | iS, _\\\\qufl_.m.z Wig A x 9 a 4 N @ 4 =1 ~ *urejunop c. AI | 1/4/11/ | 3 aA ANT «N* ~ Sm ‘Zzé, - ~ ¥: 7 s i w | gov s N vox : $Bu; bri & % JA 1x n 0 8 § 1 $v 3 0 CY & 9 1 Upy« ! 98} \ fa32|fm $50“.va :v_uow umo4,, >xo=-cw¥ __ T0 a \ I | x19 1h} kwEBV __ wy 1 CNJ ./_ XH mP - max m f 1 i \ \ § 1 $810 HRZ . se l_ a | i I" t he \ I [ J. to K guru (eve "*a gy lscl- , ~- 2. L 1 fl 21 & Yos a n N. [fill/l\. aug! 305. PA 3. A v m 0 1 f ] NR * Itumm. se-" ys rw/|)/|'! J z \I A SL ® sioug ses tlo, o os- a I . 4 Is F I 2 1 /// | ¥ £03 v a # a n , 5710498944 M 2% / V 0 x 1 \ o s M U f ¥4, Y a 1 E- NIS N 0 0 8 1 M $- = > L ~-S_N I " \ $s y a" fer s- Pej. } ~ az \ f | | y $5" ¢ w ly 4 0 xv * - \ rl v 1 0 8 & N NI \ a xsuoz. s sv x a a 1 P t og} w Addt$ a _: ~ __ ~ Ay wy 1_grgs1N 1 i \ | G & ld me < «ve [~ a s v's Nov x 9€ [~ i: .8€ [~ 3 [~ x cr [~ h «tp |" c . 9p [~ ~I ep 21 \ \ \ \ \ \ \ \ \ | 1 1 1 | | | | I I I 1 1 | | | | 1 1 4 F4 V {.s 22 ¥€ 2p 49 99 .89 04 »2L «pL 94 +84 08 .28 «v8 - .98 88 +06 26 96 .96 ° .86 .001 .2OL «POI .901 .801 .OIL .III .tIl .911 .811 .021 .¢ZL .9¢1 .82l TECTONIC SETTING OF FLUORINE DEPOSITS 25 Figure 12 shows that a number of major faults have no known fluorine mineral deposits associated with them, for example the San Andreas strike-slip fault zone along the Pacific Coast of California. Fluorspar deposits tend to occur in elongate clusters that form alined zones or belts (fig. 12; Worl, Van Alstine, and Heyl, 1974). Northwesterly and northeasterly aline- ments dominate. Some northerly alinements are evident in Nevada and Colorado and some easterly alinements show up in southern Arizona, in the Missouri-Illinois-Kentucky area, and in the northeastern United States. In places the alinements of fluorspar deposits nearly coincide with major faults even though individual deposits may not actually occur in the major faults themselves. In the Eastern United States dominant northeasterly alinements of fluorspar belts reflect the dominant north- easterly structural grain of the region. In the Western United States in the Basin-Range province, however, fluorspar belts aline northwesterly and northeasterly but structural grain alines northerly. Some pre-Basin-Range faults are oriented northwesterly and northeasterly, have been deformed by strike-slip move- ment (for example, Shawe, 1965, p. 1372-1375), and are reflected by alinement of the fluorspar belts. RELATION OF FLUORINE DEPOSITS TO INTRUSIVE IGNEOUS ROCKS The general coincidence of epigenetic fluorine mineral deposits and intrusive rocks in the United States is shown in figure 14. In some large areas in which fluorine de- posits occur no intrusive rocks are shown in the figure; numerous intrusives are known in these areas but are too small to be shown at the scale of the map. For example, many small mafic intrusives rocks are present in the Illinois-Kentucky district; intrusives are present at Magnet Cove, central Arkansas; and Precambrian intrusives (not depicted in fig. 14) have numerous associated fluorine mineral occurrences in central Texas and in central Colorado. On the basis of rather incomplete data the ages of the fluorine mineral deposits correlate with the ages of the intrusive rocks with which they are associated. Fluorite- and topaz-bearing greisens and pegmatites were formed during the igneous cycle of the Pikes Peak Granite of central Colorado and of the granite in the Llano uplift of central Texas, both of Precambrian Y age (800-1,600 m.y.-million years). The ages of fluorine-bearing greisens and pegmatites agree with those of their associated Paleozoic granitic rocks in the Eastern United States (fig. 14). Skarn or contact metamorphic deposits containing fluorine-bearing minerals are clearly associated with specific plutons of Paleozoic age in the Eastern United States and of Mesozoic and Cenozoic age in the Western United States. These deposits were formed either at the time of emplacement of the plutons or shortly thereafter as a late stage phenomenon of the igneous episode, and hence are certainly of nearly the same age as the intrusive rocks. Many fluorine mineral deposits in the Western United States in Tertiary volcanic rocks or near Tertiary intrusive rocks can be dated by geologic relations as Tertiary in age (R. G. Worl, written commun., 1972). The dated deposits are all younger than 32 m.y. and none of these are demonstrably younger than 6 m.y., with the exception of a few hot-springs apron deposits such as that at Ojo Caliente, N. Mex., that currently are being deposited. The age distribution of these 6-32 m.y. (Tertiary) deposits does not coincide, however, with the age distribu- tion of the times of transition from andesitic to funda- mentally basaltic volcanism in the Western United States shown by Christiansen and Lipman (1972, fig. 5). Their data suggest successively younger ages from western Texas northward and westward, but Worl (written commun., 1973) shows fluorspar deposits of widely differing ages randomly distributed in the Western United States. The deposits thus clearly are not related to the beginning of the episode of basaltic-rhyolitic (bimodal) volcanism, but may have formed locally any time thereafter. In fact, some fluorspar deposits are related to alkalic rocks formed during earlier predominantly andesitic volcanism. Strong evidence suggests that Tertiary deposits are related genetically in some places to inferred intrusive rocks even though the deposits are associated spatially with extrusive rocks. At Spor Mountain, Utah, for example, Staatz and Osterwald (1959, p. 59-61) demon- strated the likelihood of fluorine-rich intrusives at depth, similar in composition to the widespread topaz-bearing rhyolites at the surface near Spor Mountain, and showed that such intrusives were the likely sources of the fluorine in the fluorspar deposits. At Climax, Colo., fluorite and topaz were deposited with molybdenite in several successive episodes each related to intrusion of an alkalic-silicic porphyry stock of middle Tertiary age (Wallace and others, 1968). Fluorine-rich igneous rocks form two well-defined provinces-one in the Eastern United States and one in the Western United States (fig. 8)-that roughly accord with areas of intrusive rocks outlined on figure 14, but there are important variations in the distributions. For example, Tertiary intrusive rocks are distributed in a broad belt through the Western United States, but they are enriched in fluorine only in parts of the broad belt, and these parts generally coincide with areas of fluorine mineral deposits. The absence of fluorine-rich rocks in north-central Washington State shown in figure 8, despite the presence of fluorine mineral deposits (fig. 14), is thought to reflect lack of data rather than a real absence of such rocks in the area. The striking coincidence overall of the areas of fluorine-rich igneous rocks with areas of epigenetic fluorine mineral deposits (compare fig. 8 with fig. 14) GEOLOGY AND RESOURCES OF FLUORINE 26 'sare1g portug ayn ut s[EJoutuu aunrong; nauastdo jo uonngtnstp pue (L961 'Sury wos;) sarpoq aatsn.qut Jrozor2ueuyd zunol4 abe Are111a 1 10 $3901 JAIS -nijut pue a susodap jeaguiuw aunon}} 40 eaiy abe 210203124 40 sy 301 olliugy fl abe s10z0sa)) 40 sy 901 Q NOILVNYVT1dIX3 «pL 94 .84 .08 ze Bus .98 88 06 .26 Bus 96 86 Oot 201 «FOT 901 801 O1 L211 «pI A911 811 T T T T I I I I I I I I I I I I I I I I I I T 22 oe sqwi3W071X 008 o e _||_.|_J|.||...|.J|.||_ oM ¢ saum 00s 0 «pe I~ / J, /m \ f y as 192 [~ ~.9¢ «82 ~ ~.8e 0€ [~ ~.o€ N+ «EEF r.— ~.ze vow o k v 7 % o | ole be vided ae os oice as ie mand so une. 1 5, N.ve 9€ [~ ~.9€ .se ~ ~--£#- ~.se v X s v a a a R _ 1 P" Awan nl nc L. ov [~ T ~.or H a n 0 ”WWW 2p [~ Ner «pp [C in Paa . [" N.ov way \ A \ \ \ \ \ \ \ \ | 1 1 1 1 | 1 | I 1 I 1 1 | | 1 1 1 1 1 1 1 L S.evy PEL 199 O 109 c20l0 EL abi - «942° 186 .,08 28 Is Log Les ~ ife Lhe 196 .ge loot EOT .9OT .901~ Caofl- a¢Tt - .SH Oct .2c1 .921 .82l TECTONIC SETTING OF FLUORINE DEPOSITS 27 indicates a likely genetic relation between such igneous rocks and mineral deposits. In a number of places fluorine-rich igneous rocks are related genetically to the associated fluorine mineral deposits. This is true of the Pikes Peak Granite, Colorado, and associated topaz and greisen fluorspar deposits; the granite of the Llano uplift, Texas, and associated fluorine mineral deposits; a number of plutons in the Eastern United States and their associated greisen and pegmatite deposits; tin granites and associated fluorspar deposits at Lost River, Alaska; and the Climax, Colo., stocks and associated topaz- and fluorite-bearing molybdenite deposits. In other places well-based inference suggests a genetic relation, as in the Illinois-Kentucky district where fluorspar is spatially associated with fluorine-rich mafic alkalic rocks, at Magnet Cove, Ark., where fluorspar is spatially associated with fluorine-rich alkalic rocks, and at Spor Mountain, Utah, and the southern Wah Wah Mountains, Utah, where fluorspar deposits are spatially associated with topaz-bearing rhyolites and their inferred intrusive equivalents. The tendency for intrusive rocks of special chemical composition to correlate with fluorine mineral deposits is illustrated in another way. The quartz diorite line of Moore (1959), which marks the western margin of potassium-enriched intrusive rocks in the Western United States (fig. 14), forms an approximate western boundary of the areas of fluorine mineral deposits. Perhaps the higher potassium content of intrusive rocks east of the quartz diorite line correlates partly with a higher fluorine content of the rocks in this region. RELATION OF FLUORINE DEPOSITS TO GEOPHYSICAL PROPERTIES Figure 15 shows Bouguer gravity data for the United States in relation to the distribution of fluorine mineral deposits. Low-gravity regions, highlighted on the map, in a general way represent zones in the lithosphere-in the crust and possibly also in the upper mantle-where rocks are more silicic and hence lighter than in surrounding areas. These regions broadly represent crustal zones intruded by significant volumes of silicic igneous rocks. The low-gravity contours also roughly outline topo- graphically high regions-the mountainous parts of the country-that have risen isostatically in response to their low density. These low-gravity regions roughly correlate with areas of fluorine mineral deposits, but the correlation is imperfect or lacking in wide areas. Most significant is the absence of a low-gravity zone in the vicinity of the important Illinois-Kentucky fluorspar district. On the other hand, an area of fluorine deposits in north-central Washington State nearly coincides with a low-gravity zone; west- and east-trending lobes of fluorine mineral occurrences in western and central Montana coincide with similarly trending low-gravity lobes; the important zone of fluorine mineral deposits extending northward through central Colorado nearly coincides with a major gravity low; a string of fluorine mineral occurrences extending from eastern Kansas northeastward through Wisconsin lies along a northeasterly oriented gravity low; and fluorine mineral occurrences in northwestern Ohio are clustered on a gravity low. Moore (1962) suggested a correlation between high potassium content of Cenozoic igneous rocks and low Bouguer gravity values in the Western United States. Heat-flow values in the Western United States reported by Blackwell (1971, fig. 2) are relatively high and generally exceed 2x 10-8 cal/cm*/sec in broad regions that coincide remarkably with the distribution of fluorite deposits in this part of the country (fig. 16). Comparison of figure 16 with figure 8 also shows a striking coincidence in the distribution of high heat-flow values and the distribution of fluorine-rich igneous rocks in the Western United States. (Again, if fluorine-rich rocks are present in northern Washington, a condition unknown for lack of data, the correlation would be improved.) On the other hand, comparison of the heat-flow data with the distri- bution of Tertiary intrusive rocks in the Western United States (fig. 14) shows a generally poor correlation; numerous Tertiary intrusives occur in western Oregon and Washington, in central Montana, in western Wyoming, and in southeastern Utah where heat-flow values appear to be low and no fluorine mineral deposits are known. Present heat flow (fig. 16) may be different from that in the middle to late Tertiary when fluorine mineral deposits and fluorine-rich igneous rocks were forming. However, some of the fluorine-rich volcanic rocks are as young as 3 m.y. (Armstrong, 1970, table 3, p. 210-211), and the occurrence of fluorite in modern hot-springs aprons at Ojo Caliente, N. Mex., and elsewhere, indicates that the episode of fluorine mobility likely extends from the Tertiary into the Holocene. Probably the episode of fluorine-rich magmatism and fluorine mineral deposition that commenced about 30 m.y. ago (middle Tertiary) in the Western United States is still manifested by high heat flow throughout the region of that magmatism and mineralization. Where regional aeromagnetic data are available, some correlations between magnetic anomalies and the distribution of fluorite deposits are apparent. For example, in Colorado the positions of elongate zones (alinements) of numerous sharp positive aeromagnetic anomalies (Zietz and Kirby, 1972) nearly coincide with the positions of fluorite mineral belts (fig. 16) and are inferred to represent in part zones of buried intrusive rocks. On a global scale, residual magnetic contours based on data taken by U.S.S.R. satellite Cosmos 49 at altitudes of 155-310 miles (250-500 km) (Zietz and others, 1970, fig. 3) show pronounced northeasterly and northwesterly orientations across much of the United States, reflecting GEOLOGY AND RESOURCES OF FLUORINE 28 sarees portup ay} ut sitsodap aurion|} Jo uonngLustp pue ($961 'Sunsao{ pue preffoom 191J2) 4itae18 1nSnog-'¢| 94 84 08 28 Bus 98 .88 06 .26 v6 96 86 OOT 201 «POT 901 801 OLL »21L APL A911 T 3 I l I I I I I I I I I I I I I I I I I ze|~ sisodap r 006 0 jeauru 40 eaiy 74 ele sau 00s qgz-> Armeg E p 3 og;-> Ameig [| arl= ~, € \_ ce sjebijpiu qg-> Aine e jebyjw opg-> Ammeig [__] a @, scl *S=BL T » S5 vxyisinot / NOILVNYVT1d4X3 # 0€ [ a a \% ~ .\ v19%8 3 is < 3 oM' & «pe [C [~ 8€ [~ Ov [~ 2p [~ «pp [C »9¢ [~ cer \ \ N \ \ \ \ N 1 1 1 1 | | 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 \ 89 04 »2L «PL 94 84 08 .28 «v8 - .98 .88 «26 «v6 - .96 86 .0O01 .2OL «POI .901 .801 .OIL .2II «tll .911 .811 .O¢1 .2¢1 .921 . 2C «v2 97 Bd »2€ »v€ .9€ .8€ OP .IP «bP 99 29 TECTONIC SETTING OF FLUORINE DEPOSITS 'sypor aatsn.qut pang Jo sauoz juasaid -ordewired arre} (g/61 '4qury pue zarz uo paseq) sattewoue »nauSewmo1se aantsod dreys snorowmu jo sjuawaurfe moys opeiojoy ut saut ury '(g (34 'TL61 uo paseq) Jas/;tWw>/[E> UEY} 4912018 4fferoua8 sonfea mopp rey Yim sareig portup ay: ut (porddns) suotSar pue (saut 4aeay) sifoq perouttwu pL 94 84 08 .c8 v8 98 .88 .06 26 «v6 96 -86 001 201 «FOT 901 801 .O 211 apI 911 A811 T T I I I I I T I I I I T I I I I I I I I I T af" Nige he *s 008 o soot &. 'tlh sain 00s o «v2 J es / If, ~.v2 & a B A \ fst, AV % 192 ~ $0 ? ¥ ~.9¢ 82 |~ f 6 ~.82 f m \| \ > f. $v X a a 1 \ I sek _ 2. |- 1 14418 4 z ~ 0€ v1V N Mnlll i 0€ & _ ¢ Qa. s¥us | xx \ | JP), _ w a | \ 4 cl ~A ya's v ¥ a ti _ €. M \ -v w ory 1x o 4 KXK@L £, \ | «¥E ~ ~ve 9€ [~ ~.9€ Be |- __ ~.se » Op [~ ~.ov 2p [~ N.er «tp [7 N.vy »9¢ [~ N.or 8p L \. \ \ \ \ \ \ \ 1 1 1 1 1 | I 1 | I I 1 I I | 1 1 L. 1 1 1 1 L Sl.gp 99 99 +89 +04 »24 P4 94 +84 08 28 «v8 .98 88 +06 26 «v6 - .96 +86 .00L .2OL «POL .901 .801 .OIL .9I1 .8I1 .OZL .22L .921 .821 30 GEOLOGY AND RESOURCES OF FLUORINE the dominance of these directions in gross crustal structure. Earthquake epicenters in the United States reported through 1964 and shown by Woollard (1969, fig. 1) are characteristically disposed in elongate clusters that trend dominantly northeasterly and northwesterly. The earth- quakes undoubtedly are related to zones of active faulting which reflect a fundamental structural framework of the continent. Some of the northeasterly and northwesterly zones of seismic activity coincide with fluorite mineral belts, substantiating the conclusion that the mineral belts are in part related to major fault zones. The fault zones no doubt are of considerable age and some must reflect part of the Precambrian as well as Phanerozoic structural grain. Crustal thickness and mean crustal seismic velocity show little relation to the distribution of fluorine mineral deposits (fig. 17). For example, crustal thickness varies from more than 30 miles (50 km) to less than 20 miles (30 km) in less than 60 miles (100 km) horizontally within the fluorine province in eastern California and western Nevada, and crustal thickness ranges from slightly more than 30 miles (50 km) to less than 20 miles (30 km) from Colorado through New Mexico and into southern Ari- zona without commensurate change in the general character of fluorine mineral deposits in these regions. Mean crustal seismic velocity, which reflects primarily crustal composition and to a lesser degree structural complexity, also shows no correlation with the distribu- tion of fluorine mineral deposits. Apparently the forma- tion of fluorine deposits is not closely related to any crustal properties that influence crustal thickness and seismic velocity. A broad region of low (<8 km/sec) upper-mantle seismic velocity is present in the Western United States and this region quite closely encompasses the region of middle to late Tertiary volcanism and fluorine mineralization (fig. 17). The region of low velocity in the upper mantle likely is characterized by above-normal temperatures, and this condition generally accords with the high heat flow of the overlying crust (fig. 15). Woollard (1968) proposed that the region of low seismic velocity in the upper mantle may reflect a phase transition from normal mantle material of dunite composition, high seismic velocity, and density =3.35, into mantle material of eclogite composition, lower seismic velocity, and higher density. The compositional change resulting from the phase transition suggested by Woollard may have involved fractionation of materials from the upper mantle into the overlying crust, and this event may have taken place in the recent geologic past, from middle Tertiary time onward. Possibly these materials included high-fluorine basaltic and rhyolitic magmas and fluorine-rich fluids that penetrated upward toward the surface of the crust, to account for the fluorine- rich igneous rocks and fluorine mineral deposits of middle Tertiary and younger age in the province. On the other hand, the magmatism and fluorine mineralization may have been phenomena originating within the crust, but in response to high heat flow from the mantle and to related processes. In any case, the spatial coincidence of the region of low upper-mantle velocity with a region of high heat flow, middle to late Tertiary fluorine-rich magmatism, and middle to late Tertiary fluorine mineralization indi- cates a strong mantle control on the magmatism and fluorine mineralization. DESCRIPTIONS OF MAJOR FLUORINE DEPOSITS AND DISTRICTS AND IMPORTANT TYPES OF FLUORINE OCCURRENCES Fluorine deposits are discussed in following pages in geographic groups characterized to some degree by common form, association, mineralogy, and geologic environment. Some deposits do not fit naturally into a classification based on geographic distribution but never- theless are included in the geographic groups for convenience of discussion. The geographic divisions used are as follows: Alaska; Washington-Idaho-Montana- Wyoming-South - Dakota; - Nevada-Utah-southeastern California-western Arizona; Colorado-New - Mexico- western Texas-southeastern Arizona; and Eastern and Central United States (including the Illinois-Kentucky district). Of course not all types of fluorine deposits are present in each geographic group. Within the geographic groups, deposits are described State by State where appropriate. Fluorine-bearing phosphate deposits of the conterminous United States are discussed as an additional group because of their unique character. ALASKA By C. L. SamnsBury Fluorine has not been produced commercially in Alaska, although fluorite is reported at many localities (Cobb, 1964; Worl, Van Alstine and Heyl, 1974). Topaz and fluorine-bearing phosphate rock are also known at several localities. The largest known deposits of fluorite-at Lost River, Seward Peninsula-were drilled extensively in 1972, and large reserves of fluorite (several million tons of contained CaF,) associated with minerals of tin, tungsten, beryllium and base metals had already been announced (Northern Miner, Toronto, Oct. 1, 1970). These deposits are described in detail in this report; other deposits of smaller proved economic potential are described only briefly. DESCRIPTIONS OF INDIVIDUAL DISTRICTS AND DEPOSITS LOST RIVER Fluorspar deposits at Lost River on the Seward Peninsula constitute the largest known fluorine potential in Alaska; the main deposits are located at lat 65°28' N.; S1 ALASKA 'sares portup ay ut sitsodap fexautut auriong} Jo uornng -Lustp pue '(g961 eSSeqx, pue unuap wor) 411ojaA Jrustos apuew-1addn ((g <8 'morz pue 1aspfeg wor) 4110fai Jrwstas fersnu> Ueau pue [EASNQ-'/[ T¥N9I4 0€ «bE «IP 84 .08 28 Bus 98 88 06 «26 Bus 96 86 O01 201 «FOT 901 SOT OL A211 «PLL A911 811 T I I I I I I I I I I I I I I T I I I I I I | susodap des/wy g> Au Sis erguiu aunmon}} 40 eau ~ em f o o a s it o aed ii : o_ apuewm-addn 30 inojuop --- _| x wy 0} jeniaru; - ssau < -y9141 jeisn1a jo 1nou09 -or- LZ das wy z'9> ~l.oz Ar190jaa oiuustas jersn4o ueayy das/wuy g'9-z'9 af GPIO ~.sz T , sesjum aig< GTP FFL \\\\\\\v\\w\\\\\v\\\ P a 7 ~ §§ woe K LF y om L/G) XF \\ \X\\ .\\\n\\w\\w\ ~ 6 \\\ \\\.\ \\\\w\\ . «2p y Rf -N, [4 ff \\\\\\ N.or A 7 7 4 q NN [* ~.ov R mbt t TNMMM Z i l\m\« ksfi... Eis -fs ifs trs A- A -d. lee d -t c hfe t s _ pe se c ct p pcp te r c ge dtl Ltr Bg 22) iss «P9 99 89 04 »24 «PL 94 +84 .08 28 «v8 98 88 06 +26 «v6 96 +86 .OO .2O1 .POI .901 .801 .OIL .2IL «tl .911 «811 .0ZL .2cl «PEL .921 »82T 92 long 167°10' W. Although the main proved reserves are in and near the Lost River tin mine, other deposits of unassessed potential crop out discontinuously along the Rapid River thrust fault for a distance of several miles westward from the mine area. Although the Lost River tin deposits were found in 1903, the beryllium-bearing fluorite lodes were found much later, in 1961-64, during a detailed geologic mapping program by the U.S. Geological Survey (Sains- bury, 1969). No fluorite was produced during the mining of tin ores (1913-14 and 1949-55) from the Lost River mine, although much fluorite passed through the mill. Lost River has the potential of supplying a significant fraction of total United States fluorite production. The main fluorite-bearing deposits in the Lost River area are almost entirely replacement bodies in limestone and argillaceous limestone of Early to Middle Ordovician age, which aggregate at least 8,000 feet (2,400 m), and possibly 12,000 feet (3,700 m) in thickness. The Lower Ordovician limestones are divided into two units, a lower shallow-water facies of thin-bedded argillaceous lime- stone with minor interbeds of thick-bedded limestone, possibly 4,500 feet (1,400 m) thick, and an upper quiet- water facies of medium- to thick-bedded limestone with interbeds of thin-bedded argillaceous limestone, about 6,500 feet (2,000m) thick. Most of the known fluorite deposits are in rocks of these units although one deposit is within dolomitized limestone, below a thin black shale about 50 feet (15 m) thick, which is of early Middle Ordo- vician age. All these rocks are thrust-faulted in a major thrust province, and most ore bodies are localized in shattered dolomitized limestone and limestone beneath thrust faults, or in limestones near or above granite plutons. Two main types of large deposits are known within the district, tin-poor stockworks and tin-rich stockworks. The larger (tin-poor) of these consists of replacement veins, veinlets, small pipes, and tabular replacement bodies in limestone and along the walls of lamprophyre dikes, all of which are localized near or within shattered rocks beneath thrust faults (for more detailed information on the several known deposits of this type see Sainsbury, 1969, p. 64-84). The principal tin-poor deposit (Camp Creek) adjoins the tin deposits at Lost River mine, and was explored by diamond-drill holes in 1964 by the U.S. Bureau of Mines (J. J. Mulligan, unpub. data, 1964; Mulligan presented some diamond-drill sampling data on the fluorite-beryllium deposits in the Lost River valley; Sains- bury, 1965), and by P.C.E. Explorations, Ltd., who proved an extension of the deposit by diamond drilling in 1970-72. This deposit extends for almost 4,000 feet (1,200 m) along the shattered zone beneath the Rapid River and Camp Creek thrust faults and raggedly downdip beneath the thrusts for several hundred feet. The deposit is best GEOLOGY AND RESOURCES OF FLUORINE classed as a stockwork which must be mined on a large scale with assay boundaries-ore reserves appear sufficient to satisfy a large-tonnage operation for several years (Northern Miner, Toronto, Oct. 1, 1970). Beryllium is the most important associate of fluorite in the Camp Creek and similar deposits along the Rapid River fault, and tin is almost lacking. Four other areas, each several thousand feet long, contain stockworks of fluorite-beryllium veinlets and occur along the west extension of the Rapid River fault. The westernmost area, some 7 miles (11 km) west of Lost River mine, was drilled in 1962 by Newmont Mining Co.; this drilling showed that the mineralized zone dipped shallowly south beneath a small thrust which underlies the Rapid River fault. The mineralogy of the ores is similar to that of Camp Creek, described in the next paragraph, and tin-bearing sulfide ores were intersected in the deeper drill holes (Peter O. Sandvik, oral commun., 1963). In the Camp Creek deposit, fluorite occurs as fine- grained intergrowths and interlayers with diaspore, tour- maline, white mica, and small amounts of the beryllium minerals chrysoberyl and euclase. The general mineral- ogy of the veins and veinlets in this type of deposit is as follows, in volume percent: fluorite, 45-65; diaspore, 5-15; tourmaline, 0-10; chrysoberyl, 3-10; white mica, 0-5; and small to variable amounts of todorokite', hematite and euclase, as well as trace amounts of phenakite, bertrandite, bityite and cassiterite. These ores typically are layered rhythmically on a scale of a millimetre or two, with finer layers of fluorite interlayered with fluorite that is accom- panied by the minerals named above. The layering is parallel to the veinlets, and both veinlets and layers may cut obliquely across thin-bedded limestones without off- setting the bedding. Hence, the layers are interpreted as replacement "fronts." Studies of rocks and ores, as well as of individual minerals in the ores, reveal a common suite of trace elements such as tin, beryllium, silver, arsenic, lead, zinc, boron, lithium, and niobium, and show that the fluorite- berryllium deposits are distributed zonally above and around the tin deposits. The second main type of deposit (tin-rich) is exemplified by the Lost River tin mine, which contains fluorite and topaz in close association with the tin ores, as well as in a large body of mineralized rock above a buried granite in the Cupola area (Sainsbury, 1964b). Fluorite- rich selvages along the tin ore shoot, which lies along an altered dike that dips steeply into the granite cupola, are as much as several feet thick, and contain both tin and beryllium minerals in addition to the fluorite. The Cupola area is a crudely conical mass of marmorized limestone cut both by garnet-bearing skarn bodies, and by an intricate 'Work by Fred Hildebrand (written commun., 1970) showed that in addition to todorokite, pyrolusite and psilomelane occur in the ores, and that cryptomelane, chalcophanite, and birnes- site may occur. ALASKA ~ 33 network of veins and veinlets containing fluorite, chrysoberyl, cassiterite, wolframite, topaz, quartz and numerous other minerals. Mining of this deposit, which contains several millions tons of material, would produce fluorite and several valuable byproducts. Ore reserves in the Cupola area (number 1 deposit of P.C.E. Explora- tions, Ltd.) amount to at least 5,974,000 tons of ore grading 30.80 percent CaF,, 0.30 percent tin, 0.18 percent BeO, and recoverable amounts of tungsten and base-metal sulfide (Northern Miner, Toronto Oct. 1, 1970). Deposits which are transitional between tin lodes and beryllium lodes occur in Paleozoic limestone around the south side of the granite of Tin Creek, about 1 mile south- easterly from the Camp Creek beryllium lodes. These transitional deposits contain fluorite, chrysoberyl, cassiterite, manganese oxides, and sulfide minerals. Locally, high gold values have been reported in some of the oxidized sulfide veins (John Conrow, oral commun., 1963), a relation which was found to exist at the Rapid River beryllium -fluorite deposits (based on samples taken in 1970 by the Geological Survey). The ore minerals occur in numerous veinlets from a fraction of an inch to a few feet thick over an area of about one-half mile square; how- ever, the area has not been drilled and the depth to granite is unknown. Hence, although the potential may be large, it is unassessed. The distinguishing characteristics of the large fluorite deposits of the Lost River area are: (1) their association with tin deposits that are genetically related to biotite granites, (2) their complex mineralogy, which offers valu- able byproducts as well as giving difficulty in beneficia- tion, and (3) their irregular form, which requires assays to determine their boundaries. - The structural location of ore bodies near or above bio- tite granite stocks, and the alinement of deposits along the Rapid River thrust fault, are notable. The great extent of mineralized rock along the Rapid River fault also is exceptional. Studies of trace elements and fluid inclusions, and geo- logic mapping, show conclusively that the deposits are of hydrothermal origin, the fluorine, beryllium, boron, and other metallic elements being derived from deeper portions of the exposed granites in and around which the tin deposits are localized. OTHER DEPOSITS In addition to its occurrence in the main lodes near Lost River, fluorine occurs in fluorite, topaz or other fluorine- bearing minerals (idocrase, fluorine-micas) near other tin deposits on the Seward Peninsula, in breccia pipes cemented by fluorite in the Kigluaik Mountains, in scattered deposits of poorly assessed potential where alkalic rocks of Cretaceous age intrude carbonate rocks in the Darby Mountains, eastern Seward Peninsula, and in complex replacement bodies in interstratified marble and chloritic schist of Precambrian age north of Nome (Eakin, 1915; Brobst and others, 1971). Together, the deposits may delineate a fluorine-rich province genetically associated with silicic igneous intrusives of Cretaceous to Tertiary(?) age. Workers in this general area should at least be aware of the possibility of undiscovered commercial deposits of fluorspar. Fluorite occurs in tin deposits near granite stocks at Cape Mountain, Brooks Mountain, Black Mountain, and Ear Mountain. None of these deposits is large enough to be of commercial significance. At Black Mountain, topaz is abundant in a small part of the granite that was greisenized. Similar greisenized granite at Lost River, which contains up to 28 percent topaz, contains almost 8 percent fluorine, equivalent to about 16 percent CaF;. In the Kigluaik Mountains, about 40 miles (65 km) north of Nome, fluorite cements breccia in two pipes along a tributary of the Tisuk River. The pipes consist of angular jasperoid fragments cemented by fluorite. The largest measures at least 80 feet (25 m) long and 20 feet (6 m) wide. Tourmalinized granite crops out nearby and else- where in the vicinity, suggesting that other pipes may exist. Fluorite has long been known at scattered localities in a limestone area about 20 miles (32 km) north of Nome (Eakin, 1915; Herreid, 1966). Trenching has disclosed that fluorite accompanies barite in a brecciated and dolo- mitized marble intercalated in chloritic schist of the Pre- cambrian Nome Group. Anomalous amounts of lead, zinc, silver, and gold accompany the fluorite and barite. Because other marble beds probably occur beneath the one exposed by trenching, fluorite may be found at greater depths and may be of economic value (Brobst and others, 1971). Moreover, fluorite may be found beneath some of the many gossans known elsewhere in this general region (Eakin, 1915). In the eastern part of the Seward Peninsula, an exten- sive belt of granitic rocks is accompanied by a varied suite of alkalic rocks (Miller, 1971). Fluorite has been noted at many localities in this general region (Vivian Velette, oral commun., 1968), for instance, in a fine-grained intrusive at the southeast end of the Bendeleben Mountains. According to Thomas P. Miller (oral commun., 1970), fluorite in small amounts is widespread where alkalic rocks intrude limestones. According to A. A. Beus (oral commun., 1967), large fluorite deposits associated with alkalic rocks are known in nearby Siberia. Hence, the geo- logically similar east part of the Seward Peninsula is shown as a possible fluorine-rich province which may contain undiscovered fluorite deposits. Fluorite in minor amounts is known at many places in Alaska outside the Seward Peninsula. The most signifi- cant are shown on the map by Worl, Van Alstine, and Heyl (1974); in 1973 none were economically important. At Bokan Mountain, southeastern Alaska, a peralkalic granite and the adjoining rocks contain small amounts of 34 fluorite associated with uranium ores, but the individual deposits and the total amount of fluorite are small (E. M. MacKevett, Jr., oral commun., 1971). In Glacier and Groundhog Basins, southeastern Alaska, where substantial deposits of lead and zinc are known, sulfide-bearing veins have a gangue of quartz and fluorite, and a large breccia vein as much as 40 feet (12 m) thick con- tains notable fluorite, as do other quartz veins nearby (Gault and others, 1953, p. 23). Large breccia pipes as much as 400 feet (120 m) in diameter are known in the area. They contain some fluorite and topaz, but have not been evaluated for fluorite at depth. Fluorite in float or bedrock occurs sporadically across central and eastern Alaska from the Eagle area to west- central Alaska, along a belt which contains placer tin associated with granitic rocks of Late Cretaceous to Tertiary age. No deposits of economic promise are known. If lode tin deposits are found fluorite and (or) topaz in larger amounts are likely to be associated. In the eastern Brooks Range, near Mount Michelson in northern Alaska, fluorite occurs near and in a granitic batholith. Veins and veinlets are associated with altered rocks containing greisen or tourmaline. On Zarembo Island in southeastern Alaska, volcanic rocks of Tertiary age contain narrow veins and breccia zones which have chalcedony covered by fluorite. One breccia zone ranges from a few inches to several feet in thickness, with breccia fragments coated with chalcedony on which fluorite was deposited as incrustations (Buddington, 1923). This is the only fluorite lode in Ter- tiary volcanics reported in Alaska. Fluorine resources in Alaska in minerals other than fluorite consist principally of fluorine in topaz greisen, and in phosphate rock. The known deposits of size are at Lost River, on the Seward Peninsula, where topaz greisen forms a mass in the buried granite that measures several hundred feet across and extends vertically several hundred feet deep. However, much of this greisen was argillized, and argillization destroyed the topaz without fixing all the fluorine in late-stage fluorite; the general fluorine content probably averages not more than 3-4 percent. The granite at Tin Creek, about 2 miles southeast of the Lost River mine, also is greinsenized over an area of several thousand square feet on the northeast end; the topaz con- tent is unknown, though considerable. The phosphatic limestones of northern Alaska were dis- cussed by Patton and Matzko (1959), who concluded that high-grade phosphate rock (35.8 percent P;,0,;) occurs in the Shublik Formation of Triassic age, and lower grade phosphate rock is present in the Lisburne Group of late Paleozoic age. Fluorine is present in the phosphate rock in carbonate fluorapatite and also in fluorite. As both these units crop out for a hundred miles along the Arctic foot- hills of the Brooks Range, they unquestionably contain large resources of fluorine as well as of phosphate. GEOLOGY AND RESOURCES OF FLUORINE WASHINGTON-IDAHO-MONTANA-WYOMING- soOUTH DAKOTA By R. L. ParkER Fluorite is widely distributed in the Washington, Idaho, Montana, Wyoming, and South Dakota region; only a few deposits, however, are known to be of sufficient size and grade to be of commercial interest, and known resources are not large. The only deposits in the region with signifi- cant recorded production are the Challis and Meyers Cove deposits, Idaho, and the Snowbird, and Crystal Mountain deposits, Montana. Of these deposits only the Crystal Mountain has been a consistent producer. Fluorite deposits in the region can be grouped into several types. The most important deposits are associated either with volcanic rocks or with intrusive rocks of the Colville, Idaho, and Boulder batholiths and alkalic intrusive bodies. In addition to different volcanic and intrusive associations, the fluorite deposits in the region have different structural characteristics. Fillings of fissures and breccias by fluorite and chalcedony with vuggy, drusy, and layered features common to epithermal deposits are characteristic of fluorite deposits related to large volcanic piles in Washington and Idaho. One of the most preva- lent types of occurrence is related to the contact between intrusive and sedimentary rocks, especially between granitic and alkalic intrusive rocks and limestone or other calcareous rocks. Fluorite at such deposits is concentrated in veins or breccia zones in both intrusive and country rocks, is disseminated in the intrusive rocks, or is dissem- inated or concentrated in sedimentary rocks along the con- tact. Some deposits are actually in the contact meta- morphic aureole of the intrusives and are associated with tactite or skarn. Other deposits are in pegmatites that are in genetically related intrusive rocks or in country rocks. Certain occurrences of fluorite are obscure in their origin, and although they are found as gangue in vein deposits of metallic minerals, they may be unrelated except for the site of localization. Miscellaneous occurrences include fluorite as gangue in uranium deposits, and at least one fluorite deposit that is considered to be of evaporite origin. Deposits and occurrences of fluorite are described according to State in the order: Washington, Idaho, Mon- tana, Wyoming, and South Dakota. DESCRIPTIONS OF INDIVIDUAL DISTRICTS AND DEPOSITS WASHINGTON Fluorite occurs in a gold-silver-selenium deposit in the Republic district, Ferry County. The fluorite is locally present in chalcedony-calcite veins that are contained in Tertiary volcanics (Bancroft, 1914; Van Alstine, 1966). Fluorite is contained in quartz-calcite-chalcopyrite veins in the Zalla M, American Flag, and Silver Bell deposits in the Sheridan area, Ferry County. Some zones contain as much as 30 percent fluorite, but their extent is not known. The deposits are in Tertiary volcanics (Valen- tine, 1949; Van Alstine, 1966). WASHINGTON-IDAHO-MONTANA-WYOMING-SOUTH DAKOTA Fluorite is reported at one place in the upper part of the Ethel vein, in Snohomish County. The fluorite occurs with quartz, calcite, and bornite in granodiorite (Weaver, 1912; Van Alstine, 1966). Fluorite is contained in veins 2-12 inches (5-80 cm) thick in the Three Pinnacles-Lost River deposits, in Okanogon County. The fluorite is associated with chalcedony in the veins which cut granite (Valentine, 1949; Van Alstine, 1966). Seven parallel veins of fluorite and chalcedony occur at the Slide deposit, in Chelan County. The veins range in thickness from less than 1 to 6 inches (2.5-15 cm) and are traceable for only a few feet in a zone 300 feet (90 m) across in granite (Valentine, 1949; Van Alstine, 1966). Fluorite is a gangue mineral in a silver ore vein that cuts granite of the Colville batholith at the Apache claim, in Okanogan County. The vein is a quartz vein irregularly banded with pyrite, fluorite, and an unidentified black sulfide mineral (Valentine, 1949; Pardee, 1918; Van Alstine, 1966). Fluorite is found in veins and shear zones in Ramore, Gladora, and Wasco lead-silver-zinc-copper deposits in the Park City district, in Okanogan County. The fluorite, in minor amounts, is associated with quartz, metallic sul- fide minerals, calcite, and sericite in deposits in quartz monzonite of the Colville batholith (Pardee, 1918; Van Alstine, 1966). The Mitchem fluorspar deposit is in Ferry County. It is the only productive deposit in Washington, having produced a small tonnage of acid grade and metallurgical fluorspar as follows: 60 tons in 1918, 152 tons in 1940-42, and 170 tons in 1945-46. The fluorspar is contained in a 16-inch (40-cm) steeply dipping northwest-trending vein in granitic rock of the Colville batholith. The vein con- tains green, purple, and colorless fluorite, calcite, chalcedony, and some pyrite. Across mining widths, the material averages about 50 percent CaF,. The deposit is considered to be small (Van Alstine, 1966) and not to con- stitute a large reserve on the basis of available data. Thin veinlets of fluorite occur in limy argillite at the Montgomery deposits, in Okanogan County (Valentine, 1949). Fluorite is a minor gangue mineral associated with quartz, ferberite, scheelite, pyrite, molybdenite, and chlorite in tungsten veins at the Germania mine, in Stevens County. The veins are mostly less than 3 feet (1 m) thick and are contained in Cretaceous quartz monzonite (Becraft and Weis, 1963; Van Alstine, 1966). Fluorite occurs as a gangue mineral with wollastonite, tremolite, and quartz in the magnetite ore body at the Read iron deposit, in Stevens County. Mineralized rock is local- ized along a granite-limestone contact and fluorite is in a contact-metamorphic zone. Associated ore minerals are scheelite, chalcopyrite, and magnetite (Valentine, 1949; Van Alctine 106A\ 95 Fluorite occurs in molybdenite-bearing quartz veins and as an accessory mineral in adjacent granite expecially near schist and quartzite contacts at the Phalen Lake molybdenum deposit, in Stevens County (Valentine, 1949; Van Alstine, 1966). Fluorite is a minor constituent in the quartz veins of the Northport district, Stevens County. The quartz veins bear sulfide minerals and are contained in argillite and lime- stone cut by granite and in black slate of Paleozoic age (Van Alstine, 1966). Fluorite is reported in small lenses in gneiss and schist about 5 miles (8 km) east of Riverside, in Okanogan County (Valentine, 1949; Van Alstine, 1966). Fluorite at the Electric City deposit occurs in pegmatite lenses in granite of the Colville batholith, in Douglas County. Associated minerals are quartz, molybdenite, and pyrite (Van Alstine, 1966). Fluorite is contained in pegmatite in granite of the Col- ville batholith at the Sanpoil uranium deposit, in Okano- gan County. The fluorite is violet, radioactive, and associated with samarskite (Van Alstine, 1966). Fluorite occurs as traces in gray massive quartz segrega- tions in muscovite-rich pegmatites that cut fine-grained monzonite of the Loon Lake Granite at the Spokane molybdenum deposit, in Lincoln County (Becraft and Weis, 1963). IDAHO MEYERS COVE The Meyers Cove fluorspar district, Idaho (Anderson, 1954a; Cox, 1954), is about 38 miles (61 km) (66 miles by road) southwest of Salmon, and 26 miles (42 km) north- west of Challis, Lemhi County, on Camas Creek, a tribu- tary of the Middle Fork of the Salmon River. The deposits produced about 37,432 tons of fluorspar during 1951-53, principally from the Big Lead deposit. Other deposits in the district are North Pit, Anderson, Bear Trap, M and M, and Chamac. The fluorspar deposits are fissure fillings along northeast-trending zones which Anderson (1954a) attributed to a deep-seated shear in basement rocks. The deposits are located in Challis Volcanics (Tertiary) and Miocene porphyries, but these host rocks apparently had little effect on localization of the deposits. The deposits are exposed over a vertical range of more than 2,000 feet (610 m). The ore bodies vary widely in size, ranging from a few inches in width and a few tens of feet in length to more than a dozen feet in width and several hundreds of feet in length. They average 4-5 feet (1.2-1.5 m) in thickness with thicker local bulges and are 100-300 feet (30-90 m) long. Big Lead has a maximum stope width of 16 feet (5 m) and stope length of 300 feet (90 m). Ore bodies tend to swell where the orientation of fissures changes from northeaet tn 36 GEOLOGY AND RESOURCES OF FLUORINE The mineral assemblage of the deposits is simple, con- sisting chiefly of chalcedony, fluorite, and some barite. The ore is layered (commonly called "banded"), and chalcedony and cryptocrystalline fluorite form concentric shells on breccia fragments and form crystalline druses. The wallrocks along fluorite-bearing zones are commonly bleached and (or) stained with iron oxides. The bleaching resulted from argillic alteration, with limited sericitic alteration close to fluorite bodies. Some of the iron staining is attributed to the oxidation and decomposition of pyrite. In places, silicified rocks are closely associated with the fluorite bodies. The Meyers Cove deposits are probably epithermal as indicated by the low-temperature layered chalcedony and the open-brecciated, vuggy and drusy nature of the ore which suggests near-surface deposition. The fluorite bodies were deposited in fissures along northeast-trending zones probably from hydrothermal solutions. These deposits have high potential as a future resource of fluorspar. BAYHORSE The Challis fluorspar deposits (Bayhorse district), Custer County (Anderson, 1954b), are located in the northern part of the district about 4 miles (6.5 km) south- west of Challis at the south end of the Salmon River Mountains. Fluorite was not recognized as a gangue mineral in the silver base-metal lodes of the Bayhorse district until the early 1940's. Actual location of claims for fluorspar dates from 1947, and exploration, development, and small pro- duction was not until 1952-53 in the Keystone Mountain- Daugherty Gulch area. Production from the Chalspar No. 1 deposit amounted to 475 tons of metallurgical-grade fluorspar and 245 tons of lower grade material. The Challis fluorspar deposits are fissure fillings and breccia impregnations in which abundant open spaces and fluorite-lined druses are characteristic. All deposits reported are in the Bayhorse Dolomite of Ordovician(?) age, which forms a favorable host because it is more brecciated than adjacent formations (Anderson, 1954b, p. 7). The major structural feature of the area is the Bayhorse anticline that exposes lower Paleozoic rocks comprising several formations: Garden Creek Phyllite, Bayhorse Dolomite, Ramshorn Slate, and Kinnikinic Quartzite. Fissure veins and breccia lodes are localized along faults and relatively minor fault zones wholly within the Bay- horse Dolomite. Most veins have northerly trends and steep westerly dips. Fluorspar bodies reach widths of several tens of feet and lengths of several hundred feet. The Passed Up vein, for example, is 3-7 feet (1-2 m) thick and is exposed for more than 200 feet (60 m) along strike. The vein on Chalspar No. 2 claim is 12-18 feet (3.7-5.5 m) thick and is exposed for 900 feet (275 m). Some breccia deposits range in diameter from a few feet to 30 feet (9 m) and can be traced for a hundred feet or more at the surface. The deposits consist chiefly of fluorite with minor amounts of quartz and calcite. In some veins local masses of sulfides are relict from earlier mineralization. Some breccias have boxwork structure in which quartz forms thin partitions between cells and is coated with fluorite. Layering ("'banding") is not conspicuous, but poorly defined concentric layers of fluorite enclose breccia fragments. The grade of the fluorspar is variable and most deposits require milling to upgrade the material sufficiently for commercial use. Samples across the Passed Up vein averaged 67.1 percent and 57.1 percent CaF, over widths of 13 and 14 feet (4 and 4.3 m), respectively. At the Chalspar No. 5 the vein ran 40 percent CaF, over a width of 17 feet (5.2m). The boxwork breccias at Keystone Mountain assayed 30-60 percent CaF, across 18-30 feet (5.5-9 m). The age of fluorspar mineralization is considered to be Oligocene or younger, because probable Oligocene vol- canic rocks form the hanging wall of the Passed Up vein. The district has large potential as a future source of fluorspar. Surface exposures and limited exploration indicate fluorspar bodies sufficiently large for mining. Although the grade is lower than some currently mined deposits, future prices may be sufficiently high to allow beneficiation and shipment to centers of consumption. STANLEY AREA Fluorite has been discovered in the Stanley area, Custer County (Choate, 1962), in the following localities: (1) in epithermal gold-silver veins in quartz monzonite- granodiorite near the mouths of Big and Little Casino Creeks about 3.5 miles (5.5 km) east of lower Stanley; (2) in fluorite veins in the Challis Volcanics at the Lamb pros- pect in the northern part of the Yankee Fork district 1 mile (1.6 km) west of Jordan Creek-Mayfield Creek divide; and (3) in a gold-silver vein in Challis Volcanics on Hindman Peak Ridge 1.25 miles (2 km) south of Red Mountain about 13 miles (20 km) north of lower Stanley (no de- scription is available). The known deposits containing fluorite are the Bright Star, the Giant Spar, the Home- stake, and the Gold Chance properties, and the Hide Out and the Lamb prospects. The only production of fluorspar recorded from any of these deposits was from the Giant Spar which reportedly yielded a few truckloads of fluorite during World War II. The Giant Spar is the largest known fluorite deposit in the Stanley area. It occurs as epithermal, open-space filling in faults and breccia zones in quartz monzonite- granodiorite. Other fluorite-quartz veins of the district have been explored for gold and silver, but these metals are not present in the fluorite of the Giant Spar. Diamond drilling has indicated a composite vein system that trends northeast and has a maximum thickness of about 15 feet (4.5 m). The fluorite vein is cut by a younger fault. A chip sample across a 10.6-foot (3.2-m) width of the vein con- tained 71.5 percent CaF,. Rhyolite and basalt dikes cut the WASHINGTON-IDAHO-MONTANA-WYOMING-SOUTH DAKOTA 87 quartz monzonite in the vicinity of the deposit, but the relation of the fluorite to the dikes was not stated by Choate (1962). Presumably the fluorite mineralization was related to Challis volcanism. The fluorite is pale lavender, colorless, and pale green, and forms "banded" veins. Vugs are common and some are lined with euhedral fluorite crystals. The Bright Star deposit contains a composite fluorite vein zone 3 feet (1 m) thick composed of individual veins of fluorite up to 6 inches (15 cm) thick and stringers of quartz. The vein zone contains 60-80 percent fluorite. The extent of the deposits is unknown because of the caved condition of the workings. The Homestake, Gold Chance, and Hide Out claims contain minor fluorite in gold-pyrite-quartz-fluorite veins. The Hide Out workings are anomalously radioactive. The Lamb prospect in the northern part of the Yankee Fork district contains fluorite in a series of veins, stringers, and breccia fillings along a northeast-trending zone that is at least a mile (1.6 km) long and several hundred yards wide. The zone cuts the Challis Volcanics. The largest vein in the zone is traceable for 600 feet (180 m) and ranges in thickness from a few inches to 4 feet (1.2 m). A sample across a 4-foot (1.2-m) vein width contained 39.4 percent CaF; and showed fluorite, calcite, chalcedony, quartz, and breccia; no metallic minerals were detected. Too little information exists on the Stanley fluorite deposits to assess their potential. Some veins are high enough in grade and are wide enough to merit investiga- tion of their vertical and longitudinal extent in order to determine the presence or absence of minable ore bodies. BIG SQUAW CREEK The Big Squaw Creek fluorspar deposit, previously known as the Smothers property or the Noussan mine, is located on the Salmon River between Big Squaw Creek and Smith Gulch in the Salmon River Breaks Primitive Area. It is 18 miles (29 km) downriver from the end of the road at Corn Creek. The description of the deposit presented here is based on Weis and others (1972). The deposit was prospected for gold as early as 1860, but was not claimed until 1937. Properties have passed through considerable reassignment and relocation to the present time. Exploration under a contract with Defense Minerals Exploration Administration was conducted in 1957-58 including geologic mapping, trenching, and diamond drilling as well as additional work by the owners. No production has been recorded from the claims although inferred reserves are considered to be about 100,000 tons of fluorspar containing 70 percent CaF;. The fluorspar deposit consists of three ore shoots on the hanging wall (or east side) of a persistent quartz vein that strikes N. 10° W. and dips 60° E. The vein is contained in biotite gneiss that is intruded by quartz monzonite about a mile (1.6 km) to the west. The vein is one of many quartz veins that occur in a zone trending N. 10°-15° W., and so far as known, contains the largest bodies of fluorite. The fluorite is present as small stringers and as crystals lining vugs in the quartz vein, and as ore shoots along the hanging wall of the vein. It is both massive and coarsely granular and ranges from transparent through green to purple. Three ore shoots have lengths of 600, 260, and 220 feet (180, 80, and 67 m) and average widths of 7.2, 7.9, and 7.1 feet (2.2, 2.4, and 2.2 m), respectively. Average grades of the shoots as determined from surface samples are 67.50, 73.93, and 70.50 percent CaF, respectively. Diamond drilling on the 600-foot-long shoot mentioned above indicates that it extends to a depth of at least 300 feet (90 m) but may decrease in grade in depth. The deposit is apparently related to the Idaho batholith because of its proximity thereto, and has certain features in common with the Crystal Mountain, Spar, and Snowbird deposits of Montana. Although 100,000 tons of ore with a grade of 70 percent CaF, is inferred for the deposit, it remains submarginal because of its remoteness. OTHER DEPOSITS A narrow vein of high-grade fluorite is reported to occur in the Pungo Creek area, Valley County (Anderson and Van Alstine, 1964). The vein is in porphyritic diorite cut by mafic dikes. Fluorite occurs in a northeast-trending vein at the Simer prospect, 6 miles (9.7 km) north-northwest of Salmon. The vein is composed mostly of radioactive jasperlike rock as much as 2-3 feet (1 m) thick. It cuts quartz monzonite and contains monazite, variable amounts of barite, and purple and colorless fluorite (Anderson, 1959, p. 89). Fluorite occurs in a number of gold lodes in the Edwardsburg district, Valley County, but is most abundant in upper Smith Creek at the Independence mine. The fluorite is disseminated in coarse grains in irregular quartz veins 3 feet (1 m) or more in width which occur at intervals in a silicified zone that is several thousand feet long and 200 feet (60 m) wide (Clyde P. Ross, in Burchard, 1933, p. 14). The deposit is in the Pre- cambrian Y Yellowjacket Formation that in the vicinity of the mine consists mainly of schistose argillites and quartz- ites and a layer of crystalline limestone at least a hundred feet (30 m) thick. The silicified rock contains scattered sul- fides: pyrite, sphalerite, tetrahedrite, and galena (Shenon and Ross, 1936, p. 28). Fluorite is a gangue mineral in tungsten-bearing quartz veins in quartzite in the Blue Wing district, Lemhi County. According to Callaghan and Lemmon (1941) it is prominent only in a zone extending 600 feet (180 m) from the granite contact on the intermediate level of the Ima mine where it bears a definite zonal relation to the granite. Associated minerals in the veins are orthoclase, rhodochrosite, huebnerite, pyrite, tetrahedrite, chalcopy- rite, bornite, galena, sphalerite, molybdenite, and scheelite. Some fluorite is also disseminated in the granite. 38 GEOLOGY AND RESOURCES OF FLUORINE Fluorite occurs as a minor accessory mineral associated with aquamarine disseminated in albitized quartz monzo- nite in certain parts of the Sawtooth batholith in central Idaho (Reid, 1963). MONTANA CRYSTAL MOUNTAIN The Crystal Mountain fluorspar deposit is located on the Rye Creek-Sleeping Child Creek divide, in the south- western part of Ravalli County. It is 26 miles (42 km) by road east of Darby and lies at an elevation of about 6,800 feet (2,100 m). The deposit consists of two separate groups of outcrops about 3,000 feet (900 m) apart. An eastern group, called the Retirement claims, consists of two small ellip- tical outcrops of fluorspar 60 feet (18 m) wide and 150 feet (45 m) long. The fluorspar bodies are more resistant to erosion than the enclosing rocks and hence form the crests of knolls (Taber, 1953). Sahinen (1962) suggested that the bodies are erosion remnants of a flat-lying vein and described the enclosing rock as granite and fine-grained gray gneiss with an overlying mass of white quartz in the hanging wall and veinlets of glassy quartz in the footwall. A larger western group, called the Lumberjack claims, is the site of an open pit mine. This group consists of three main bodies of fluorspar 100-200 feet (30-60 m) wide and 200-400 feet (60-120 m) long which are tabular and dip easterly 10°-35°. These bodies, according to Weis and others (1958), are underlain by coarse-grained biotite granite that contains inclusions of hornblende- plagioclase gneiss, biotite-quartz-plagioclase gneiss and pegmatitic granite. Foliation in the inclusions strikes about north and dips 20°-30° E. nearly parallel to the fluorspar bodies. Small dikes of granite cut both gneiss and fluorspar. The fluorspar at the Crystal Mountain deposit is excep- tionally high grade and remarkably uniform in quality. The average grade (Sahinen, 1962) is more than 96 percent CaF,. Weighted averages of nine samples taken by the U.S. Bureau of Mines (Taber, 1952) contained 97.2 percent | CaF,, 1.44 percent SiO,, and 0.13 percent Fe. The ore con- sists of massive crystalline fluorite and local minor amounts of biotite, quartz, feldspar, sphene, rare-earth- bearing apatite, amphibole, fergusonite, thorianite(?), and thortveitite (Parker and Havens, 1963). The fluorite ranges in crystal size from 0.1 inch to at least 1 inch (2.5 mm-2.5 cm) across and ranges in color from white through pale green to deep purple. The minor minerals are concen- trated in deep-purple fluorite aggregates that are radio- active and which occur mostly at the margins of the fluorspar bodies. Spectrographic analysis of run-of-the- mine fluorite shows the presence of minor aluminum, magnesium, barium, nickel, titanium, and niobium; analysis of the dark-purple fluorite indicates amounts of iron, potassium, aluminum, silicon, uranium, thorium, niobium, rare earths, yttrium, phosphorus, and scandium higher than the average mined fluorite. The Crystal Mountain deposit is distinguished by several characteristic features. The ore bodies are nearly pure fluorite and are sharply delineated from the enclosing host granite that is considered to be part of the Idaho batholith. Younger granite(?) dikes cut both fluorite and enclosing host granite. The host granite also contains ordinary quartz-feldspar pegmatites. Radioactive fluorite which contains an anomalous content of rare-earth elements, niobium, thorium, and scandium forms zones roughly bordering the fluorite bodies. Some minerals enclosed in radioactive fluorite are found commonly in pegmatites. Indeed, thortveitite has been found only in pegmatites. The origin of the Crystal Mountain deposit is not clear. The characteristics mentioned above point to a plutonic origin related to pegmatites. The deposits of fluorite might be considered to be monomineralic pegmatites. SNOWBIRD The Snowbird deposit, also known as the F and S fluorspar mine, is located on the Montana-Idaho Boundary at the head of Cedar Log Creek, just below the crestline of the Bitterroot Range in southwest Mineral County. The deposit, which is covered by 19 unpatented claims, was mined for fluorspar in 1956-57 by the F and S Mining Co., producing approximately 6,500 short tons of metallurgical-grade fluorspar. The ore reportedly has been exhausted (Sahinen, 1962). The Snowbird has certain features that suggest genetic similarities with the Crystal Mountain and Spar deposits. The deposit is a zoned quartz-calcite-fluorite body that Clabaugh and Sewell (1964) described as a carbonatite "pegmatite." It is discordantly intruded into the Precam- brian Wallace Formation of the Belt Supergroup which at this locality consists of argillite, argillaceous limestone, slate, and limy quartzite. These rocks are intruded by rocks of the Idaho batholith south of the deposit, and in the vicinity of the deposit they have been sufficiently thermally metamorphosed to produce metacrysts of scapo- lite in the argillaceous beds (Sahinen, 1962). The fluorspar deposit lies in a pegmatite pod that is at least 600 feet (180 m) long and 100 feet (30 m) wide and has an east-west long axis. The major part of the pod is massive white quartz that forms the hanging wall of the main fluorspar body. The quartz mass contains no inclusions, but euhedral crystals of quartz can be recognized protruding from the mass and some crystals are as much as 20 feet (6 m) long (Clabaugh and Sewell, 1964). The quartz forms an incomplete envelope around the cal- cite body, and fluorite is located in pockets in the calcite at the calcite-quartz contact. Two main bodies of fluorite were 120 feet (37 m) and 40 feet (12 m) long. The fluorspar assayed 96.2-96.5 percent CaF, (Sahinen, 1962). The deposit contains-in addition to quartz, calcite, and - fluorite-gersdorffite (NiAsS) and parisite [Ca(Ce,La,Y),(CO,),F,]. These rare minerals have not WASHINGTON-IDAHO-MONTANA-WYOMING-SOUTH DAKOTA 39 been noted in other fluorspar deposits in the region, but rare-earth elements, especially yttrium, are associated with fluorite at the Crystal Mountain deposit near Darby. The Snowbird deposit is of pegmatitic origin and is probably related to the Idaho batholith. SPAR The Spar mine is located on the ridge between Dry Creek and Bear Creek, in central Mineral County about 12 miles (19 km) west of Superior (Ross, 1950; Sahinen, 1962). The Spar claim was originally located in 1943 by Mr. Joseph Brooks and was developed in 1944, during which time an unknown amount of ore was mined. Additional fluorspar deposits were discovered nearby in 1948 in the Wilson Creek drainage south of Dry Creek and in 1949 in nearby Lime Gulch. The Spar and Wilson Creek proper- ties, under control of various subsequent operators, to- gether produced 318 tons of fluorspar in 1948, 422 tons in 1949, and 41 tons in 1950 (last reported production). Individual property production is not known. The Spar deposit is localized in the Wallace Formation of the Precambrian Belt Supergroup which in the area consists of quartzite, argillite, calcareous quartzite, dolo- mite, and ferruginous limestone; most of the rocks strike northwest. The rocks in the vicinity of the deposit have undergone moderately intense and somewhat irregular folding, and locally contain steep, sheeted zones and zones of intense brecciation. Conspicuous lenticular milky- white quartz bodies cut the Wallace Formation, and at least one of them contains fluorite (Spar mine). In the general area of Ann Arbor and Bear Creeks quartz lenses exceed 100 feet (30 m) in length, and some crop out along a nearly straight line. The lenses are discordant and show no parallelism or obvious relation to the trends of the beds of the Wallace Formation that they cut. The larger lenses of quartz trend N. 70° E. to S. 70° E. and have northwest- trending joints. Igneous rocks are not exposed in the area. The Idaho batholith which lies about 35 miles (55 km) to the south and a diabase sill about 10 miles (16 km) to the west are the closest known occurrences of igneous rocks. The fluorspar at the Spar mine occurs in irregular pods and lenses ranging in diameter from several inches to several feet. A sample cut by Clyde P. Ross (1950, p. 208) across a 45-inch (1.5-m) face in the main pit at the Spar property assayed 98.57 percent CaF,, 1.18 percent SiO,, and 0.14 percent carbonate. The fluorite at the Spar property occurs with calcite and ankerite and minor galena, tetrahedrite, and pyrite. A large body of milky-white quartz 30 feet (9 m) wide and 150 feet (45 m) long forms the hanging wall. Some large crys- tals of quartz a foot or more across are exposed in the Spar prospect. The Spar fluorspar deposit resembles the Snowbird deposit in some respects. Both deposits lie in the cal- careous Wallace Formation of the Belt Supergroup and both are associated with calcite and minor sulfides and with massive milky-white quartz bodies containing large quartz crystals suggestive of pegmatitic orgin. Neither deposit can be related directly to visible igneous rocks although the Snowbird deposit is close to exposed rocks reported as part of the Idaho batholith. OTHER DEPOSITS The Jetty mine, also known as the Balkan lode, is on the north slope of Garrity Hill, 6 miles (10 km) west of Ana- conda, Deer Lodge County. Fluorite occurs as a gangue mineral in a metalliferous vein at an intermediate level in the mine. The vein is composed of quartz, fluorite, barite, and the metallic sulfides-boulangerite, pyrite, sphalerite, tetrahedrite, and stibnite. Host rock of the deposit is Madison Limestone (Mississippian) and Quadrant Quartzite (Pennsylvanian). Intrusive rocks of Boulder batholith age occur in the region. The extent of fluorite in the mine is not known but the mineral might be recovered as a byproduct in the milling of the sulfide ore (Sahinen, 1962). The Weathervane Hill prospect lies south of the Smelter stack at Anaconda, in Deer Lodge County. Fluorite occurs in a fault zone between upper Paleozoic limestone and upper Mesozoic strata. The fluorite is distributed irreg- ularly in pockets a few inches to a few feet in maximum dimension. The potential for commercial fluorspar is low for anything but very small-scale mining (Sahinen, 1962). The Silver Bow deposits are 6 miles (10 km) west of Butte along Silver Bow Creek, in Silver Bow County. Interstate Highway 90 (U.S. Highway 10) and four railroads cross the property. The Bull Moose, Merrimac, and Wrong Font are claims in this group of deposits. Fluorite occurs in shoots within a gold-silver(?)-bearing fracture zone along a fault(?) contact between andesite and quartz monzonite and aplite. The mineralized zone extends more than 2,000 feet (600 m) in a northwest direction and in places is more than 100 feet (30 m) wide. The rocks in the zone are silici- fied and stained with iron oxides. The principal bodies of fluorspar are on the east end of the Bull Moose claim where layers of fluorite range from an inch (2.5 cm) to 3.5 feet (1 m) in width and are tens of feet in length. An average of six samples gave 83.84 percent CaF,, 13.02 percent SiO,, 2.02 percent R;0,, and 1.40 percent CaCO, (Sahinen, 1962; Ross, 1950). Exploration has been insufficient to define the extent of fluorspar at the deposits; the grade and size of bodies reported, however, would indicate potential for small-scale mining. Actual operations might be hampered by the highway right-of-way 'that runs through the properties. The Bald Butte mine lies within the Marysville mining district, 17 miles (22 km) northeast of Helena, Lewis and Clark County, and 3.5 miles (5.5 km) southwest of the old mining camp of Marysville. Fluorite occurs with quartz as the gangue of metallic ore that was originally developed and mined by underground workings that are no longer 40 GEOLOGY AND RESOURCES OF FLUORINE accessible. The deposit is in limestone of the Helena Formation of the Precambrian Belt Supergroup which has been intruded and hornfelsed by dikes and sheets of micro- diorite and later cut by diorite porphyry dikes of Belmont Porphyry. All these rocks are shattered, and infiltrated with quartz and fluorite. The deposit may be related to a porphyry of middle Tertiary age. Sahinen (1962) con- cluded from the examination of dump specimens and Barrell's (1907) description of the deposit that no fluorite ore could be marketed from the deposit without extensive beneficiation. The Boeing prospect is located about 3.5 miles (5.5 km) west of Austin in the Austin mining district, Lewis and Clark County. Fluorspar is contained in small irregular pods in the Madison Limestone adjacent to an altered and mineralized biotite rhyolite dike. The limestone is intruded by quartz monzonite a short distance south of the deposit. The principal fluorite body was 3-6 feet (1-2 m) wide and exposed over a height of 7 feet (2.1 m). A sample taken near the surface across a 41l-inch (1.1-m) width yielded 81.76 percent CaF,, 10.09 percent SiO, and 2.52 percent carbonates (Ross, 1950). The extent of the fluorspar has not been determined, but reports indicate that fluorite bodies are small and irregularly distributed. In the Sweetgrass Hills in northwestern Liberty County, just south of the Canada-United States boundary the principal fluorspar bodies are in Tootsie Creek (Ross, 1950; Sahinen, 1962); they consist of veins, replacement seams, and minor disseminations in or adjacent to marble- ized Madison Limestone near contacts with porphyritic syenite of the laccoliths and stocks of the Sweetgrass Hills. Most of the fluorspar masses closely follow bedding or sheeting in the limestone. Some masses are irregular, some nearly flat, and others roughly spherical. Some mineral- ized zones attain widths of 10 feet (3 m) and some can be followed continuously along strike for as much as 50 feet (15 m). Most individual bodies are much smaller and are scattered in the Tootsie Creek area through a zone in lime- stone that is well over 2,000 feet (600 m) long and at most is several hundred feet wide. The fluorite is intimately inter- grown with other minerals; a 200-mesh grind is required to release the fluorite and obtain a 95-percent CaF, con- centrate. The CaF, content of the mineralized material ranges from 3.06 to 87.01 percent averaging somewhat less than 40 percent in bodies of minable size. The silica content ranges from 1.68 to 43.83 percent. The area has potential for minable fluorspar bodies, but exploration has been insufficient to define the extent of fluorite miner- alization. Minor fluorite is reported in the South Moccasin Mountains, a satellite uplift of the Judith Mountains intrusive center (Miller, 1959; Sahinen, 1962). The fluorite occurs as a gangue mineral in a gold-silver vein, in a dickite clay body, and in an area of breccia and altered rock adjacent to a large fault. All the deposits are related to alkalic rocks-porphyritic syenite and leucorhyol- ite-which intrude the Madison Limestone. Fluorite is also reported in at least one zone within the intrusive rock. At the American Flag prospect near the summit of Judith Peak, Fergus County, fluorite and quartz form a 10- foot (3-m) vein that cuts an altered zone in a syenite intrusive in the Judith Mountains alkalic intrusive center (Sahinen, 1962). The vein also contains lead, silver, and gold. Another nearby occurrence of fluorite is in fine seams and disseminated grains in a 5-foot (1.5-m) tinguaite dike. The fluorite-quartz vein contains 22.8-34.0 percent CaF,, 52.8-66.9 percent SiO,, 2-4 percent lead, 0.1-0.5 ounce silver per ton, and a trace of gold. Other, per- haps major, fluorspar deposits are veins and replacement seams in Madison Limestone in and near the mine. The property was being developed in 1971 (R. G. Worl, oral commun., 1971). Fluorite at Antone Peak in the Little Rocky Mountains, in Phillips County (Sahinen, 1962), is sparsely disseminated in porphyritic syenite that intrudes sedi- mentary rocks of varied lithology ranging in age from Middle Cambrian to Late Cretaceous. An occurrence of fluorite is reported in the Pryor Mountains at the Old Glory mine, 12 miles (19 km) east of Warren, Carbon County (Sahinen, 1962). Fluorite is associated with uranium minerals in the Madison Lime- stone. The fluorite occurs as ovules and spheroids that have concentric structure and are considered to be replace- ments of oolitic structures in the limestone. WYOMING Several deposits of fluorite in the Bear Lodge Mountains about 10 miles (16 km) north of Sundance occur in a north-trending zone about 0.5 mile (0.8 km) wide and about 3 miles (5 km) long. Production of fluorspar from the district has been less than a ton. Fluorite is localized in limestone and sandstone that are intruded by Tertiary hypabyssal alkalic rocks (phonolites and trachyte por- phyries). Bodies of fluorite for the most part are only a few feet long and are in the form of lenses parallel to bedding planes. Some fluorite is disseminated in the Minnelusa Sandstone and some is found in veins in a breccia zone in the intrusive rocks. The fluorite is associated with calcite, quartz, siderite, and orthoclase. CaF, content of the fluorite lenses ranges from 5 to 90 percent (Osterwald and others, 1959, p. 69). The deposits are similar in some respects to the central Montana occurrences in which alkalic and carbonate rocks are the hosts of the deposits. Fluorite is reported in a pegmatite in granite at the Bear deposit, in Laramie County. Twenty tons of fluorite was stockpiled in 1944 as a byproduct of feldspar mining (Osterwald and others, 1959). In the northern part of the Bighorn Basin a carbonate rock sequence of Permian age is reported by De Koster (1960, p. 53-55) to contain 12-15 percent fluorite. This fluorite is considered to be of primary sedimentary origin NEVADA-UTAH-OREGON-CALIFORNIA-ARIZONA 41 as a precipitate from saline waters at a stage in evapora- tion shortly preceding that at which gypsum is formed (Gulbrandsen and Reeser, 1969). The extent of the Big- horn Basin fluorite occurrence is not known, but the large size of many marine evaporite deposits would suggest the possibility of large low-grade new-type fluorite deposits. SOUTH DAKOTA Fluorite in minor amounts has been found in drill core from the Minnelusa Formation in the Pass Creek area, in Custer County. The mineral has also been identified in minor quantities in the Minnekahta Limestone about 25 miles (40 km) southwest of Custer (Roberts and Rapp, 1965). Fluorite is reported as a gangue mineral in several gold mines of the northern Black Hills: Ross Hannibal, Ulster, Hardscrabble, Pennsylvania, Cutting, Boscobel, Double Standard, Hidden Treasure, and Homestake mines, Lawrence County. It occurs with autunite and limonite at the Davier mine (Roberts and Rapp, 1965). Some of the fluorite may be of Tertiary age and related to volcanism, as it impregnates rhyolite at the Hardscrabble mine. NEVADA-WESTERN UTAH-SOUTHEASTERN OREGON-SOUTHEASTERN CALIFORNIA-WESTERN ARIZONA By R. G. Wort and W. R. Grirritts The region made up of Nevada, western Utah, south- eastern Oregon, southeastern California, and western Arizona contains numerous fluorspar deposits and fluorite occurrences. Most of the deposits are small, undeveloped, and poorly described, as compared to the numerous well-documented fluorspar deposits, many with substantial production, in Colorado, New Mexico, and trans-Pecos Texas. Nevertheless, this region contains three major fluorspar districts: Spor Mountain, Utah; Fluorine, Nev.; and Broken Hills, Nev., and one of inter- mediate size: Indian Peak Range (Cougar Spar and other mines), Utah (localities shown by Worl and others, 1974). Most fluorspar deposits in this region have been dis- covered and developed since 1935. Burchard in 1983 (p. 24) listed only four important fluorspar deposits in this region: the Castle Dome district, Arizona, the Fluorine dis- trict, Nevada, the Afton Canyon deposits, California, and the Silver Queen mine, Utah. Lindgren (1983b, p. 164) listed several fluorite gangue occurrences in this region, but very few of the now-known fluorspar deposits. In con- trast, most of the known fluorspar districts in Colorado and New Mexico were listed by Lindgren (1933b, p. 165) or Burchard (1933, p. 24), suggesting that Colorado and New Mexico were more intensely prospected for fluorspar in the early days than were the States in this region. Fluorspar in this region occurs in veins, mantos, and pipelike bodies in and associated with volcanic rocks, in contact zones and stockworks of hypabyssal and plutonic intrusive rocks, in veins and mantos of indefinite associa- tion, in volcaniclastic sediments, and in pegmatites. Other potential fluorine resources are topaz-bearing volcanic rocks, and fluorapatite in replacement iron ore bodies. Most fluorspar deposits in western Arizona and south- eastern California are layered ("banded") and crustiform veins composed of fluorite, barite, black calcite, calcite, manganese oxides, and in places argentiferous galena, pyrite, and gold-bearing quartz. Fluorite is uncommonly the major mineral present. Most of these deposits are well- defined veins in nearly vertical fracture systems, although in the Bouse district, in the Trigo Mountains, and in the Harquahala Mountains, Ariz., fluorite, barite, calcite, and black calcite occur as breccia cement and space filling in small concordant tabular bodies in volcanic agglomerates and flows (Van Alstine and Moore, 1969). Most of the fluorspar veins are in rhyolitic to andesitic volcanic rocks, pelitic metamorphic rocks, and granitic plutonic rocks. Several have had small fluorspar production, less than 10,000 tons, whereas many have had considerable lead, silver, gold, and manganese-oxide production. Examples of deposits with appreciable metal and manganese oxide production are those in the Vulture Mountains, Harquahala Mountains (Snowball, etc. deposits), Trigo Mountains, and Castle Dome district, Arizona, and the Orocopia, Afton Canyon, Red Bluff, and Warm Springs deposits, California. The Castle Dome district, Yuma County, Ariz., is described in more detail in later pages. In Nevada, Utah, and the Clark and Ivanpah Mountains, Calif., fluorspar veins and manto deposits are composed of fluorite, calcite, barite, quartz, and chalcedony and in a few places base- and precious-metal minerals. Layering and crustification are common and fluorite is generally the major mineral present. Most of these deposits are in volcanic rocks, rhyolite to andesite in composition, and in carbonate sedimentary rocks. Some are in granitic plutonic rocks. Many of the deposits are believed related genetically to Tertiary volcanism although not all are layered-vein types occurring in volcanic rocks such as at the Baxter (Kaiser) mine in the Broken Hills district, Nevada, to be described in later pages. Other examples of volcanic-related deposits, lying in Horton's (1961, p. 8) "western epi- thermal fluorspar belt" in Nevada, occur in Pershing, Churchill, Mineral, and Nye Counties. Similar deposits elsewhere in the region are in the Indian Peak Range (Cougar Spar and other mines), Utah, and Wah Wah Mountains, Utah. Some deposits are similar to volcanic related types but lack evidence of volcanic connections; these occur in eastern Nevada and in the Clark and Ivanpah Mountains, Calif. Several major structural zones appear to have localized fluorspar deposits in the region. The "western epithermal 42 fluorspar belt" of Horton .(1961) is a vaguely defined structural zone. Its northerly trend, about 50 miles (80 km) wide and 200 miles (320 km) long, through western Nevada is defined partly by the dominant northerly orien- tation of ranges in this part of the Basin-Range province. Other significant structures appear to have been more important in localizing districts, however, and these extend out of the western belt and have controlled local- ization of other districts in central and southeastern Nevada. For example, the Boulder Hill mine and Mount Montgomery Pass district in western Nevada and the Fluorine district and Nipton deposit in southern Nevada lie within the prominent northwesterly oriented Walker Lane structural zone. Similarly, farther north, the Broken Hills district, Union district, Quinn Canyon Range district, Lincoln mine, and Wells Cargo deposit lie along a northwesterly structural zone at the northeast edge of the Walker Lane. Deposits in eastern Nevada and western Utah are largely within two prominent northeast- trending structural belts, one extending from White Pine County, Nev., to the vicinity of Park City, Utah, and the other from Lincoln County, Nev., to the vicinity of Marys- vale, Utah. In western Arizona and southeastern California, fluorite deposits are largely restricted to a northwesterly structural zone that terminates near Phoenix. This area is conspicuous on the tectonic map of the United States because of the shortness and orientation of its northeast- trending mountain ranges, which contrast with the longer ranges having rather consistent northerly or north westerly trends. Part of this aberrant structural grain probably resulted from the abundant, very young, north- easterly faults with right-lateral movement, of which the Garlock fault of California is an example, that may post- date the fluorspar mineralization. Part of the structure certainly is older than these faults and is part of the set of northwest-trending lineaments that are found through much of the Western Cordillera. This northwest-trending set of structures is perhaps related to the Texas lineament that is parallel to it and about 100 miles (160 km) north of it; the Texas lineament weakens northwestward as the southwestern Arizona lineament strengthens. The south- western Arizona lineament may curve southward as it weakens, merging into north-trending structures that pre- vail near the Arizona-New Mexico State line. Fluorspar occurs in carbonate rocks throughout the "western epithermal fluorspar belt" of Nevada (Horton, 1961, p. 6), in the Union district, Nevada, at several localities in White Pine County, Nev., at the Wells Cargo mine, Lincoln County, Nev., in the Clark Mountains, Calif., in the Star district, Utah, and in minor amounts in carbonate rocks at numerous other localities in these States. At most of these localities only veins have been recognized; replacement bodies as mantos along poorly defined fluorspar veins in carbonate sedimentary rocks GEOLOGY AND RESOURCES OF FLUORINE have been noted at the Wells Cargo mine, in the Clark Mountains (Crosby and Hoffman, 1951, p. 630), and in the Quinn Canyon Range, Nev. Fault-controlled replace- ment bodies in limestone are present in the Daisy mine, Fluorine district, Nevada (Cornwall and Kleinhampl, 1961, p. J20). The fluorspar deposits in the Quinn Can- yon district, described below, are spatially related to intru- sive rocks. Areas underlain by carbonate rocks containing either fluorspar veins or intrusives with high fluorine con- tent should be excellent prospecting targets for large re- placement bodies. Fluorspar replacement bodies may not be readily noticeable to the naked eye. For example, fine- grained fluorspar in replacement bodies at the Wells Cargo mine is similar in texture and color to unaltered limestone, and at some locations in the Clark Mountains, Calif., abundant sericite tends to mask the fluorite. Fluorspar veins and mantos in this region are commonly in, or related to, Basin-Range bounding faults of Tertiary age and occur scattered along some fault systems, commonly as isolated pods. Many of the deposits in the Nevada-western Utah- southeastern - Oregon-southeastern - California-western Arizona region have had minor production, less than 50,000 tons; the Indian Peak Range and nearby districts (Cougar Spar, Staats, and other mines), Utah, have pro- duced more than 50,000 tons; the Spor Mountain district, Utah, has produced more than 100,000 tons; and the Fluorine district (which contains other deposits besides veins and mantos) and the Broken Hills district (Baxter mine), Nevada, have each produced more than 150,000 tons. Pipelike fluorspar bodies are present at Spor Mountain, Utah (Staatz and Osterwald, 1959), and in the Fluorine dis- trict, Nevada (Cornwall and Kleinhampl, 1961). In both areas the bodies occur in fractured dolomite and are fault controlled. At Spor Mountain, described in more detail below, most of the fluorspar occurs in the pipes, veins being of minor importance. The pipes in the Fluorine dis- trict are part of a complex of fluorspar fissure veins, breccia zones, mantos, tabular replacement bodies, and pipelike bodies in dolomite which has been intruded by hypabyssal bodies of rhyolite. Fluorspar production and estimated reserves at Spor Mountain, all from the pipes, exceed 350,000 tons (Thurston and others, 1954, p. 48). Three occurrences of fluorite in volcaniclastic sedi- ments have been recorded: Spor Mountain, Utah (Staatz and Griffitts, 1961; Staatz, 1963), Honey Comb Hills, Utah (McAnulty and Levinson, 1964), and Rome, Oreg. (Sheppard and Gude, 1969). At Spor Mountain fluorite occurs in a water-laid tuff on the flats just west of the area of the fluorspar pipes described above. Microscopic fluorite occurs dispersed through altered tuff and in a fine- grained intergrowth with opal and chalcedony in nodules up to 12 inches in diameter. These nodules, probably replaced limestone pebbles and cobbles, are dispersed NEVADA-UTAH-OREGON-CALIFORNIA-ARIZONA through the altered tuff and make up several percent of large bodies of mineralized rock. X-ray analyses of 16 samples of altered tuff containing no large fluorspar nodules indicate an average of about 3 percent fluorite with a range from less than 1 percent to 8 percent fluorite (D. A. Lindsey, oral commun., 1971). The Honey Comb Hills deposits are similar to those at Spor Mountain but smaller in scale. Fluorite content of near-surface samples from the Honey Comb Hills averages 6.3 percent and is as much as 10 percent (McAnulty and Levinson, 1964, p. 773). Both the Honey Comb Hills and Spor Mountain de- posits are thought to have formed from laterally spreading hydrothermal solutions (McAnulty and Levinson, 1964, p. 773; Staatz and Griffitts, 1961, p. 948-949). Both deposits contain anomalous amounts of beryllium. The Spor Mountain deposits, which constitute the World's largest known resource of beryllium, have been mined since 1969 for their beryllium content; fluorine is not being recovered at present (1973). Near Rome, Oreg., fluorite occurs as submicroscopic, nearly spherical grains in tuff, tuffaceous mudstone, and mudstone of a lacustrine deposit of Tertiary age. Fluorite content is generally less than 5 percent, but is as much as 16 percent in some zones (Sheppard and Gude, 1969, p. D69). According to Sheppard and Gude (1969, p. D69) the "fluorite probably formed during diagenesis in sediments that had been deposited in an alkaline, saline lake." No fluorspar has been produced from any of the deposits in volcaniclastic sediments. Several tactite bodies in the Nevada-western Utah- southeastern - Oregon-southeastern - California-western Arizona region, many mined for their tungsten content, contain appreciable fluorite. Examples are deposits in the Tushar and Mineral Ranges and Star district, Utah, the Tem Piute mine, Nevada, and the White Mountains, Calif. Fluorite was obtained as a byproduct of tungsten mining at the Tem Piute mine by treating tailings from the milling operations (Horton, 1961, p. 13). Replacement iron ore deposits in the Iron Springs district, Iron County, Utah, contain fluorapatite. Although the fluorine content is less than 1 percent, fluorine was extracted in 1957 during the processing of this ore for its iron content. Similar deposits of magnetite containing fluorapatite occur in Nevada but tonnages are not large and fluorapatite contents are not high (Reeves and Kral, 1955; Shawe and others, 1962); fluorine was not recovered from these deposits. Fluorine content of Tertiary igneous rocks, although too low to be considered as a potential resource, may indicate fluorine-rich zones worthy of exploration. Shawe (1966, p. C210, table 2) listed several areas in this region where Tertiary volcanics contain unusual amounts of fluorine. The rocks are porphyritic rhyolite and topaz- bearing rhyolite from hypabyssal bodies, rhyolite and topaz-bearing rhyolite flow rocks, and rhyolitic tuffs. The H fluorine content of these rocks ranges from 1,200 to 17,500 ppm (0.12 to 1.75 percent). Coats, Goss, and Radar (1963, p. 942, fig. 1) outlined several areas in this region where Cenozoic silicic volcanic rocks contain fluorine in the range of 1,000-5,000 ppm. DESCRIPTIONS OF INDIVIDUAL DISTRICTS AND DEPOSITS NEVADA BROKEN HILLS The Broken Hills fluorspar district is in the wedge- shaped northeastern part of Mineral County and is 70 miles (110 km) southeast of Fallon, the nearest railroad point. All rocks exposed in the Broken Hills district are Ter- tiary volcanics, mainly rhyodacite to andesite flows, tuffs, and breccias and rhyolite and quartz latite welded tuffs (Ross, 1961). Limestone beds that constitute a possible host for fluorspar deposits are present in the dominantly volcanic Excelsior Formation of Triassic(?) age just west of the district. The district is cut by a series of northeast- to north-trending faults, most of which have accompanying zones of hydrothermal alteration. Few, however, have been mineralized with fluorite, and only one fault zone is sig- nificantly mineralized. For more detailed descriptions of the district see Thurs- ton (1946), Matson and Trengove (1957), Ross (1961), and Archbold (1966). One mine, the Baxter or Kaiser, has yielded significant fluorspar, and the following discussion mainly concerns this mine. The Baxter mine was in almost continuous operation between 1928 and the time of its exhaustion in 1957, producing more than 200,000 tons of acid- and metallurgical-grade fluorspar (Archbold, 1966, p. 10). The mine is closed and is considered to be mined out (Horton, 1961, p. 14). ' The volcanic rocks in the vicinity of the Baxter mine are strongly faulted, brecciated, silicified, and hydro- thermally altered. Much of the wallrock is altered beyond recognition of its original composition. According to Archbold (1966, p. 12), "All the fluorite veins are in Tertiary rhyolite that unconformably overlies weakly metamorphosed sedimentary and volcanic rocks presumed to be of Mesozoic age. The rhyolite is overlain by Tertiary volcanic rocks of intermediate composition." Small amounts of unaltered porphyry (quartz mon- zonite?) can be found on some dumps of the Baxter mine. This porphyry may be from narrow dikes that Thurston (1946, p. 2) reported to be common in the area. Thurston (1946, p. 2) also reported that there is no apparent relation between these dikes and the fluorspar mineralization. The fault zone that has been significantly mineralized is called the main vein; it strikes N. 40° E., dips an average of 50° northwesterly, and has numerous bifurcations which trend north to northeast. The fault zone bearing the main vein is undulating and pinches and swells within short 44 GEOLOGY AND RESOURCES OF FLUORINE distances, varying in width from 2 feet (0.6 m) at places on the 400-level to 18 feet (5.5 m) at the surface. The main vein is an aggregate of independent seams and masses of fluorspar 3 inches (7.5 cm) to 1 foot (30 cm) wide. From published descriptions the main vein apparently is made up of a series of connected lenticular pipelike bodies in the plane of the fault (fig. 18). Fluorspar did not extend far below the 700-level of the Baxter mine, and along much of the vein it did not extend below the 400-level. Ore mined near the surface averaged 5 feet (1.5 m) in width and was as much as 9 feet (2.7 m) in width. On the 300-level ore mined averaged 1.5 feet (0.5 m) in width and was as much as 5 feet (1.5 m) (Thurston, 1946, p. 3). Early mining was for ore containing more than 85 percent CaF,, and Thurston (1946) considered only material of this grade as ore. Kaiser Corporation's portion of the total production, on the other hand, averaged 46 percent CaF. Fluorspar occurs as lenses, veinlets, and breccia cement within the fault zone. Layering and crustification with botryoidal and mammillary structures are normal. Judged from material left on the dumps, the fluorite was either a clear, pale-green or light-violet fibrous and columnar variety or a gray fine-grained variety intimately inter- grown with chalcedony. Brecciated fluorite cemented by later fluorite or chalcedony is common, and slickensides along fluorspar seams are widespread. Intercalated with the fluorite are layers of chalcedony and minor layers of manganese oxide and "mountain leather," a fibrous mineral reported by Thurston (1946, p. 3) to be a member of the palygorskite group. Laser probe analyses of pure fluorite from the Baxter mine show only trace amounts of barium, manganese, and strontium common to almost all fluorite. Veinlets of brown chalcedony in silicified and altered country rock and minor gray breccias in the fluorspar contain trace amounts of zinc, arsenic, antimony, lead (up to 1,500 NE ppm), and silver (up to 30 ppm). Manganese oxides pre- sent in the fluorspar also contain trace amounts of lead and silver. The Broken Hills mine, in the eastern part of the Broken Hills district, in addition to production of fluorspar, has had minor production of silver and lead that came from quartz-filled fissure veins and stockwork lenses in vol- canic rocks (Ross, 1961, p. 81). QUINN CANYON RANGE Fluorite occurs at several locations in the Quinn Can- yon Range, Lincoln and Nye Counties (fig. 19). Past pro- duction of fluorspar has been small, 200 tons from the Rainbow prospect in 1945-46 and about 14 tons that contained 50 percent CaF, from the Nyco mine by Wah Chang Corp. prior to closure of the Tem Piute mill (Horton, 1961, p. 18). Two areas are currently (1972) being mined-the Steel property by Lee Fennison, Panaca, Nev., and the Mammoth property by the Adaven Mining Company. Both operators are trucking their ore to the mill at Caliente, Nev., 100 miles (160 km) to the east. For more detailed descriptions of the geology and mineral deposits of the Quinn Canyon Range see Sainsbury and Klein- hampl (1969). In the Quinn Canyon Range fluorspar occurs in lime- stone and volcanic rocks in the vicinity of intrusive rocks. The deposits in the volcanic rocks are similar to the deposits in the Broken Hills district, Nevada, described above. The fluorspar deposits in limestone are diverse types of replacement bodies, fracture fillings, and veins characterized by presence of chalcedony and by layering. Sedimentary rocks of early to middle Paleozoic age and tuffaceous sedimentary and volcanic rocks of Tertiary age make up the country rock of the Quinn Canyon Range (fig. 19). These rocks are intruded by a small hypabyssal quartz latite porphyry pluton in the south-central part of the area, by a medium- to coarse-grained granitic stock in SW 0 500 FEET ok 0 100 METRES FicurE 18. -Longitudinal projection along "main vein" of the Baxter (Kaiser) fluorspar mine, Broken Hills district, Nevada (from Matson and Trengove, 1957, fig. 4). Black indicates mined-out areas. Horizontal lines show mine levels. s 4 NEVADA-UTAH-OREGON-CALIFORNIA-ARIZONA 45 ; i 38°15 |- \ Nyala «-.. Cte tl 2 B 4 2 t < NyECouNnty -._._ ___ \ LINCOLN COUNTY \_ § Rainbow? 2D 4 38°00' R 55 E 115°45' R56 E R 57 E 116° EXPLANATION I: Quaternary alluvium _A+** - Tertiary dike Silicified rocks 3g m Tertiary intrusive rocks Li Fluorspar deposit Fault - Dashed where concealed Paleozoic sedimentary rocks Tertiary volcanic and tuffaceous sedimentary rocks Las Vegas o 5 MILES {.~ ] 1 1 R 1 I 0 5 KILOMETRES Ficure 19.-Map of the Quinn Canyon Range, Nevada, showing locations of fluorspar deposits (from Sainsbury and Kleinhampl, 1969, PI. 1). 46 the north-central part of the area, and by numerous dikes and sills of fine- to medium-grained dacitic to rhyolitic porphyry and andesite (fig. 19) (Sainsbury and Klein- hampl, 1969, p. C3). The Paleozoic formations, composed mainly of limestone and dolomite, generally correlate with those assigned to the miogeosynclinal carbonate assemblage of the eastern Great Basin (Sainsbury and Kleinhampl, 1969, p. C3). The oldest Tertiary unit, of local extent only, consists of conglomerate and minor interbedded freshwater limestone. Volcanic rocks of Oligocene and Miocene age cover an extensive area of the Quinn Canyon Range (fig. 19). These rocks are rhyolite, rhyolitic to quartz latitic welded tuffs, and minor ande- sitic flows and breccias intercalated with some tuffaceous and sedimentary strata. "Many of the mineral deposits are localized along the contacts of light-colored rhyolitic to dacitic porphyry dikes" (Sainsbury and Kleinhampl, 1969, p. C4). Wall- rock alteration consisted of silicification of the carbonate sedimentary rocks and silicification and pyritization of volcanic rocks. Several large altered zones are shown in figure 19. Not all these silicified zones contain fluorspar. The Quinn Canyon Range is made up of several structural blocks bounded in part by Basin-Range faults. The structural blocks in which Paleozoic sedimentary rocks are exposed consist of gently folded or homoclinal sedimentary sequences cut by high-angle faults and, in some areas, thrust faults. The volcanic rocks are not as complex structurally but nevertheless probably overlie and conceal major structural units in the older rocks (Sainsbury and Kleinhampl, 1969, p. C4). The major dikes of the area trend east to northeast and cut across the strike of Paleozoic and Tertiary units and across the trend of thrust faults and folds. Discrete mineralized alteration zones in parts of the Quinn Canyon Range also trend east to northeast. The disparate trend and also the relatively late development of alteration, mineralization, and dike formation suggested to Sainsbury and Kleinhampl (1969, p. C6) that the fluorite deposits may be related to a late structural and igneous event, not yet clearly identifiable. Fluorspar in the Quinn Canyon Range occurs in a variety of deposits. The following classification of fluor- spar deposits in the Quinn Canyon Range is taken from Sainsbury and Kleinhampl (1969, p. C7). (Individual deposits mentioned are shown on fig. 19). 1. Relatively small but high-grade replacement deposits in limestone along dike walls (Spar claims). 2. Mixed fine-grained silica (jasperoid) and fluorite that have replaced shattered limestone and dolomite, commonly forming tabular, flat-lying deposits in which the bulk of the fluorite cements breccia frag- ments or coats vugs (Valley View and Bonanza claims, Crystal group). Some deposits are localized in shattered jasperoid zones along thrust faults (Jumbo and Horseshoe claims, Spar group). GEOLOGY AND RESOURCES OF FLUORINE 3. Veins along fractures, in which fluorite and quartz form relatively coarse-grained intergrowths. 4. Irregular, relatively high-grade replacement deposits in limestone in which fluorite is intergrown with silica, both irregularly and rhythmically, to form layered ores ("coontail'"' or "banded" ores), as at the HiGrade and Mammoth deposits. Locally these re- placement bodies are separated from or extend well out from dikes and are thus distinguishable from the other class of high-grade replacement deposits described under 1, above. The deposits in volcanic rocks are less diverse, and only two main types are recognized: (1) well-defined fissure veins filled with nearly pure coarsely crystalline fluorite. These veins are bordered by argillized wallrocks contain- ing noticeable to abundant disseminated pyrite (Nyco mine); (2) large tabular fluorite deposits. Fluorite occurs irregularly as discrete but discontinuous veins and vein- lets, commonly intergrown with jasperoid, throughout large tabular strongly silicified zones and fractured zones in volcanic rocks. The rocks in the outer portions of the altered zones contain disseminated pyrite. Fluorspar grade is related to type of deposit and type of fluorspar. A small breccia pipe, less than 50 feet (15 m) in diameter, on the Valley View claim of the Crystal group contains about 35 percent CaF, all fluorspar occurring as breccia cement (Sainsbury and Kleinhampl, 1969, p. C11). At the Mammoth property, a large silicified zone in limestone-fine-grained siliceous "jasperoid"-contains 16-22 percent CaF, and 70-74 percent SiO, whereas layered "coontail" ore contains 22-35 percent CaF, and 50-65 percent SiO; (Sainsbury and Kleinhampl, 1969, p. C13). Small fluorspar replacement bodies at the Spar group contain more than 90 percent CaF, (Sainsbury and Kleinhampl, 1969, p. C15). An area comprising about 28,000 ft? (2,600 m?) on the HiGrade group contains approximately 65 percent fluorite, the richest zones con- taining about 75 percent. Another area, about 6,000 ft? (560 m?) on the same group, contains by assay 72 percent CaF, (Sainsbury and Kleinhampl, 1969, p. C17). The HiGrade group consists of silicic replacement deposits in lime- stone and at least one shoot of layered "coontail" ore. Mineralogy of the fluorspar deposits is simple: fluorite, quartz, jasperoid, and pyrite, and, in limestone replace- ment bodies, abundant calcite. Color and form of fluorite differ among different types of fluorite deposits. Fluorite in replacement bodies in limestone is either coarsely crystalline, nodular, and grayish white, or fine grained, compact, and light purple. The fluorite in large pure veins in volcanic rocks is mostly light green to colorless; the fluorite in the "coontail" ore and that cementing brecciated jasperoid is colorless or white, and the fluorite that forms small cubes is very faintly green (Sainsbury and Kleinhampl, 1969, p. C10). NEVADA-UTAH-OREGON-CALIFORNIA-ARIZONA UTAH SPOR MOUNTAIN The Spor Mountain fluorspar district, described in detail by Staatz and Osterwald (1959), is in the southwest part of the Thomas Range in central Juab County about 100 miles (160 km) southwest of Salt Lake City, Utah. Initial fluorspar production was in 1948, but large pro- duction did not occur until the late 1940's. More than 100,000 short tons of fluorspar was produced in the period 1943-56, (Staatz and Carr, 1964, p. 143). Following inactivity, the district recently (1971) has been revived as a result of the increased fluorspar market. Exposed rocks range in age from Early Ordovician to Pleistocene. Most of the district is underlain by a thick sequence of apparently conformable rocks, chiefly carbo- nates, of Paleozoic age. Silurian and Ordovician dolo- mites are the chief fluorspar-bearing units. Tertiary vol- canic rocks that locally overlie the Paleozoic rocks are latite, dacitic tuff, quartz latite, volcanic breccia, lapilli tuff, quartz latite tuff, rhyolitic water-laid tuff, and topaz- bearing rhyolite. Dikes and plugs of breccia, rhyolite, and quartz latite intrude the Paleozoic rocks, commonly along faults. Gravels and marls of the Lake Bonneville Group were deposited during Pleistocene time in the region sur- rounding the Thomas Range. The Spor Mountain fluorspar deposits, though occurring dominantly in Paleozoic sedimentary rocks, are arbitrarily classified as associated with Tertiary volcanic rocks because of evidence of genetic relation to the topaz- bearing rhyolite. Near the fluorspar deposits, very large beryllium-bearing deposits in Tertiary volcaniclastic rocks contain about 5 percent fluorite (Staatz, 1963; Shawe, 1968). The Paleozoic sedimentary rocks and the volcanic rocks have been tilted and now strike northeast and dip north- west. These consistently dipping rocks are cut by hundreds of faults belonging to at least five sets: (1) northeast- trending thrusts; (2) northeast-trending normal and high- angle reverse faults; (3) northwest-trending faults; (4) east- trending faults; and (5) north-trending faults. According to Staatz and Osterwald (1959, p. 1-2), "Fluorspar deposits are of three types: oval to irregular pipes, veins, and disseminated deposits. The pipes, which show considerable range in shape and size with depth, have produced more than 99 percent of the ore* **. [They] show evidence of two chief types of structural control: faults and intrusive breccia bodies." Ore pipes formed mainly by replacement along shattered zones in dolomite or volcanic rocks. The ore consists of 65-95 percent of fluorite with montmorillonite, dolomite, quartz, chert, calcite, and opal as impurities. The fluorspar closely resembles a brown, white, or purple clay and forms either pulver- ulent masses or boxworks. With depth the grade of the ore commonly decreases, and masses of montmorillonite, chert, or quartz and dolomite have been found in increasing abundance in some deposits. 17 The fluorspar ore contained 0.003-0.33 percent U, according to analyses of numerous samples, and uranium grade varied considerably from place to place. Staatz and Osterwald (1959, p. 2) concluded that the fluorspar ore formed from fluorine-rich fluids, containing minor amounts of uranium, which were derived from the magma that formed the topaz-rich rhyolites of the Thomas Range during the last stages of volcanism. These fluids rose along faults and replaced shattered zones in the dolomite. Introduced elements other than fluorine and uranium were probably obtained from rocks underlying the deposits. ARIZONA CASTLE DOME The Castle Dome district lies along the southwestern edge of the Castle Dome Mountains, Yuma County, in the southwest corner of the Kofa Game Range, and is bounded on the west and south by the Yuma Proving Grounds. From the time of its organization in 1863 until 1949 the district produced 17,726,937 pounds of lead and 459,321 ounces of silver, mostly during the early part of the period. According to Wilson (1933, p. 87): The first important mining of fluorspar, which is a plentiful gangue mineral at Castle Dome, began in 1902. During 1902, 1903, 1904, 1908, 1909, and 1913 carefully hand-sorted crystals and pure screening from the De Luce claims were shipped to the Riverside Portland Cement Company's plant at Riverside, California, to be used as a flux in pro- ducing cement clinker. The area was also active during World War I; fluorspar production from the district in 1902-17 exceeded 1,000 tons (Wilson, 1950). Since then production has been small and intermittent. For detailed descriptions of the district see Wilson (1933; 1951). Country rocks in the Castle Dome district are well- bedded, weakly metamorphosed shale of probable Cretaceous age and diorite porphyry of Cretaceous(?) age (Wilson, 1951, p. 99). The shale rests unconformably upon granite and is faulted against schist, both of probable Precambrian age (Wilson, 1951, p. 99). These lithologies make up only a small part of the Castle Dome Mountains, the major part being volcanic rocks of probable Tertiary age (Wilson, 19383, p. 81). This volcanic pile includes about 2,000 feet (600 m) of rhyolites, andesites, tuffs, and obsidians and in places is overlain by several hundred feet of Quaternary(?) basalt. Locally, hypabyssal bodies of rhyolite porphyry have intruded the extrusive rocks; the fluorspar deposits appear to be related to these. The Cretaceous sedimentary rocks, host for argentiferous galena-fluorite veins, are dominantly greenish-gray shales and impure cherty limestones. Wilson (1983, p. 80) also reported locally abundant maroon shales, fairly pure lime- stones, brown arkosic sandstones, quartzites and conglomerates. Crosscutting the Cretaceous sedimentary | rocks is a remarkable and pervasive swarm of dikes of diorite and rhyolite composition. Wilson (1983, p. 82) indicated that the diorite dikes are older than the lavas and | the rhyolite dikes are younger. Alteration consisted of the 48 GEOLOGY AND RESOURCES OF FLUORINE development of sericite and pyrite metacrysts throughout the rhyolite porphyry and in the Cretaceous sedimentary rocks, and in diorite in contact with the rhyolite porphyry dikes. Almost all structural elements in the Castle Dome dis- trict strike north to northwest. These include range-front faults and other faults, mineralized veins, the diorite and rhyolite dikes, and the folded Cretaceous sedimentary rocks. In the vicinity of the Castle Dome district the vol- canics form a broad anticline plunging east. Fluorspar in the Castle Dome district occurs in argenti- ferous galena-fluorite veins. Gold-quartz veins and silver- bearing manganese oxide veins are common in sur- rounding parts of the Castle Dome Mountains. The argentiferous galena-fluorite veins occupy north- to northwest-trending steeply dipping fault zones that cut the Cretaceous sedimentary rocks and the diorite and rhyolite porphyry dikes. Most mineralization occurred in altered zones around the rhyolite porphyry, and the veins are best developed where diorite porphyry makes up one or both of the vein walls. Although most are less than 5 feet (1.5 m) in width, veins as much as 12 feet (3.7 m) in width have been reported (Wilson, 1983, p. 83), and most veins are continuous for great lengths, some as much as 5,000 feet (1,500 m). The argentiferous galena-fluorite veins are largely fissure fillings and cemented breccias in which layering and crustification of the gangue minerals are common. Gangue consists mainly of layers of coarsely crystalline varicolored fluorite, coarsely crystalline calcite, and black calcite that are cut by veins and veinlets of bladed to massive barite and chalcedony. Several veins in the district are composed entirely of the gangue minerals occurring as described above. Where present, galena forms sheetlike masses or irregular veinlike bodies in the fluorite or calcite. In places, veins consist of nearly solid masses of galena up to 8 feet (2.5 m) thick (Wilson, 1933, p. 83). Other vein minerals are hydrozincite, smithsonite, wulfenite, vanadinite, mimetite, aragonite, gypsum, anglesite, cerussite, and other lead oxides. Galena is the only sulfide mineral that has been noted. No silver minerals have been recognized even though the galena contains as much as 30 ounces silver per ton. Fluorspar deposits like those at Castle Dome are char- acterized by coarsely crystalline fluorite of high purity. Silica contamination does not affect the quality of the mined product, although lead contamination might. Most of these deposits may be related to Tertiary volcanism and many appear to be physically restricted to the volcanic pile with which they are associated. However, in districts like Castle Dome the source of the mineralizing solutions evidently was much deeper than the surface volcanics, and ascent of these solutions was controlled by major regional structures. Although fluorite is considered to be a gangue mineral at Castle Dome, as it is in most deposits of this type, it is plentiful enough to be considered as a source of fluorspar, either as the major product or as a coproduct. OTHER DEPOSITS Several fluorite, barite, and manganese oxide deposits occur in an east-northeast-trending zone along the southern end of the Harquahala Mountains, the northern end of the Big Horn Mountains, and through the Vulture Mountains. Production of fluorspar has been minor and mainly from the Snowball deposit in the Harquahala Mountains and from deposits in the eastern part of the Vulture Mountains (Van Alstine and Moore, 1969). Varying amounts of quartz, chalcedony, calcite, black calcite, barite, fluorite and manganese oxides are present in the deposits. Iron oxides, malachite, and trace amounts of gold are widespread. Layering, crustification, druses, and vugs characterize the deposits. Country rocks are granite, gneiss, schist, limestone, basalt, and volcanic agglomerate. The deposits are generally localized along northwest- or east-northeast-trending fracture systems as pods, short veins, and breccia fillings; deposits in basalt are made up of numerous stringers and pods spread through diffuse breccia zones. In the vicinity of the Snow- ball deposit, fluorite-rich deposits form horizontal layers consisting of cavity fillings and breccia cement in volcanic agglomerate. Two trench samples across a fluorspar breccia zone at the Snowball deposit indicate a grade of 72 percent CaF, across a width of 4 feet (1.2 m), and 36 percent CaF, across a width of 12 feet (8.7 m). COLORADO-NEW MEXICO-WESTERN TEXAS-SOUTHEASTERN ARIZONA By R. G. Wort, W. R. GrirFITts, and R. E. Van Arstin® Fluorspar deposits are plentiful in Colorado and New Mexico, States that rank third and fourth, respectively, in total fluorspar production, following Illinois and Kentucky. Fluorspar production in Colorado began about 1873 in the Jamestown district, Boulder County, and near Evergreen, Jefferson County. Almost 90 percent of the sub- sequent production has come from six districts: Northgate (Steven, 1960), Jamestown (Goddard, 1946), Browns Canyon (Van Alstine, 1969), Poncha Springs (Van Alstine, 1964; Russell, 19477; 1950), Wagon Wheel Gap (Van Alstine, 1964; Steven, 1968), and St. Peters Dome (Steven, 1949). Production has been intermittent in most districts and reached a peak in 1944 when 65,209 tons of concen- trates was shipped. Total shipments of fluorspar from Colorado through 1970 have amounted to about 1.3 million tons, or about 10 percent of the fluorspar produced in the United States since records of production were started in 1880. Most of the concentrates have been shipped to the steel industry at Pueblo, Colo., to chemical, aluminum, and ceramic plants in other States; and to U.S. Government stockpiles. By 1963 shipments decreased to COLORADO-NEW MEXICO-WESTERN TEXAS-SOUTHEASTERN ARIZONA less than 10,000 tons, and only the Jamestown district was active. During 1969 an enlarged Northgate mine and mill began operation and produced concentrates for a new pelletizing plant at Cowdrey, Colo. The grade of Colorado deposits that have been worked ranges from about 20 percent to more than 75 percent CaF; most deposits were opened by underground workings, but some of the lower grade but larger ore bodies were mined in opencuts. Deposits are as much as 45 feet (14 m) thick and 2,600 feet (800 m) long in the Browns Canyon district, and opencuts on a vein zone in the North- gate district have been developed over a length of about 4,400 feet (1,350 m) (Steven, 1960, p. 395). Fluorspar is mined below a depth of 1,400 feet (430 m) in the Jamestown district. Many other localities in Colorado have produced a little fluorspar, mostly from small veins (Cox, 1945) A pegmatite in the South Platte district, Jefferson County, yielded a little fluorspar as a byproduct of feldspar mining. Fluorite is present as a gangue mineral in various metal- liferous deposits of gold, lead, zinc, tungsten, molybdenum, beryllium, and thorium (Van Alstine, 1964), and extensive mine dumps and piles of mill tailings, such as those with fluorite in and near the Cripple Creek district, represent a possible future source of fluorspar. Topaz has not been an industrial product in Colorado or New Mexico, and most of the known deposits are too small or of too low grade to offer much promise. The mineral is a prominent component of hydrothermally altered rock in the molybdenum mines near Climax (Wallace and others, 1968) and Henderson (White and MacKenzie, 1973), Colo., and might thus become a bypro- duct of molybdenum mining at those or similar deposits in the region. The large tonnages of rock processed in the concentrators could yield significant amounts of topaz, even with a low content in the raw ore. Fluorspar mining began in New Mexico in the early 1880's in the Burro Mountains and Gila districts. Production reached a peak in 1944 when 43,000 tons of fluorspar was marketed. Fluorspar has been produced in at least 10 counties of New Mexico, with a total yield of about 650,000 tons. Of the total, about one-third came from the Zuni Mountains district, Valencia County (Williams, 1966, p. 4), nearly half originated in Grants County, coming from deposits in the Burro Mountains district, the Gila district, and the Cooks Peak district, and most of the rest came from the Fluorite Ridge district, Luna County. Most of the fluorspar deposits in Colorado, New Mexico, western Texas, and southeastern Arizona are of Oligocene or younger age-younger than the latitic and rhyodacitic rocks of the great Sierra Madre Occidental volcanic field that occupies much of western Mexico and extends northward into New Mexico. General aspects of the types of fluorspar deposits in the region are discussed below. Fluorite-bearing deposits of 49 Precambrian age were formed in a plutonic environment and have as their most widespread representatives the fluoritic pegmatites of the Llano area, Texas, the Petaca and other districts of New Mexico, a broad area in Douglas, Teller, and Park Counties, Colo., and an area east of Laramie, Wyo. Most of the pegmatites contain rather coarse-grained purple, green, and pink or cream- colored fluorite in the inner parts of moderately well developed zones. Associated minerals are red microcline, gray quartz, plagioclase, a black mica or amphibole, topaz, beryl, and a variety of rare-earth minerals including yttrian fluorite. A carload of fluorspar was reported to have been shipped from pegmatites in Douglas County, Colo., as a byproduct of feldspar mining, and another car- load was a byproduct of pegmatite mining operations in Wyoming, but the potential of pegmatitic fluorite deposits is small. The pegmatite deposits in the Petaca district, New Mexico, contain coarse green fluorite in their inner parts associated with pale microcline, plagioclase, and muscovite mica; the mica has been the main economic product. Rare-earth minerals, though present, are much less common than in the other fluoritic pegmatites. Fluorite deposits of Laramide or early Tertiary age are less widespread than either the Precambrian or the Oligocene and younger Tertiary fluorite deposits. Laramide deposits apparently are restricted largely to the Front Range mineral belt in central Colorado, where the younger Tertiary fluorite deposits also are found. Environments of formation, structures, textures of mineral assemblages, and accessory mineral and metal contents differ among deposits of different ages and among deposits within age groups. This diversity of type is not so pronounced if only the major districts are considered. The Jamestown, Colo., deposits of middle Tertiary age contain fine-grained fluorite associated with minerals of lead, silver, gold, copper, and uranium in stockworks, pipelike bodies, and mineralized breccia zones, and represent a mesothermal type of mineral- ization (Lovering and Goddard, 1950, p. 48, 64, 258). Similar fluorspar deposits and associated alkalic intrusives occur in the Eagle Mountains, Tex.; Gallinas Mountains, N. Mex.; Iron Mountain, N. Mex.; Bear Lodge Mountains, Wyo.; and the Little Rocky Mountains, Sweet- grass Hills, Judith Mountains, and South Moccasin Mountains, Mont.; and fluorite occurs in trace amounts in and around many calc-alkalic intrusives throughout the West. These deposits are found in many rock types, although deposits of significant size occur in limestone or in shattered granitic plutonic rocks. Fluorite-rich deposits at Iron Mountain, N. Mex., are tactites in the contact zones of small silicic intrusives. Most other major districts in Colorado and New Mexico are of epithermal type and of later Tertiary age, commonly layered and crustified veins, but also mantos in sedi- mentary and volcanic rocks. They are localized by 50 geologically young steep faults and breccia zones in Tertiary volcanic rocks, Paleozoic and Mesozoic sedi- mentary rocks, and Precambrian silicic igneous and meta- morphic rocks near the edges of uplifts or depressions. Most deposits of this type in Colorado are well-defined veins in plutonic and metamorphic rocks or in volcanic rocks, whereas those in New Mexico are commonly diffuse vein systems and mantos in calcareous country rocks, such as those at Sierra Cuchillo, described elsewhere in this report. Deposits in southeastern Arizona generally are veins in plutonic, metamorphic, or volcanic rocks. Deposits in Colorado include those in the Northgate, Browns Canyon, and St. Peters Dome fluorspar districts, and the deposits at Crystal, Dillon, Poncha Springs, and Wagon Wheel Gap. Notable deposits in New Mexico are those in the Gila, Burro Mountains, Anderson, Gold Hills, Steeple Rock, Sierra Caballos, Zuni Mountains, Fluorite Ridge, and Cooks Peak fluorspar districts, and deposits at Tonuco, Tortugas Mountain, Bishop Cap, and Sierra Cuchillo. In southeastern Arizona, similar deposits occur in the Tonto Basin, Aravaipa and Stanley areas, the Whetstone, Sierrita, and Chiricahua Mountains, and the Duncan and Copper Dome copper districts. Most formed at rather low temperatures and pressures near the earth's surface; warm or hot springs are associated with the deposits in some districts (Van Alstine, 1947, p. 461, 465). In the Browns Canyon and Poncha districts, Chaffee County, Colo., fluorspar deposits of late Tertiary age formed under near-surface hot-spring conditions from very dilute fluids (Van Alstine, 1969). Study of fluid inclusions indicates temperatures of deposition at about 119°-168°C. The heat and fluoride are regarded as volcanic contributions to the mineralizing fluid. The epithermal fluorspar deposits are aggregates of fluorite, quartz, chalcedoney, barite, and various minor components. Many contain pale-green or white fluorite in grains 1-5 inches (2.5-12.5 cm) long. Some of these are equant, but columnar or fibrous structures and successions of layers with different grain size, color, or structure are common. Most vein fluorspar shows slight to strong brecciation. Blocks of country rock and early fluorspar are cemented by late fluorspar or gangue minerals. Fine-grained fluorspar is present in large amounts in some districts in which coarse fluorspar also is found. Several Tertiary deposits in New Mexico are epithermal replacement deposits in upper Paleozoic sedimentary rocks, and some in Texas are in Cretaceous limestone. The deposits in Cretaceous limestone in Texas (Gillerman, 1953; Underwood, 1963) are among the few in the Western United States that are bedding replacement deposits like those in the Illinois-Kentucky district. Fluorite is dispersed in Tertiary granite and occurs with topaz in small open veins in the Mount Antero area, Colorado. Like the vuggy veins in the Precambrian Pikes GEOLOGY AND RESOURCES OF FLUORINE Peak Granite, these fissure deposits have yielded excellent crystals to mineral collectors but no industrially important fluorite or topaz. The intrusive rocks at Jamestown, Colo., at Mount Antero, Colo., and at similar occurrences are not closely associated with volcanic rocks but they yield radiometric ages of about 35 m.y. and are thus nearly synchronous with the volcanism of the region; the intrusives may occupy deep parts of former (now eroded) volcanic vents. Even where deposits and volcanic rocks are widely separated, as at Northgate, Colo., and the Big Bend area of Texas, they nevertheless may be related genetically. A general discussion of the structural and geologic controls of fluorine mineral deposits in this broad region follows; the discussion encompasses northern Mexico as well as Colorado, New Mexico, western Texas, and south- eastern Arizona in order to provide a more comprehensive review of the environment of the deposits. In the southern mountain region of Texas, New Mexico, southeastern Arizona and northern Mexico, the distribution of fluorite districts is related both to the great Sierra Madre Occidental volcanic field of Mexico and to regional tectonic features. Three major structures that are involved are (1) the Texas lineament, a belt of faults, folds, and intrusives which extends southeast from east-central Arizona, through El Paso and the Big Bend area to a terminus northeast of Monterrey, Mexico; (2) the Rio Grande trough, a series of structurally depressed basins that extends north from El Paso probably to the vicinity of Northgate, at the Colorado-Wyoming State line; and (3) a rather ill-defined zone of intense faulting that crosses the volcanic field in a northeast direction near the Inter- national boundary where the volcanic field enters south- eastern Arizona. The Texas lineament is a complex structure with a history of instability extending onward in time at least from the Carboniferous. A submerged Paleozoic uplift is indicated by the transgression of formations of Carboni- ferous age across much older formations. The zone was also important in late Mesozoic time, as the Lower Cretaceous formations of Mexico terminate abruptly northward along a hinge line that lay along the lineament in southern New Mexico. Early Tertiary deformation resulted in tight folds and thrust faults, which appear to be the dominant structures in the eastern third of the lineament between the Big Bend and Monterrey, and have been mapped northwestward in Mexico to the Inter- national boundary near southwestern New Mexico. West of the Big Bend, these compressional structures are best developed south of the Border but they have been mapped in western Texas. There is a marked contrast near El Paso, where east-trending compressional structures pre- dominate south of the Rio Grande and north-trending fault blocks predominate in the Franklin Mountains north of the Rio Grande. COLORADO-NEW MEXICO-WESTERN TEXAS-SOUTHEASTERN ARIZONA Superimposed upon these older folds and thrust faults are many high-angle faults which bound blocks of both sedimentary and volcanic rock that may be raised or dropped several hundreds or even thousands of feet. Many of these, like the older structures, trend northwest, parallel to the lineament as a whole. Block faulting is best developed north of the border and west of the Big Bend. The faults farther southeast are persistent both laterally and vertically, but the displacements are commonly only a few feet and seldom exceed 100 feet (30 m) (Temple and Grogan, 1963, p. 1041). The trends of these eastern fractures are less consistently northwest than are the western fractures, and most commonly are north- northeast, east-northeast, west-northwest, or north- northwest. Small stocks and dikes of rhyolite are rather common in the lineament (as in the Sierra Blanca area) and are accompanied by intrusives of alkalic rocks between Sierra Blanca and the Big Bend. Fluorite-bearing deposits are found along most of the length of the Texas lineament, and are most numerous or largest near the two ends. The deposits in Mexico, east of the Big Bend, form a rather coherent though wide- spread group. Temple and Grogan (1963, p. 1051-1052) succinctly summarized the regional history that followed deposition of the Georgetown Limestone in Early Creta- ceous time: 1. Broad warping, development of joints, and develop- ment of an undulating surface with low relief deter- mined by joint-controlled erosion. 2. Deposition of the Del Rio Clay in Late Cretaceous time, largely in depressions in the erosion surface. 3. Deposition of Buda Limestone and the Eagle Ford Shale. s 4. Folding along northwest-trending axes and renewed jointing. 5. Block faulting, most prominent in the north-north- west-trending Boquillas Valley graben; intrusion of rhyolite and other siliceous rocks. 6. Widespread fluorspar mineralization. The Mexican fluorspar deposits include crosscutting veins and mantos, or replacement deposits parallel to beds in limestone (Van Alstine and others, 1962). Both types are restricted to the Georgetown Limestone of Early Cretaceous age. The largest crosscutting veins are along contacts between rhyolite dikes and the limestone, or between segments of dikes and in the same structures as the dikes. The very large mineralized pipe at the Cuatro Palmas mine, Coahuila, lies in such a gap in a rhyolite dike (McAnulty and others, 1963, pl. 1). According to Temple and Grogan (1963) the vein deposits are largely in the upper part of the Georgetown Limestone, extending through a vertical range of several hundred feet. The mantos are found in the uppermost part of the George- 91 town Limestone. They are tabular in vertical section and are elongated parallel to joints in the limestone. The same restriction of fluorite mineralization to the upper part of the Devils River Limestone (equivalent to Georgetown Limestone) is found in the quicksilver veins of the Terlingua district, Texas. Yates and Thompson (1959, p. 72) reported that the veins contain accessory fluorite only where the vein walls are Devils River Lime- stone. However, Gillerman (1953, p. 61-62) reported 800 feet of Georgetown Limestone in the fluorite district in the Eagle Mountains, Tex., but the best host rock is the Finlay Limestone, which is below the Georgetown Limestone and separated from it by the intervening Kiamichi Forma- tion. The mineralizing solutions must have had access to the Georgetown Limestone inasmuch as the reefs in the Grayson Formation, still higher in the section, are mineralized. There is little mention of systematic explora- tion of the limestone below the Georgetown in other districts, although McAnulty and others (1963) showed several small fluorite deposits in the Aguachile area in Edwards Limestone and one in the Kiamichi Formation; and in section they showed the large Cuatro Palmas deposit to extend downward through the Edwards Lime- stone (1963, pls. 1, 2). Farther northwest along the Texas lineament, forma- tions of the Lower Cretaceous Series are poorly developed and are not known to contain fluorite. Toward the north, along the northwestern end of the Cretaceous Mexican geosyncline, owing to the change in sedimentary facies, the Georgetown Limestone is apparently absent from the stratigraphic section. Thus the only fluorite prospect reported by Albritton and Smith (1965) in the Quitman Mountains of Texas in this northern area is in rhyolite. Between El Paso, Tex., and Silver City, N. Mex., the structural relations of the fluorite deposits are rather complex, as the deposits are influenced by fractures of the Rio Grande trench system as well as those of the Texas lineament and the northeasterly fracture system. Structure is simpler still farther to the northwest. There are two rather ill-defined northwest-trending belts of fluorite mineralization in this westernmost area, separated from one another by 12-25 miles (20-40 km) of barren ground consisting of gravel-filled basins and hills of volcanic rocks. In the northeastern belt, the most westerly deposits commonly strike northwest along faults that parallel the lineament. Similarly, most of the deposits in the western part of the southwestern branch of the Texas lineament are in the northwest-trending fractures that are so common in the lineament. Farther eastward along the northeastern belt north-trending fractures become prominent as con- trolling structures for the fluorite mineralization, reflecting the fracturing associated with the Rio Grande trough. Northeast- and east-trending veins also are present and probably resulted from left-lateral movement along the Texas linecament. The northeast-trending veins lie 52 GEOLOGY AND RESOURCES OF FLUORINE mainly in the area between the Lordsburg and Silver City districts, New Mexico, an area in which fluorite deposits are particularly plentiful and several moderately large mines have been opened. Such enhancement of mineral- ization at intersections of structures is well known to geologists both on this regional scale and on a smaller scale, affecting individual ore bodies, as at the Burro Chief fluorspar mine, Burro Mountains, N. Mex., one of the largest mines in the northeastern belt. Much farther south in Mexico, major fluorspar districts are also proximal to the Sierra Madre Occidental volcanic field, but they are restricted to regions in which the continuity of the volcanic rocks is broken by younger structures that bound the elongate areas of various rock units protrayed on geologic and tectonic maps of the Republic of Mexico. These areas are elongate in a north- westerly direction in east-central Durango, southwestern Zacatecas, and southwestern San Luis Potosi, and in a northerly direction in northern Guerrero. There is thus a southward persistence of the northwest-trending regional structures that are widespread in the Western United States; the structures in Mexico have controlled mineral- ization in a similar fashion. The fluorspar district in northern Guerrero is associated with high-angle faulting in a region of north-trending structures and at the south end of the great volcanic field. Situated as it is south of the northwest-trending transform fault zone, marked by recent volcanic rocks, this region may not be closely related structurally to the region farther north. The fluorspar district near Parral, southern Chihuahua, is in a rather small area of northeast-trending structures that lie athwart the northwest-trending structures that prevail regionally. The northeasterly trend at Parral is parallel to, and possibly similar in origin to, the broad structure that crosses the Sierra Madre volcanic field in and near south- eastern Arizona. The Rio Grande trough is a general term for a series of basins that begins abruptly at the edge of the Texas lineament at El Paso and extends northward probably to Wyoming. Although the name gives prominence to one north-trending series of basins, the tectonic unit that concerns us here includes a much broader group of north- trending fault-block mountains, with their intervening basins, the whole reflecting major vertical movements. This tectonic unit, its parts very likely related in time and origin, is about 100 miles (160 km) wide in southwestern New Mexico, narrows to about 50 miles (80 km) at the latitude of Albuquerque, N. Mex., and then broadens to about 100 miles (160 km) again north of Santa Fe. In Colorado the tectonic relations are more obscure, as the Rio Grande trough cuts across the volcanic field of the San Juan Mountains and strongly developed northwest- trending fault blocks of the Front Range farther north- east. The two largest basins at the southern end of the tectonic unit, the Tularosa Valley to the east and the Jornada del Muerto to the west, die out northward. In part this termination resulted from displacement along faults in the set that passes northwest through the Guadalupe and southern Sacramento uplifts. As the San Andres Range dies out northward, the broader structural high of the Sacramento uplift east of the Tularosa Valley continues northward to join the southern end of the Sangre de Cristo Range. Fluorite deposits apparently are scarce along the eastern flank of the Tularosa Valley and the Sacramento uplift, but the deposit reported from Torrance County, N. Mex., is in the broad structural high that lies between the Sacramento uplift and the Sangre de Cristo Range. A belt of north-trending faults, nearly 20 miles (30 km) wide, lies along the western side of this part of the Rio Grande valley and is the westernmost part of the broad tectonic system. Faults of this belt form the eastern boundary of the Black Range and the southwestern boundary of the San Mateo Mountains and are found throughout the length of the intervening Sierra Cuchillo south nearly to the Cooks Range. Fluorite deposits are widespread in this belt, especially in the part north of Hillsboro. Fluorspar has been mined in the Sierra Cuchillo from typical epithermal veins in volcanic rocks and from replacement deposits in limestone of Paleozoic age, some of which are mantos without silicate gangue minerals and some are bedding-controlled replacement deposits with silicate gangue minerals (pyrometasomatic deposits). The northern part of the Rio Grande trough in Colorado is very complex structurally and has the basins in an en echelon series of northwest-trending fault blocks that make up the Front, northern Sangre de Cristo, Sawatch, Mosquito, and Gore Ranges, and the entire complex is crossed in a northeasterly direction by intrusives and ore deposits of the Colorado mineral belt. As at the south end of the trough in southern New Mexico, this intersection of tectonic units is the locus of a large number of fluorite deposits. Here also, high-angle faults have localized fluorspar deposits. The largest deposits are along the Arkansas River graben, at the faulted north end of the North Park basin, near caldera walls in the San Juan Mountains, and near northwest-trending breccia reef faults, as at Jamestown. Possible northwest-trending belts of deposits may reflect structural control by faults with that trend, or may reflect greater exposure of older geologic features that resulted from uplift along the faults. The northeast-trending zone of intense faulting that crosses the Sierra Madre Occidental volcanic field near the International boundary is one of the major structures that influenced fluorite mineralization; it is both structurally complex and varied in its effects. It is rather sharply marked from the southwestern corner of New Mexico southwestward almost to Hermosillo, Sonora, Mexico, by the abrupt ends of segments of the volcanic field to the COLORADO-NEW MEXICO-WESTERN TEXAS-SOUTHEASTERN ARIZONA 53 south. North of this northeast-trending boundary the mountain ranges are rather short and commonly oriented north-northwesterly. Fluorite deposits have been found in this region on both sides of the International border. To the northeast, along the trend of the zone, there is much less evidence of faulting, but fluorite deposits are parti- cularly abundant where this zone crosses the Texas lineament and the major structures related to the Rio Grande trough between Lordsburg and Silver City. Some of the fluorite-bearing veins in these places strike north- east, an orientation that is less common than the north- west orientation prevalent elsewhere in the southwestern United States. The only important fluorspar district in New Mexico that is not in either the Rio Grande trough system or the Texas lineament is in the Zuni Mountains. This range is an upfolded and upfaulted block in which rocks of Pre- cambrian age have been brought to the surface. It is elongated northwesterly and very likely is not closely related to other structures of the region, although a cluster of fluorspar deposits, the frontal fault of the Sacramento uplift, and a constriction in the main chain of basins of the Rio Grande trough in the Socorro area are on the south- east projection of the Zuni Mountain trend. The Zuni Mountains are part of a late Paleozoic high area; possibly their trend reflects the older structure. Most of the fluorspar districts in Colorado, New Mexico, western Texas, and southeastern Arizona have been idle during the last 15 years, because of the low price paid for the commodity and the competition of foreign mines. Few deposits have been exhausted. With the increased prices of recent years, however, the known deposits will be reopened and new ore bodies will be found in most of the districts. The discovery in 1968 of a large deposit in the central Sierra Cuchillo, N. Mex., a range with very little earlier production, indicates that the minor districts cannot yet be disregarded as important sources of ore. Fluorspar is currently (early 1973) being mined in New Mexico in the Winkler anticline, Hidalgo County, and exploration in the State is under way in the Gila and Cooks Peak districts, in the Sierra Caballos, at Iron Moun- tain, at Holt and Whitewater Canyons, Catron County, and in the Bitter Creek area, Grant County. Good possibilities exist for the discovery of large low- grade deposits of the replacement or manto type in sedi- mentary rocks in this region. Deposits of this type have been described in New Mexico and Texas. In New Mexico, various Paleozoic limestones are the host rock in the northern Sierra Caballos, Sierra County; the Magdalena Group of Pennsylvanian age is found at the Tortugas and Bishop Cap deposits, Dona Ana County, and at deposits in the Sierra Cuchillo, Sierra County (Rothrock and others, 1946). Fluorspar is in Permian sandstones and siltstones of the Yeso Formation in the Gallinas district, Lincoln County (Perhac and Heinrich, 1964) and accompanies uranium minerals in the Todilto Limestone of Jurassic age near Grants, Valencia County. Lower and Upper Cretaceous limestones were replaced along three favorable horizons in the Eagle Mountains, Hudspeth County, Tex. (Gillerman, 1953) and in the Christmas Mountains- Corazones Peak district, Brewster County, Tex. (McAnulty, 1967). Large, low-grade fluorspar deposits may form in certain sedimentary materials, for Tweto and others (1970, p. C69-C74) have described the Hammer deposit, Summit County, Colo., where Tertiary colluvium was mineralized by hot springs and then was broken up and diffused by Pleistocene landslides to form disseminated deposits. DESCRIPTIONS OF INDIVIDUAL DISTRICTS AND DEPOSITS COLORADO JAMESTOWN The Jamestown, Colo., fluorspar deposits are in the Front Range, in the central part of Boulder County 17 miles (27 km) northwest of Boulder. Mining activity in the district began in 1874, and until 1940 the major commodi- ties were gold, lead, and silver. Fluorspar was a byproduct of lead and silver mining as early as 1874, with the first appreciable production coming during the period 1903 to about 1920. In 1940 two companies, Allied Chemical Corporation and Harry Williamson and Son, began large- scale production of metallurgical: and acid-grade fluorspar. Allied Chemical Corporation's total produc- tion to 1972 of about 700,000 tons of acid-grade fluorspar (data published with permission) has been almost entirely from one mine, the Burlington, which has been in continuous operation since 1942. Harry Williamson and Son produced some fluorspar, mainly from the Emmett mine, and some from the Argo and Blue Jay mines, during 1940-56. All deposits in the district have been developed by underground workings, several to depths of 400 feet (120 m) or more; the Emmett mine, now closed, is developed to a depth of 1,000 feet (300 m) and the Burlington to a depth of 1,400 feet (430 m), the present (1972) lower working level. For detailed geologic descriptions of the district see Goddard (1946), Lovering and Goddard (1950), and Kelly and Goddard (1969). Rock associations in the Jamestown district are char- acteristic of the Front Range, a Precambrian igneous and metamorphic complex containing numerous areas of high-level Laramide and middle Tertiary intrusions and related mineral deposits. The Precambrian rocks are Silver Plume Granite with scattered inclusions of metamorphic rocks, mainly biotite schist and hornblende gneiss (fig. 20). These rocks are intruded by a Laramide stock of horn- blende granodiorite composition about 4 mi? (10 km?) in area, and a middle Tertiary stock of alkalic granite-quartz monzonite composition about 1 mi" (2.6 km?) in area (Goddard, 1946, p. 12). The alkalic stock, intruded at the 54 40°10 GEOLOGY AND RESOURCES OF FLUORINE 105°25" | 7 taa Area containing numerous breccia zones // 3 - xf \ Jif # Yop \ ,/ A/ G7 yh l// r/’ / / Pras ar" / 0 1 MILE 0 1 KILOMETRE EXPLANATION Argentiferous galena vein s : Tertiary alkalic granite-quartz monzonite Fluorspar vein Pyritic gold vein Tertiary granodiorite Gold telluride vein Precambrian rocks Breccia reef Ficure 20.-Jamestown district, Colorado (modified from Lovering and Goddard, 1950, fig. 77, and Goddard, 1946, pl. 1). COLORADO-NEW MEXICO-WESTERN TEXAS-SOUTHEASTERN ARIZONA 55 north end of the granodiorite stock, is a composite intrusive. Fluorspar deposits at Jamestown are related to this alkalic stock. Alkalic dikes are common in the district; they were intruded late in the igneous cycle and in places are younger than the fluorite (Goddard, 1946, p. 12). Fluorite is a common mineral in the alkalic granite-quartz monzonite and also occurs in crosscutting silicified zones within the stock. Zones in which all rock types were altered to clay minerals, sericite, chlorite, and carbonates are common, and extend beyond known mineralized zones. Similar altered rock is common in the contact zone of the alkalic stock. Country rock alteration was minor in the contact zone of the granodiorite stock. Jamestown is at the northeastern end of the Colorado mineral belt, a northeast-trending zone in which Pre- cambrian basement structures and Laramide structures controlled the location of Laramide and Tertiary intrusives and mineralization. This part of the Front Range is characterized by a series of strong northwest- trending faults, called breccia reefs, of Late Cretaceous age. These were important in the localization of ore deposits in several districts (Lovering and Goddard, 1950, p. 237). Two significant breccia reefs cut the Jamestown district just west of the major fluorspar deposits. They contain minor amounts of clear fluorite; at places in the district pyritic gold veins and gold telluride veins radiate from breccia reefs. Localization of the major fluorspar deposits, however, probably was controlled by fractures forming breccia zones different from the breccia reefs and created by the forceful intrusion of the alkalic stock (Goddard, 1946, p. 13); the deposits were not localized by the breccia reefs. Fluorspar bodies in the Jamestown district are localized mainly in breccia zones and fissures that form radiating and ring fracture zones around the alkalic stock; fluorspar bodies also occur in fractures of regional northeast and northwest sets. The Jamestown mining district contains four types of mineral deposits characterized by argentiferous galena, by fluorspar, by pyritic gold, and by gold telluride; all are about the same age, and all are genetically related. Pitch- blende is present in some deposits. Lovering and Goddard (1950, p. 260) pointed out a rough zonal arrangement of these deposits around the alkalic stock; the sequence of zones outward from the stock is argentiferous galena, fluorspar, pyritic gold, and gold telluride. Argentiferous galena occurs mainly in massive pipelike bodies, broken blocks, and individual grains scattered through fluorspar-bearing breccia zones, but also it occurs as massive lenses and disseminated grains in fluorspar and pyritic gold veins. Most of the fluorspar is found in veins and large breccia zones around the western and southern edges of the alkalic stock. Minor fluorite also occurs in most pyritic gold and gold telluride veins within 1 mile (1.6 km) of the alkalic stock. Several generations of fluorite formed in the James- town district and most are separated by periods of brecciation. Pyritic gold veins occur throughout the district and cut all rock types, including alkalic granite-quartz monzonite. They are best developed outside the fluorspar zone and as far as 2 miles (3 km) from the alkalic stock. In mines where both pyritic gold and gold telluride veins are present, the gold telluride veins are demonstrably younger (Goddard, 1946, p. 12). Most gold telluride veins are 1.5-3 miles (2.5-5 km) from the alkalic stock. However, the Buena mine, mainly on gold telluride veins, is just outside the major fluorspar zone on the west side of the alkalic stock. Tellurides have been found in many fluorspar deposits and in pyritic gold veins. Gold telluride veins are not confined to the James- town district; they have wide distribution in the northeast part of the Colorado mineral belt (Kelly and Goddard, 1969). The major fluorspar ore bodies are in breccia zones, best developed in Silver Plume Granite along the western contact of the alkalic stock. The breccia zones are hundreds of feet in diameter, irregular in shape, and not related to known fault structures. Fluorspar in the breccia zones occurs in nearly vertical pipelike bodies that in plan are irregular in shape but tend toward lenticular, and are 10-100 feet (3-30 m) wide and as much as 500 feet (150 m) long (fig. 21). In the Burlington mine a vertical exposure EXPLANATION Fluorspar vein containing 60-90 percent CaF, Fluorspar breccia containing. 30-60 percent CaF, Breccia containing less than 30 percent CaF, Silver Plume Granite, fractured and altered 0 100 FEET 0 25 METRES FicurE 21.-Fluorspar deposit at the tunnel level of the Burlington mine, Jamestown district, Colorado (workings not shown). From Goddard (1946, pl. 8). 56 GEOLOGY AND RESOURCES OF FLUORINE of one pipelike fluorspar body is more than 1,400 feet (420 m). Commonly the fluorspar ore bodies consist of a lenti- cular central core of a breccia composed of 60-90 percent CaF, and minor wallrock, surrounded by an irregular zone of fluorite-bearing breccia containing 5-60 percent CaF, (Goddard, 1946, p. 20). Large areas of breccia containing 5-30 percent CaF, are common. The fluorspar mineral- ized bodies are surrounded in most places by unmineral- ized but silicified breccia composed of altered wallrock. Fluorspar veins several inches to 20 feet (6 m) wide and 150-1,000 feet (45-300 m) long are common around the southern and western edges of the alkalic stock, mostly in granodiorite. Goddard (1946, p. 20) reported that the veins commonly contain 60-85 percent CaF,. Silica content of Jamestown fluorspar, especially of low-grade ore, is relatively high. A high-grade concen- trate is readily obtained by flotation methods as the fluorite is generally not intimately intergrown with quartz or chalcedony. Most silica present is in wallrock fragments. Many of the fluorspar ore bodies contain enough lead-, gold-, and silver-bearing sulfides to make saving of a sulfide concentrate potentially profitable. Goddard (1946, p. 20) reported two instances of Jamestown fluorspar yielding sulfide concentrates. At the Lehman mill in 1933 a galena-pyrite concentrate containing 30 percent lead and 0.28 ounce of gold and 3.6 ounces of silver to the ton was recovered from jigs and tables. At the Wano flotation mill in 1942 and 1943 a sulfide concentrate containing about 30 percent lead and 20-30 ounces of silver to the ton was recovered. Copper has also been recovered as a byproduct from Jamestown ore (Goddard, 1946, p. 20-21). Semi- quantitative spectrographic analyses of 38 samples of fluorspar from the Burlington ore body, a pipelike body in a breccia zone, indicate an average of 33 ppm Ag, 3,700 ppm Pb, 4,850 ppm Ce, and 1,800 ppm Nd. Semiquanti- tative spectrographic analyses of 35 flourspar samples from the rest of the district, mainly from veins, indicate an average of 9 ppm Ag, 2,600 ppm Pb, 1,150 ppm Ce, and 580 ppm Nd. Fluorspar from breccia zones and veins alike is every- where brecciated, and it exhibits at least two phases of fluorite and evidence of several events of brecciation. Generally the fluorspar ore consists of angular crystalline fluorite, sulfide fragments, and angular wallrock frag- ments, all less than an inch (2 cm) in diameter, set in a very fine grained matrix of rounded and corroded fluorite, clay minerals, carbonate minerals, and quartz. In breccia zones the fluorspar material is as described above, but it encloses large blocks of altered wallrock as much as several feet in diameter. Most of the fluorite is crystalline, light violet to dark purple, and is commonly zoned. Fluid-inclusion studies (J. T. Nash, oral commun., 1970) of fluorite from the Jamestown district indicate homogenization temper- atures in the range 350%-400°C. Fluid salinities are commonly high, 40 weight percent or more. Many fluid inclusions contain several "daughter" minerals. These data suggest that fluorite deposition took place from a hot, highly saline solution at high pressure. Pyrite, gal- ena, sericite, and books of euhedral biotite, all fine grained, are disseminated through the fine-grained matrix. Sphalerite, chalcopyrite, tennantite, and enargite are locally present. Sulfides also occur as massive pipelike bodies in breccia zones and as massive lenses in veins. Chalcedony and hematite are locally present as a late breccia cement. Some low-grade fluorspar breccias con- tain abundant celestite, which also occurs with green fluorite in vugs. Ankerite, arfvedsonite, calcite, biotite, adularia, and purple fluorite crystals that display coarse- grained pegmatitic texture compose some fluorspar veins southwest of the alkalic stock. Minute grains of pitch- blende have been identified in fluorspar from several de- posits (Goddard, 1946, p. 19). The Jamestown fluorspar deposits are characterized by fluorite crystals, both massive and granular, in aggregates intergrown with the other minerals. Conversely, columnar-fibrous fluorite, monomineralic layering, and crustification are not present. The depositional sequences were complex; more than one mineral phase was deposited from each hydrothermal surge, in contrast to the mono- mineralic depositional phases of low-temperature layered fluorspar deposits. The alkalic magma was genetically related to, if not the ultimate source of, fluorine-bearing hydrothermal solu- tions that deposited the fluorite. The alkalic granite- quartz monzonite contains 0.05 to 2.3 percent F (average 0.27 percent) based upon analyses of 36 samples collected throughout the stock. Similar calc-alkalic or alkalic intrusives in receptive host rocks would be excellent exploration guides elsewhere, especially if they contained similar anomalous amounts of fluorine. The potential of the Jamestown fluorspar deposits is enhanced by the presence of large areas of mineralized breccia, albeit low grade, and possible byproducts of lead, silver, gold, and of rare-earth elements that are present in the fluorite. NORTHGATE The Northgate fluorspar district lies in the north- eastern end of the North Park intermontane basin near the Colorado-Wyoming boundary line. Walden, Jackson County, Colo., is 10 miles (16 km) south of the district and Laramie, Wyo., is about 50 miles (80 km) east. A major paved highway between Walden and Laramie (Colorado 127-Wyoming 230) passes through the district, as does a Union Pacific Railroad branch line originating in Laramie. Although prospectors for copper noted veins in the district as early as 1900, no fluorspar claims were staked until 1918 and there was no recorded production until 1922. Colorado Fluorspar Corporation produced about 15,000 tons of metallurgical-grade fluorspar between 1922 COLORADO-NEW MEXICO-WESTERN TEXAS-SOUTHEASTERN ARIZONA 57 and 1927 from the Fluorspar vein system. The district was then idle until 1941 when the Western Fluorspar Corpora- tion leased the Colorado Fluorspar Corporation holdings, and a 450 ton-per-day capacity sink-float mill was con- structed by Defense Plants Corporation. This operation ceased in 1945 after about 115,000 tons of fluorspar had been produced from underground and open-pit opera- tions on both the Fluorine and Fluorspar vein systems. During this same period about 700 tons of fluorspar was mined from the Camp Creek mine on the northern end of the Fluorine vein system by Kramer Mines, Inc. In 1950 the Aluminum Company of America acquired the Camp Creek property, but to date (1972) has produced no fluorspar. In 1951, Ozark-Mahoning Co. acquired control of most property on both vein systems, with the exception of the Camp Creek and Penber mines. Following construction of a 300-ton-per-day flotation concentrator in 1952, until operations ceased in 1959, Ozark-Mahoning produced fluorspar from three sources in the following proportions: 70 percent from open pit, 18 percent from sink-float tailings, and 12 percent from mine development. Ozark- Mahoning Co. expanded their Northgate facilities in 1968 to 650 tons per day capacity and also added a concentrate briquetting facility. Since this enlarged facility began operation, Ozark-Mahoning has been producing fluorspar from underground mine work along the Fluorspar vein system and from underground mine and open pit work along the Fluorine vein system. The follow- ing geologic description of the Northgate district is large- ly from Steven (1960). Country rocks at Northgate are quartz monzonite and hornblende-biotite gneisses of Precambrian age; shales, sandstones and minor limestone of Mesozoic age; and arkose, tuffaceous siltstones and clays of the White River Formation of Oligocene age (Steven, 1960, p. 336). The bulk of the fluorspar occurs in veins within Precambrian quartz monzonite and within arkose of the White River Formation, and only minor fluorite occurs in faults in Mesozoic sedimentary rocks and in veins in Precambrian metamorphic rocks. Silicification and pyritization of the quartz-monzonite wallrock were common along the Fluorspar vein system, and argillization and silicification of the White River Formation were common along the Fluorine vein system. The dominant structural feature of the district is the Independence Mountain fault, an east-trending and northward-dipping reverse fault with a hanging wall of Precambrian rocks and a footwall of Mesozoic sedi- mentary rocks. This fault, which predates the White River Formation (Steven, 1960, p. 369), transects earlier Laramide structures and forms the abrupt north boundary of the North Park intermontane basin. Fluorspar occurs as veins in two parallel northwest- trending fault zones about 2 miles (3 km) apart that post- date the White River Formation. These fault zones are in the hanging wall of the Independence Mountain fault. The western (Fluorspar) vein system extends slightly more than a mile (1.6 km) northwest from the Independence Mountain fault and is mineralized throughout most of its length; the eastern (Fluorine) vein system extends more than 2 miles (3 km) and is mineralized also along most of its length, though sparsely in places. Fluorite was deposited through a vertical range of more than 1,050 feet (320 m). In the Fluorspar vein system fluorspar occurs as veins and veinlets a fraction of an inch to a few feet (millimetre to a metre) in width, as breccia "cement'' coating breccia fragments, and as breccia fragments, all in coarse-textured fault breccia. Commonly, ore-grade fluorspar occurs in a zone, up to 20 feet (6 m) wide, of broken and cemented blocks of wallrock and fluorspar (breccia) containing veins of nearly pure fluorspar along or near one or both edges. Assays of samples of the fluorite stringers run as high as 97 percent CaF, (Steven, 1960, p. 399). The average for the vein zone for widths of 8-11 feet (2.4-3.4 m), however, runs between 40 and 50 percent CaF,, and the range is from 10 to 90 percent CaF, (Steven, 1960, p. 399). Mineralogy of the fluorspar in the Fluorspar vein system is simple. Fluorite and chalcedony are the major minerals with some quartz and very minor barite, manganese oxides, and pyrite. Clay is common but is thought to be a weathering product of fine-grained fault gouge. Monomineralic layers ("bands'"), crusts, and botryoidal layers with numerous drusy cavities are the characteristic forms of fluorite occurrence. Watercourses lined with fluorite are common. Columnar green fluorite, columnar honey-colored fluorite, crystalline purple fluorite, dense black fluorite, milky-white chalcedony, bladed barite, and manganese oxides occur in distinct layers where each layer is composed almost entirely of one mineral or form of mineral (fig. 22). A depositional sequence for the Fluorspar vein system is shown in the following table. Only the events marked "major" on the table occurred throughout the vein system; the others were of a local nature. Depositional sequence of fluorspar in the Fluorspar vein system, Northgate, Colorado [Listed in order of increasing age. X, period of brecciation; m, major] Thin coatings of chalcedony, honey-colored columnar-fibrous fluorite, and manganese oxides. X - Gray to black siliceous fluorite, mainly as a breccia cement. Contains stringers of pyrite. Part of the wallrock alteration occurred during this stage. (Major.) mX Honey-colored columnar-fibrous fluorite in veins and as breccia cement. (Major.) X - Alternating bands of gray siliceous fine-grained fluorite, white chalcedony and small stringers of barite and manganese oxide. X - Siliceous fine-grained breccia with some stringers of barite. (Occurs throughout, but in places is very minor.) 58 mX Alternating bands of clear and purple crystalline fluorite, gray siliceous fluorite, and minor barite. X Gray siliceous breccia with stringers of pyrite. Mainly a dense black siliceous fluorite. Part of the wallrock alter- ation occurred during this stage. (Major.) mX X - Columnar-fibrous green fluorite, white chalcedony, fine- grained gray siliceous fluorite, and gray siliceous breccia in alternating layers. Fluid inclusion studies (J. T. Nash, oral commun., 1970) of fluorite from the Fluorspar vein indicated homo- genization temperatures of 130°-175° and salinities of 0.1-3 weight percent. Fluid inclusions here are small and sparse. The data suggest that the mineralizing solution was a hot springs type consisting of predominantly meteoric water. The Fluorine vein system is larger than the Fluorspar vein system and is as much as 70 feet (20 m) wide-but tends to be more diffuse and branching. The highest assays quoted by Warne (1947, p. 15-20) for sample widths of about 5 feet (1.5 m) are 70-80 percent CaF,. Mineralized rocks in the White River Formation contain about 30 percent CaF, for 15-foot (4.5-m) widths and between 20 and 25 percent CaF, for 35-foot (10.5-m) widths. Where quartz monzonite constitutes the wallrock, vein material averages about 45 percent CaF, for widths of 25-30 feet (7.5-9 m), and vein material averages about 30 percent CaF, for widths of 35-45 feet (10.5-14 m) (Warne, 1947, p. 15-20). In the Fluorine vein system most of the fluorspar EXPLANATION 1 Layered columnar-fibrous green fluorite 2 Gray siliceous breccia with wallrock and fluorspar fragments and siliceous fluorite cement 3 Columnar-fibrous honey-colored fluorite 4 Gray fine-grained siliceous fluorite Ficur® 22.-Slab of fluorspar from the Fluorspar vein system, North- gate, Colorado, showing different types of fluorite. GEOLOGY AND RESOURCES OF FLUORINE occurs as earthy masses and stringers in fine-grained fault breccia, and as fracture and space filling in arkosic beds of the White River Formation. Drusy cavities are common but massive veins of pure fluorite are not. The fluorspar consists mainly of fluorite and chalcedony with some quartz and manganese oxides. Montmorillonite, in many places intergrown or mixed with the fluorite and fault gouge, is common throughout, and crystalline calcite is common in mineralized arkose of the White River Formation. Pure fluorite from both vein systems contains trace amounts of barium, manganese, strontium, and yttrium as do fluorites elsewhere. Laser probe analyses also indicate that trace amounts of zinc, copper, and silver occur in manganese-oxide zones, in scattered places in some fine- grained gray breccias, and in some minor columnar green fluorite layers. Molybdenum occurs in anomalous amounts in the wallrocks surrounding the Fluorspar vein and in places within the vein as described by R. G. Worl (U.S. Geol. Survey, 1970, p. A5). IIsemannite, reported in parts of the Fluorspar vein system (Goldring, 1942), may be jarosite stained blue by molybdenum trioxide. The source of the molybdenum anomaly is molybdenite, widely disseminated through siliceous black material that is a submicroscopic intergrowth mainly of fluorite, chalcedony, and pyrite. This molybdenum-bearing material occurs as minor stringers and breccia fragments within the vein zone and as a pervasive stockwork of veinlets in the wallrocks around the fluorspar vein. The molybdenum-bearing material was the only vein material that penetrated the wallrocks away from major fractures. The Northgate fluorspar district is not closely associated with Tertiary volcanic rocks, but such rocks are present in the region and probably the deposits formed during the episode of middle Tertiary volcanism that was responsible for the volcanic rocks of the region. OTHER DEPOSITS Fluoritic pegmatites of Colorado are in or near the Pre- cambrian Pikes Peak batholith. Pikes Peak Granite contains 4,040 ppm fluorine (average of 38 samples; Fred Barker, written commun., 1972), much of it as fluorite. Although the total tonnage of fluorite in the batholith may be very large, the concentration of the mineral is so low that little, if any, will ever be recoverable. This dispersed fluorite increases the fluorine content of ground water, in places to objectionable levels, causing mottled teeth in the Colorado Springs area. The Pikes Peak Granite contains, in addition to feldspars, quartz, and biotite, small amounts of the cerium minerals allanite and bastnaesite. The Pikes Peak Granite locally also contains veins of quartz or of quartz and feldspar that contain crystals of topaz, fluorite, amazonite, or other minerals that are of value to mineral collectors but not to industry. The main economic importance of the veins and pegmatites is as indicators of mobile metal and fluorine COLORADO-NEW MEXICO-WESTERN TEXAS-SOUTHEASTERN ARIZONA that somewhere in the area may be concentrated into minable deposits. A Precambrian topaz-quartz-sillimanite gneiss in the central part of the Front Range, in Jefferson and Clear Creek Counties just west of Denver, contains major amounts of topaz. This unit, part of a high-grade meta- morphic gneiss and granitic plutonic complex, is 11-100 feet (3.4-30 m) thick and crops out along strike for 7,000 feet (2,150 m). Three composite chip samples of this unit, taken by Sheridan, Taylor, and Marsh (1968, p. 3-6), contain 67, 23, and 36 weight percent topaz and, respectively, 4.2, 3.9, and 2.2 weight percent rutile. Modal analyses of four samples of the topaz-bearing gneiss indicate by volume 5.5-68.2 percent quartz, 30.1-71.9 percent topaz, 1.7-3.1 percent rutile, 18.2 percent sillimanite in one sample, 13.0 percent apatite in one sample, and trace amounts of zircon and muscovite. Rutile concentrates contain 0.05-0.15 percent Nb and 0.3-0.7 percent Fe as determined by semiquantitative spectro- graphic analyses. Sheridan, Taylor, and Marsh (1968, p. 7) suggested that the gneiss warrants investigation as an ore of rutile, especially because it might also yield a topaz or topaz-and-sillimanite product, which is salable as raw material for refractories and ceramics. NEW MEXICO BURRO MOUNTAINS The Burro Mountains district is in southwestern Grant County, 9-30 miles (14.5-48 km) southwest and south of Silver City. Fluorite was recognized in the mineral deposits in the area when mining of copper and turquoise began in the 1870's. Fluorspar was first mined there in the 1880's for use as flux in the copper smelters. More . significant fluorspar mining began at the Burro Chief mine in 1913 and continued with interruptions until the 1950's. Information on the geology of the district was obtained from Gillerman (1953; 1964) and Rothrock, Johnson, and Hahn (1946). The Burro Mountains are geologically complex. An up- lifted fault-block mass of crystalline rocks of Precambrian age is intruded by dikes and stocks and overlain by vol- canic rocks of Cretaceous and Tertiary age. The crystal- line rocks were separated into four blocks by three persis- tent northwest-striking faults that may be elements of the Texas lineament. In the fluorspar areas these faults are accompanied by northeast- and east-trending cross-faults and north-trending splits of the main faults. The fluorite is found along these north-trending minor faults. One deposit is in volcanic rock and the others are in granite and schist; the deposits formed during the region's Tertiary volcanism. Precambrian granite is the most favorable host rock; hence the deposits are associated spatially with granite and temporally with volcanism. The largest deposit, the Burro Chief, now within the Tyrone open-pit copper mine, is in a northeast-trending 59 breccia zone that is 10-100 feet (3-30 m) wide in granite. The deposit earlier had been opened by a shaft to a depth of 700 feet (210 m). Four types of ore have been distin- guished: (1) high-grade massive vein fluorspar contain- ing 75-90 percent CaF; (2) soft fluorspar breccia contain- ing 35-75 percent CaF»; (3) hard fluorspar breccia con- taining 50-90 percent CaF;; and (4) breccia and other low- grade ore containing as much as 35 percent CaF,. Where unbrecciated, the fluorspar is clear, green, and moderately coarse grained. Clay and quartz are the most common impurities. The hard fluorspar breccia consists of an earlier variety with a green fluorite cement that is found in the upper part of the mine and of a later variety with a purple fluorite cement that is found throughout the ore zone. GILA The Gila fluorspar district in Grant County encompasses about 5 mi? (13 km?) along the Gila River 5 miles (8 km) upstream from the town of Gila. It was the locus of early fluorspar mining in New Mexico when the Foster mine was opened in the 1880's. Though small in area, the district has provided more than 47,000 tons of fluorspar (Williams, 1966). About 29,000 tons of this came from the Clum mine, the biggest producer during World War II. The fluorspar deposits are in latitic and andesitic rocks of the Datil Formation of Tertiary age. The ore deposits were formed along faults and breccia zones, mainly as open-space fillings, and minor replacement of the volcanic rocks. The rocks were strongly argillized in some places and quartz was deposited with fluorspar in and near the veins. Most veins strike north to north-northwest. The Foster vein strikes northeast and the Cedar Hill, Clum East, and Green Spar veins strike northwest. Persistent northwest-striking faults separate the volcanic rocks from the younger unmineralized Gila Conglomerate farther southwest and cut the volcanic rocks northeast of the mining district. The ore deposits have not, in general, been completely delimited. The Clum North ore body was at least 510 feet (155 m) long, 150 feet (45 m) deep, and 6 feet (1.8 m) thick. The main ore body mined at the Foster mine was at least 280 feet (85 m) long, 130 feet (40 m) in vertical extent, and 3-8 feet (1-2.5 m) thick. The prospects for dis- covery of additional ore seem good. Mining may be pre- vented, however, if the canyon of the Gila River is flooded by a proposed dam. COOKS PEAK-FLUORITE RIDGE The Cooks Peak district in Luna County is about 30 miles (50 km) northeast of Deming. The district is credited with a production of 58,011 tons of which about 56,780 came from the White Eagle mine and the rest from four other deposits (Williams, 1966). The district is underlain by a variety of rocks, including Precambrian granite, upper Paleozoic limestone, and Ter- 60 GEOLOGY AND RESOURCES OF FLUORINE tiary volcanic rocks. These have all been cut by faults that strike northwest or north. The White Eagle fluorspar deposits, like others in the Cooks Peak district, are in minor irregular, branching, brecciated, and overlapping fault zones of generally north- westerly trend (Rothrock and others, 1946, p. 98-102). The ore bodies have a total length of 850 feet (260 m) and have been mined through a vertical range of 175 feet (55 m). The veins are nearly all in Precambrian granite, but the western 100 feet (30 m) of one is in Tertiary rhyolite, and the fluorspar mineralization is believed to be related to the volcanic rocks. The fluorspar consists of white to pale- green fluorite with quartz and jasperoid impurities. Fluorspar in the north part of the Cooks Peak district (which includes the Linda Vista deposit; Williams, 1966, p. 35) occurs as stringers, pods, and veins in slightly brecciated volcanic rocks. These fluorspar bodies are generally small, less than 5 feet (1.5 m) wide, but are numerous and sporadically distributed over a large area. The Fluorite Ridge area in Luna County is near the southern end of the Cooks Range, about 15 miles (25 km) south of the Cooks Peak district and about 10 miles (15 km) north of Deming. The mining claims cover an area of nearly 1 m? (2.6 km?). Ten deposits are known in the district, which has yielded about 65,400 tons, 10 percent of the total of the State. The district is underlain by sedimentary rocks of Paleo- zoic and Tertiary ages, and monzonite porphyry that intrudes them. All these rocks were cut by faults, basalt dikes, and veins of fluorspar and quartz. The veins and dikes strike north to north-northeast. All but the deposits of the Tip Top and Hilltop Spar prospects are fissure veins, which are cavity fillings in limestone. The deposits have been most productive in broad breccia zones and at vein intersections. The veins contain coarse-grained pale- green fluorite and coarse- to fine-grained white, green, or red fluorite. Quartz and clay are the main gangue min- erals. SIERRA CUCHILLO Fluorite has been found in many places in the northern 20 miles (30 km) of the Sierra Cuchillo in Sierra County, the part of the range extending south from Monticello Box to a point about east of Winston. In most places the fluorite is present only in small amounts, but fluorspar prospects have been opened and shipments have been made from a few of them. Tactite deposits on Iron Mountain, Sierra County, have been known for many years and have been investigated as possible sources of iron, beryllium, tungsten, and fluorite (Jahns, 1944). The only ore shipped from the deposits has been fluorspar; less than 100 tons was shipped in the early 1950's. On the west side of Iron Mountain tactite replaces Carboniferous limestone that dips eastward into the mountain. The rocks that are richest in valuable minerals are found only in certain beds and near silicic intrusive masses. Most of these beds are truncated at a shallow depth by a large mass of intrusive igneous rock. The total ton- nage of valuable tactite therefore is much less than the extensive outcrops suggest. The tactite contains many silicate minerals including the beryllium mineral helvite, and much scheelite and magnetite. Fluorite in dispersed small grains makes up 5-35 percent of parts of the silicated rock. In one area fluorspar forms irregular tabular layers several feet thick and a few tens of feet long that yielded the ore that was shipped. The fluorite constitutes at least 75 percent of these layers, as white translucent grains several inches across. Several small fluorite-bearing veins in Tertiary vol- canic rock have been prospected a few miles north of State Highway 52 and near the center of the Sierra Cuchillo. The fluorspar is moderately coarse grained, brecciated, and mixed with quartz, country rock, and calcite. None of the veins has been an economic source of fluorspar. The veins, together with fluorite in dry-wash sediments in this part of the range, attest to widespread fluorite mineraliza- tion; deposits of economic size and grade may be present in sedimentary rocks below the volcanic rocks. The Victorio prospect is about 2 miles (3 km) east of Chise in the Sierra Cuchillo. It is a shallow shaft and several trenches opened in a vein that cuts limestone of Carboniferous age. The vein strikes N. 25° E. and is almost vertical. It consists of brown jasperoid, finely crystalline quartz, and fluorite and is 3-3.5 feet (about 1 m) thick. The largest known fluorspar deposit in the Sierra Cuchillo was discovered in 1969 by Ira Young, of Winston, N. Mex. It is about 1 mile (1.5 km) south of the Victorio prospect and 2 miles (3 km) east-southeast of Chise. Fluorite has replaced several beds in limestone and forms mantos that dip gently to the west. The strike length of the mineralized beds exceeds 1,000 feet (300 m) and mineral- ized rock extends at least 500 feet (150 m) downdip. At the surface the total minable thickness of fluorspar is at least 20 feet (6 m), and some mantos are more than 5 feet (1.5 m) thick. The individual mantos are separated from one another by sufficient unmineralized limestone to require the mining of each manto separately. The mantos contain both coarsely crystalline white to pale-green fluorite and very fine grained friable white "sugar spar." Barite is uncommon. The most abundant impurity is unreplaced limestone. The Ira Young deposit and the Victorio prospect are near the south ends of two major north-trending faults that occur in Canada Rancho de los Chivos and Montoya Canyon (Alminas and others, 1972). Fluorite can be found in concentrates panned from gravel from most tributaries COLORADO-NEW MEXICO-WESTERN TEXAS-SOUTHEASTERN ARIZONA 61 that drain into these valleys as well as from the two main valleys. The wide extent of the favorable limestone forma- tion and of fluorite mineralization suggests that addi- tional deposits may be found in the area. ZUNI MOUNTAINS The Zuni Mountains in Valencia County form a low ridge that extends about 50 miles (80 km) southeastward from the vicinity of Gallup. Fluorite was discovered in them in about 1908 and mining began in 1937. About 224,000 tons of fluorspar has been shipped. The Zuni Mountains consist of a broad dome of Pre- cambrian crystalline rocks and Permian sandstone and limestone flanked by Mesozoic sedimentary rocks (Goddard, 1966). The rocks of Precambrian and Permian ages are cut by many steep dip-slip faults that trend north- west. In most places such faults separate the Precambrian rocks from younger rocks. A persistent north-trending fault cuts the Precambrian rocks and offsets the northwest- trending faults. The northwest-trending faults are strongly silicified fracture zones that are stained red by fine-grained hematite. Fluorite is present in them in a few places. The fluorspar veins in the Zuni Mountains are found in three groups, one near each end of the area of Precam- brian rocks and one in its interior. All veins are within 1,000 feet (300 m) of northwest-trending faults. The veins strike northeast or east and are mineralized to depths as much as 800 feet (240 m) and through widths as great as 15 feet (4.5 m). They contain coarse-grained green and fine- grained purple fluorite; much of the ore is high-grade fluorspar containing 85-93 percent CaF;. Calcite, barite, and rare sphalerite, galena and chalcopyrite are accessory minerals. Three mines in the district have yielded important amounts of fluorspar. In addition to the mine worked to a depth of 800 feet (240 m) two were worked to depths of 350 and 600 feet (105 and 180 m), with no diminution of either size or grade in the veins. SIERRA CABALLOS Fluorspar deposits have been explored along a distance of about 20 miles (30 km) in the Sierra Caballos, Sierra County. The deposits are clustered into a northern group, near and north of Palomas Gap, and a southern group. The northern group has been most productive, providing about 25,000 tons of fluorspar; the Illinois and Indepen- dence or Blue Jacket mines yielded about 21,000 tons of this production (Williams, 1966). The northern Sierra Caballos group of deposits is in an area of Precambrian granite and Paleozoic sedimentary rocks. The rocks were much faulted on the west side of the range, and open fractures and breccia zones formed, in which fluorite was deposited. The veins near Palomas Gap strike northeast and those farther north strike north- west. The Illinois mine achieved its large production from three parallel veins, the Universal, the Oakland, and the White Star. They are mainly in Paleozoic sedimentary rocks (Johnston, 1928; Rothrock and others, 1946). The Oakland vein has been traced for 1,600 feet (500 m) and continues as a vein of quartz for an additional 600 feet (180 m). Fluorspar was mined along 350 feet (105 m) of the vein, with an average width of 11 feet (3.3 m) and average CaF, content of 50 percent. Barite and calcite are gangue minerals. The other two veins are shorter, but the Univer- sal vein has been mined through a distance of 400 feet (120 m) and a vertical range of 200 feet (60 m). The average width was 9 feet (2.7 m) and the average CaF, content about 50 percent. The fine- to medium-grained white and pale-green fluorite is intergrown with calcite and quartz. The Independence or Bluejacket mine was opened on a breccia zone that strikes N. 15°-20° W., between granite and sandstone. Fluorspar was mined from an opencut about 100 feet (30 m) long. The ore, like that in the Illinois mine, contains quartz and minor amounts of barite. Most fluorspar deposits in the southern group are on the west side of the Sierra Caballos and a few are on the east side. The total fluorspar production has been about 9,000 tons, of which 5,800 tons came from the Alamo (Nakaye) mine and 2,500 tons from the Lyda-K mine (Williams, 1966). The Alamo or Nakaye mine was opened in fluorspar veins that strike northeast and northwest and occupy fractures in limestones of the Magdalena Group (Rothrock and others, 1946). The veins, opened by shafts and adits, contain brecciated limestone cemented by pale- green fluorite and calcite, as well as fluoritic fissure fillings. The Lyda-K mine exploits a fault-controlled vein in granite of Precambrian age; the vein is at least 2,000 feet (600 m) long. Recurrent fault movement has brecciated the early quartz-fluorite mixture, which was then cemented with purple fluorite. Shafts in the vein exposed it to a depth of 260 feet (80 m). GRANTS According to Peters (1958, p. 669-670), small bedded re- placement deposits of fluorite are closely associated with uranium deposits in Jurassic limestone near Grants, N. Mex. "In some of the bodies, fluorite and calcite are the dominant minerals, with fine-grained uraninite incorporated in the fluorite crystals The fluoritic uranium bodies lie within the Todilto limestone and locally within limestone lenses in the overlying shales of the Summerville formation." They are mostly in the upper half of the Todilto, which is generally about 15-25 feet (4.5-7.5 m) thick. According to Peters (1958, p. 669-670): The fluorite bodies within the Todilto limestone are irregular in out- line. Some of the bodies have no apparent structural control; others 62 exhibit a relation to the regional fold and fracture pattern. Some of the bodies have a tendency to follow the local joint pattern as mantos or as a linear series of tabular lenses. An association exists between broad gentle folds and some of the mineralized zones in that ore is most commonly found in crest or trough areas. A more direct association is noted in the occurrence of sharply broken and crumpled synclines within some of the orebodies. SOUTHEASTERN ARIZONA DUNCAN Several fluorspar deposits occur on both sides of the State boundary between Arizona and New Mexico about 20 miles (30 km) northeast of Duncan, Ariz., in the foot- hills of the Mule Creek Mountains. There are two main groups of deposits: those in the vicinity of the Fourth of July mine, which is in Arizona in Daniels Camp, China Camp, and Goat Camp canyons, and those along Bitter Creek farther north, which are on the State line. The area around the Fourth of July mine is underlain by por- phyritic and vesicular olivine basalt and fine-grained rhyolitic tuff intruded by dikes of rhyolite porphyry. The rocks are fractured and brecciated along northwest- trending zones. Fluorspar occurs as veins, pods, and stringers in brecciated zones. The fluorspar consists of coarsely crystalline green fluorite, dense milky quartz, and minor amounts of psilomelane. Quartz, chalcedonic quartz, and calcite stringers are common throughout fractured and brecciated basalt; fluorite, however, seems to be mainly along major northwest-trending fractures where it forms lenticular zones within quartz-rich veins (Trace, 1947). Total production of fluorspar from mines in Daniels Camp, China Camp, and Goat Camp canyons has been less than 8,000 tons CaF;, most came from the Fourth of July mine prior to 1946. Mined fluorspar shoots averaged about 3 feet (1 m) wide and 50 feet (15 m) long; the largest was about 400 feet (120 m) long. The fluorspar averaged about 65 percent CaF, and 25-30 percent silica (Wilson, 1950). Fluorspar deposits near Bitter Creek are concentrated along or between the Steeple Rock and East Camp faults, which are parallel northwest-trending faults about 2 miles (3 km) apart and 14 miles (23 km) long (Gillerman, 1964, pl. 1). The faults separate a horst of predominantly dacitic volcanics of Tertiary age from younger Tertiary pre- dominantly andesitic volcanics on either side. Small bodies of rhyolite intrude the other rock types. One mine, the Mohawk, which is in New Mexico, produced about 3,000 tons of fluorspar grading 65-70 percent CaF, prior to 1945 (Gillerman, 1964, p. 195). Fluorspar was found in pods, shoots, stringers, and disseminations in highly silicified breccia zones. Mined ore had widths as much as 7 feet (2.1 m), but average widths were less. Ore consisted of pale-green fluorite, quartz, pyrite, and blocks of highly silicified country rock. Much of the fluorite is massive and crystalline or columnar and occurs in layers with quartz; some is very fine grained and intimately intergrown with GEOLOGY AND RESOURCES OF FLUORINE chalcedonic quartz. Many fluorspar prospects in the vicinity of the Mohawk mine were originally gold or silver mines. Part of the Bitter Creek fluorspar area is the Steeple Rock mining district, New Mexico, where fluorite is a common gangue mineral in the gold and silver deposits that occur along the Steeple Rock and East Camp faults (Gillerman, 1964, p. 186). OTHER DEPOSITS The Lonestar mine at the eastern base of the Whetstone Mountains, Cochise County, ranks as Arizona's most pro- ductive fluorspar mine. During 1946-51 the grade of produced fluorspar was 85 percent CaF»; silica content was 0.02-2.7 percent (Wilson, 1950). Coarse-grained green fluorite and quartz occur as layers in a northwest-trending fissure vein in Precambrian schist. Ore shoots average about 2 feet (0.6 m) wide, 25 feet (8 m) long, and 30 feet (9 m) high (Wilson, 1950). Several small fluorspar deposits are located in the foot- hills east of the Tonto Basin post office, Gila County. Fluorite, quartz, muscovite, and clay minerals occur in narrow lenses, about 3 feet (1 m) in width, along fault zones in granite and granite gneiss. A bulk sample taken by the U.S. Bureau of Mines contained 72 percent CaF», 20 percent silica, and trace amounts of gold, silver, barite, iron, and calcite (Cummings, 1946, p. 4). Tonto Mining and Milling Company was reported to be producing fluor- spar from the area in 1972 (Guccione, 1972, p. 70). CENTRAL AND EASTERN UNITED STATES By A. V. Heyt and R. E. Van Austin® Fluorine deposits in the Central and Eastern United States vary widely in composition, size, and geologic type. Fluorine-enriched igneous rocks are common and wide- spread, and they consist mainly of granites, granodiorites, pegmatites, nepheline syenites, and mafic alkalic igneous rocks. Marginal varieties, considered magmatic by some, are nelsonite dikes composed of ilmenite and fluorapatite in central Virginia, and magnetite deposits that contain fluorite and fluorapatite at and near Iron Mountain, Mo. High-temperature deposits enriched in fluorine occur in the Appalachian structural belt, in the Ozark dome of Missouri, and in the Llano uplift of Texas. They include four types: 1. Veins, lodes, and associated greisens that contain flu- orite and topaz. These have been mined mainly for tungsten, tin, silver, lead, and gold. One group of deposits in Connecticut has produced a small quant- ity of fluorite. 2. Disseminations of fluorite and topaz, commonly associated with small veins and tactites. Scheelite, gold, and cassiterite are common associated min- erals. CENTRAL AND EASTERN UNITED STATES 63 3. Complex skarns among which are iron and zinc de- posits in 'the Precambrian of New Jersey, Pennsyl- vania, and southeastern and northern New York, that contain fluorite, fluorapatite, rare-earth fluocar- bonates, and in places abundant chondrodite, nor- bergite, and humite. 4. Fluorine-bearing tactites along granite and pegmatite contacts, and contact iron deposits next to Triassic diabase in Pennsylvania. Fluorite and topaz are the main fluorine-bearing minerals. Magnetite, scheel- ite, wolframite, molybdenite, helvite, chalcopyrite, garnet, and rare-earth and thorium minerals are associated with the fluorine minerals. One contact zone on Long Island, Maine, has yielded a small quantity of fluorite for apothecary shops. Most deposits are veins which were formed at inter- mediate temperature. A few fluorite-quartz veins in the Cheshire County-Westmoreland district of southwestern New Hampshire have produced fluorspar. Complex fluorite-galena-sphalerite-quartz veins in the Faber, Va., area have also been productive. Other intermediate- temperature veins are silver-lead-copper-zinc-fluorspar- barite veins of the Connecticut Valley in Connecticut and Massachusetts and near Phoenixville, Pa. Somewhat more complex veins occur on Deer Isle, Maine, and at Thomaston Dam, Litchfield, Conn. Most of the fluorspar deposits in the Central and East- ern parts of the United States are low-temperature veins, mantos, breccia bodies, stockworks, pipes, and dissemi- nated deposits of several types of origin. The most abundant deposits and some of the largest known economic concentrations of fluorspar are low- temperature vein and manto deposits in the Mississippi Valley and in the Appalachian Valley and Ridge province near related zinc deposits. Of these, the largest concentration of fluorspar in the Nation is in the Illinois- Kentucky district which contains Mississippi Valley-type (low-temperature vein and manto) deposits Other Mississippi Valley-type deposits that have produced or may contain important resources are in the Central Tennessee district, the Cumberland River vein area, the Central Kentucky district, and the Rossie-Lowville, N. Y., area. Appalachian Valley and Ridge province deposits are numerous and occur from Alabama to Pennsylvania (Van Alstine and Sweeney, 1968). Most of these deposits (Appalachian type) are irregularly shaped to blanketlike breccia bodies or are clusters of discontinuous veins, in both Cambrian and Ordovician carbonate rocks. A few deposits are fissure veins and disseminations in quartzite, shale, and carbonate rocks. By far the largest and poten- tially most important district in the Appalachian belt is the Sweetwater district in eastern Tennessee south of Knoxville. Farther east near the Tennessee-North Carolina State boundary are the adjoining Del Rio, Tenn., district and similar Hot Springs, N. C., district; the former produced barite and a small quantity of fluorspar from breccia bodies, disseminations, and fissure veins in quartzite and crystalline rocks. A small tonnage has been produced from barite-fluorite breccia bodies that consti- tute the Gilley deposit in northeastern Alabama. Disseminations of fluorite distributed widely in vugs and thin veinlets along with calcite, dolomite, barite, celestite, and sphalerite are found in central Pennsyl- vania, western New York along the south shore of Lake Ontario, northeastern Iowa, northwestern Ohio, and in the Serpent Mound cryptoexplosion structure in south- eastern Ohio (Heyl, 1968). Such deposits may have been formed by simple warm connate brines or cold meteoric waters, or they may in places indicate undiscovered districts of the geologically more complex Mississippi Valley type or Appalachian type. Residual or placer deposits derived from underlying bedrock deposits are known in a few places. Residual gravel spar deposits have been mined in western Kentucky, central Kentucky, and central Tennessee, but are unknown in the Appalachian region. The Streeter topaz placer in the Llano uplift of central Texas was described by Paige (1911), although a bedrock source has not been found. Several of the districts and deposits in the Central and Eastern United States are described in more detail below, State by State, because of their resource potential or geo- logic interest. For each State, deposits with the greatest commercial potential are described first, followed by deposits of geologic interest, or of potential byproduct production of fluorite. Other possible fluorine sources such as apatite, rare-earth fluocarbonates, and fluo- silicates are described last. ILLINOIS-KENTUCKY ILLINOIS-KENTUCKY FLUORSPAR DISTRICT By R. D. Trace More than three-fourths of the fluorspar produced in the United States has come from the Illinois-Kentucky district. Substantial quantities of zinc and some lead and barite also have been produced, mostly as a byproduct of fluorspar mining. The district is in Hardin and Pope Counties in southeasternmost Illinois, and adjacent Crittenden, Livingston, and Caldwell Counties of western Kentucky (fig. 23). Nearly all the mines are within the boundaries of 16 74-minute quadrangles aggregating nearly 1,000 mi (2,500 km). The Illinois Geological Survey has published geologic maps of the Illinois part of the district, and the U.S. Geological Survey, in coopera- tion with the Kentucky Geological Survey, in 1972 has published geologic maps for most of the Kentucky quad- rangles. 64 GEOLOGY AND RESOURCES OF FLUORINE The earliest mining in the district was for lead at the Columbia mine, Crittenden County, Ky., in 1835 (Ulrich and Smith, 1905, p. 115). Lead mining started shortly thereafter in Illinois (Norwood, 1866, p. 366-372). From 1835 to the early 1870's, little fluorspar was mined. Only small amounts were produced from the early 1870's to about 1890, when an expanded market was created by the development of the basic open-hearth steel furnace in 37° 37/30" which fluorspar was used for flux. Production since 1890 has been erratic but in general has risen. Production during World Wars I and II and the Korean conflict rose sharply, but in 1953-58, because of rising imports, pro- duction in the district decreased. Production has risen in recent years and probably will continue to increase in the near future. Production figures are summarized in the following table. 37° 30/00" 372230" 37° 15/00" 37°07 30" EXPLANATION Fault x Mafic dike or sill 3 & Intrusive breccia Structural arch Outline of area of known fluorspar deposits 14 | $ x lion LJ 9 < TABB FAULT ”ST?” T s ss" 30/00" ss 1500" 0 0 4 ss" 00/00" 87° 5230" se 07 30" 4 8 MILES 1 8 KILOMETRES FicurE 23.-Major structural features, igneous rocks, and distribution of known fluorspar deposits, Illinois-Kentucky fluorspar district. Compiled from published and unpublished geologic maps, 1962-72, by Illinois Geological Survey and U.S. Geological Survey. CENTRAL AND EASTERN UNITED STATES 65 Fluorspar production from Illinois and Kentucky [From U.S. Bur. Mines Minerals Yearbooks unless otherwise noted] Years Tons Illinois 0685 ::?.3.;:1....1 : 0.100... ere ended 6,163,345 1964 127,454 1965 .... 159,140 1966 .... 176,175 1967 .... 210,207 1968 .... 188,325 1969 .... 88,480 1970 .... 148,208 NOTA HTH HeR Y Ped cord eta ranna nat Che Pe sade '141,000 FOLAI: :s: .fi enero aa darn da ire da weir 7,402,334 Kentucky Through 1963 ..... 3,016,980 1904 revel eres o cone erie ve vss aree bea ien i Te Phe s ie eee saa 38,214 1965 EEE H Aenon enc anus a e re VE rr Waa e 31,992 1960 ceri eir e nevis s cnn inns n hbr rer teres 28,725 1IOT errr rise sre ads fence 32,952 NIOSH YEA TEETH Eerie ere sr reer 17,050 1909 :::: s 37! HARE Er ier e. 0.0 aeons nes rer 210,000 iu (Lis .- Tot js nicee. 000. wined " na 25,000 POMAIE GYH Heist s: sizin Lor ren oes » ndp if devoid 3,180,913 Remar by ROB: Until the early 1930's almost the entire production was from vein ores. Since then, the amount of ore produced from bedding-replacement deposits near Cave in Rock, Ill., increased and so by the middle to late 1960's bedding- replacement ore probably constituted more of the Illinois production than did vein ore. The rise in the bedding- replacement to vein-ore ratio was in part due to the depletion of some of the large vein deposits near Rosi- clare. Within the last few years the Illinois production from veins has increased and may soon equal the production from bedding-replacement deposits. In Kentucky, nearly all production has been from veins, although a substantial deposit of bedding-replacement ore near Joy has been known since the 1950's and was put into production in 1970 and was mined until 1973. GEOLOGIC SETTING Sedimentary formations that crop out in the district range in age mostly from Late Devonian through Early Pennsylvanian; where mining has been concentrated, the rock units at the surface are the Meramecian and Chesterian Series of Mississippian age. A small amount of unconsolidated Cretaceous and Tertiary sand, silt, clay, and gravel and Quaternary loess and alluvium is present locally. The geologic column of exposed formations in the Illinois part of the district was summarized by Grogan and Bradbury (1968, fig. 2) (fig. 24). A slightly different column applicable only to the Kentucky part of the district is shown in figure 25. The lithologies in Illinois and Kentucky are generally similar but some differences in nomenclature are evident. Most of the vein ores occur where the wallrocks are of either early Chesterian or late Meramecian age (early Chesterian and late Valmeyeran in fig. 24). The three principal bedded ore horizons also are either in the lower part of the Chesterian Series or in the upper part of the Meramecian Series. Numerous intrusive mafic-alkalic dikes and a few sills are present in the district, and in addition in Illinois breccia forms dikelike or pluglike bodies (fig. 23). The mafic dikes and sills have been called mica-peridotites and lamprophyres; most are so highly altered as to defy precise classification. The dikes and sills are mostly dark gray to dark greenish gray, finely crystalline, and commonly por- phyritic. They are composed mainly of dolomite, calcite, serpentine, chlorite, and generally phlogopitic biotite. Small quantities of magnetite, leucoxene, marcasite, fluorapatite, garnet, perovskite, and goethite are commonly present. Olivine and pyroxene are primary constituents but in a majority of the dikes these minerals mostly have been altered to serpentine (Clegg and Brad- bury, 1956; Koenig, 1956; Trace, 1962b). Rarely, the mafic rocks are medium to coarsely crystalline (sample no. 197 in Zartman and other, 1967), light gray, and contain no chlorite or serpentine. Oesterling (1952, p. $24) described light-green dike material composed "nearly entirely of fine-grained calcite. The phlogopite has been completely replaced by calcite; there is no indication that olivine or pyroxene phenocrysts were ever present." Chemical and - semiquantitative - spectrographic analyses of samples of the Illinois dikes and sills were re- proted by Bradbury (1962) and chemical analyses of sam- ples of Kentucky rocks were given by Koenig (1956). The highly weathered Robinson dike near Hicks dome, Illinois, (Bradbury and others, 1955, p. 8) and similarly altered material from the Fowler dike in the east-central part of the Salem quadrangle, Kentucky (Trace, 19622) were sampled in 1955. Semiquantitative spectrographic analyses of the samples showed that the Robinson dike contains much larger quantities of rare earths than does the Fowler dike (see fig. 23), as shown in the following table. Semiquantitative spectrographic analyses of weathered mafic dikes, Illinois and Kentucky [Analyses by Mona Frank, U.S. Geol. Survey, Sept. 1955] Fowler dike Percent Robinson dike (East-central part of Salem (Near Hicks dome, Illinois) quadrangle, Kentucky) PIO ..s anns Si Si 5-10. .. Fe Al Fe Al 1-D esd sr chas Ti Mg Mg Ca Q .5-1 ..: Ca K Na Ti K 0.1-0.5..... Ni Y Ba Ce Ni Cr Mn Nd! .s k 5... 0.01-0.05 :::;. is:- La Sr V Co Cr Zr Cu _ Mn Ba Na Sr V Cu 0.005-0.01 .............. Nb Zn Ga Co Ga 0.001-0.005 . .. Yb Pb Mo Sc Be La Nb Y Pb Nd Zr Mo Sc 0.0005-0.001 eure 6. o perte ir h the " Hel cs Pag 0.0001-0.0005 ......... Ag Yb Ag Be 640 66 GEOLOGY AND RESOURCES OF FLUORINE RANGE OF 25 |9 LITH- | THICK: FLUORITE t |&| FORMATION MEMBER |OLOGY | NESS LITHOLOGIC - DESCRIPTION DEPOSITS > | (F T) VEINS |BEDDED «-! a as 888 Sandstone; shale; thin coals | 22 a 4 f t' > || ~ [KINKAID f= 0 = 80] Gray_cherty limestfone;_shale | DEGONIA ( 0 - 30| Shale and thin - bedded sandstone CLORE =3100-120| Shale; limestone; thin- bedded sandstone ___| ZAELNEASJANE vesi-: ~> ]50 - 60| Sandstone, silty shale z | WALTERSBURG Fine-grained limestone; shale Shole-l shaly sandstone a ¥L§a~gsamcs N IG - Z| Limestone: shaly Iimestons _ & GLEN DEAN (H Sandstone; shale, thin coal t ~f Fossiliferous partly oolitic limestone, shale z, :::2|$SBURG Sre Sandstone; shale u F ~B Fossiliferous limestone < FRAILEYS 105-140[ Shale; thin limestone Uo gsgggngEEK Silty limestone RIDENKHOWER n: B0-110| Sandstone; shale BETHEL : 25- 65] Shale, shaly sandstone [+f -I ngston z DowNnEySs BLUFF \ l Tocally, oolific limestone | wage -I 4 YANKEE TOWN s === 30-45) Shole; siltstone (Yankeetown); limestone; shale (Shetierville) hee RENAULT Shetierville 15 - 35|Light-colored oolitic limestone (Levias) wage a AUX VASES “figment ¢ \15 - 35| Calcareous sandstone, shale at base T C STE. GENE VIE VE [Spor Mim. : 120-4160|Light- colored largely oolific limestone; sandstone lenses : A re t T+ 8 ST. LOUIS Zz 250,3- Fine - grained cherty limestone «-~ III * ® *~* z £ a TT <4 y- x $ 3 2 $ u aas >| SALE M 500: |Dork-colored, fine-grained limestone; foraminiferal calcorenite w yaz b 3 pel- C z= J I I <4 > TG ULLIN E5 52650- Crinoidal, bryozoan limestone; dark-gray fine-grained limestone FORT PAYNE 225-| siitstone; silty cherty limestone be 5& [SPRINGVILLE --------I Gray and greenish-gray shale = a 98 w/ |new acgany cRouP 395: |Groy to black shale aB g LINGLE 5 GRAND TOWER l ee aso. | Limestone and chert I I 1 a ne [=] CLEAR CREEK footer FicurE 24.-Stratigraphic column of exposed formations in the Illinois-Kentucky fluorspar district. From Grogan and Bradbury (1968, fig. 2). CENTRAL AND EASTERN UNITED STATES FORMATION system | series AND LithoLogy TH'C'ENESS' DESCRIPTION MEMBER IN FEET a SE a < har 60-765 Sandstone; shale; thin coals & s a tid a. Kinkaid Limestone E 0-165 Limestone, sandstone, and shale Degonia Sandstone ~~ 5-38 Shale and sandstone Clore Limestone 70-125 Shale and limestone; thin sandstone Palestine Sandstone 30-75 Sandstone and shale Menard Limestone 80-145 Limestone and shale; thin sandstone locally Waltersburg Sandstone \ =+ 20-60 Shale, siltstone, and sandstone Vienna Limestone " 15-35 Limestone, cherty & | Tar Springs Sandstone 70-120 Sandstone and shale; thin coal locally g Glen Dean Lim 40-95 Limestone and shale & & | Hardinsburg Sandstone 80-150 Sandstone and shale Golconda Formation 90-165 Shale and limestone; thin sandstone common Cypress Sandstone 45-140 Sandstone and shale; thin coal locally Paint Creek Formation 5-100 Shale, limestone, and sandstone Bethel Sandstone 20-120 Sandstone Renault Limestone 70-125 Limestone and shale riers < Ste. Genevieve ST I + | Warsaw Limestone o eta 230+ Limestone, large Echinocrinus spines at top (i, i T : h ‘r + ll y I x I ¢ I z I 9] a pI € OQ. in pj e 0 |= 5 Fort e| «| e Limestone, mostly dark gray and very cherty or silty; =r Payne aL s | s 600+ is I § Formation »| «|« locally upper part is light gray O e| «| e lo *] o Ld LJ o DEVONIAN | Chattanooga Shale Dark-gray shale Ficure 25.-Generalized stratigraphic column of exposed formations in the western Kentucky fluorspar district. 67 68 GEOLOGY AND RESOURCES OF FLUORINE In the last two decades, many more dikes and sills have been found principally because of the large amount of diamond drill exploration. A few new occurrences also have been found from surface mapping (Amos, 1966; Trace, 1966); mafic rock is difficult to find in surface outcrop, however, because a freshly exposed outcrop decomposes within a few years and becomes difficult to recognize. Also, the mafic rock material appears fresh and superficially resembles dark-gray limestones where exposures are continuously washed clean such as on Clay- lick Creek near the eastern edge of the Salem quadrangle (Trace, 19622). Numerous breccias and breccia "dikes" have been mapped in the Illinois part of the district (Baxter and Des- borough, 1965; Baxter and others, 1967) and have been described in some detail (Clegg and Bradbury, 1956; Brown and others, 1954). According to Baxter and Des- borough (1965, p. 287), "'The breccias consist of angular to subrounded fragments of sedimentary, metamorphosed sedimentary, and igneous rocks in a matrix of finely ground rock and mineral fragments. The mineral frag- ments include quartz, pyroxene, augite, hornblende, apatite, mica, and feldspar." Grogan and Bradbury (1968, p. 376) identified fine-grained nepheline-feldspar rock in one of the breccias. Many of these breccias are inferred to be pipes or small stocks, roughly circular or elliptical in plan and as much as 1,000 feet (300 m) in diameter. More commonly the breccias are vertical or steeply dipping dikelike bodies. Unusual mineralization occurred in many of the breccias, both on the surface and to depths of over 2,000 feet (600 m) (Brown and others, 1954, p. 897-900; Heyl and others, 1965, p. B11; Bradbury, 1962, p. 6-8; Trace, 1960; Hall and Heyl, 1968). Minerals present are rare-earth- bearing fluorite, sphalerite, galena, barite, monazite, bertrandite, florencite, calcite, quartz, pyrite, brookite, yttroparisite(?), - biotite, rutile, xenotime, apatite; beryllium, niobium, and zirconium are present in more than normal concentrations. Until the early 1960's, the dikes, sills and breccias could be dated from stratigraphic relations only as post-Middle Pennsylvanian. In 1967 Zartman and others (p. 860-861) studied samples of biotite and hornblende from the dikes, sills, and breccias; on the basis of radio-isotope data they reported that "All the ages lie within experimental uncer- tainty of each other and give an average Early Permian age of 267 million years." Wallrock alteration associated with vein deposits occurred along faults. Silicification of wallrock was the most striking change. Where Chesterian sandstone comprises the wallrock, sandstone close to or within the fault zone is extremely dense and commonly makes out- crops which are called "quartz reefs" or quartzite. Indeed, fragments of quartzite in places provide the only evidence of the surface trace of faults. The alteration of limestone wallrocks generally was slight to moderate. In places numerous small doubly terminated quartz crystals and small masses of chalcedony have been reported where limestone is adjacent to a vein (Hardin, 1955, p. 24; Trace, 1962b, p. 17; Trace, 1954, p. 70). Fragments of dolomitic limestone, which is relatively common in the country rocks of the area, are present within the veins. No secondary dolomitization of the wall- rocks of vein deposits has been reported. Hall and Heyl (1968. p. 661) described visible wallrock alteration as much as 100 feet (30 m) laterally from bedding-replacement deposits. According to them, altera- tion included solution thinning of strata, silicification, dolomitization, and clay mineral alteration. Brecke (1962, p. 525-530) described the alteration of country rock, mostly within the mineralized zones, as decalcification, dolomitization, and silicification. According to Heyl and Brock (1961, p. D3), The Illinois-Kentucky mining district is centered in the most complexly faulted area in the central craton of the United States. Structural studies suggest that the mining district lies within a collapsed, block-faulted, sliced, and partly rotated domal anticline that is located at and near the intersections of several major fault lineaments. (See fig. 26.) Prominent fault zones that intersect at the dis- trict are the east-trending Cottage Grove-Rough Creek- Shawneetown system-coinciding with Heyl's (1972) 38th-Parallel lineament-and the northeast-trending New Madrid fault zone. Within the fluorspar district (fig. 23) the rocks are complexly broken by a series of steeply dipping to vertical normal or gravity faults that trend dominantly northeast and divide the area into a series of elongated northeast- trending blocks. Many cross-faults that trend north to northwest are occupied in places by mafic dikes. Near the east edge of the Kentucky part of the district, the faults trend more nearly east; and radial and arcuate faults are present around Hicks dome. Hicks dome is a structural high with about 4,000 feet (1,200 m) of vertical uplift that is centered in the south- west part of the Karbers Ridge quadrangle (Baxter and Desborough, 1965). Brown, Emery, and Meyer (1954) described the dome as an incipient cryptovolcanic struc- ture, on the basis of breccias (in an oil test hole and nearby at the surface), arcuate faults, and unusual mineral- ization. As a result of an aeromagnetic survey, McGinnis and Bradbury (1964, p. 11) stated that: "Hicks Dome is not underlain by a laccolithic intrusion. * * * A broad magnetic anomaly of relatively low intensity centered about 5% miles [9 km] northeast of Hicks Dome and the featureless configuration of contour lines around the anomaly suggest a large igneous body overlain by a thick sedimentary rock section." The southeasterly extension of Hicks dome is partly obscured by a fault zone 4-5 miles (6.5-8 km) wide known as the Rock Creek graben (fig. 23). Southeast of the graben, the structural high reappears as the Tolu dome or arch, CENTRAL AND EASTERN UNITED STATES 69 Gl is 1 St | ~ ap f < CROOKED CREEK IZe "a- Pix DISTURBANCE /* € } S */ sertha knos ND DISTURBANCE © OZARK O *4 DOME * T _ KENTUCKY ___ s AT TENNESSEE. £ / CKY MISSOURI / C KENTU s* FLYNN CREEK -# o as mua aa a -- / U- - ~~ WELLS CREEK DISTURB ARKANSAS £ é? / Tennessee " § . _ @Bistursmnce pence { / m k A MISSISSIPPIIaRIVE EMBAYMENT \ é NASHVILLE," DOME J / ¢ i.’ 36° “4L“ tarsi i S I: 1. o 25 50 75 100 MILES 0 - 25 50 75 100 KILOMETRES E X P LA N A T 1 O N [\-] «g- L ___ Base of Cretaceous rocks Fault Monocline Showing dip and direction of eorer cree displacement where known; dashed where inferred or Base of Pennsylvanian rocks concealed -_ Mississippian through Cambrian rocks hoon aced a fir?“ mia ¥entucky Precambrian rocks Anticline Ficure 26.-Major structural features in the region surrounding the Illinois-Kentucky fluorspar district. Modified from Heyl and Brock (1961, fig. 294.1). which is centered about a mile east of Tolu, Ky., in the Cave in Rock quadrangle (Baxter and others, 1963). Southeast of the Tolu dome is another series of grabens (Weller and Sutton, 1951) bounded on the south by the Tabb fault system (Rogers and Hays, 1967); in this area a structural high is difficult to trace because of numerous faults. Possibly an extension of the high is shown on the Salem quadrangle (Trace, 1962a) between the Levias- Crittenden Springs fault system and the Moore Hill fault system. Farther south in the Dycusburg quadrangle, pre- liminary mapping suggests three more possible segments of the arch (Dewey Amos, oral commun., 1971) (see fig. 23). Nearly all the faults of the district are normal and dip generally 75° to vertical; rarely are they inclined as low as 45°. Locally the dip of a fault or vein is reversed. Along the edges of many of the larger grabens, fault zones consist of several subparallel and sinuously intersecting faults. The total width of these zones may be more than 1,000 feet (300 m); more commonly it is a few hundred feet. Total dis- placement across the fault zone is commonly distributed irregularly among the several individual faults-in places as step faults and also small grabens within the fault zone, such that all faults either are vertical or dip toward the hanging wall of the fault zone. In some places, small grabens are formed by antithetic faults that dip toward the footwall side of the fault zone (fig. 27). Minor anti- thetic faults and fractures are present in places for a few hundred feet from the margins of the major graben. Vertical displacement generally has been considered to have been the principal direction of movement along the major northeast-trending faults in the district. Clark and Royds (1948) and Heyl and Brock (1961), however, suggested that a horizontal component of movement may have been significant. Field mapping has shown that as much as 3,000 feet (900 m) of vertical displacement may have taken place, but movement generally has been much 70 GEOLOGY AND RESOURCES OF FLUORINE FAULT 10 UBLEVE : 200 LEVEL 0 260 LEVEL 0 | I I 0 E <-> All workings projected to plane of section NW C Al FAULT 11 Overburden AFAULT 1 300 FEET 50 75 METRES EXPLANATION Fluorspar Ficure 27.-Cross section of Davenport mine, Moore Hill area, Kentucky. Cc, Cypress Sandstone; Cpc, Paint Creek Formation; Cb, Bethel Sandstone; Cr, Renault Limestone; Cl, Levias Limestone Member of Ste. Genevieve Limestone; Crc, Rosiclare Sandstone Member of Ste. Genevieve Limestone; Cf, Fredonia Limestone Member of Ste. Genevieve Limestone; CsI, St. Louis Limestone. Modified from Thurston and Hardin (1954, pl. 14). less. Evidence of strike-slip movement on the faults is conflicting; there is no evident horizontal offset of the northwest-trending mafic dike swarm where these elements are cut by northeast-trending faults, but there is some possible horizontal offset of the northwest-trending structural high (fig. 23) along these faults. Data from a few drill holes in the Crittenden Springs and Commodore areas (Salem quadrangle) suggest that the amount of vertical displacement on faults at depths of nearly 2,000 feet (600 m) below the surface is substantially the same as at the surface. At least some and probably most of the major northeast- trending faults are younger than Hicks dome and the mafic dikes, and therefore are post-Early Permian in age. The major displacement along these faults may have been completed by middle Cretaceous, although some move- ment has continued through Cretaceous and Tertiary time and up to the present (Rhoades and Mistler, 1941; Ross, 1963; McGinnis, 1963; Amos, 1967). Some faulting may have begun as early as Early Pennsylvanian time. FLUORSPAR ORE DEPOSITS Most of the fluorspar ore deposits are along northeast- trending faults; a few veins are found along the northwest- CENTRAL AND EASTERN UNITED STATES 71 and north-trending faults and in places are associated with mafic dikes (Trace, 1962b). The distribution of known deposits is shown on figure 23. D. R. Shawe observed (written commun., 1972) that the ore deposits tend to be concentrated in areas of mafic dikes and breccia pipes, as is suggested from figure 23; Trace is doubtful of the spatial association, however. Fluorspar ore bodies occur dominantly either as steeply dipping to vertical vein deposits along faults or as low- dipping to nearly flat bedding-replacement deposits along certain stratigraphic horizons in Mississippian lime- stones (fig. 24; Currier and Wagner, 1944). A few deposits are a combination of vein and bedding-replacement type ores and so-called gravel deposits that resulted from a fluorite concentration in the residuum overlying vein deposits. Protore amounts of disseminated fluorite are present in the breccias of Hicks dome (Brown and others, 1954). The fluorspar veins commonly are fissure fillings along faults and in fault breccias, accompanied by replacement of vein calcite and some wallrock. A typical vein is lenticular, pinching and swelling erratically, and is commonly a mixture of fluorite and highly variable quantities of calcite and country rock fragments. Locally the entire vein may be either calcite or fluorite. Commonly the vein walls are sharp but in places veinlets of fluorite and calcite extend into the slightly broken wallrocks for a few hundred feet. The fluorite varies from commonly purple, fine- and medium-grained crystalline in small veinlets to commonly brown, white, or colorless, coarsely crystalline in more massive veins. In many places, the veins are complexly brecciated and (or) sheeted. Rude layering ("banding") parallel to vein walls is present locally. Shaly fault gouge, presumably dragged in along the fault from the more shaly of the Chesterian units, is common in places. Most fissure-vein deposits average 3-10 feet (1-3 m) in width (Hardin and Trace, 1959; Trace, 1962b); a width of as much as 45 feet (14 m) is known. Mined ore shoots commonly range in strike length from 200 to 400 feet (60-120 m) though greater lengths are common in many places, particularly in the Rosiclare area. The average height of the ore shoots is 100-200 feet (30-60 m) but some ore shoots extend from the surface to a depth of nearly 800 feet (250 m) (Grogan and Bradbury, 1968, fig. 10). The stratiform or bedding-replacement deposits are elongate bodies that trend northeast and less commonly southeast (fig. 28), and "are localized along a network of fractures and minor faults parallel to and half a mile dis- tant from the Peters Creek fault * * **" (Grogan and Brad- bury, 1967, p. 41-42). In cross section, they appear as flat wedge-shaped or crescentic (concave upward) forms with thicker parts adjacent to individual faults: A few nearby pipelike breccia zones are mineralized and may have been EXPLANATION §\\\\ Bethel horizon Rosiclare horizon Sub-Rosiclare bed Hill mine Breccia pipe Possible breccia pipe Fluorspar-sulfide é}— it < ~£ 0 DEEP SOLUTIO SYNCLINE (D I R.9 E. [f 0 2000 4000 FEET 0 1000 METRES Ficure 28.-Cave in Rock fluorspar district, Hardin County, III., showing elongate character of northeast-trending ore bodies. Modified from Brecke (1967, fig. 2. p. 381). conduits for ore solutions (Brecke, 1962, p. 511-514; Grogan and Bradbury, 1968, p. 393-894). The bedding-replacement ore bodies in Illinois are commonly 50-200 feet (15-60 m) wide, and 5-20 feet (1.5-6 m) thick; their length is highly variable, from 200 feet (60 m) to 2 miles (3 km). The grade of the ore also is highly variable, but the general grade of mine-run ore ranges from 20 to 35 percent of CaF3. The bedding-replacement deposit being mined in Kentucky is believed to be approximately 2 miles (3 km) long, 150-250 feet (45-75 m) wide, and averages 7 feet (2.1 m) in thickness, containing about 30 percent of CaF, and small quantities of zinc and barite (John S. Tibbs, oral commun., 1971). Texture of the bedding-replacement deposits was described by Brecke (1962, p. 515-525) as "banded," imperfectly "banded," and brecciated. The latter two types characterize the majority of the deposits. The deepest minable fluorspar in the Illinois-Kentucky district is in a vein in the Rosiclare area and at a depth of about 800 feet (235 m) (Grogan and Bradbury, 1968, p. 380). In the Kentucky part of the district the deepest mine working on a vein is about 700 feet (210 m). A substantial number of holes has been drilled in the Rosiclare area (Muir, 1947) to intersect the veins more than 1,300 feet (400 m) below the surface, but at that depth they are 72 dominantly calcite and wallrock breccia. In Kentucky a few holes have been drilled to almost 2,000 feet (600 m) below the surface, with virtually the same results. The average grade of ore mined from the deposits has varied. Twenty to thirty years ago, the ore mined commonly contained at least 45 percent of CaF, but gradually the ore cutoff grade has dropped, probably to somewhere between 25 and 40 percent of CaF; in 1972. Various factors have influenced the ore grade in addition to the selling price of the finished product, such as the amount of byproduct zinc and lead, and the varying costs of mining that depend primarily on the type of deposit and also on mine depth, amount of water, and degree of stability of mine walls or roof. The most abundant minerals in the vein deposits are calcite and fluorite. Concentrations of sphalerite or barite are present in places, although more commonly sphalerite, barite, and galena occur in small quantities with the fluorite. Minor quantities of quartz, pyrite, marcasite, and chalcopyrite are present and locally zinc, lead, or copper alteration - products (cerussite, pyromorphite, smithsonite, hemimorphite, greenockite, cuprite and malachite). Celestite has been reported in one locality in Kentucky (Hardin and Thurston, 1945). The bedding-replacement deposits contain virtually the same suite of minerals. The principal difference is the quantity of calcite-abundant in the vein deposits, and common, though not abundant, in the bedding- replacement deposits. Fluorite is the principal mineral in the bedding-replacement deposits. Locally sphalerite, galena, and barite are abundant but generally they are uncommon. Calcite and quartz are widespread but sparse. Chalcopyrite, marcasite, pyrite, witherite, strontianite, smithsonite, cerussite, malachite, and pyromorphite have been reported. Paragenesis of the vein and of the bedding-replacement deposits appears to be similar. Of the more abundant minerals, calcite is the earliest, overlapped and followed by fluorite. Sphalerite and galena generally appear to overlap and follow fluorite, and barite is the youngest common mineral. Hall and Friedman (1963, p. 891) reported on the paragenesis of the Cave in Rock deposits, and Pinckney and Rye (1972) reported on oxygen and carbon isotope variations in minerals deposited at different times in the wallrocks. Fluid-inclusion studies of the temperature of deposition of the fluorspar have been done by Freas (1961) and Grogan and Shrode (1952). Their studies indicate a range in temperature from 90° to 140°C. The composition of fluid inclusions in the ore and gangue minerals of the Cave in Rock area was studied by Hall and Friedman (1963). According to them, the general composition of the fluid inclusions suggests that the fluorite was deposited from a concentrated Na-Ca-Cl brine, probably of connate origin. Fluid inclusions in the quartz, however, have lower salt content, have a high ratio GEOLOGY AND RESOURCES OF FLUORINE of K/Na, Ca/Na, Cl/ Na, and SO4/Na, and a lower relative deuterium concentration than that found in the fluorite itself. These fluids may represent a magmatic water. No apparent districtwide zoning of fluorite, sphalerite, and galena has been reported. Brecke (1964a) suggested that barite is more common on the fringes of the district, but exceptions in the Lola, Ky., area tend to obscure a broad picture of this zoning. Heyl (1969) suggested a zonation pattern for silver and antimony in galena, around Hicks Dome (see also Heyl and others, 1966), and a systematic change in the lead isotopic ratios from Hicks dome southeast into Kentucky. Close dating of the age of mineral deposition has been difficult. Heyl and Brock (1961, p. D6) reported an age of 90-100 m.y. (middle Cretaceous) for monazite in fluorite- bearing breccia found at Hicks Dome, but this age is in doubt because of the weathered character of the monazite (Heyl, written commun., 1972). No other direct evidence exists for the age of the deposits. Minor elements in the ore deposits have been reported by Hall and Heyl (1968). According to them, the fluorite is remarkably pure, except for the fluorite in breccias at Hicks dome which contains abundant rare earths. The sphalerite of the district is relatively high in cadmium, germanium, and gallium, but low in silver. The galena contains varying quantities of silver. SUMMARY The geologic history and genesis of the Illinois- Kentucky fluorspar district have been discussed by Amstutz and Park (1967), Brecke (1962, 1964b, 1967), Brown (1970), Grogan and Bradbury (1968), Hall and Friedman (1963), Hall and Heyl (1968), Heyl, Brock, Jolly, and Wells (1965), Heyl (1969), Oesterling (1952), Park and Amstutz (1968), and Snyder and Gerdemann (1965). The origin and chronological age of many of the significant geologic features of the district are still not established. Many structural and mineralogical events must have both overlapped and recurred, and therefore many a geologist has been confused and frustrated in attempting to establish the correct geologic history. In general, however, the history can be divided into three broad phases, all of which occurred principally within the interval of post- Middle Pennsylvanian to Cretaceous. The oldest events were the structural arching and formation of Hicks and Tolu domes, the intrusion of dikes and sills, and the formation of breccias. From the data of Zartman, Brock, Heyl, and Thomas (1967) these features appear to be early Permian in age. Next came most of the movement along the northeast-trending faults, followed by the third major phase, that of mineral deposition. The general location of the Illinois-Kentucky district may be structurally controlled by the intersection of the northeast-trending faults of the New Madrid system with mafic dikes, and the structural high trending north- westerly through the district, and possibly with an east- CENTRAL AND EASTERN UNITED STATES 73 trending lineament of igneous activity described by Snyder and Gerdemann (1965) that is part of Heyl's (1972) 38th- Parallel lineament. In the district, the primary localizing elements, particularly for the vein deposits, were faults, which served both as ore solution channelways and as space for the ore bodies. Fractures and minor faults apparently were aslo a significant factor in the local- ization of the bedding-replacement deposits although the control exerted by these structures on the shape of the replacement deposits is somewhat less obvious than their control of shape of the vein deposits. In addition, relatively impermeable blanketing lithologic units, breccia pipes, and unconformities have been suggested as contributing to the localization of the bedding-replacement deposits. These factors were discussed by Brecke (1962) and by Grogan and Bradbury (1968). The veins contain highly variable quantities of calcite and fluorite, and so far as the writer is aware, no guides are known to predict the calcite/fluorite ratio of a given vein, and thus to indicate the favorableness of the vein. Vein widths appear to be directly related to the lithology of the fault walls (Grogan and Bradbury, 1968; Thurston and Hardin, 1954; Hardin, 1955; and Trace, 1962b, 1954). Most of the ore has been mined from veins where one wall and commonly both walls are made up of the Mississippian rocks from the St. Louis Limestone up through the Cypress Sandstone totaling about 1,000 feet (300 m) in stratigraphic thickness. Possibly this control was effective because these units are dominantly competent and massive in contrast to the post-Cypress units that contain many beds of incompetent shale, silt- stone, and thin-bedded sandstone. Openings formed by faults where the walls are Cypress or older likely remained open for vein fillings, and incompetent shaly material was not as likely dragged into the fault zones. Also, in the more competent beds, the faults may steepen in dip, which could have allowed more space for mineral deposition. Similarly, where competent beds were faulted, the total displacement of the fault system was taken up by many faults (thus creating more space) rather than by a few faults with steeper drag-folded walls, such as where post-Cypress units were faulted. Moderate fault displacement appears to have favored vein formation (Thurston and Hardin, 1954, p. 98); Grogan and Bradbury (1968, p. 390-392) suggested 25-500 feet (7.5-150 m) as the optimum displacement along faults occupied by veins. A change in overall strike of a fault system may have influenced vein width in places. Possible examples are the Babb (Trace, 1962a) and Dyers Hill (Amos, 196) areas in Kentucky and the Rosiclare area (Baxter and Desborough, 1965; Amos, 1966) in Illinois. Several writers have noted the spatial association locally of vein deposits that are principally zinc-bearing (and also of fluorspar deposits that contain substantial byproduct zinc and lead) either along northwest-trending mafic dikes (Hutson and Old Jim mines) or where dikes intersect northeast-trending faults (Hickory Cane mine). Recently a new deposit containing abundant sphalerite was found in Livingston County, Ky., near a mafic dike (The Mining Record, Denver, Colo., Dec. 23, 1970). On the other hand, the Nine Acres mine, Kentucky, contains principally sphalerite, but no mafic rocks are known in the immediate vicinity. As the dikes probably were intruded prior to mineral deposition, the genetic relationship between them and the mineralization is obscure. Brecke (1967, p. 387) suggested that zinc sulfide deposition may have been related to iron sulfide present in the mafic rocks. Oesterling (1952, p. 332) suggested that sphalerite was deposited where post-fluorite faulting took place along the dikes. Hardin (1955, p. 23) stated that the distribution of sphalerite was controlled to some extent by the cross-faults and subparallel faults of the fault zone; that is, by implication, the sphalerite was related to late- stage faulting. The deposits have been considered by most authors (Brown, Grogan and Bradbury, Heyl and associates) as epigenetic, with the ore elements mostly carried by hot connate water that was heated by deep-seated igneous activity, and carried upward along the faults to their present location. Some of the ore elements presumably were contributed to the connate water from deep-seated igneous activity. One group (Amstutz and associates) suggested a syngenetic origin, and one writer (Brecke) suggested that connate water picked up the ore elements from the sedi- ments and moved them laterally into the district from the southwest, without recourse to igneous activity. Fluorspar reserves in the Illinois-Kentucky district are substantial. No comprehensive survey of reserves has been made since 1956 (''Fluorspar Reserves of the United States Estimated," press release, Office of Minerals Mobilization and U.S. Geol. Survey, November 28, 1956). However, a brief survey of the major producers in the district suggests that, under present economic conditions, new reserves found have at least been equal to production since 1956, and total reserves are probably adequate for at least 15-20 years. New finds of near-surface gravel spar deposits probably will be rare, and there has been only slight to moderately encouraging results from deep exploration (deeper than 1,000 feet-300m). Large numbers of faults within the district have not been adequately tested at intermediate depths or even shallow depths in places, suggesting that the district has much additional potential and will be a source of fluorspar for many years. Currently, explora- tion is increasing, and several new discoveries have been made in the last several years. Large quantities of low-grade fluorite in the breccias at Hicks Dome are too lean to be mined now; under more favorable economic conditions and with possible byproduct production, large-scale mining operations on 74 GEOLOGY AND RESOURCES OF FLUORINE the breccias might be undertaken in the future. Mining of the breccia ores at Hicks dome could increase sub- stantially the resources of the Illinois-Kentucky fluorspar district. CENTRAL KENTUCKY The Central Kentucky district, covering more than 2,000 mi: (5,000 km?) in Boyle, Clark, Fayette, Garrard, Jessamine, Lincoln, Madison, Mercer, and Woodford Counties, has yielded about 30,000 tons of fluorspar over a period of many years; the largest production was made in 1947 (Sutton, 1953, p. 140), and the district in 1972 is inactive. Veins cut Middle Ordovician limestones along strike-slip and normal faults (Currier, 1923; Robinson, 1931; Earl, 1959; Jolly and Heyl, 1964). The veins are localized along the Kentucky River fault zone astride the Lexington dome and the Cincinnati arch. Most of the nearly vertical veins trend approximately northerly and range in width from less than an inch (2.5 cm) to about 6 feet (2 m). A few veins trend easterly or northeasterly. They have been mined to a depth of 220 feet (67 m). Lateral zoning in this district (Jolly and Heyl, 1964) consists of an inner two zones with fluorite and an outer zone of barite- galena-sphalerite. Textures range from coarse comb crystals in the inner zone to fine colloform layers in the outer zone. Fluorite, sphalerite, galena, barite, strontian barite, and calcite are major minerals. Celestite, strontianite, marcasite, pyrite, chalcopyrite, quartz, witherite, and ferroan dolomite are minor minerals. The vein walls are almost unaltered. The veins are epithermal fissure fillings and breccia replacements that formed in late Paleozoic time. Exploration in progress in 1972 in the Central Kentucky district was searching for sulfide-bearing breccia deposits similar to those in the Lower Ordovician carbonate strata of the Cumberland River vein district and the Central Tennessee district. The carbonate rock stratigraphy changes westward in central Kentucky across the Cincinnati arch and so lithologies are more like those in central Tennessee, making the probabilities of deposits similar to those in central Tennessee better on the eastern side of the arch. Widespread faults in the Central Kentucky district probably provided channels for easy flow of ore- bearing solutions and allowed formation of fissure veins. Bedded and breccia deposits more probably occur in areas sparsely faulted and still capped by unbroken semi- permeable shaly beds. Such geologic conditions exist southward along the Cincinnati arch toward Tennessee and to a lesser extent northwestward toward Louisville, Ky. Probabilities are good for the development of new commercial deposits of fluorspar on the known veins, and for discovery of new fluorspar veins in the central part of the district. CUMBERLAND RIVER, KENTUCKY A northward extension of the Central Tennessee district, described below, lies along the Cumberland River in Monroe, Cumberland, and Russell Counties, Ky. (Jolly and Heyl, 1968), where small veins containing traces of purple fluorite, but otherwise mineralogically identical with those in central Tennessee, have been prospected. These veins in Middle Ordovician limestones lie 30-50 miles (50-80 km) northeast of the main Central Tennessee district in an area of good outcrops. Unquestionably they indicate an extension northeastward of the district, and a possible connection of the Central Tennessee district with the Central Kentucky district along the crest of the Cincinnati arch. Breccia fluorspar deposits similar to those in the Central Tennessee district may occur in quantity in the Lower Ordovician and lower parts of the Middle Ordovician carbonate strata in this Cumberland River vein district. TENNESSEE-NORTH CAROLINA-VIRGINIA CENTRAL TENNESSEE The Central Tennessee zinc-fluorspar-barite-lead district (Jewell, 1947; Maher, 1970) covers about 2,400 mit (6,200 km?) in Smith, DeKalb, Putnam, Trousdale, Wilson, Davidson, Williamson, Rutherford, and Cannon Counties, and was the source of about 1,500 tons of hand- picked flourspar in 1902-06, 1914-18, 1925-28, 1985-36, 1942-43, and 1956. This district is on the Nashville dome, which is structurally connected by the Cincinnati arch to the similarly mineralized Lexington dome farther northeast. After the Illinois-Kentucky fluorspar district, the Central Tennessee district may have the largest fluorspar resource potential in the United States, occurring in still- unmined breccia blankets of fluorspar, barite, sphalerite, and calcite in Lower and Middle Ordovician carbonate rocks. Past production of fluorspar nevertheless was entirely from nearly vertical veins cutting Lower Ordovidian to Mississippian rocks; most deposits are in limestone of Middle Ordovician age. Igneous rocks are lacking in the district. The veins formed largely by fissure filling and replacement of breccia fragments; there is little evidence for replacement of the limestone walls. Mineral- ized faults, commonly having horizontal displacement, generally trend northeast; most northwest-trending faults are post-ore. Most veins are less than 3 feet (1 m) thick but veins range in thickness from a few inches to 6 feet (2 m) where the vein material consists of well-layered material, and to 15 feet (4.5 m) where the vein is a mineralized breccia zone. A few small bedded replacement bodies or pipes of fluorite have been mined or prospected here in the Middle Ordovician limestone. Nearly all the known deposits exposed at the surface were described by Jewell (1947). The vertical extent of some deposits is at least 200 feet (60 m). A few veins may extend for 1,000 feet (300 m) in length. The veins consist chiefly of strontian barite, multicolored fluorite, sphalerite, galena, calcite, quartz, dolomite, and a little chalcopyrite and pyrite and their oxidation products. A few veins consist almost entirely of sphalerite and CENTRAL AND EASTERN UNITED STATES 75 calcite. Analyses of vein material (Jewell, 1947, p. 21) show the following ranges: 3-86 percent CaFz, 1-79 percent BaSQO,, 3-60 percent CaCO;, less than 1-3 percent SiO». As on the Lexington dome, fluorite is the prominent mineral near the center of the Nashville dome, and the mineral zoning and paragenesis in these two districts are quite similar (Jolly and Heyl, 1964, p. 618). Some veins were worked also for lead, zinc, or barite, and possibly some of them may yield coproducts in the future. Most veins, however, are much too small to be important future producers. According to Jewell, the veins have never been adequately explored. They are regarded as low- temperature hydrothermal deposits (Jolly and Heyl, 1964, p. 621; Jewell, 1947, p. 37) formed from heated saline brines in post-Carboniferous time. Large blanket and podlike breccia deposits of fluorspar, barite, sphalerite, and calcite have been found at depth in recent years in the Central Tennessee district during extensive exploration programs by many companies searching for zinc deposits similar to the large ones dis- covered by the New Jersey Zinc Company near Carthage, Tenn. In contrast to the overlying veins the deeper deposits are lead-free or lead-lean. In these deposits dolomite is sparse and replacement features uncommon, unlike the somewhat similar deposits in eastern Tennessee near Sweetwater. Fluorite, most of it colorless, cements breccia fragments of limestone, dolomite, and gray jasperoid. Several ages of breccia are known, but most of the mineralized breccia bodies are in the middle and upper parts of the Lower Ordovidian Kingsport Dolomite. Some bodies extend through the unconformity at the top of the Lower Ordovician rocks up into overlying Middle Ordovician limestones. Fluorite-bearing aureoles of potential byproduct value may exist around some of the zinc-rich ore bodies as well as around the fluorite breccia bodies. Large parts of the Central Tennessee district that are marked by small veins are still scarcely prospected,; notable are the southern and southwestern parts along the crest of the Cincinnati arch. SWEETWATER, TENNESSEE According to Brobst (1958, p. 102), the Sweetwater district in parts of McMinn, Monroe, and Loudon Counties has yielded most of the barite mined in Tennessee. The barite occurs with fluorite, sphalerite, and pyrite in veins or shatter zones in the upper part of the Knox Group of Cambrian and Ordovician age and it is concentrated in the overlying residual clays. Only the residual barite deposits have been mined. Fluorite is generally absent in the residual deposits and no gravel spar is known, a fact that has largely obscured the true composition of the deposits. In bedrock the deposits contain almost as much fluorite as barite, and in places considerable sphalerite (Laurence, 1939, 1960). D. A. Brobst (oral commun., 1972) confirms this widespread abundance of fluorite throughout the district wherever the unweathered primary deposits are exposed. Many of the bodies in the shattered zones are large, and except for the main minerals in them are geologically similar to the collapse breccia and disseminated zinc ore bodies of eastern Tennessee. In the main zinc deposits of Tennessee, barite and fluorite are sparse late minerals, but elsewhere in the zinc districts of the Appalachian Valley and Ridge province between Alabama and Pennsylvania all gradations of deposits occur from nearly pure sphalerite to barite (or barite- fluorite). Deposits in the large Sweetwater district occur in three parallel northeast-trending belts of limestone and dolomite separated by zones devoid of deposits. According to Brobst (1958, p. 102) the three belts of gently east- dipping Knox Group strata lie in nearly parallel fault blocks bounded by southeast-dipping thrust faults. Workable bodies of residual barite ore occur at irregular intervals along the three belts. Unlike fluorspar in the Central Tennessee district to the west, the fluorspar here was apparently dissolved during weathering and did not accumulate in the residual clays as did the barite (D. A. Brobst, oral commun., 1972). Thus the absence of fluorspar in the residual barite deposits does not indicate either its absence or its possible abundance in the bedrock deposits. In fact a possibility for exploration would be to test the bedrock beneath the leaner barite residual deposits to determine whether sparsity of barite indicates more abundant fluorite. The eastern belt is 20 miles (30 km) long and the other two belts are each 40 miles (65 km) long. Individual commercial deposits of barite extend over many acres, and zinc deposits are known in the district. The marked geologic similarity (including the same favorable strata) between the primary unweathered deposits and the huge zinc deposits of eastern Tennessee farther north suggests that a major fluorspar-barite (and in places zinc) resource potential exists here; it may be one of the larger deposits in the country. Careful exploration and evaluation of the primary deposits are required to determine whether they are economic in grade and adaptable to known milling techniques. DEL RIO, TENNESSEE-HOT SPRINGS, NORTH CAROLINA The Del Rio district in Cocke County of easternmost Tennessee and the Hot Springs district in Madison County of westernmost North Carolina are two parts of one mineralized area of barite and fluorite (Stuckey, 1942; Oriel, 1950; Ferguson and Jewell, 1951; Brobst, 1958). A moderate tonnage of barite has been produced from both districts and a small tonnage of fluorspar from the Del Rio district (Ferguson and Jewell, 1951). The mineralized area forms a belt that trends east for about 20 miles (30 km) across the crest of the Blue Ridge Mountains. The belt is about 10 miles (15 km) wide at its widest, which is in the Del Rio district. Many mines and prospects are in the 100- mi (260 km#) area. Most of the mining has been in residual deposits in deeply weathered zones. The rocks in the Del Rio district are gneiss, schist, and quartzite of the Snowbird Group, and the Sandsuck 76 GEOLOGY AND RESOURCES OF FLUORINE Formation, all of Precambrian age (Brobst, 1958, p. 90), and clastic rocks of the Unicoi and Hampton Formations of Early Cambrian(?) age, the Erwin Formation of Early Cambrian and Early Cambrian(?) age, and the Shady Dolomite and shale of the Rome Formation of Early Cambrian age (Ferguson and Jewell, 1951, p. 12). These rocks occur in two overthrust sheets that have been thrust from the southeast. Minerals in the deposits are barite, fluorite, ankerite, dolomite, sericite, specular hematite, quartz (including jasperoid), pyrite, and copper minerals; in some places all these minerals are present, in others, only a few. Fluorite is found in most deposits, and in places exceeds 20 percent of the total. The ore bodies were probably deposited at higher temperatures than those that prevailed at Sweetwater, Tenn. Most of the primary deposits in the Hot Springs district in North Carolina are fissure veins in quartzite. In the Del Rio district, the primary deposits are of two other types: (1) replacement of finely granulated sheared gouge and breccia in thrust faults, and (2) cementation and replace- ment within tension fracture zones and along bedding planes and permeable beds. Some of the ore bodies of the second type are large and constitute an important resource of many tens of thousands of tons of barite and as much as 20 percent fluorite that could be recovered by concentration mills. HAMME, NORTH CAROLINA-VIRGINIA The Hamme tungsten district is in the Piedmont province, almost centered along the North Carolina- Virginia State line between Vance County, N. Car., and Mecklenburg County, Va. High-temperature fissure veins occur near the granite-schist contact, mostly in granite (Espenshade, 1947; Parker, 1963). Tungsten-bearing veins lie in a northeast-trending belt about 8 miles (13 km) long. Individual veins range in length from a few feet to at least 1,000 feet (300 m) and they extend to depths exceeding 1,000 feet (300 m). The principal minerals are quartz, huebnerite, scheelite, fluorite, garnet, apatite, sericite, rhodochrosite, topaz, pyrite, argentian galena, calcite, tetrahedrite, sphalerite, and chalcopyrite; a little gold is present. Fluorite is the most abundant mineral aside from quartz, but it has not been recovered as a byproduct even though it occurs in nearly every tungsten-bearing quartz vein. The fluorite is pale green, blue, and purple where fresh, but bleaches to white upon exposure to sunlight. Espenshade (1947, p. 7) reported that about 4 percent of fluorite was noted in the mill tailings in the early years of the Hamme mine. The deposits are large, and for about 20 years the Hamme mine was the largest tungsten mine in the United States. A substantial fluorspar resource is present in this district and fluorite could be recovered from the mill tailings and as a byproduct of mining when the mine is reopened. FABER, VIRGINIA A long vein of quartz, fluorite, argentian galena, sphalerite, and chalcopyrite produced some zinc, lead, and a small quantity of fluorite during 1905-19 in Albermarle County, 2 miles (3 km) east of Faber (Luttrell, 1966, p. 46-47). The fissure vein trends northeast and lies in the western part of the Piedmont province in a heavily wooded area where outcrops are scarce. The vein occurs in a shear zone within biotite-chlorite-muscovite gneiss. Little effort, as far as is known, has been made to determine if other veins occur in the vicinity. OTHER DEPOSITS, VIRGINIA Appalachian-type deposits of barite, sphalerite, and fluorite are common in three parts of Virginia: (1) the Gate City-Lebanon area in southwestern Virginia, (2) the Marion-Austinville area in south-central Virginia, and (3) the northern part of the Shenandoah Valley of Virginia, where there are little-known fluorite veins and zinc- fluorite breccias near Woodstock and Lebanon Church at the northeastern end of the Timberville zinc district. The Gate City-Lebanon area includes a long discon- tinuous line of deposits of sphalerite, barite, and fluorite that trends east-northeast for at least 50 miles (80 km) through Scott, Russell, Tazewell, and Bland Counties. Fluorite, locally forming as much as 20 percent of a deposit, occurs with the sphalerite and barite in mineral- ized breccia zones in Kingsport Dolomite of Early Ordovician age similar to those at Sweetwater, Tenn. Residual barite deposits occur in places and the entire belt needs re-examination in terms of a possible fluorspar potential similar to that in the Sweetwater district, Tennessee. An area of similar deposits extends at intervals from Marion, Virginia, to the Austinville-Ivanhoe zinc-lead district farther east. Some of the western deposits near Marion contain abundant fluorite. They consist of barite- fluorite-sphalerite breccia bodies. West of Marion, for example, is the Myers deposit of coarse platy white barite, colorless and purple fluorite, pale-yellow sphalerite, and a little pyrite, in the upper part of the Kingsport Dolomite. Fine-grained gray dolomite and black and gray jasperoid are the main gangue minerals. The mineralized zone is in a vertical body of collapse breccia alined along the crest of an anticlinal nose. In the eastern parts of this area south of Austinville and Ivanhoe similar barite-fluorite deposits are known, but little prospected, in the Shady Dolomite and in the Unicoi Formation, of Cambrian and Cambrian(?) age respectively; deposits in Grayson County are in Pre- cambrian gneiss, schist, granite, and volcanic rock. At the Poole barite prospect on the east bank of the New River 5 miles (8 km) east-northeast of Independence, a fissure vein as much as 16 inches (40 cm) wide of fine-grained barite, epidote, chlorite, fluorite, and quartz has been prospected by three shallow shafts (Edmundson, 1938, p. 48-49). CENTRAL AND EASTERN UNITED STATES 77 Another locality is the Simmerman prospects, 5 miles (8 km) east of Ivanhoe Ferry, in the Cambrian Shady Dolomite. There, in old pits, where the iron ore is undoubtedly a limonite gossan, brecciated dolomite outcrops contain sphalerite, galena, fluorite, barite, cherty quartz, and calcite (Luttrell, 1966, p. 119). Fluorite occurs in small quantities throughout the Timberville zinc district of northern Virginia in Augusta, Rockingham, and Shenandoah Counties. The mineral is abundant at the Tusing prospect 4.9 miles (7.9 km) S. 82° W. of Mount Jackson (Herbert and Young, 1956, p. 36). Here coarse purple fluorite occurs in disseminations associated with yellow sphalerite and calcite throughout a bed of crystalline dolomite. Between 1956 and 1965 a prospect shaft was dug on the property, and a bin was filled with a few tens of tons of sorted fluorite-sphalerite ore (C. H. Maxwell, 1972, oral commun.) Maxwell estimated that the sorted coarse-grained ore contained about 50 percent purple, colorless, and pale-yellow fluorite, 30 percent honey-yellow sphalerite, and the rest barite and calcite. About 8 miles (13 km) northeast of the Tusing prospect is the most probable location of the Barton fluorspar vein, which is reported to be many feet wide and to trend northerly (Silliman, 1821, p. 243-244). This vein, which was described and then forgotten in the early 19th century, is reported to be in limestone at the very foot of the southeast slope of Little North Mountain, northwest of Woodstock. The vein is reported to consist of purple fluorite and calcite. It is on a small ridge and between two walls of limestone. At the time of the discovery of the vein prior to 1820, fluorspar had little or no value and the exact location of this vein is now lost. About 10 miles (15 km) northeast of the Barton vein, and 0.5 mile (0.8 km) northwest of the village of Lebanon Church in northernmost Shenandoah County, a similar deposit of fluorspar was found in the 1960's by C. H. Maxwell (oral commun., 1972). It lies at the southeast foot of Little North Mountain on a small knoll of Paleozoic limestone. Coarse pieces of residual colorless and yellow fluorite are abundant in the yellow clay soil atop the knoll between two low outcrops of northeast-trending limestone. The above three deposits of fluorspar found at intervals along a 20-mile (30-km) trend at the foot of Little North Mountain may indicate a fluorspar district of commercial importance. Similarly occurring and equally little-known occurrences of sphalerite near Timberville, farther south, led to the exploration and development of the Timber- ville zinc district in the early 1950's. ALABAMA Deposits notably similar to those in the Sweetwater district, Tennessee, are known in slightly older carbonate rocks in Alabama at the Sinks district, Bibb County and at the Gilley (or Gilly) deposit in northeastern Alabama in Calhoun and Cherokee Counties. The Gilley deposit has produced a small quantity of fluorspar from veins and other concentrations in Conasauga Limestone of Cambrian age (O'Neill, 1950). This deposit and several other residual barite deposits and fluorspar occurrences in Alabama and in nearby Georgia (Brobst, 1958, p. 106; Worl and others, 1974) need reevaluation in respect to possible similarities to the Sweetwater, Tenn., deposits. MISSOURI-ARKANSAS-CENTRAL TEXAS SOUTHEAST MISSOURI The Precambrian magnetite deposits being mined in 1972 at Pilot Knob and Pea Ridge, Mo., and the recently closed mine at Iron Mountain, all contain fluorite and fluorapatite. Both fluorine minerals are locally abundant in these large deposits. For example, the initial new shaft put down in recent years at Pilot Knob cut a fluorite- chalcopyrite-magnetite body of sizable tonnage (P. W. Guild, oral commun., 1967). Elsewhere in the district the abundance of purple and colorless fluorite and associated apatite is apparent and in parts of the Pilot Knob ore body the mineral may constitute as much as a few percent of the ore. The same potential for byproduct fluorite and copper concentrates exists here as in the Adirondack Mountains. The partly developed Boss iron-copper deposit, as well as the Bourbon and Sullivan magnetite deposits, should also be examined for fluorite potential. OTHER DEPOSITS, MISSOURI AND ARKANSAS Fluorite occurs elsewhere in the Precambrian of south- east Missouri. Fluorite is found with calcite and apatite in a veinlet in a granite quarry at Granite, north of Ironton; disseminated fluorite is reported in this same granite. Fluorite occurs in high-temperature Precambrian veins with quartz, topaz, cassiterite, wolframite, and silver-rich galena at the Einstein silver-tin mines southwest of Fredricktown. Coarse yellow fluorite and pink platy barite occur in large gas cavities in a Devonian kimberlite breccia on the O.K. Cash farm southeast of Avon, Mo. (Zartman and others, 1967). The kimberlite, which contains many lime- stone fragments, is fluorine rich. Fluorite is abundant enough to suggest that the 70 or so other Devonian intrusives near Avon should be examined for fluorite. Fluorite is locally abundant in the Magnet Cove and Fourche Mountain, Arkansas, alkalic igneous rock complex, and in small veins in the southern Quachita Mountains, Arkansas. A search might be warranted on the margins of these complexes and near the contacts of the other alkalic syenites in central Arkansas for fluorite deposits of commercial size, especially where the wallrocks are limestone or dolomite. LLANO UPLIFT, CENTRAL TEXAS Fluorite deposits and occurrences are common in Pre- 78 GEOLOGY AND RESOURCES OF FLUORINE cambrian granites and gneisses in the Llano uplift west of Austin (Paige, 1911), particularly northwest of Austin at Burnet where a small quantity of fluorspar was produced in 1943. The deposits are mostly disseminations of fluorite and clusters of small fluorite veins in granite and in gneiss. The main deposits form a small cluster 4-7 miles (6.5-11 km) west of Burnet along the north fork of Spring Creek. Fluorite and topaz are also common in the complex pegmatites of the Llano uplift such as those on Barringer Hill. Fluorite is disseminated in the granite of Opaline in Llano County and in the Oatman Creek Granite of Stenzel (1932) along Honey Creek in Mason County. The Llano uplift appears to be a favorable area in which to search for commercial fluorite deposits. A topaz placer occurs in the Llano uplift of central Texas (Paige, 1911). The bedrock source of the coarse- grained white topaz has not been found, but it is probably in pegmatite. The deposit, which has been prospected but not mined, is near Streeter in Mason County. The occurrence of topaz in pegmatites and granites elsewhere in the uplift suggests that other such topaz placers may exist. NEW YORK ADIRONDACK MOUNTAINS Nearly all the Precambrian magnetite deposits of the Adirondack Mountains of New York contain fluorite, but in most deposits fluorite is a sparse mineral. Certain types of skarn magnetite deposits, however, contain locally ' abundant fluorite; in places the mineral forms 5-60 percent of the deposit. Types of skarns that may contain locally abundant fluorite (according to Leonard and Buddington, 1964, p. 38-35) are clinoamphibole skarn, biotite skarn called "skol," (in which the mica is in contorted streaks, masses, and selvages), and a modified scapolitic skarn, which is derived by further meta- morphism from the other types. All types except the scapolitic fluorite-bearing skarn are associated with magnetite deposits. Some of the larger fluorite-bearing deposits are the Parish, Trembley Mountain, and Jayville, all in the St. Lawrence magnetite district, the Palmer Hill iron deposits in the northeastern Adirondack Mountains, and the Mineville deposits in the southeastern Adirondack Mountains. f Many of these skarn magnetite deposits are large. They range in length from 800 to 4,000 feet (250-1,200 m); generally they are 1,000-2,000 feet (300-600 m) long and 30-65 feet (10-20 m) wide and extend to depths greater than 500 feet (150 m) (Leonard and Buddington, 1964, p. 39-40). Deposits of this size may contain up to several million tons of iron ore. Many deposits are much smaller. An example of a skarn magnetite deposit that is rich in fluorite is the Jayville deposit, 1.25 miles (2 km) east of Kalurah, St. Lawrence County, in the western Adirondack Mountains. The mine has yielded several tens of thousands of tons of magnetite ore; associated with the magnetite ore are supergene hematite, vonsenite (ferrous- ferric borate), fluorite in dark mica skarn (skil), pyroxene- amphibole skarn, and quartz-bearing amphibole skarn. Parts of the amphibole skarn and the iron ore are rich in fluorite; this type of skarn is as much as a few feet thick and ranges from 20 to 60 percent fluorite (Leonard and Buddington, 1964, p. 134, 142). If the iron deposit is mined again, fluorite and vonsenite (borate) should be considered as coproducts. The Parish fluorite-bearing iron deposit is about 6 miles (10 km) east of Degrasse in the central part of St. Lawrence County. The deposit is probably 4,000-5,000 feet (1,200-1,500 m) long, has an average thickness of 55 feet (17 m), and has been explored by drilling to depths of 600 feet (180 m). The deposit consists of magnetite-hematite ore with spessartite garnet, chalcopyrite, bornite, pyrite, and fluorite. In places the fluorite constitutes several percent of the ore (Leonard and Buddington, 1964, p. 223). Potentially the deposit has several valuable coproducts including the fluorite. If methods now applied elsewhere to recover the magnetite and hematite were used, and in addition copper sulfide and fluorite concentrates were recovered, this large deposit would have more value than deposits from which iron alone is recovered. At Mineville, rare-earth fluorapatite, rare-earth fluo- carbonates (including bastnaesite), and fluorite occur in several deposits (McKeown and Klemic, 1956). Large quantities of these minerals are already partly concen- trated in the mill tailings, making a potential ready source of fluorine and thorium- and yttrium-bearing rare earths. North of Mineville, at Palmer Hill, fluorite and fluorapatite are abundant in the iron deposits, and in one mine an undeveloped vein of fluorite is reported adjacent to the magnetite ore. The magnetite deposits of the Adirondack Mountains © should be considered a major potential source of fluorine, mostly as a coproduct or byproduct of iron mining. To date the fluorite has not been recovered. The potential tonnage of fluorite in the magnetite deposits is very large, probably exceeding a million tons. OTHER DEPOSITS Fluorspar has been produced from only two of several fissure veins in the Rossie vein district of northern New York. The production has not been large, and the vein system only locally contains commercial concentrations of green and colorless fluorite. More commonly, galena and sphalerite are the main ore minerals. The veins, which are large and wide, are in vertical faults, mostly of north- westerly trend on the south flank of the Frontenac arch. The faults have large strike-slip components, and are post- Precambrian in age. The deposits are low temperature and contain coarse-grained minerals. Galena in the veins is radiogenic. In these respects and in all others, the veins can be classified as Mississippi Valley-type. The veins in the district are not well known and have not been prospected CENTRAL AND EASTERN UNITED STATES 79 extensively; a belt may extend northwestward from Lowville well into southern Ontario to the Madoc fluorspar district. The Rossie district, though of secondary resource potential, should be reevaluated in terms of its fluorspar possibilities. NEW ENGLAND CHESHIRE COUNTY-WESTMORELAND, NEW HAMPSHIRE The most productive fluorspar district in New England is about 2.5 miles (4 km) southwest of Westmoreland, Cheshire County, where about 8,000 tons (7,262 t) was produced in 1911-23 and 1935-38 (Bannerman, 1941). The deposits are 2-3 miles (3-5 km) southeast of the Connecticut River valley in granite gneiss overlain by hornblende schist, quartzite, and staurolite-mica schist. Lenticular fluorite-quartz veins are found near the contact with overlying hornblende schist. Blue-green fluorite occurs in lenses, pockets, and fine-grained intergrowths with quartz, calcite, and barite; streaks of sulfide minerals are rare. The sulfide minerals are chalcopyrite, galena, pyrite, sphalerite, and bornite. The small production came from shallow workings on seven veins. An ore sample from a dump contains 35 percent CaF, and 60 percent SiO; (Bardill, 1946, p. 5). A vein intersected by a drill hole 150 feet (45 m) deep contained only about 2 percent fluorine and less than 1 percent barium. These deposits and others in New Hampshire may be related genetically to Devonian granites, as Cox (1970, p. D3-D4) suggested for fluorite-bearing deposits associated with the Conway Granite. Possibly, however, the structural position of the Cheshire County-Westmoreland district in the Connecticut Valley graben, on trend with mineralogically similar Triassic or younger veins in Massachusetts, indicates a Triassic or younger age for the veins in southern New Hampshire. The small veins of fluorite nearby at Putney, Vt., are probably genetically related. The Cheshire County-Westmoreland veins were productive even without milling; the fluorspar was concentrated by hand sorting. Thus only the coarsest and purest ore could be shipped. Bannerman's (1941) descrip- tion of the veins indicates that only the richest parts of the known veins were mined, and only to shallow depths. Some lower grade veins amenable to milling were not mined, and he considered indications good that the veins extended to greater depth. Thus the district may have a future as a potential resource of moderate size. NORTH CHATHAM, NEW HAMPSHIRE Granite pegmatites near the village of North Chatham contain fluorite and topaz along with beryl, lepidolite, and muscovite. Locally, the fluorite is so abundant that a few tons was mined many years ago (Morrill, 1951). The pegmatites are possibly related to the Conway Granite of the White Mountain Series. Many nearby pegmatites in western Maine contain fluorite and topaz. LONG HILL (TRUMBULL)-MONROE, CONNECTICUT One of the first producers of fluorspar and tungsten in the United States is north of Bridgeport in southwestern Connecticut. In 1837 fluorspar was mined from a vein on the southern crest of Long Hill in the Township of Trumbull and used as a flux to smelt copper ores (Hatmaker and Davis, 1988, p. 17). Coarse-grained green, colorless, and pale-purple fluorite occurs in a few large east-trending fissure veins of fluorite, topaz, quartz, margarite, margarodite, wolframite, scheelite, and arsen- opyrite. Pegmatites a few hundred feet north of the veins contain coarse fluorite in miarolitic cavities, and fluorite is reported also to occur with scheelite in marble and in a skarn contact zone between the Straits Schist and the Collinsville Formation of Crowley (1968). Fluorite occurs north of Long Hill in smaller quantities at Lanes mine, 1 mile (1.6 km) west of Monroe village. There it is in a quartz vein with sparse tungsten and silver-lead, bismuth, and arsenic minerals. A few miles southeast of these deposits is the Pinewood Adamellite of a quartz monzonite (or adamellite) stock, which contains disseminated fluorite (Crowley, 1968). Intrusives of this type at depth may have been the source of the fluorite on Long Hill. Although this area is favorable for other fluorspar deposits, and fluorite is still available in the known veins, Long Hill is now a State Park, and both it and Lanes mine are being engulfed by the suburban area of southern Connecticut. LONG ISLAND, BLUE HILL BAY, MAINE A hundred-foot-wide tactite zone of eastward trend more than a mile (1.5 km) long crosses the northern third of Long Island, southeast of Blue Hill, Maine. The zone is part of the northern tactite-bearing contact between biotite granite and Ellsworth Schist. The tactite contains appreciable scheelite, molybdenite, apatite, a beryllium mineral, and coarse crystals of green fluorite. A very small quantity of the fluorite was mined in 1838 and sold to apothecary shops (Hatmaker and Davis, 1938, p. 17). The contact zone potentially may be an important resource of the coproducts fluorspar, tungsten, molybdenum, and beryllium. OTHER DISTRICTS AND DEPOSITS OF THE CENTRAL AND EASTERN UNITED STATES Parts of the Midwest region, other than those already described, that should be examined for possible new fluorspar districts are: (1) northern Ohio in the vicinity of Toledo where fluorite is known in limestone quarries on a structural dome; fluid-inclusion studies by Edwin Roedder (oral commun., 1970) indicate that the fluorite was deposited by hot concentrated brines and therefore may be present in substantial deposits; (2) Serpent Mound cryptoexplosion structure in Adams and Highland Counties, southeastern Ohio, where some fluorite and barite cement the breccias in places; (3) southern Indiana, west of Louisville, Ky., where a few minor fluorite 80 occurrences are known in limestone quarries-possibly indicative of an undiscovered district; (4) many minor occurrences of fluorite, barite, sphalerite, and galena in and near the Ste. Genevieve fault zone of southeast Missouri and southern Illinois, which also may be an undiscovered Illinois-Kentucky-type fluorspar and sulfide district. Fluorite is a common mineral in many of the other southwestern Virginia barite and zinc deposits, including a cluster in the vicinity of Roanoke. Indeed, Edmundson (1938, p. 12) stated that "fluorite was identified in all of the barite areas [in Virginia), being particularly abundant in the deposits of the Appalachian Valley and Ridge province." Thus a reexamination of all the barite deposits (and some of the zinc deposits) in terms of fluorspar potential in the Valley and Ridge province may be warranted. Similarly, the Falls Branch barite-zinc district and the Green County residual barite area (Brobst, 1958, p. 115) might also be reexamined for fluorite in the bedrock deposits. Some similar possibilities exist in the same province in northwestern Georgia and northeastern Alabama. Fluorite occurs in some abundance in at least two of several small veins in Paleozoic carbonate rocks in the Great Valley extension of the Shenandoah Valley in the triangular area between Waynesboro, Chambersburg, and Mercersburg, Pennsylvania. Geologically these little- known deposits resemble those to the south in Virginia. Some granitic and volcanic rocks in Virginia and North Carolina contain abundant disseminated fluorite. These igneous rocks may indicate undiscovered fluorite deposits nearby. Notable examples are the granites of Browns Mountain and Beach Mountain in western North Caro- lina, the Grayson Granodiorite Gneiss and the Mount Rogers Formation in the southern Blue Ridge province of Virginia, and the Petersburg Granite in the eastern Virginia Piedmont (W. C. Overstreet, oral commun., 1972). In the Piedmont many complex pegmatites are en- riched in fluorine minerals, either fluorite or topaz; examples are known in the Rockford area, Alabama, at Amelia, Va., near Sykesville, Md., in the lower Connecti- cut Valley, at Long Hill, Conn. (p. 79), and in western Maine and eastern New Hampshire. These pegmatitic occurrences are generally not economic. Contact zones of the Triassic diabases of Virginia, Mary: land, Pennsylvania, Connecticut, and Massachusetts, and genetically related magnetite, copper, and silver-lead-zinc deposits, in places contain small quantities of fluorite. Locally, as in the northern part of the Connecticut Valley of Massachusetts, some veins contain sufficient fluorite and barite to be of commercial interest. Felsic and mafic syenites of northern New Jersey also contain dissemin- ated fluorite, and some fluorite has been deposited in ad- jacent limestones and in the Franklin zinc ore body along the dike contacts. GEOLOGY AND RESOURCES OF FLUORINE Fluorite is a common mineral in many of the skarn iron and zinc deposits of the New Jersey Highlands province, which is underlain by Precambrian gneisses, schists, skarns, and marbles (Worl and others, 1973). The high- lands extend from west of Reading in eastern Pennsyl- vania northeastward through New Jersey into southeast- ern New York in the vicinity of West Point. In no part of this province are commercial fluorite deposits known, but in places fluorite associated with other fluorine-bearing minerals makes major bodies of fluorine-enriched rocks. Such rocks are most abundant in the marble and skarn belts on the northwest side of the highlands. Barite deposits of the Piedmont province of Virginia, as well as similar deposits in the Kings Mountain district of North Carolina, contain some fluorite (Edmundson, 1938). As fluorite has no doubt been dissolved out of the residual deposits which were mined for barite, the possi- bilities of fluorite in quantity in the unmined bedrock deposits should be examined. Fluorite is a common mineral in the high-temperature small hydrothermal deposits of the Pawtucket area of northeastern Rhode Island. These deposits were mined long ago for copper and perhaps gold. They contain-in addition to fluorite-garnet, magnetite, scheelite, molybdenite, chalcopyrite, galena, and sphalerite (Quinn, 1971, p. 58). Similar small quantities of fluorite are found in the gold-silver-lead-zinc-copper deposits of Deer Isle, Penobscot Bay, Maine. Other potential sources of fluorine in the Central and Eastern United States are topaz, fluorapatite, rare-earth fluocarbonates, and magnesium-iron silicate-fluoride- hydroxides (chondrodite, humite, and norbergite), as well as commercial phosphate deposits. Topaz most commonly is associated with complex pegmatites, but not in commercial quantities. Such complex pegmatites occur throughout the Appalachian metamorphic belt from Maine to Alabama, and in central Texas. Notable occurrences are the pegmatites in Oxford and Kennebec Counties, Maine, and in New Hampshire on Baldface Mountain in Carroll County, where topaz occurs with beryl and lithium mica. Similar pegmatites are known near Amelia, Va., where very large white crystals of topaz occur in the cores of the Morefield and Rutherford complex pegmatites. Coarse tin-bearing quartz veins that contain topaz are found at Rockford, Coosa County, Ala., and in the Llano uplift of central Texas. Topaz most commonly is associated with complex gran- ite pegmatites, but not in commercial quantities. Such kyanite, andalusite, pyrophyllite, and diaspore, and in high-temperature veins containing gold, tin, and tungsten minerals. In the Central and Eastern United States large veins and replacement masses of topaz in potentially com- mercial quantities occur in the Appalachian meta- morphic belt, southeast Missouri, and the Llano uplift of Texas. CENTRAL AND EASTERN UNITED STATES Appreciable topaz occurs in fissure veins, some consis- ting of nearly pure, coarse, gray topaz and others consisting of pale-blue topaz crystals, with fluorite and scheelite, on Long Hill near Trumbull in southwestern Connecticut (Crowley, 1968). A similar but less-known occurrence is near Willimantic in eastern Connecticut, where topaz associated with fluorite, columbite, and scheelite is disseminated and in small veins in gneiss. One of the largest known deposits of topaz is at the Brewer gold mine, 1.5 miles (2.4 km) northwest of Jeffer- son, Chesterfield County, S.C. More than 700 tons of topaz was mined there before August 1941 (Fries, 1942) for experimental refractory purposes. At the Brewer mine, fine-grained rhyolite tuff and breccia have been metamor- phosed to quartz-sericite schist. In places the schist is completely silicified and contains pyrite, gold, and topaz. The topaz forms irregular replacement bodies ranging in width from less than an inch (2.5 cm) to 10 feet (3 m). Some topaz is in large veins and in veinlets; other topaz is dis- seminated in the schist associated with pyrite and gold (Pardee and Park, 1948, p. 106-111). The topaz contains 13-14 percent fluorine. The district contains probable reserves of 106,000 tons of rock aver- aging 15 percent topaz (Fries, 1942) or about 16,000 tons of topaz; some byproduct gold is present. In addition, possible reserves of 194,000 tons of similar grade topaz- bearing rocks might contain 0.05 ounce of gold per ton. The topaz very closely resembles slightly stained vein quartz and can be distinguished only with difficulty from it by a specific gravity greater than that of quartz and by the dull appearance of its weathered surfaces. A placer of topaz occurs downslope from the replacement deposits. Similar topaz deposits occur with pyrophyllite that is mined at Bowlings Mountain, Granville County, and at Corbett, Caswell County, N.C. At Bowlings Mountain large veins and bodies of massive topaz occur with pyro- phyllite and andalusite in Paleozoic sericite schist. At Cor- bett large veins and disseminations of topaz occur with kyanite in metavolcanic schist. Possibly some of the North Carolina topaz could be recovered profitably from material mined for pyrophyllite if there were a market for the topaz. Near Rockford, Coosa County, Ala., massive fine- grained topaz occurs in tin greisens and quartz veins in the extensive but not very productive Rockford tin district. A substantial tonnage of topaz could be recovered by concentration. Very fine grained gray topaz is abundant in greisens in vein walls of several tin, tungsten, and silver-lead veins of the Einstein silver district at Silver mines, southwest of Fredricktown, Madison County, Mo. Fluorite (p. 77) occurs in these veins also, and the two minerals together may constitute a potential fluorine resource especially if recovered as a byproduct of the high-value metallic minerals of the veins. Fluorine-bearing minerals fluorapatite (nonsedimen- 81 tary), chondrodite, humite, norbergite, doverite, and bastnaesite, in some places occurring alone but commonly with two or more associated together, form a large- tonnage potential source of fluorine, if the need should arise for such lower-grade materials or if extractive tech- nology improves. A few deposits of these minerals are very large and could supply substantial quantities of fluorine. Recently fluorapatite has been separated and discarded from metalliferous deposits at Mineville, N. Y.; Dover, N. J.; Piney River, Va.; Pea Ridge and Pilot Knob, Mo.; and, to a lesser extent, at Sterling Hill, N. J. Some of the larger deposits and the main minerals present are listed in the following table. Deposit Location Main minerals New York Palmer Hill mines ....... AUSADIC z.... ssrers d. Magnetite, _ fluorite, t> fluorapatite. Mineville iron mines ... Mineville .................... Magnetite, _ fluorite, rare-earth fluorapa- tite, bastnaesite. Trembley Mountain.... St. Lawrence County.. Magnetite, - flourapa- tite, fluorite. Tilly Foster mine......... Brewster, Putnam Magnetite, - chondro- County. dite, humite, fluor- ite, norbergite. Mahopac mines............ Mahopac, Putnam Do. County O'Nie!l mine...... Monroe, Orange Magnetite, fluorapa- County. tite, chondrodite. Amity-Edenville marble quarries. Edenville, Orange County. Chondrodite, fluorapa- tite. New Jersey Rudeville marble quarries. McAfee marble Sterling Hill zinc mine. Rudeville, Sussex County. McAfee, Sussex County. Sterling Chondrodite, _ fluor- apatite. Norbergite, fluora- patite. Chondrodite, norberg- ite, fluorapatite, fluorite, franklinite. Franklin zinc mine ...... Do. Scrub Oaks iron mine.. Dover ......................... Magnetite, zircon, _ * $ doverite. Canfield apatite mine .....dO.............................. Fluorapatite, magne- tite. Hurdtown apatite Hurdtown, Morris Fluorapatite. mine County. Pennsylvania Harvey's marble Brinton's Ford near _ Chondrodite. quarry. West Chester. Virginia Nelsonite dikes............. Piney River, Nelson County. IImenite, apatite. All these occurrences except Piney River, Va., are in Pre- cambrian skarns. Almost all are in the Adirondack Mountains, in the New Jersey Highland province, or in the Piedmont province. Many others undoubtedly exist and could be located by a detailed field or literature search. Some of the quarry occurrences, such as those in the mar- ble quarry at McAffee, N. J., contain substantial tonnages of minerals such as norbergite that could be quarried, milled, and concentrated on a large scale. 82 FLUORINE IN PHOSPHATE DEPOSITS By J. B. CatTHcaRT Fluorine in sedimentary phosphate deposits is con- sidered here separately as a type distinct from the epigene- tic fluorine deposits described in detail in preceding pages. Economic phosphate deposits in the United States are confined to sedimentary phosphorites of Ordovician age in Tennessee and Alabama, of Permian age in the Western States, of Miocene age in North Carolina, Georgia, and South Carolina, and of Miocene and Pliocene ages in Florida (fig. 11). Total reserves are billions of tons of phos- phate rock containing 28-38 percent P;0, and more than 3 percent F. Some fluorine is being produced in Florida as a by- product of manufacture of chemical fertilizer from phos- phate rock, and probably much more will be produced in the future (Sweeney, 1971). In January 1971, at the time of a field investigation in Florida by the author, phosphate companies reported that all the phosphate chemical plants there were recovering at least part of the fluorine produced during chemical or thermal treatment of phos- phate rock. Production of phosphate rock in the United States in 1970 was about 37.6 million short tons, a decline for the second straight year. Production data, and tonnages of phosphate rock by major uses, are shown in the following table. Production and uses of phosphate rock, 1970, in millions of short tons [Data from U.S. Bureau of Mines Mineral Industry Surveys-Phosphate Rock, Crop Year GEOLOGY AND RESOURCES OF FLUORINE phosphate rock so treated contained about 875,000 short tons of fluorine, assuming an average fluorine content of 3.5 percent in the rock, not all of which was liberated in gaseous effluent. Siems (1951, p. 409) indicated that about one-third of the fluorine is liberated during the manu- facture of ordinary superphosphate, and phosphate companies indicate that this figure is approximately correct for the process used in chemical plants in Florida. About two-thirds of the fluorine evidently combines with calcium and does not go off in the gaseous effluent. The total amount of fluorine currently being produced from phosphate rock is not known. Effluent gases con- taining fluorine are treated to remove as much of this element as possible (to avoid air pollution), mainly by passing the effluent through powdered limestone. The fluorine in the gas combines with calcium in the lime- stone to form CaF;, but this fluorite was so finely divided that early attempts to recover it were not successful, and the material was discarded as waste. In 1972 several com- panies were making aluminum fluoride (AIF,) and arti- ficial cryolite (Na;,AIF,) from the hydrofluosilicic acid derived from the fluorine in the gas (see p. 84). Partial chemical analyses of phosphate rock given in the following table represent deposits that were being mined in 1972. Partial chemical analyses, in percent, of phosphate rock from deposits in the United States [U.S. Geological Survey analysts: -P;0; by G. D. Shipley and Winona Wright, F by Johnnie Gardner, AliOs by H. H. Lipp and Wayne Mountjoy, Fe and CaO) by Violet Merritt and Johnnie Gardner, and SiO; by W. D. Goss, H. H. Lipp, and Elsie Rowe] Annual, November 1970] w. ¥7I¢§am E* (aanndd ble digg aséefn 50311}! em nglina! Florida land-pebble district 'Tenriessect PRODUCTION Georgia® Pebble _ Concentrate} Florida (Includes land-pebble, Northern Florida field, and 30.5 33.2 29.7 31.4 34.5 32.8 NOFUA reset 42s s 29.7 44.0 47.0 46.2 46.3 49.7 46.1 Tennessee (Includes mines in northern Alabama)....................... 3.0 §.1 3.8 3.6 3.6 3.8 3.6 Western States (Montana, Wyoming, Idaho, Utah, and 11.9 6.6 2.9 9.3 5.0 8.6 CaOMAIQ) ...s. eis Cen vere 4.9 1.7 1.5 .5 1.3 .9 2.2 Tome MER. oll aad Sem an 37.6 12 8 6 1.2 ' al Average Phosphoria Formation whole-rock analysis (Gulbrandsen, 1966, table 1). *Average of three samples of concentrate. Analyses by U.S. Geological Survey. MAJOR USES SAverage of 40 samples of concentrate. Analyses by phosphate companies and U.S. Geological Survey. z , ; fa, *Average of 20 samples of +1 mm material. Analyses by U.S. Geological Survey. Chemical or thermal processes in which fluorine is liberated; total 67 percent = "Average of 20 samples of concentrate (<1 mm+0.1 mm). Analyses by U.S. Geological h urvey. PhOSDhOFiC CIG (WEt DOCES$) ...... ..... rere rrr 10.1 of 6 samples (he?!) and gihors 1928, table 1 Electric furnace phosphorus.... . 6.8 The analyses listed are all averages; those from the Triple superphosphate .......... - 38 | western field and from Tennessee are for total rock; the Ordinary SUPETDPROSDRAE ........ 4.5 Uses in which no fluorine is liberated; total 33 percent EXDOFES ..:: ov nes ass rec anar cern rr rer nena ronin vies 10.9 All others.. Fluorine is liberated during acidulation and electric furnace reduction of phosphate rock as a mixture of poly- meric forms of hydrogen fluoride, which attack silica and are released as silicon tetrafluoride gas (Siems, 1951). About two-thirds of the total production of phosphate rock currently is treated chemically or thermally to produce superphosphate and a gaseous effluent that contains fluorine. In 1970, the 25 million tons of others are for beneficiated material, either flotation con- centrate or coarse (pebble) fraction of the land-pebble dis- trict of Florida. The average analyses are thought to be rep- resentative of the large tonnages shown in the following table of reserves. The fluorine content ranges from 3.1 to 3.8 percent. Individual samples of phosphate rock from the western field contain as much as 4.2 percent F (Gulbrandsen, 1966), and individual samples of Florida concentrate contain as much as 4.3 percent F; the highest grade samples of Florida pebble and of North Carolina concentrate contain about 3.8 percent F. Reserves of phosphate in the United States total about 10 billion tons as shown in the following table, but these PRESENT AND FUTURE RESOURCES OF FLUORINE 83 figures represent only a fraction of the total resources. For example, reserves in Wyoming are listed as 2 billion tons, whereas total resources in Wyoming (Sheldon, 1963, table 20) are about 19 billion tons, and the total resources of phosphate in North Carolina are estimated to be about 10 billion tons. Not all the total resources are minable, and the chemical composition of the material is not well known, but it is certain that the additional resources will contain lower percentages of P,0, and fluorine than the reserves shown. Phosphate reserves, United States, in millions of long tons MAap No: Reserves References (Fig. 11) Eastern United States 1. Florida, land-pebble............ 1,000 McKelvey and others (1953). 2. Northern Florida-southern 150 Sever, Cathcart, and Patterson Georgia. (1967); Olson (1966). 3. Northern Georgia-southern _ 800 Furlow (1969); Heron and South Carolina. Johnson (1966). 5. North Carolina.................... 2,000 U.S. Geol. Survey (unpub. data). 6. 80 Jacob (1953). Total.. ..that niin 4,030 Western United States (Includes reserves above entry level and within 100 feet (30 m) below entry level that contain more than 24 percent P,0,. Preliminary figures.) i - breast 2,700 W. C. Gere (written commun., 1972). 8. Wyoming.... ,000 Do. 9. Utah........ 900 Do. 10. Montana.... 400 Do. Total 6,000 PRESENT AND FUTURE RESOURCES OF FLUORINE By D. R. Suaws, R. G. Wort, R. E. Van Arstin®, and A. V. Heyr Present and future resources of fluorine in the United States will be examined largely in the light of past exploi- tation. Until recently fluorspar was virtually the only source of fluorine, and probably will remain the major source in the near future. A relatively minor amount of the fluorine consumed is presently being recovered from pro- cessing phosphate rock, and an insignificant amount has been acquired in the past from topaz deposits. Phosphate rock and other byproduct and coproduct sources of fluorine will become increasingly important in the future. Byproduct recovery from fluorspar mining throughout the United States in the past has not been extensive; sul- fide concentrates containing mainly lead, zinc, and silver are recovered in some fluorspar operations, and in a few barite is a coproduct. During the period 1900-70 the United States has changed from the major world producer of fluorspar to about seventh or eighth rank, although its production has generally risen. According to Worl, Van Alstine, and Shawe (1973, fig. 1), during 1900-20 annual fluorspar pro- duction grew from about 18,500 short tons to an average of about 200,000 short tons of CaF;; during 1920-40 it averaged somewhat more than 100,000 short tons; and during 1940-70 average annual production increased from about 300,000 short tons to a maximum of more than 400,000 tons during World War II, and then decreased to about 200,000 tons. Until about 1950 United States pro- duction of fluorspar nearly equalled its consumption, but since 1950 consumption has increased dramatically (to about 1,500,000 short tons CaF; in 1970) and so domestic production now supplies less than a quarter of the needed fluorspar, and imports have been required for the balance. Furthermore, since 1950 worldwide fluorspar consump- tion has increased spectacularly to more than 4,500,000 short tons CaF,, emphasizing the fact that the United: States faces strong competition for foreign supplies of fluorspar. SUMMARY OF FLUORINE RESOURCES By R. E. Van Arstin® and R. G. Wort World reserves of fluorspar appear to be adequate to meet fluorine requirements in the foreseeable future. Similarly, United States reserves of fluorspar appear to be adequate for domestic needs for the foreseeable future. Worl, Van Alstine, and Shawe (1973, table 1) have esti- mated that present world fluorspar reserves outside the United States are about 165,000,000 short tons of crude ore and United States reserves are about 25,000,000 short tons of crude ore (crude ore containing more than 35 percent CaF,). Hypothetical fluorspar resources (defined as fluor- spar that is undiscovered in known districts, or is exploit- able only under more favorable economic or technologic conditions-the latter termed "conditional" resources by Brobst and Pratt, 1973, p. 4) have been estimated as almost 530,000,000 short tons in the world outside the United States and as about 45,000,000 short tons in the United States (Worl and others, 1973, table 1). Note that the ratio of identified reserves to hypothetical resources is about 1:2 for the United States whereas it is about 1:3 for foreign deposits, reflecting the fact that United States resources generally have been more extensively explored and exploited than have foreign resources. Since 1927 several estimates of the fluorspar reserves of the United States have been made by government geolo- gists and mining engineers, and each successive estimate except the 1968 estimate has been a larger figure. In 1927 a reserve of about 5 million tons was estimated for the Illinois-Kentucky district (Iron Age, 1927). In 1987 (Bur- chard, 1987) this figure was increased to 6 million tons because of the discovery of more replacement deposits in the Cave in Rock area of Illinois (Currier, 1937), and by 1 million tons for western deposits (Burchard, 1933). A joint estimate in 1947 of the U.S. Geological Survey and U.S. Bureau of Mines (Davis and others, 1947) showed 14.7 million tons of fluorspar; a 1952 estimate (Paley, 1952) 84 GEOLOGY AND RESOURCES OF FLUORINE gave the reserve at 15 million tons. A more detailed esti- mate by Van Alstine (U.S. Geol. Survey, 1956) of about 250 fluorspar deposits in the United States showed about 22% million tons (see the following table); 50 deposits con- tained 77 percent of the ore reserves. A similar estimate of 21 million tons was made by a geologist in private industry (Sutton, 1954). Estimated reserves of higher grade fluorspar in the United States as of October 1956, by regions, in terms of short tons of crude ore [Higher grade fluorspar contains at least 35 percent CaF; or equivalent] Measured and Approximate Location indicared Inferred Total percent Illinois, Kentucky .. 8,100,000 _ 4,000,000 _ 12,100,000 54 Colorado, Idaho, Montana, Utah, Wyoming............ 4,500,000 _ 3,700,000 8,200,000 36 New Mexico, Arizona, Texas.... 800,000 700,000 1,500,000 7. California, Nevada, Washington ........ 400,000 350,000 750,000 3 New Hampshire, Tennessee, 2 < loving 10,000 10,000 : ...... Fotal :::...... 13,800,000 _ 8,760,000 _ 22,560,000 100 In 1968 Van Alstine (U.S. Geol. Survey, 1968) estimated that fluorspar reserves were decreased to about 10 million tons because about 8 million tons of crude ore had been mined, new ore had been discovered, and about 7.5 million tons was transferred to the potential resource category. This quantity was no longer regarded as ore because it became more profitable to use cheaper imports of fluor- spar from Mexico, Italy, and Spain than to mine from many of our domestic deposits. The current estimate of 25,000,000 tons takes into account the improved market situation as well as new fluorspar discoveries (Worl and others, 1973, p. 230). Nearly all these reserves are believed amenable to bene- ficiation into acid-grade and metallurgical-grade concen- trates. Even ores in which fluorite is intimately intergrown with gangue minerals might be made into concentrates acceptable to the steel industry with the use of sink-float methods, blending of ores or concentrates, and the sintering, nodulizing, or pelletizing of material that meets the chemical specifications of metallurgical-grade fluor- spar. In recent years increased use of heavy-media precon- centrators, the pelletizing of flotation concentrates into metallurgical-grade fluorspar, and cost-cutting mining practices have permitted the working of lower grade ores, even though salary scales and material costs have mounted. However, the price of fluorspar has increased above previous prices so that in 1972 all grades averaged about $70/ton. Deposits included in the estimate range in size from less than 10,000 tons to at least 500,000 tons. Replacement bodies of fluorspar in the Cave in Rock area, Illinois, average about 9 feet (3 m) in thickness and about 125 feet (38 m) in width. Dimensions of typical veins in New Mexico are given by Rothrock and others (1946, p. 21). Potential resources of fluorspar in the United States total about 45 million tons of material containing 15-35 percent CaF,. Such resources, located chiefly in Illinois, Kentucky, Colorado, Montana, Nevada, New Mexico, Utah, Idaho, Texas, and Alaska, are incompletely known because few marginal or submarginal fluorspar deposits have been explored. A substantial addition to our poten- tial resources resulted from the recent (The Northern Miner, Toronto, Feb. 18, 1971) estimates for the Lost River area, Alaska, where about 10 million tons of material con- tain 28.5 percent CaF, and 0.2 percent Be; 5.3 million tons of this material also contain 0.42 percent Sn and 0.1 per- cent WO (Sainsbury, 1964b; 1969). Sedimentary phosphate rock which contains 3-4 per- cent F constitutes a largely untapped fluorine resource; fluorine has been recovered from phosphate rock only in recent years and it does not yet constitute an appreciable fraction of the market. On the basis of a fluorine content of about 3 percent in marine phosphate rock, the phosphate deposits of Florida, North Carolina, South Carolina, Tennessee, Alabama, Utah, Wyoming, Idaho, and Mon- tana contain about '/; billion tons of fluorine in known reserves and about 2 billion tons of fluorine in identified resources. This indicates that marine phosphate rock con- stitutes by far the United States' and the world's largest fluorine resource. Phosphate rock soon could become an important source of artificial fluorite, artificial cryolite, fluorine, or fluorine chemicals; two new fluosilicic acid recovery units were put into operation in recent years (Everhart, 1963) by phos- phate producers in Florida to supply raw material to a new fluorine chemical plant in the area. Aluminum Company of America has announced (American Metal Market, 1969) construction of a new plant in central Florida to make aluminum fluoride and cryolite from fluosilicic acid from phosphate rock. The production of fluosilicic acid and sodium fluosilicate from phosphate rock for water fluo- ridation is increasing, especially with greater efforts to reduce air pollution around phosphate plants. Topaz, found relatively rarely in large concentrations, occurs in probable commercial quantity and grade in at least two localities in the United States. Reserves at the Brewer gold mine in South Carolina were estimated at about 100,000 tons of schist averaging approximately 15 percent topaz and about 1,200 tons of topaz in adjacent placer deposits that contain 1-24 percent topaz. The hypo- thetical resources at this deposit are estimated at about 800,000 tons of topaz rock. In the Front Range, Colo., a lens in Precambrian gneiss containing about 15 percent topaz was estimated to constitute a reserve of about 600,000 tons of topaz-bearing rock. The hypothetical resources of this topaz-bearing rock are estimated to be about 250,000 tons for every 100 feet (30 m) of depth. Topaz also occurs in significant amounts (along with fluorite) in the major molybdenum porphyry deposits at PRESENT AND FUTURE RESOURCES OF FLUORINE Climax and Henderson, Colo., and Questa, N. Mex. Reserves of fluorine at these deposits have not been estimated but byproduct resources in molybdenum por- phry deposits may constitute as much as several million tons of fluorine. Bastnaesite in carbonatite at Mountain Pass, Calif., con- tains a potential fluorine resource. This rare-earth mineral constitutes 5-15 percent of the carbonatite rock and con- tains about 7 percent fluorine. Based on an estimate of 100 million tons of potential ore resources at Mountain Pass (Olson and others, 1954), nearly 1,000,000 tons of fluorine exists as a potential byproduct of rare-earth extraction. KNOWN FAVORABLE AREAS FOR FLUORINE EXPLORATION By D. R. Saws, R. E. Van Arsting®, R. G. Wort, and A. V. Heyt Search for fluorspar naturally should first be made close to areas where the mineral already has been found. In the productive districts extensions of known fluorspar bodies should be explored along strike and in depth. Additional ore bodies may be found along nearby faults in mineral- ized districts. Search should be directed especially for stratiform deposits in limestone which have been sus- ceptible to replacement by fluorite; these deposits tend to be large and amenable to large-scale low-cost mining operations, and in the important southern Illinois- western Kentucky district sizable areas having economic potential have not yet been adequately explored. Geochemical investigation of fluorine and related metal anomalies may reveal commercial deposits. Disseminated, lower grade deposits marginal to known bodies may be profitable; some drab, fine-grained, earthy material is difficult to recognize visually as fluorite. An increasingly important source of fluorspar will be as a coproduct from the mining of iron, base-metal, barite, and rare-earth ores. As in Mexico, fluorite may be recovered also from mine dumps and tailings ponds, where it is already ground and partly concentrated. More fluorine analyses should be made routinely on earth materials; thus a new environment for fluorspar may be discovered, as recently described in Tertiary tuffaceous lake beds of Oregon (Sheppard and Gude, 1969). Most fluorspar districts and localities in the United States remain as areas favorable for further exploration for fluorine resources. The recent discovery of large fluorspar reserves in the Lost River tin district, Alaska, suggests that similar geologic settings along the Rapid River thrust fault elsewhere in the York Mountains are favorable for the occurrence of such carbonate-replacement and pipe- like fluorspar deposits. Other tin districts on the Seward Peninsula, fluorite-mineralized carbonate rocks north of Nome, breccia pipes in the Kigluaik Mountains, and carbonate rocks surrounding alkalic rocks in the Darby Mountains, all offer potential for undiscovered fluorspar deposits. Elsewhere in Alaska, areas where peralkalic 85 granites are known or where fluorite and topaz are associated with sulfides are favorable for fluorine mineral concentrations. Large fluorine resources are available although currently inaccessible in phosphate rock in northern Alaska. In the northwestern part of the United States important fluorspar deposits that may have additional potential are the Challis, Big Squaw Creek, and Meyers Cove deposits, Idaho, where fluorite veins occur in dolomite, biotite gneiss, and volcanics and porphyries, respectively; and the Snowbird and Crystal Mountain deposits, Montana, which are pegmatitelike bodies. Numerous other fluor- spar deposits associated with volcanic rocks, granitic intrusive rocks, and alkalic rocks, and occurring as peg- matites and disseminated in granite are widespread in the Northwest, specifically in north-central Washington, cen- tral and southwestern Idaho, western to central Montana, several areas in Wyoming, and the Black Hills of South Dakota, and the areas of these deposits may be favorable for further exploration for fluorspar resources. The States of Idaho, Montana, Wyoming, and Utah contain extensive fluorine-bearing phosphate rock; as the distribution of these deposits is well known, large unknown resources of fluorine in phosphate rock are not likely in this region. In the southwestern part of the United States substan- tial fluorspar potential is likely in the vicinity of major known districts as well as around districts of only minor production. Additional replacement bodies in limestone in the Fluorine district, Nevada, vein deposits in andesite and rhyolite at Broken Hills, Nev., pipe deposits in carbo- nate rocks at Spor Mountain, Utah, and veins in volcanic rocks in the Indian Peak Range, Utah may still be found. Throughout much of Nevada, southeastern California, western Utah, and western Arizona numerous minor occurrences of fluorspar in veins, mantos, and pipes in and associated with volcanic rocks, in contact zones and stock- works of hypabyssal and plutonic igneous rocks, in vol- caniclastic sediments, and in pegmatites, mark possible targets for exploration for additional fluorspar resources. Of the localities having had only minor fluorspar pro- duction, perhaps the most promising are the Quinn Canyon Range, Nev., and the Castle Dome district, Arizona. The southern Rocky Mountain region of the Western United States in Colorado, New Mexico, western Texas, and southeastern Arizona offers numerous favorable areas where additional fluorspar resources may be discovered in the vicinity of known districts and deposits. Among the deposits characterized as layered and crustified veins in young steep tensional faults and breccia zones occurring in Tertiary volcanic rocks, Paleozoic and Mesozoic sedi- mentary rocks, and Precambrian silicic igneous and meta- morphic rocks are the Northgate and Browns Canyon districts, and deposits at Crystal, Poncha Springs, and Wagon Wheel Gap, Colo., and the Gila, Burro Moun- 86 GEOLOGY AND RESOURCES OF FLUORINE tains, Anderson, Gold Hills Steeple Rock, Sierra Caballos, Zuni Mountains, Fluorite Ridge, and Cooks Peak districts, and deposits at Tonuco, Tortugas Mountain, Bishop Cap, and Sierra Cuchillo, N. Mex. The Sierra Caballos, Sierra Cuchillo, and Tortugas and Bishop Cap deposits in New Mexico and the Eagle Mountains and Christmas Mountains-Corazones Peak districts in Texas also offer potential for manto-type deposits. Significant stockworks, pipelike shoots, and mineralized breccia zones near a middle Tertiary alkalic-silicic stock and contain- ing fluorite, lead, silver, gold, copper, and uranium minerals may still be found in the Jamestown district, Colorado. Other minor occurrences in the Southern Rocky Mountain region also may mark favorable targets for significant undiscovered fluorspar deposits. Numerous favorable structures in the Illinois- Kentucky district should be explored further for fluorspar. A very large potential for undiscovered deposits still exists in the district. Large low-grade tonnages of fluorspar-mineral- ized breccia are known near Hicks dome but have not yet been developed. Elsewhere in the Central and Eastern United States a large potential also exists for the discovery of new fluorine resources in the vicinity of known districts and deposits. All districts and deposits warrant further evaluation for additional resources, but only those with the most prom- ise are mentioned here. The Central Tennessee district may have the largest fluorspar potential in the United States outside of the Illinois-Kentucky district. Breccia fluorspar deposits are known at depth in the Central Ten- nessee district, and similar deposits may occur in quant- ity in Ordovician carbonate strata in the Cumberland River vein district in Kentucky. The Central Kentucky district offers good possibilities for undiscovered fluor- spar-bearing bedded and breccia deposits and veins. Fluorspar-bearing veins in or near the Rossie district, New York, promise additional potential. The Sweetwater district of eastern Tennessee and similar zinc-barite dis- tricts with fluorite throughout the Appalachian Valley and Ridge province between Alabama and Pennsylvania offer fluorspar potential. Districts where fluorite occurs with quartz and in mineralogically more complex veins are known in the Blue Ridge, Piedmont, and New Jersey Highland provinces from Alabama to southern New York and in New England. The most favorable for exploration are the Del Rio district, Tennessee, and the Hot Springs district, North Carolina. Ordovician carbonate rocks in the subsurface in the vicinity of the alkalic intrusive cen- ter at Magnet Cove, Ark., may have potential for fluorspar deposits. Careful geologic mapping where fluorine minerals are known but where adequate geologic maps are not now available will be essential to the development of addi- tional fluorine resources in many areas, for example, as demonstrated at Lost River, Alaska. SUGGESTIONS FOR LIKELY FUTURE FLUORINE PROVINCES: SPECULATIVE FLUORINE RESOURCES By D. R. Saws, R. G. Wort, and A. V. Exploration for and evaluation of speculative fluorine resources-those occurring in districts and environments that are yet to be identified-must be based on knowledge of the geologic environment and the geochemical cycle of fluorine. Geologic mapping in areas not now adequately covered by geologic maps will be essential to evaluation of geologic environment and to successful exploration in such areas. Major geologic parameters that partly define the geologic environment of fluorine mineral deposits are: (1) extensional tectonism related to deepseated north- westerly and northeasterly conjugate strike-slip struc- tures; (2) high heat flow from the mantle; and (3) magma- tism of alkalic affinity characterized by abundant fluorine. Regions of extensional tectonism are clearly the locales of fluorine mineral deposition in the recent geologic past and presumably have been in older geologic periods. Fluorine deposits of middle Tertiary and younger age are widespread in regions of rift and block faulting also of the same age, and in fact numerous fluorine deposits occur in tension faults in these provinces. Tension fractures in regions of extension appear to be a phenomenon of the upper part of the crust, but no geologic evidence suggests that most of the fluorine itself was derived from the upper part of the crust to be concentrated into tension fractures. Furthermore, the patterns of distribution of fluorine mineral deposits within tensional provinces, that is, the tendency for deposits to form northwesterly and north- easterly alinements, reflects the similar alinement of deeper-seated regional fracture zones of strike-slip character that probably penetrate downward through the crust and constitute the basic structural framework of regions of tension. The general distribution of fluorine mineral belts in extensional provinces thus suggests derivation of fluorine from deep within the crust or from the upper mantle. Broad regions of high heat flow in the United States coincide with regions of extensional tectonism that is now going on or has occurred in the recent geologic past. Such regions in the Western United States are further dis- tinguished by low seismic velocity in the upper mantle, suggesting abnormally high temperature in the upper mantle that in turn accounts for the high heat flow in the overlying crust. Of course such regions are also character- ized by abundant hydrothermal activity, including present-day hot springs. Low seismic velocity in the upper mantle also may indicate a phase change from dunitic to eclogitic composition, implying chemical transfer between mantle and crust, perhaps fractionation of mobile components from the mantle into the crust. Regions of high heat flow thus provide two significant factors essential to the formation of fluorine mineral deposits: an PRESENT AND FUTURE RESOURCES OF FLUORINE 87 original source of fluorine (from the upper mantle), and the energy necessary to mobilize fluorine into upper levels of the crust. The distribution of fluorine-rich igneous rocks of alkalic affinity in the Western United States coincides with regions of extensional tectonism and high heat flow. Again, the ages of the igneous rocks generally coincide with the periods of tectonism and of inferred correlative high heat flow. The foregoing generalization needs elaboration, however, in order to account for the occur- rence and distribution of alkalic rocks of different ages and petrologic character with which fluorine mineral deposits are associated. According to one model, during early to middle Tertiary time in the Western United States imbricate subduction took place as a result of differential crustal plate movements, such that an eastward-dipping oceanic plate underrode the continent in two north- trending zones parallel to the continental margin, one underlying the western segment of the Cordillera and the second underlying the present Rocky Mountain trend. In each zone potassium content of the andesitic igneous rocks that were generated by subduction increases eastward (Lipman and others, 1972). Rocks formed along the eastern edge of the Rocky Mountain zone are typically of alkalic character-including trachytes, trachybasalts, phonolites, and other alkalic types-suggesting that their derivation from the subduction zone was at greater depth and under stronger influence of mantle conditions than was the derivation of calc-alkalic types of similar age. Our colleague T. A. Steven suggested orally to us that deriva- tion from a mantle environment also accounted for the higher fluorine content of these rocks and the association with them of fluorine mineral deposits. * In middle Tertiary time extensional tectonism was in- itiated in the Western United States, marked by a cessation of subduction in the regions of extension (Lipman and others, 1972). Onset of extensional tectonism in the Western United States was marked by transition from andesitic to fundamentally basaltic or basaltic-rhyolitic volcanism (Christiansen and Lipman, 1972). The cause of the generation of fluorine enriched magmas of alkali rhyolite and associated basaltic types during the episode of extension is problematical. We infer it to be likely that such magmas originated in the upper mantle, probably as molten fractions of components with high vapor pressure, that were transported upward into the lower-pressure en- vironment of the crust. According to our belief, distention of the continental lithospheric plate due to differential plate motion resulted in at least incipient dilatant zones in the crust and in the underlying mantle, such that fluorine- enriched magma and fluids moved into the crust from the mantle, and the system of northwesterly and northeasterly strike-slip faults in the crust which implemented the dis- tention and accounted for the block faulting in the upper part of the crust allowed passage of magmas and fluids from the mantle farther upward through the crust. The suggestion that change in state within the upper mantle resulted from dilatancy due to differential plate motion, rather than the converse, was also proposed by Christiansen and Lipman (1972, p. 276). Thus we see that although fluorine-enriched magmas are spatially associated with regions of extensional tectonism they are in minor part older than the episode of extension and related to other processes. We believe the significant fact of their origin is that it was influenced by mantle processes, and the material of the magmas was derived from the mantle. The generalizations above, drawn on the relations observed in the Western United States, can be extrapola- ted to other regions. We propose, for example, that fluorine-enriched magmas of late Paleozoic age in the Eastern United States were derived from the lower zones of subduction-and hence strongly influenced by the mantle-in a tectonic belt that then bordered the eastern edge of the continent. These magmas were responsible for formation of some but not all of the fluorine mineral deposits in the Appalachian belt of the Eastern United States. Also probably in late Paleozoic time deep- penetrating strike-slip faulting along the 38th parallel of latitude in the Eastern United States (Heyl, 1972) allowed mantle-derived magmas and fluids to move upward into the upper part of the crust in the area of southern Illinois and western Kentucky to account for the extensive development of fluorine mineral deposits there. In Africa, the notable restriction of halogen-rich igneous rocks and fluorine mineral deposits to the rift-valley zones reflects the deep penetration by those structures. Fluorine apparently is very mobile in the geologic en- vironments described in preceding paragraphs. Fluorine- bearing solutions genetically related to the major geo- logic parameters specified have permeated country rock, such as altered and mineralized volcaniclastic sediments of the Western United States; have filled fractures in and replaced receptive country rock, as in the Illinois- Kentucky district; and have filled subsidiary fractures in places some distance from igneous activity and major structures, as in many fluorspar districts of the Western United States. Fluorine also moved into other geologic en- vironments associated in time and space with the major geologic parameters to form deposits in the other en- vironments. To use an example from another continent, fluorine mobilized by exhalations or weathering has been transferred from alkalic igneous rocks and carbonatites along African rift structures into lake brines, to be con- centrated in evaporites of the rift valleys. Formation of the marine sedimentary carbonate fluoraptite deposits is an important example of reconcentration of fluorine in the supergene cycle. Older concentrations of fluorine minerals may be re- worked virtually in place by younger processes that tend to 88 GEOLOGY AND RESOURCES OF FLUORINE mobilize fluorine. In such cases it is impossible to tell on the basis of current knowledge whether younger deposits associated with older deposits consist entirely of "primary" introduced fluorine, or entirely of remobilized older fluorine, or a combination of these. One example is fluorite in the Tertiary gold ores of Cripple Creek, Colo., that are associated with fluorine-rich phonolites and related rocks of Tertiary age (Lindgren and Ransome, 1906, p. 122, 219, 100) and also occur within fluorine-rich Pikes Peak Granite of Precambrian age. Another is fluorite in faults of post-Precambrian age at Mountain Pass, Calif., that are associated with dikes of Tertiary age and also are close to bastnaesite-bearing carbonatite of Precambrian age. These occurrences recall the point emphasized by Peters (1958, p. 667, 683-684) of recurrent mobilization of fluorine from older deposits in fluorine- rich regions. The major geologic parameters previously discussed de- fine environments and processes in which fluorine was intrinsically available in abnormal amounts in some part of the system, and could be mobilized into concentrations of economic interest. On the other hand, geologic processes in other environments also have acted to concen- trate minor (crustal abundance) amounts of fluorine into deposits of economic interest, for example, the precipita- tion of fluorapatite from seawater. The geochemical cycle of fluorine can be summarized to show that the inferences are valid regarding the geologic environments in which fluorine is concentrated. Despite the fact that fluorine-rich alkalic rocks appear to be derived from the mantle, and perhaps from deeper in the mantle than its top, rocks irrupted through oceanic crust and clearly derived from the mantle are low in fluorine. In the oceanic environment only magmas that have differ- entiated in near-surface chambers have tendencies toward alkalic affinities and fluorine enrichment. In the continental environment, however, some fluorine-rich magmas likely were derived from the mantle by a process that resulted in fluorine enrichment, or were differ- entiated in higher crustal levels to account for fluorine en- richment; some fluorine-rich magmas may have been generated within the crust in a local fluorine-enriched en- vironment, but we do not believe that this was a dominant process, and it probably occurred in a sporadic fashion. The tendency for nigh mobility of fluorine within the crust is exhibited by the occurrence of fluorine mineral halos around fluorine-rich intrusives, by the fact that ex- trusive rocks tend to contain less fluorine than do intru- sive rocks, showing a loss of fluorine under lower pressure conditions, and by the fact that volatile components emitted by volcanoes formed from fluorine-rich magmas include substantial fluorine. Geochemical data also show that crystallization of magma tends to drive off fluorine if minerals that can take up fluorine are not formed. Weathering of rocks at the earth's surface transfers large amounts of fluorine into surface and ground waters; some of this may be reconcentrated in the near-surface environ- ment, such as in lake brines, but much is transported to the oceans where it is redeposited in sedimentary rocks that generally contain amounts of fluorine comparable to those of igneous rocks. Under special conditions in the oceans large amounts of fluorine become concentrated in sedimentary phosphate deposits, yet these amounts are still only a small percentage of the total fluorine deposited in marine sediments having lower fluorine content. Metamorphism of sedimentary rocks may cause segre- gation of fluorine, but evidence of this process is fragmen- tary. In this respect, Cannon and Pierce (1967) stated that most stratiform lead-zinc-barite-fluorite deposits "in the central lowland of North America contain J-lead of variable isotopic composition, part or all of which originated by mobilization of rock-lead from the Precam- brian basement or enclosing marine sedimentary rocks." Whether the source of the fluorine in these deposits is the same as that of the lead is a moot point. Ultrametamor- phism of sedimentary and other rocks having concentra- tions of fluorine may result in development of fluorine- rich magmas and associated fluorine mineral deposits, but again more evidence is needed to prove that such a process has occurred. Negative fluorine-silica correlations in some igneous suites suggest this mechanism as they indicate formation of the igneous rocks by other than magmatic differentiation. The principles discussed in preceding pages assure us that numerous possibilities exist for new fluorine resources. At present it is impossible to give a quantitative indication of speculative resources other than that they are probably very large. Generally, prospecting for fluorine mineral deposits in regions of relatively young tectonism and magmatism should have greatest success where regional north- westerly and northeasterly strike-slip structural zones have been intruded by fluorine-rich magmas, or are underlain by large volumes of alkalic volcanic rocks. In regions where older tectonism and magmatism have been obscured by younger geologic events these major geologic features should still be sought. Of course, other deeply penetrating structural zones, particularly major east-west fault zones, where fluorine-rich mantle-derived magmas have been irrupted, also should be favorable locales for fluorine mineral deposits. In the United States new fluorspar districts may be found northwest of the important Illinois-Kentucky dis- trict in carbonate rocks beneath Pennsylvanian rocks, and also northeast of this district toward Ohio and the Michi- gan basin, beneath Pleistocene cover where fluorite occur- rences are known in Silurian and Devonian dolomites. Northern Ohio in the vicinity of Toledo, the Serpent Mound cryptoexplosion structure in southeastern Ohio, southern Indiana west of Louisville, Ky., and the Ste. Genevieve fault zone of southeast Missouri and southern Illinois, may all contain undiscovered fluorspar deposits. GEOCHEMICAL AND GEOPHYSICAL PROSPECTING METHODS 89 Similarly, the Paleozoic carbonate rocks of central Tennessee and central Kentucky are widely mineralized with fluorite, as shown in recent drilling, and may con- tain other commercial deposits. Carbonate rocks in sub- surface in the vicinity of alkalic intrusives, at Magnet Cove, Ark., may have potential for new fluorspar districts. Large but lower grade stratiform fluorspar deposits in carbonate rocks, like those worked in the United States, France, USSR, and southern Africa, probably will be popular targets for exploration in the United States. In the Western United States fluorspar deposits un- doubtedly underlie parts of the widespread cover of Ter- tiary and Quaternary sediments and volcanic rocks. Cenozoic calcareous tuffaceous lake beds in the Western United States may be fluorite-bearing, similar to those known in Oregon, Italy, and southern Africa, some of which are of commercial interest. These and other occur- rences of this type of fine-grained fluorite may offer po- tential for large fluorine resources. Some fluorspar deposits of the Western United States are related genetically to areas of silicic-alkalic volcanic rocks, namely the Big Bend igneous subprovince of Texas; the Nacimiento igneous subprovince of Arizona, New Mexico, and Colorado; and the Shoshone igneous province of Nevada, Utah, Idaho, and Montana. Other areas with relatively high fluorine contents in the vol- canic rocks might merit further prospecting, especially southeastern Idaho, northwestern Wyoming, and north- western Utah. In the Thomas Range of west-central Utah, topaz is disseminated in volcanic rocks and commercial fluorspar deposits are distributed around the western periphery of a caldera. Only a third of the caldera is exposed; the remainder, which is covered by rhyolite flows or alluvium, may contain undiscovered deposits (Shawe, 1972, p. B76-B77). A tectonic unit consisting of the Rio Grande trough in New Mexico and its northern extension through Colorado is especially rich in fluorspar deposits and should be explored further. Deposits of syngenetic fluorite associated with gypsum and limestone are reported from Permian marine evaporites in the Bighorn Basin, Wyo. These resemble extensive deposits known in the Permian of USSR; similar evaporites elsewhere in the United States should be investigated for fluorite deposits. Chao and others (1961) reported a new fluorine mineral, neighborite (NaMgF;), from the Green River Formation, Utah; the economic potential of this mineral here and else- where is unknown. Large deposits of fluorapatite in marine phosphate rock probably exist beneath the sea in Tertiary (and younger?) coastal plain sediments extending east from phosphate deposits in the Carolinas and Florida. Although fluorine is recovered from huge deposits of fluorapatite in alkalic rocks of the USSR, the possibility is not great that deposits of this type may be found and profitably explored in the United States. Probably millions of tons of topaz-bearing rock are yet to be discovered in widespread metamorphic terranes throughout the United States as well as in still-unknown porphyry deposits. Because cryo- lite is a rare mineral, it may not be an important source of fluorine; the world's only known commercial deposit of natural cryolite, in Greenland, has been depleted, but similar deposits elsewhere are conceivable. Sodium fluoride has been found in saline lake beds in Tangan- yika; it may be present in similar rocks in the United States and may some day be a source of fluorine. Some brines, volcanic gases, fumaroles, and hot springs with relatively high fluorine contents may suggest fluorine-rich provinces in which deposits of fluorspar, phosphate rock, or other fluorine materials should be sought. Deep-sea sediments, which have more fluorine than most other marine sediments except phosphate rock, contain about 700 ppm F; these types of sediments and sea water itself, which has about 1 ppm F, may be sources of fluorine in the distant future. Other possible new resources of fluorine, not in the cate- gory of speculative resources but classified as conditional resources, are known small but high-grade deposits, and very large probably low-grade multicommodity ores. Numerous high-grade but small deposits of fluorspar, such as those of the Western and Eastern United States might be utilized. To do so would require better tech- niques than now used of locating and evaluating small pods of fluorspar, and it would also require the use of portable extraction and processing equipment. Even though mining of this type of deposit may be feasible else- where in the world it is not likely to be in the United States, however, because of high labor costs and land problems. Very large low-grade multicommodity ores offer poten- tial fluorine resources. These include deposits in which fluorine, mainly in the form of fluorite, is associated with metallic minerals and those in which it is associated with or contained in other minerals of commercial interest such as barite, bastnaesite, fluorapatite, zeolites, and feldspars. It may be possible in the future to process such multicom- modity ores in which no single commodity is of high enough grade to be commercial by itself. GEOCHEMICAL AND GEOPHYSICAL PROSPECTING METHODS FOR FLUORINE By W. R. GrirFitts and R. E. Van Austin® Fluorspar deposits have commonly been found by iden- tification of fluorite in outcrops and surficial materials (Grogan, 1960, p. 374; 1964). The surficial materials include soils, anthills, spoil piles of burrowing animals, landslides, and alluvium. The persistence of fluorite during weathering under certain conditions, and the high fluorine content of the mineral, suggest that geochemical methods might be used in prospecting for fluorspar. 90 GEOLOGY AND RESOURCES OF FLUORINE Nevertheless, geochemistry has not been used much, in part because of the lack of rapid analytical precedures that could be used to determine the fluorine contents of samples with the necessary reliability and low cost. Recent work in western North America has shown that fluorite deposits may effectively be sought there by heavy mineral and geochemical techniques. The mineral per- sists for long distances during transport in the sediment of streams, and can be concentrated therefrom with the gold pan. Thus, the fluorite of the Jamestown district, Colo- rado, can be detected in stream sediment as far down- stream as the South Platte River, a distance of about 35 miles (55 km). In southwestern New Mexico fluorite has similarly been traced in the sediment of dry washes for dis- tances of several miles. The softness and excellent cleav- age of the mineral thus do not cause rapid destruction of the grains during transport, nor does the rather low specif- ic gravity prevent its recovery by panning. The occurrence of fluorite in stream and seashore sand was summarized a decade ago by Milner (1962, p. 107). A systematic heavy-mineral survey near Winston, N. Mex., during 1970 by H. V. Alminas and K. C. Watts (oral commun., 1971; Alminas and others, 1972) showed that fluorite is plentiful in sediment near two major subparal- lel north-trending faults. Near the intersection of the two faults Ira Young of Socorro, N. Mex., had independently found a large fluorspar deposit. The fluorite that occurs without associated metals or with barite in wash sedi- ments is colorless or pale green, whereas many of the fluorite grains found in sediments with anomalous base- metal contents are purple. It may be possible in some areas therefore to distinguish accessory fluorite of metal deposits from fluorite of higher grade fluorspar deposits. Preliminary chemical analyses show that silts of dry washes contain most fluorine near known fluorspar deposits and where fluorite was panned from the sediment. Geochemical exploration investigations in the semi- arid Browns Canyon district, Colorado, by Van Alstine (1965) showed that fluorine is found in abnormal concen- trations in residual soil directly above and downslope from the principal vein and in alluvium downstream from the vein. The highest anomalous values were about 6-20 times the background value in the soil and 3 times the background in the alluvium. Fluorine thus can have rather low mobility in soils and alluvium in semiarid regions. Van Alstine also pointed out the problems intro- duced where samples contain abundant biotite and horn- blende that increase the background fluorine values-a problem which is avoided by the determining of contents of fluorite instead of fluorine. Neither heavy mineral nor geochemical prospecting methods for fluorite have been widely applied in fluor- spar districts in humid regions. Sang Kyu Yun and others (1970) found geochemical studies of stream sediment and soils to be useful in prospecting the Hwanggangri region of Korea. Nackowski (1957) found that fluorine was tech- nically unsuited as an indicator element for preliminary exploration for fluorspar in the Illinois-Kentucky dis- trict; evidently the humid climate there is unfavorable for fluorine to be highly concentrated in the soil. Development of thick or well-zoned soils in humid areas is commonly accompanied by removal of fluorine from near-surface materials. Hence, soil samples taken at shallow depth might show subdued anomalies or no anomalies over fluoritic bedrock. Therefore, where soils are well developed, samples should be taken as deep as possible. Soils now being eroded generally have a low fluorine content and may result in stream sediments diluted sufficiently to be ineffective indicators of mineral- ized rocks. The great concentration of elements of interest that is provided by heavy-mineral separations may be par- ticularly useful in seeking fluorspar. Friedrich and Pliiger (1971) found that lead, zinc, and mercury contents of soils were useful guides in exploring for fluorite in Spain and that mercury was useful in the Wolsendorf district, Germany. They also found that the fluorine content of Black Forest stream water increased by factors of 2 to 5, exceptionally even higher, in flowing through fluorite districts. The use of pathfinder elements instead of fluorine avoids the problems of low-level fluorine analysis but is successful only where the path- finder metal is consistently associated with fluorspar. Ineson (1970) found fluorine to be widely and uniform- ly dispersed in the limestone wallrocks of the Derbyshire district, England, and suggested that the element might be used as a guide to fluorite-galena-barite ore. Heavy-mineral methods and geochemistry-using zinc as a pathfinder metal-both have been tried in the Illinois- Kentucky district, but without completely overcoming the problems posed by the cover of loess and the segregation of fluorite and sphalerite within separate ore bodies along the mineralized faults (Keighin, 1968). In the Illinois-Kentucky district silver and antimony are concentrated in galena, and cadmium, gallium, and ger- manium are in sphalerite (Hall and Heyl, 1968). These ele- ments, as well as elements like mercury or vanadium that may have been introduced with the mineralizing solu- tions, may be clues to fluorspar bodies. In the Northgate district, Colorado, Worl (U.S. Geol. Survey, 1970) found anomalous molybdenum values as high as 800 ppm centered on one fluorspar vein. If a fluorspar contains uranium minerals-as in the Thomas Range, Utah -scintillation counters or Geiger counters might be helpful in prospecting for fluorspar. Districtwide studies, intermediate between regional re- connaissance surveys and detailed surveys of individual vein zones, are still needed. It is obvious, of course, that neither geochemical surveys nor heavy-mineral surveys alone can reliably show the presence of economic fluor- REFERENCES 91 spar deposits, regardless of the scale or detail of the work. Such surveys should be used in conjunction with evalua- tion of the regional geology and of the known structural and stratigraphic controls of mineralization. The widespread presence of accessory fluorite in metal deposits of several types can result in fluorine anomalies that are useful primarily as guides to metal ore. J. H. McCarthy and G. B. Gott (oral commun., 1971) observed that a fluorine anomaly correlates with the area of highest gold values in the Cripple Creek gold district, Colorado. A fluorine anomaly extends from the Tertiary rocks into the Precambrian granite and metamorphic rocks north and south of the caldera and closely follows an aeromagnetic high. The fluorine values were mostly between 1,000 and 2,000 ppm; and a few were greater. J. E. Gair (oral commun., 1971) reported that fluorine anomalies in stream sediments are associated with quartz- fluorite-huebnerite veins in the Hamme tungsten district, North Carolina. Fluorine forms broad anomalies around tungsten- bearing skarn in the Pyrenees, according to Cachau- Herreillat and Prouhet (1971). These authors therefore concluded that fluorine would be a useful pathfinder ele- ment in a regional search for additional tungsten dis- tricts. An introduction of fluorine during skarn mineral- ization is not universal, however. Darling (1971) found that biotite from intrusive rock contains less fluorine near tungsten-bearing skarn than at some distance from such ore. Geochemical investigation (Kesler and others, 1973) of fluid inclusions in certain igneous rocks, using a simple and economical method of analyzing water-extractable fluoride in rock powders, may be a helpful tool in search- ing for fluorine-bearing rocks possibly genetically asso- ciated with fluorspar deposits. The study by Kesler and others (1973) of nine Laramide granodiorite intrusive complexes in the Greater Antilles and Central America shows fluoride values in intrusives mineralized with base metals to be higher than values in barren intrusives. In summary, the fluorite contents of concentrates panned from creek and wash sediments are useful in re- connaissance in hilly areas of the semiarid southwest and of humid southern Korea. At some places, as in south- western New Mexico, a visual examination followed by determination of the total heavy metal content of concen- trates may distinguish between fluorspar and fluoritic base-metal deposits. The fluorine contents of soil over- lying and downhill from outcrops of faults can be used to determine the presence of fluorite mineralization, as demonstrated by Van Alstine and by Sang and associates. Bardovskiy and Kolpakov (1961), Bardovskiy (1964), and Ingerman (1965) indicated that fluorite-bearing zones can be effectively distinguished in boreholes by the induced activity method, in which the short-lived isotope F* is activated by neutrons from a polonium-beryllium source. Percentage of CaF, determined from the induced activity logs agrees well with that determined by chemical analysis (error +1-2percent). Faults in the Illinois fluorspar district (Hubbert, 1944) were studied by the earth-resistivity method. The main faults were clearly shown by anomalies; elsewhere the causes of observed anomalies were not evident. Many of the faults are mineralized with fluorite, but it is imprac- ticable to try to detect minable bodies by direct methods of geophysical prospecting. The refraction seismic method has been used as an indirect means of exploring for fluor- spar in southern Illinois (Johnson, 1957). Faults and dikes are detectable by this method but the presence of fluorspar is not. Self-potential investigations also have been made by the Illinois Geological Survey (Weller and others, 1952, p. 36-43). Alteration zones surrounding the main fluorspar vein at Northgate, Colo., were outlined by induced potential methods, but the veins in unaltered rock were not (Worl, oral commun., 1972). REFERENCES Abramovich, Yu. M., and Nechayev, Yu. A., 1960, Authigenic fluorite in Kungurian deposits of the Permian Priurals [in Russian]: Akad. Nauk SSSR Doklady, v. 135, no. 2, p. 414-415; English translation in Acad. Sci. USSR Doklady, Earth Sci. Sec., v. 185, nos. 1-6, p. 1288-1289, 1961. Abreu, S. F., 1960, Recursos minerais do Brasil-Materials n4o metalicos, V. 1: Rio de Janeiro, Inst. Nac. de Tecnologia, 471 p. Albritton, C. C., Jr., and Smith, J. F., Jr., 1965, Geology of the Sierra Blanca area, Hudspeth County, Texas: U.S. Geol. Survey Prof. Paper 479, 131 p. ' Allen, R. D., 1952, Variations in chemical and physical properties of fluorite: Am. Mineralogist, v. 37, nos. 11-12, p. 910-980. Alminas, H. V., Watts, K. C., and Siems, D. L., 1972, Maps showing fluorite distribution in the Winston and Chise quadrangles and in the west part of the Priest Tank quadrangle, Sierra County, New Mexico: U.S. Geol. Survey Misc. Field Studies Map MF-402. American Metal Market, 1969, Fluorides plant set: p. 11, Oct. 8. Amos, D. H., 1966, Geologic map of the Golconda quadrangle, Ken- tucky-Illinois, and the part of the Brownfield'quadrangle in Ken- tucky: U.S. Geol. Survey Geol. Quad. Map GQ-546. 1967, Geologic map of part of the Smithland quadrangle, Livingston County, Kentucky: U.S. Geol. Survey Geol. Quad. Map GQ-657. Amstutz, G. C., and Park, W. C., 1967, Stylolites of diagenetic age and their role in the interpretation of the southern Illinois fluorspar deposits [with German abs.]: Mineralium Deposita, v. 2, no. 1, p. 44-53. Anderson, A. L., 1954a, Fluorspar deposits near Meyers Cove, Lemhi County, Idaho: Idaho Bur. Mines and Geology Pamph. 98, 34 p. 1954b, A preliminary report on the fluorspar mineralization near Challis, Custer County, Idaho: Idaho Bur. Mines and Geo- logy Pamph. 101, 13 p. 1959, Geology and mineral resources of the North Fork quad- rangle, Lemhi County, Idaho: Idaho Bur. Mines and Geology Pamph. 118, 92 p. Anderson, A. L., and Van Alstine, R. E., 1964, Mineral resources- Fluorspar in Mineral and water resources of Idaho: U.S. 88th Cong. 2d sess., Comm. Print, p. 79-84. Archbold, N. L., 1966, Industrial mineral deposits of Mineral County Nevada: Nevada Bur. Mines Rept. 14, 32 p. 92 GEOLOGY AND RESOURCES OF FLUORINE Armstrong, R. L., 1970, Geochronology of Tertiary igneous rocks, eastern Basin and Range Province, western Utah, eastern Nevada, and vicinity, U.S.A.: Geochim. et Cosmochim. Acta, v. 34, no. 2, p. 203-232. Bancroft, Howland, 1914, The ore deposits of northeastern Washing- ton, including a section on The Republic mining district, by Waldemar Lindgren and Howland Bancroft: U.S. Geol. Survey Bull. 550, 215 p. Bannerman, H. M., 1941, The fluorite deposits of Cheshire County, New Hampshire: New Hampshire Planning and Devel. Comm. Mineral Resources Survey, pt. 5, 11 p. Bardill, J. D., 1946, Exploration of the Stoddard fluorite mine, Cheshire County, New Hampshire: U.S. Bur. Mines Rept. Inv. 3937, 6 p. Bardovskiy, V. Ya., 1964, Quantitative determination of fluorspar by the method of induced activity [in Russian]: Razved. Geofizika, no. 1, p. 66-72. Bardovskiy, V. Ya., and Kolpakov, A. V., 1961, Use of the method of induced activity for bringing out ore zones in boreholes in fluorite deposits: Ministerstvo Geologii I Okhrany Nedr SSSR, Voprosy Rudnoy Geofiziki, no. 3, p. 85-86. Barrell, Joseph, 1907, Geology of the Marysville mining district, Montana, a study of igneous intrusion and contact metamorphism: U.S. Geol. Survey Prof. Paper 57, 178 p. Barth, T. F. W., 1947, On the geochemical cycle of fluorine: Jour. Geology, v. 55, no. 5, p. 420-426. Baxter, J. W., and Desborough, G. A., 1965, Areal geology of the Illinois fluorspar district-Pt. 2, Karbers Ridge and Rosiclare quadrangles: Illinois Geol. Survey Circ. 385, 40 p. Baxter, J. W., Desborough, G. A., and Shaw, C. W., 1967, Areal geol- ogy of the Illinois fluorspar district-Pt. 3, Herod and Shetler- ville quadrangles: Illinois Geol. Survey Circ. 413, 41 p. Baxter, J. W., Potter, P. E., and Doyle, F. L., 1963, Areal geology of the Illinois fluorspar district-Pt. 1, Saline mines, Cave in Rock, Dekoven, and Repton quadrangles: Illinois Geol. Survey Circ. 342, 43 p. Becraft, G. E., and Weis, P. L., 1963, Geology and mineral deposits of the Turtle Lake quadrangle, Washington: U.S. Geol. Survey Bull. 1131, 73 p. Blackwell, D. D., 1971, Heat flow in U.S. National Report, 1967-1971, 15th General Assembly, IUGG: EOS (Am. Geophys. Union Trans.), v. 52, no. 5, p. IUGG 135-I1UGG139. Blanchard, F. N., 1966, Thermoluminescence of fluorite and age of deposition: Am. Mineralogist, v. 51, nos. 3-4, p. 474-485. Bradbury, J. C., 1962, Trace elements, rare earths, and chemical com- position of southern Illinois igneous rocks: Illinois Geol. Survey Circ. 330, 12 p. Bradbury, J. C., Finger, G. C., and Major, R. L., 1968, Fluorspar in Illinois: Illinois Geol. Survey Circ. 420, 64 p. Bradbury, J. C., Ostrom, M. E., and McVicker, L. D., 1955, Preliminary report on uranium in Hardin County, Illinois: Illinois Geol. Sur- vey Circ. 200, 21 p. Brecke, E. A., 1962, Ore genesis of the Cave-in-Rock fluorspar dis- trict, Hardin County, Illinois: Econ. Geology, v. 57, no. 4, p. 499-535. 1964a, Barite zoning in the Illinois-Kentucky fluorspar district: Econ. Geology, v. 59, no. 2, p. 299-302. 1964b, A possible source of solutions of the Illinois-Kentucky fluorspar district: Econ. Geology, v. 59, no. 7, p. 1293-1297. 1967, Sulfide and sulfur occurrences of the Illinois-Kentucky fluorspar district: Econ. Geology, v. 62, no. 3, p. 376-389. Brobst, D. A., 1958, Barite resources of the United States: U.S. Geol. Survey Bull. 1072-B, p. 67-130. Brobst, D. A., Pinckney, D. M., and Sainsbury, C. L., 1971, Geology and geochemistry of the Sinuk River barite deposit, Seward Peninsula, Alaska, in Geological Survey research 1971: U.S. Geol. Survey Prof. Paper 750-D, p. D1-D8. Brobst, D. A., and Pratt, W. P., eds., 1973, United States mineral re- sources: U.S. Geol. Survey Prof. Paper 820, 722 p. Brown, J. S., 1970, Mississippi Valley type lead-zinc ores-a review and sequel to the "Behre Symposium" [with French abs.]: Miner- alium Deposita, v. 5, no. 2, p. 103-119. Brown, J. S., Emery, J. A., and Meyer, P. A., Jr., 1954, Explosion pipe in test well on Hicks Dome, Hardin County, Illinois: Econ. Geology, v. 49, no. 8, p. 891-902. Buddington, A. F., 1923, Mineral deposits of the Wrangell district: U.S. Geol. Survey Bull. 739-B, p. 51-75. 1934, Geology and mineral resources of the Hammond, Antwerp, and Lowville quadrangles, with a chapter on the Paleozoic rocks of the Lowville quadrangle, by Rudolf Ruedemann: New York State Mus. Bull. 296, 251 p. Burchard, E. F., 1983, Fluorspar deposits in Western United States: New York, Am. Inst. Mining Metall. Engineers, Tech. Pub. 500, 26 p. 1937, Fluorspar and cryolite, Chap. 15, in Industrial minerals and rocks (nonmetallics other than fuels): New York, Am. Inst. Mining Metall. Engineers, p. 283-8302. Cachau-Herreillat, F., and Prouhet, J. P., 1971, The utilization of metalloids (arsenic, phosphorus, fluorine) as pathfinders for skarn tungsten deposits in the Pyrénées, in Geochemical exploration- Internat. Geochem. Explor. Symposium, $d, Toronto 1970, Proc.: Canadian Inst. Mining Metallurgy Spec. Vol. 11, p. 117-120. Callaghan, Eugene, and Lemmon, D. M., 1941, Tungsten resources of the Blue Wing district, Lemhi County, Idaho: U.S. Geol. Survey Bull. 931-A, p. 1-21. Cannon, R. S., and Pierce, A. P., 1967, Isotopic varieties of lead in stratiform deposits, in Genesis of stratiform lead-zinc-barite- fluorite deposits (Mississippi Valley type deposits)-A symposium, New York, 1966: Econ. Geology Mon. 3, p. 427-433. Carpenter, Roy, 1969, Factors controlling the marine geochemistry of fluorine: Geochim. et Cosmochim. Acta, v. 33, no. 10; p. 1153-1167. Chao, E. C. T., Evans, H. T., Jr., Skinner, B. J., and Milton, Charles, 1961, Neighborite, NaMgF,, a new mineral from the Green River Formation, South Ouray, Utah: Am. Mineralogist, v. 46, nos. 3-4, p. 379-393. Choate, Raoul, 1962, Geology and ore deposits of the Stanley area: Idaho Bur. Mines and Geology Pamph. 126, 122 p. Christiansen, R. L., and Lipman, P. W., 1972, Late Cenozoic [Pt.] 2 of Cenozoic volcanism and plate-tectonic evolution of the West- ern United States: Royal Soc. London Philos. Trans., ser. A, v. 271, no. 1213, p. 249-284. Clabaugh, S. E., and Sewell, C. R., 1964, Snowbird deposit, Montana- A carbonatite "pegmatite"?, in Abstracts for 1963: Geol. Soc. America Spec. Paper 76, p. 268-269. Clark, S. K., and Royds, J. S., 1948, Structural trends and fault systems in Eastern Interior Basin [U.S.]: Am. Assoc. Petroleum Geologists Bull., v. 32, no. 9, p. 1728-1749. Clegg, K. E., and Bradbury, J. C., 1956, Igneous intrusive rocks in Illinois and their economic significance: Illinois Geol. Survey Rept. Inv. 197, 19 p. Coats, R. R., 1956, Distribution of uranium and certain other trace elements in felsic volcanic rocks of Cenozoic age in the Western United States, in Page, L. R., Stocking, H. E., and Smith, H. B., compilers, Contributions to the geology of uranium and thorium by the United States Geological Survey and Atomic Energy Commission for the United Nations International Con- ference on Peaceful Uses of Atomic Energy, Geneva, Switzerland, 1955; U.S. Geol. Survey Prof. Paper $00, p. 75-78. Coats, R. R., Goss, W. D., and Rader, L. F., 1963, Distribution of fluorine in unaltered silicic volcanic rocks of the western conter- minous United States: Econ. Geology, v. 58, no. 6, p. 941-951. Cobb, E. H., 1964, Industrial minerals and construction materials occurrences in Alaska: U.S. Geol. Survey Mineral Inv. Resource Map MR-41. REFERENCES Cornwall, H. R., and Kleinhampl, F. J., 1961, Geology of the Bare Mountain quadrangle, Nevada: U.S. Geol. Survey Geol. Quad Map GQ-157. 1964, Geology of Bullfrog quadrangle and ore deposits related to Bullfrog Hills caldera, Nye County, Nevada, and Inyo County, California: U.S. Geol. Survey Prof. Paper 454-J, 25 p. Correns, C. W., 1956, The geochemistry of the halogens, [Chap.] 7, in V. 1 of Ahrens, L. H., Rankama, Kalervo, and Runcorn, S. K., eds., Physics and chemistry of the earth: New York, McGraw- Hill Co., Inc., p. 181-233. Cox, D. C., 1945, General features of Colorado fluorspar deposits: Colorado Sci. Soc. Proc., v. 14, no. 6, p. 263-285. 1954, Fluorspar deposits near Meyers Cove, Lemhi County, Idaho: U.S. Geol. Survey Bull. 1015-A, p. 1-21. Cox, D. P., 1970, Lead-zinc-silver deposits related to the White Moun- tain Plutonic Series in New Hampshire and Maine: U.S. Geol. Survey Bull. 1312-D, 18 p. Crosby, J. W., 3d, and Hoffman, S. R., 1951, Fluorspar in California: California Jour. Mines and Geology, v. 47, no. 4, p. 619-638. Crowley, W. P., 1968, The bedrock geology of the Long Hill and Bridgeport quadrangles, with maps: Connecticut Geol. and Nat. History Survey Quad. Rept. 24, 81 p. Cummings, J. B., 1946, Exploration of the Packard fluorspar property, Gila County, Arizona: U.S. Bur. Mines Rept. Inv. 3880, 7 p. Currier, L. W., 1923, Fluorspar deposits of Kentucky; a description and interpretation of the geologic occurrence and industrial im- portance of Kentucky fluorspar: Kentucky Geol. Survey, ser. 6, v. 13, 198 p. 1937, Geologic factors in the interpretation of fluorspar reserves in the Illinois-Kentucky field: U.S. Geol. Survey Bull. 886-B, p. 5-14. Currier, L. W., and Wagner, O. E., Jr., 1944, Geology of the Cave In Rock district, Pt. 1, of Geological and geophysical survey of fluorspar areas in Hardin County, Illinois: U.S. Geol. Survey Bull. 942, p. 1-72. Darling, R., 1971, Preliminary study of the distribution of minor and trace elements in biotite from quartz monzonite associated with contact-metasomatic . tungsten-molybdenum-copper ore, Cali- fornia, U.S.A., in Geochemical exploration-Internat. Geochem. Explor. Symposium, $d, Toronto, 1970, Proc.: Canadian Inst. Mining and Metallurgy Spec. Vol. 11, p. 315-322. Davis, H. W., Johnson, A. C., and Williams, J. S., 1947, Fluorspar, in Investigation of the factors affecting minerals, fuels, forestry, and reclamation projects: U.S. 80th Cong., lst sess., Senate hearings before subcommittee on Public Lands, p. 242-245. Deans, T., 1966, Economic mineralogy of African carbonatites, in Tuttle, O. F., and Gittins, J., eds., Carbonatites: New York, Interscience Publishers, p. 885-416. Deans, T., Sukheswala, R. N., Sethna, S. F., and Viladkar, S. G., 1972, Metasomatic feldspar rocks (potash fenites) associated with the fluorite deposits and carbonatites of Amba Dongar, Gujarat, India: Inst. Mining and Metallurgy Trans., Sec. B, v. 81, Bull. 783, p. Bl-B9. De Koster, G. R., 1960, Petrology of the (Permian) Phosphoria Forma- tion carbonates, eastern Big Horn Basin, Wyoming: Ames, Iowa State Univ. Sci. and Technology M. S. thesis, 112 p. Drewes, Harald, 1963, Geology of the Funeral Peak quadrangle, Cali- fornia, on the east flank of Death Valley: U.S. Geol. Survey Prof. Paper 413, 78 p. 1967, Geology of the Connors Pass quadrangle, Schell Creek Range, east-central Nevada: U.S. Geol. Survey Prof. Paper 557, 93 p. Eakin, H. M., 1915, Iron-ore deposits near Nome: U.S. Geol. Survey Bull. 622-I, p. 861-865. Earl, K. M., 1959, John Burdette barite-fluorite deposit, Garrard County, Kentucky: U.S. Bur. Mines Rept. Inv. 5480, 14 p. 93 Edmundson, R. S., 1988, Barite deposits of Virginia: Virginia Geol. Survey Bull. 53, 85 p. Elston, W. E. 1970, Volcano-tectonic control of ore deposits, south- western New Mexico, in Guidebook of the Tyrone-Big Hatchet Mountains-Florida Mountains region-New Mexico Geol. Soc. 21st Field Conf., 1970: New Mexico Bur. Mines and Mineral Re- sources, p. 147-153. Espenshade, G. H., 1947, Tungsten deposits of Vance County, North Carolina, and Mecklenburg County, Virginia: U.S. Geol. Survey Bull. 948-A, p. 1-17. Espenshade, G. H., and Potter, D. B., 1960, Kyanite, sillimanite, and andalusite deposits of the southeastern States: U.S. Geol. Survey Prof. Paper 336, 121 p. Everhart, D. L., 1963, Fluorine reserves and potential recovery from domestic phosphate rock [abs.]: Mining Eng., v. 15, no. 11, p. 56. Fenner, C. N., 1983, Pneumatolytic processes in the formation of minerals and ores, in Ore deposits of the Western States (Lindgren volume): New York, Am. Inst. Mining Metall. Engineers, p. 58-106. Ferguson, H. W., and Jewell, W. B., 1951, Geology and barite deposits of the Del Rio district, Cocke County, Tennessee: Tennessee Div. Geology Bull. 57, 235 p. Fleischer, Michael, 1969, The lanthanide elements in fluorite: Indian Mineralogist, v. 10, p. 36-39. Fleischer, Michael, and Robinson, W. O., 1963, Some problems of the geochemistry of fluorine, in Studies in analytical geochemistry: Royal Soc. Canada Spec. Pub. 6, p. 58-75. Freas, D. H., 1961, Temperatures of mineralization by liquid inclu- sions, Cave-In-Rock fluorspar district, Illinois: Econ. Geology, v. 56, no. 3, p. 542-556. Friedrich, G. H., and Pliiger, W. L., 1971, Geochemical prospecting for barite and fluorite deposits, in Geochemical exploration- Internat. Geochem. Explor. Symposium, $d, Toronto 1970, Proc.: Canadian Inst. Mining and Metallurgy Spec. Vol. 11, p. 151-156. Fries, Carl, Jr., 1942, Topaz deposits near the Brewer mine, Chester- field County, South Carolina: U.S. Geol. Survey Bull. 936-C, p. 59-78. Fuge, R., and Power, G. M., 1969, Chlorine and fluorine in granitic rocks from S. W. England: Geochim. et Cosmochim. Acta, v. 33, no. 7, p. 888-893. Furlow, J. W., 1969, Stratigraphy and economic geology of the eastern Chatham County phosphate deposit: Georgia Geol. Survey Bull. 82, 40 p. Gault, H. R, Rossman, D. L., Flint, G. M., Jr., and Ray, R. G., 1953, Some zinc-lead deposits of the Wrangell district, Alaska: U.S. Geol. Survey Bull. 998-B, p. 15-58. Gerasimovskiy, V. I., and Savinova, Ye. N., 1969, Fluorine contents of volcanic rocks in the Rift zone of East Africa [in Russian]: Geokhimiya, no. 12, p. 1466-1471; English translation in Geo- chemistry Internat., v. 6, p. 1124-1128, 1969. Gilbert, G. K., 1928, Studies of Basin-Range structure: U.S. Geol. Survey Prof. Paper 153, 92 p. Gillerman, Elliot, 1952, Fluorspar deposits of Burrow Mountains and vicinity, New Mexico: U.S. Geol. Survey Bull. 973-F, p. 261-289. 1953, Fluorspar deposits of the Eagle Mountains, Trans-Pecos, Texas: U.S. Geol. Survey Bull. 987, 98 p. 1964, Mineral deposits of western Grant County, New Mexico: New Mexico Bur. Mines and Mineral Resources Bull. 83, 213 p. Gilluly, James, 1963, The tectonic evolution of the western United States: Geol. Soc. London Quart. Jour., v. 119, pt. 2, no. 474, p. 133-174. Gilluly, James, and Masursky, Harold, 1965, Geology of the Cortez quadrangle, Nevada, with a section on Gravity and aeromagnetic surveys, by D. R. Mabey: U.S. Geol. Survey Bull. 1175, 117 p. 94 GEOLOGY AND RESOURCES OF FLUORINE Goddard, E. N., 1946, Fluorspar deposits of the Jamestown district, Boulder County, Colorado: Colorado Sci. Soc. Proc., v. 15, no. 1, 47 p. 1966, Geologic map and sections of the Zuni Mountains fluor- spar district, Valencia County, New Mexico: U.S. Geol. Survey Misc. Geol. Inv. Map 1-454. Goldring, E. D., 1942, An occurrence of ilsemannite [Colorado]: Am. Mineralogist, v. 27, no. 10, p. 717-719. Grametbaur, A. B., 1959, Selected bibliography of andalusite, kyanite, sillimanite, dumortierite, topaz, and pyrophyllite in the United States: U.S. Geol. Survey Bull. 1019-N, p. 973-1046. Griffitts, W. R., and Powers, H. A., 1963, Beryllium and fluorine con- tent of some silicic volcanic glasses from Western United States, in Short papers in geology and hydrology: U.S. Geol. Survey Prof. Paper 475-B, p. B18-B19. Grogan, R. M., 1960, Fluorspar and cryolite, in Industrial minerals and rocks (nonmetallics other than fuels) [3d ed.]: New York, Am. Inst. Mining Metall. Petroleum Engineers, p. 363-382. 1964, Finding sources of fluorine [abs.]: Mining Eng., v. 16, no. 1, p. 65. Grogan, R. M., and Bradbury, J. C., 1967, Origin of the stratiform fluorite deposits of southern Illinois, in Genesis of stratiform lead- zinc-barite-fluorite deposits (Mississippi Valley type deposits)- A symposium, New York, 1966: Econ. Geology Mon. 3, p. 40-51. 1968, Fluorite-zinc-lead deposits of the Illinois-Kentucky mining district, in Ridge, J. D., ed., Ore deposits of the United States, 1933-1967 (Graton-Sales Volume), V. 1: New York, Am. Inst. Mining, Metall. and Petroleum Engineers, p. 370-399. Grogan, R. M., and Shrode, R. S., 1952, Formation temperatures of southern Illinois bedded fluorite as determined from fluid in- clusions: Am. Mineralogist, v. 37, nos. 7-8, p. 555-566. Guccione, Eugene, 1972, What's going on in the fluorspar industry: Eng. and Mining Jour., v. 173, no. 12, p. 64-73. Gulbrandsen, R. A., 1966, Chemical composition of phosphorites of the Phosphoria Formation: Geochim. et Cosmochim. Acta, v. 30, no. 8, p. 769-778. Gulbrandsen, R. A., and Reeser, D. W., 1969, An occurrence of Permian manganese nodules near Dillon, Montana, in Geological Survey research 1969: U.S. Geol. Survey Prof. Paper 650-C, p. C49-C57. Hall, W. E., and Friedman, Irving, 1963, Composition of fluid in- clusions, Cave-in-Rock fluorite district, Illinois, and Upper Mississippi Valley zinc-lead district: Econ. Geology, v. 58, no. 6, p. 886-911. Hall, W. E., and Heyl, A. V., 1968, Distribution of minor elements in ore and host rock, Illinois-Kentucky fluorite district and Upper Mississippi Valley zinc-lead district: Econ. Geology, v. 63, no. 6, p. 655-670. Hamilton, Warren, 1972, The Hallett volcanic province, Antarctica: U.S. Geol. Survey Prof. Paper 456-C, 62 p. Hardin, G. C., Jr., 1955, Babb fault system, Crittenden and Livingston Counties, in Pt. 1 of Fluorspar deposits in western Kentucky: U.S. Geol. Survey Bull. 1012-B, p. 7-37. Hardin, G. C., Jr., and Thurston, W. R., 1945, Celestite from Livingston County, Kentucky: Am. Mineralogist, v. 30, nos. 9-10, p. 639-640. Hardin, G. C., Jr., and Trace, R. D., 1959, Geology and fluorspar deposits, Big Four fault system, Crittenden County, Kentucky: U.S. Geol. Survey Bull. 1042-8, p. 699-724. Hatmaker, Paul, and Davis, H. W., 1988, The fluorspar industry of the United States with special reference to the Illinois-Kentucky district: Illinois Geol. Survey Bull. 59, 128 p. Hawley, C. C., 1969, Geology and beryllium deposits of the Lake George (or Badger Flats) beryllium area, Park and Jefferson Counties, Colorado: U.S. Geol. Survey Prof. Paper 608-A, 44 p. Hawley, C. C., Huffman, Claude, Jr., Hamilton, J. C., and . Rader, L. F., Jr., 1966, Geologic and geochemical features of the Redskin Granite and associated rocks, Lake George beryllium area, Colo- rado, in Geological Survey research 1966; U.S. Geol. Survey Prof. Paper 550-C, p. C138-C147. Herbert, Paul, Jr., and Young, R. S., 1956, Sulfide mineralization in the Shenandoah Valley of Virginia: Virginia Div. Geology Bull. 70, 58 p. Heron, S. D., Jr., and Johnson, H. S., Jr., 1966, Clay mineralogy, stratigraphy, and structural setting of the Hawthorn Formation, Coosawhatchie district, South Carolina: Southeastern Geology, v. 7, no. 2, p. 51-63. Herreid, Gordon, 1966, Preliminary geology and geochemistry of the Sinuk River area, Seward Peninsula, Alaska: Alaska Div. Mines and Minerals Geol. Rept. 24, 19 p. Herrin, Eugene, and Taggart, James, 1962, Regional variation in Pn velocity and their effect on the location of epicenters: Seismol. Soc. America Bull., v. 52, no. 5, p. 1037-1046. Hewett, D. F., 1964, Veins of hypogene manganese oxide minerals in the southwestern United States: Econ. Geology, v. 59, no. 8, p. 1429-1472. Hewett, D. F., and Fleischer, Michael, 1960, Deposits of manganese oxides: Econ. Geology, v. 55, no. 1, p. 1-55. Hewett, D. F., Fleischer, Michael, and Conklin, Nancy, 1963, Deposits of manganese oxides-Supplement: Econ. Geology, v. 58, no. 1, p. 1-51. Heyl, A. V., 1967, Some aspects of genesis of stratiform zinc-lead-barite- fluorite deposits in the United States, in Genesis of stratiform lead-zinc-barite-fluorite deposits (Mississippi Valley type de- posits)-A symposium, New York, 1966: Econ. Geology Mon. 3, p. 20-29. 1968, Minor epigenetic, diagenetic, and syngenetic sulfide, fluorite, and barite occurrences in the central United States: Econ. Geology, v. 63, no. 6, p. 585-594. 1969, Some aspects of genesis of zinc-lead-barite-fluorite deposits in the Mississippi Valley, U.S.A.: Inst. Mining and Metallurgy Trans., Sec. B, v. 78, Bull. 756, p. B148-B160. 1972, The 38th parallel lineament and its relationship to ore deposits: Econ. Geology, v. 67, no. 7, p. 879-894. Heyl, A. V., and Brock, M. R., 1961, Structural framework of the Illinois-Kentucky mining district and its relation to mineral de- posits, in Short papers in the geologic and hydrologic sciences: U.S. Geol. Survey Prof. Paper 424-D, p. D3-D6. Heyl, A. V., Brock, M. R., Jolly, J. L.., and Wells, C. E., 1965, Regional structure of the southeast Missouri and Illinois-Kentucky mineral districts: U.S. Geol. Survey Bull. 1202-B, 20 p. [1966]. Heyl, A. V., Delevaux, M. H., Zartman, R. E., and Brock, M. R., 1966, Isotopic study of galenas from the upper Mississippi Valley, the Illinois-Kentucky, and some Appalachian Valley mineral dis- tricts: Econ. Geology, v. 61, no. 5, p. 933-961. Horton, R. C., 1961, An inventory of fluorspar occurrences in Nevada: Nevada Bur. Mines Rept. 1, 31 p. " Hubbert, M. K., 1944, An exploratory study of faults in the Cave In Rock and Rosiclare districts by the earth-resistivity method, Pt. 2, of Geological and geophysical survey of fluorspar areas in Hardin County, Illinois: U.S. Geol. Survey Bull. 942, p. 73-147. Industrial Minerals, 1970, Steelmakers face possible shortages of fluor- spar: Industrial Minerals, no. 33, p. 9-13, June. 1970, Europe and N. America (excluding Mexico), Pt. 1 of World fluorspar industry: Industrial Minerals, no. 33, p. 15-29, June. 1970, Latin America, Africa and the Far East, Pt. 2 of World fluorspar industry: Industrial Minerals, no. 34, p. 21-28, July. 1970, The adequacy of acid-grade fluorspar supplies questioned: Industrial Minerals, no. 34, p. 11-19, July. Ineson, P. R., 1970, Trace-element aureoles in limestone wallrocks adjacent to fissure veins in the Eyam area of the Derbyshire ore field: Inst. Mining and Metallurgy Trans., Sec. B, v. 79, Bull. 768, p. B238-B245. REFERENCES 95 Ingerman, V. G., 1965, Detecting fluorite mineralization in borehole sections by the activation gamma-logging method: Razved. Geofizika, no. 7, p. 78-80. Iron Age, 1927, p. 1835, June 23. Jacob, K. D., 1953, Phosphate resources and processing facilities, in Jacob, K. D., ed., Fertilizer technology and resources in the United States: New York, Academic Press, Inc., p. 117-165. Jacob, K. D., Hill, W. L., Marshall, H. L., and Reynolds, D. S., 1933, The composition and distribution of phosphate rock with special reference to the United States: U.S. Dept. Agriculture Tech. Bull. 364, 89 p. Jahns, R. H., 1944, Beryllium and tungsten deposits of the Iron Moun- tain district, Sierra and Socorro Countiés, New Mexico, with a section on The berylium minerals, by J. J. Glass: U.S. Geol. Sur- vey Bull. 945-C, p. 45-79. Jewell, W. B., 1947, Barite, fluorite, galena, and sphalerite veins of Middle Tennessee: Tennessee Div. Geology Bull. 51, 114 p. Johnson, R. B., 1957, Goose Creek area, Pt. 1, of Refraction seismic investigations, Rosiclare fluorspar district, Illinois: Illinois Geol. Survey Circ. 231, 15 p. Johnston, W. D., Jr., 1928, Fluorspar in New Mexico: New Mexico Bur. Mines and Mineral Resources Bull. 4, 128 p. Jolly, J. L., and Heyl, A. V., 1964, Mineral paragenesis and zoning in the central Kentucky mineral district Econ. Geology, v. 59, no. 4, p. 596-624. 1968, Mercury and other trace elements in sphalerite and wall- rocks from central Kentucky, Tennessee, and Appalachian zinc districts: U.S. Geol. Survey Bull. 1252-F, 29 p. Kazakov, A. V., and Sokolova, E. I., 1950, Usleviya obrazovaniya flyuorita v osadochnykh porodakh (flyooritovaya sistems) [Condi- tions of the formation of fluorite in sedimentary rocks (The fluorite system)]: Akad. Nauk SSSR Geol. Inst. Trudy, no. 114, Geol. Ser. 40, p. 22-64; unpub. English translation in U.S. Geol. Survey TEI-386, 76 p., 1951. Keighin, C. W., 1968, Geochemical exploration in the southern Illinois fluorspar district-A progress report [abs.]: Mining Eng., v. 20, no. 8, p. 25. Kelly, W. C., and Goddard, E. N., 1969, Telluride ores of Boulder County, Colorado: Geol. Soc. America Mem. 109, 237 p. Kesler, S. E., Van Loon, J. C., and Moore, C. M., 1973, Evaluation of ore potential of granodioritic rocks using water-extractable chloride and fluoride: Canadian Mining and Metall. Bull., v. 66, no. 730, p. 56-60. King, P. B., compiler, 1967, National Atlas-Sheets 69-70, Tectonic features: U.S. Geol. Survey, scale 1:7,500,000. Koenig, J. B., 1956, The petrography of certain igneous dikes of Ken- tucky: Kentucky Geol. Survey Bull. 21, ser. 9, 57 p. Konovalov, G. S., 1959, Removal of microelements by the principal rivers of the USSR [in Russian]: Akad. Nauk SSSR Doklady, v 129, no. 4, p. 912-915; English translation in Acad. Sci. USSR Doklady, Earth Sci. Sec., v. 129, nos. 1-6, p. 1034-1037, 1960. Koritnig, S., 1972, Fluorine, in Wedepohl, K. H., ed., Handbook of geochemistry, v. II/3: Springer-Verlag, Berlin. Laurence, R. A., 1989, Origin of the Sweetwater, Tennessee, barite deposits: Econ. Geology, v. 34, no. 2, p. 190-200. 1960, Geologic problems in the Sweetwater barite district, Tennessee (Bradley Volume): Am. Jour. Sci., v. 258-A, p. 170-179. Lee, D. E., and Van Loenen, R. E., 1970, Biotites from hybrid granitoid rocks of the southern Snake Range, Nevada, in Geological Survey research 1970: U.S. Geol. Survey Prof. Paper 700-D, p. D196-D206. 1971, Hybrid granitoid rocks of the southern Snake Range, Nevada: U.S. Geol. Survey Prof. paper 668, 48 p. Leonard, B. F., and Buddington, A. F., 1964, Ore deposits of the St. Lawrence County magnetite district, northwest Adirondacks, New York: U.S. Geol. Survey Prof. Paper 377, 259 p. Lindgren, Waldemar, 1983a, Mineral deposits [4th ed.]: New York and London, McGraw-Hill, 980 p. 1933b, Differentiation and ore deposition, Cordilleran region of the United States, in Ore deposits of the Western States (Lindgren volume): New York, Am. Inst. Mining Metall. Engi- neers, p. 152-180. Lindgren, Waldemar, and Ransome, F. L., 1906, Geology and gold deposits of the Cripple Creek district, Colorado: U.S. Geol. Survey Prof. Paper 54, 516 p. Lipman, P. W., Prostka, H. J., and Christiansen, R. L., 1972, Early and Middle Cenozoic [Pt.] 1 of Cenozoic volcanism and plate- tectonic evolution of the Western United States: Royal Soc. Lon- don Philos. Trans., ser. A, v. 271, no. 1213, p. 217-248. Livingstone, D. A., 1963, Chemical composition of rivers and lakes, Chap. G of Fleischer, Michael, ed., Data of geochemistry [6th ed.]: U.S. Geol. Survey Prof. Paper 440-G, 64 p. Lovering, T. S., 1957, Halogen-acid alteration of ash at Fumarole No. 1, Valley of Ten Thousand Smokes, Alaska: Geol Soc. America Bull., v. 68, no. 12, pt. 1, p. 1585-1604. Lovering, T. S., and Goddard, E. N., 1950, Geology and ore deposits of the Front Range, Colorado: U.S. Geol. Survey Prof. Paper 223, 319 p. Luttrell, G. W., 1966, Base- and precious-metal and related ore deposits of Virginia: Virginia Div. Mineral Resources, Mineral Resources Rept. 7, 167 p. McAnulty, W. N., Sr., 1967, Fluorspar in Brewster County, Texas: Texas Univ. Bur. Econ. Geology Geol. Circ. 67-2, 16 p. McAnulty, W. N., and Levinson, A. A., 1964, Rare alkali and beryllium mineralization in volcanic tuffs, Honey Comb Hills, Juab County, Utah: Econ. Geology, v. 59, no. 5, p. 768-774. McAnulty, W. N., Sewell, C. R., Atkinson, D. R., and Rasberry, J. M., 1963, Aguachile beryllium-bearing fluorspar district, Coahuila, Mexico: Geol. Soc. America Bull., v. 74, no. 6, p. 785-743. McGinnis, L. D., 1963, Earthquakes and crustal movement as related to water load in the Mississippi Valley region: Illinois Geol. Sur- vey Circ. 344, 20 p. McGinnis, L. D., and Bradbury, J. C., 1964, Aeromagnetic study of the Hardin County area, Illinois: Illinois Geol. Survey Circ. 363, 12 p. McKelvey, V. E., Cathcart, J. B., Altschuler, Z. S., Swanson, R. W., and Buck, K. L., 1953, Domestic phosphate deposits, Chap. 11, in Pierre, W. H., and Norman, A. G., eds., Soil and fertilizer phosphorus in crop nutrition: New York, Academic Press, Inc., p. 347-376. McKeown, F. A., and Klemic, Harry, 1956, Rare-earth-bearing apatite at Mineville, Essex County, New York: U.S. Geol. Survey Bull. 1046-B, p. 9-23. MacMillan, R. T., 1970, Fluorine, in Mineral facts and problems, 1970: U.S. Bur. Mines Bull. 650, p. 989-1000. Maher, S. W., 1970, Barite resources of Tennessee: Tennessee Div. Geol- ogy Rept. Inv. 28, 40 p. Matson, E. J., and Trengove, R. R., 1957, Investigation of fluorspar deposit, Kaiser mine, Mineral County, Nevada: U.S. Bur. Mines Rept. Inv. 5344, 38 p. Miller, R. N., 1959, Geology of the South Moccasin Mountains, Fergus County, Montana: Montana Bur. Mines and Geology Mem. 37, 44 p. Miller, T. P., 1971, Petrology of the plutonic rocks of west-central Alaska: U.S. Geol. Survey open-file report, 136 p. Milner, H. B., 1962, Principles and applications, V. 2 of Sedimentary petrography [4th ed.]: New York, Macmillan Co., 715 p. Moore, J. G., 1959, The quartz diorite boundary line in the western United States: Jour. Geology, v. 67, no. 2, p. 198-210. 1962, K/Na ratio of Cenozoic igneous rocks of the western United States: Geochim. et Cosmochim. Acta, v. 26, p. 101-130. 96 GEOLOGY AND RESOURCES OF FLUORINE Morrill, Philip, 1951, New Hampshire mine and mineral localities: Dartmouth College Museum, 46 p. Mudd, H. T., 1949, Fluorspar and cryolite, in Industrial minerals and rocks (nonmetallics other than fuels) [2d ed.]: New York, Am. Inst. Mining Metall. Engineers, p. 381-403. Muir, N. M., 1947, Daisy fluorite mine, Rosiclare Lead & Fluorspar Mining Co., Hardin County, Illinois: U.S. Bur. Mines Rept. Inv. 4075, 16 p. Nackowski, M. P., 1957, Geochemical prospecting applied to the Illinois-Kentucky fluorspar area, in Erikson, J. E., Geochemical prospecting abstracts, July 1952-December 1954; U.S. Geol. Survey Bull. 1000-G, p. $78. National Materials Advisory Board, 1970, Trends in the usage of fluor- spar: Washington, D. C., Natl. Research Council, NMAB-269, 71 p. Noble, D. C., Smith, V. C., and Peck, L. C., 1967, Loss of halogens from crystallized and glassy silicic volcanic rocks: Geochim. et Cosmochim. Acta, v. 31, no. 2, p. 215-223. Norton, J. J., 1970, Composition of a pegmatite, Keystone, South Dakota: Am. Mineralogist, v. 55, nos. 5-6, p. 981-1002. Norton, J. J., Page, L. R., and Brobst, D. A., 1962, Geology of the Hugo pegmatite, Keystone, South Dakota: U.S. Geol. Survey Prof. Paper 297-B, p. 49-127. Norwood, J. G., 1866, Report on the Rosiclare lead mines, in Hardin County [Illinois], Chap. 13 of Worthen, A. H., Whitney, J. D., Les- quereux, Leo, and Englemann, Henry. Geology: Illinois Geol. Survey, v. 1, p. 866-372. Oesterling, W. A., 1952, Geologic and economic significance of the Hutson zinc mine, Salem, Kentucky-its relation to the Illinois- Kentucky fluorspar district: Econ. Geology, v. 47, no. 3, p. 816-338. Olson, J. C., Shawe, D. R., Pray, L. C., and Sharp, W. N., 1954, Rare-earth mineral deposits of the Mountain Pass district, San Bernardino County, California, with a foreword on History of the discovery at Mountain Pass, California, by D. F. Hewett: U.S. Geol. Survey Prof. Paper 261, 75 p. [1955]. Olson, N. K., compiler and ed., 1966, Geology of the Miocene and Pliocene series in the north Florida-south Georgia area-Atlantic Coastal Plain Geol. Assoc., 7th, and Southeastern Geol. Soc., 12th Ann. Field Conf., 1966: 94 p. O'Neill, J. F., 1950, Investigation of Gilley fluorspar deposits, Cherokee County, Alabama: U.S. Bur. Mines Rept. Inv. 4660, 6 p. Oriel, S. S., 1950, Geology and mineral resources of the Hot Springs window, Madison County, North Carolina: North Carolina Div. Mineral Resources Bull. 60, 70 p. Osterwald, F. W., Osterwald, D. B., Long, J. S., Jr., and Wilson, W. H., 1959, Mineral resources of Wyoming: Wyoming Geol. Survey Bull. 50, 259 p. Paige, Sidney, 1911, Mineral resources of the Llano-Burnet region, Texas, with an account of the pre-Cambrian geology: U.S. Geol. Survey Bull. 450, 103 p. Pakiser, L. C., and Zietz, Isidore, 1965, Transcontinental Crustal and upper-mantle structure: Rev. Geophysics, v. 3, no. 4, p. 505-520. Paley, W. S., chm., 1952, Fluorspar, Chap. 18, in The outlook for key commodities, V. 2 of Resources for freedom: [U.S.] President's Materials Policy Comm., p. 88-91 (Washington, U.S. Govt. Print- ing Office). Panov, B. S., 1962, Fluorite in carbonates at the southwestern margin of the Donets basin [in Russian]: Akad. Nauk SSSR Doklady, v. 147, no. 5, p. 1172-1174; English translation in Acad. Sci. USSR Doklady, Earth Sci. Sec., v. 147, nos. 1-6 p. 160-164, 1964. Pardee, J. T., 1918, Geology and mineral deposits of the Colville Indian Reservation, Washington: U.S. Geol. Survey Bull. 677, 186 p. Pardee, J. T., and Park, C. F., Jr., 1948, Gold deposits of the Southern Piedmont: U.S. Geol. Survey Prof. Paper 213, 156 p. Park, W. C., and Amstutz, G. C., 1968, Primary "cut and fill" channels and gravitational diagenetic features: Mineralium Deposita, v. 3, p. 66-80. Parker, J. M., $d, 1963, Geologic setting of the Hamme tungsten dis- trict, North Carolina and Virginia: U.S. Geol. Survey Bull. 1122-G, 69 p. Parker, R. L., and Havens, R. G., 1963, Thortveitite associated with fluorite, Ravalli County, Montana, in Short papers in geology and hydrology: U.S. Geol. Survey Prof. Paper 475-B, p. B10O-BI11. Patton, W. W., Jr., and Matzko, J. J., 1959, Phosphate deposits in northern Alaska: U.S. Geol. Survey Prof. Paper 302-A, p. 1-17. Perhac, R. M., and Heinrich, E. W., 1964, Fluorite-bastnaesite deposits of the Gallinas Mountains, New Mexico, and bastnaesite para- genesis: Econ. Geology, v. 59, no. 2, p. 226-239. Perret, F. A., 1924, The Vesuvius eruption of 1906-study of a vol- canic cycle: Carnegie Inst. Washington Pub. 339, 151 p. Peters, W. C., 1958, Geologic characteristics of fluorspar deposits in the western United States: Econ. Geology, v. 53, no. 6, p. 663-688. Pinckney, D. M., and Rye, R. O., 1972, Variation of 013/016» Ciy Caz, texture, and mineralogy in altered limestone in the Hill mine, Cave-in-Rock district, Illinois: Econ. Geology, v. 67, no. 1, p. 1-18. Quinn, A. W., 1971, Bedrock geology of Rhode Island: U.S. Geol. Survey Bull. 1295, 68 p. Ratte, J. C., Landis, E. R., Gaskill, D. L., and Raabe, R. G., 1969, Mineral resources of the Blue Range primitive area, Greenlee County, Arizona, and Catron County, New Mexico, with a section on Aeromagnetic interpretation, by G. P. Eaton: U.S. Geol. Survey Bull. 1261-£, 91 p. Ratte, J. C., and Steven, T. A., 1967, Ash flows and related volcanic rocks associated with the Creede caldera, San Juan Mountains, Colorado: U.S. Geol. Survey Prof. Paper 524-H, 58 p. Reeves, R. G., and Kral, V. E., 1955, Geology and iron ore deposits of the Buena Vista Hills, Churchill and Pershing Counties, Nev- ada, Pt. A, in Iron ore deposits of Nevada: Nevada Bur. Mines Bull. 53, p. 1-82. Reid, R. R., 1963, Reconnaissance geology of the Sawtooth Range: Idaho Bur. Mines and Geology Pamph. 129, 37 p. Rhoades, R. F., and Mistler, A. J., 1941, Post-Appalachian faulting in western Kentucky: Am. Assoc. Petroleum Geologists Bull., v. 25, no. 11, p. 2046-2056. Roberts, W. L., and Rapp, George, Jr., 1965, Mineralogy of the Black Hills: South Dakota School Mines and Technology Bull. 18, 268 p. Robinson, L. C., 1931, Vein deposits of central Kentucky: Kentucky Geol. Survey, ser. 6, v. 41, p. 3-127. Rogers, W. B., and Hays, W. H., 1967, Geologic map of the Fredonia quadrangle, western Kentucky: U.S. Geol. Survey Geol. Quad. Map GQ-607. Ross, C. A., 1963, Structural framework of southernmost Illinois: Illinois Geol. Survey Circ. $51, 27 p. Ross, C. P., 1950, Fluorspar prospects of Montana: U.S. Geol. Survey Bull. 955-E, p. 173-224. Ross, D. C., 1961, Geology and mineral deposits of Mineral County, Nevada: Nevada Bur. Mines Bull. 58, 98 p. Rothrock, H. E., Johnson, C. H., and Hahn, A. D., 1946, Fluorspar resources of New Mexico: New Mexico Bur. Mines and Mineral Resources Bull. 21, 239 p. Ruch, R. R., 1968, Neutron activation analysis at the Illinois State Geological Survey: Illinois Geol. Survey Indus. Minerals Note 34, 8 p. Russel, R. T., [no date, 1947?], The Poncha fluorspar deposits, Chaffee County, Colorado: U.S. Geol. Survey Strategic Minerals Inv. Prelim. Rept. 3-210, 12 p. 1950, The geology of the Poncha fluorspar district, Chaffee County, Colorado: Univ. Cincinnati unpub. Ph. D. thesis, 70 p. Sahinen, U. M., 1962, Fluorspar deposits in Montana: Montana Bur. Mines and Geology Bull. 28, 38 p. Sainsbury, C. L., 1964a, Association of beryllium with tin deposits rich in fluorite: Econ. Geology, v. 59, no. 5, p. 920-926. REFERENCES 97 1964b, Geology of Lost River mine area, Alaska: U.S. Geol. Survey Bull. 1129, 80 p. 1965, Planetable maps and drill logs of fluorite and beryllium deposits, Lost River area, Alaska: U.S. Geol. Survey open-file report, 38 p. 1969, Geology and ore deposits of the central York Mountains, western Seward Peninsula, Alaska: U.S. Geol. Survey Bull. 1287, 101 p. [1970]. Sainsbury, C. L., and Kleinhampl, F. J., 1969, Fluorite deposits of the Quinn Canyon Range, Nevada: U.S. Geol. Survey Bull. 1272- C, 22 p. [1970]. Sang Kyu Yun, Hyo Joon Seo, Sahng Yup Kim, Jae Sand Cho, and Nam Iue Hong, 1970, Detailed geochemical exploration for Mo- W and CaF, in Hwanggangri region, Republic of Korea, in Second seminar on methods and techniques of geochemical exploration in the humid tropical environment, Peradeniya, Ceylon: Docu- ment GSM (2)/107, 9 p. Sever, C. W., Cathcart, J. B., and Patterson, S. H., 1967, Phosphate deposits of south-central Georgia and north-central peninsular Florida: Georgia Dept. Mines, Mining and Geology, South Georgia Minerals Program, Proj. Rept. 7, 62 p. Shawe, D. R., 1961, Rhyolites in the Egan range south of Ely, Nevada, in short papers in the geologic and hydrologic sciences: U.S. Geol. Survey Prof. Paper 424-B, p. B178-B181. 1965, Strike-slip control of Basin-Range structure indicated by historical faults in western Nevada: Geol. Soc. America Bull., v. 76, no. 12, p. 1361-1378. 1966, Arizona-New Mexico and Nevada-Utah beryllium belts, in Geological Survey research 1966: U.S. Geol. Survey Prof. Paper 550-C, p. C206-C213. 1968, Geology of the Spor Mountain beryllium district, Utah, in Ridge, J. D., ed., Ore deposits of the United States 1933-1967, (Graton-Sales Volume), V. 2: New York, Am. Inst. Mining Metall. and Petroleum Engineers, p. 1148-1161. 1972, Reconnaissance geology and mineral potential of Thomas, Keg, and Desert calderas, central Juab County, Utah, in Geol- ogical Survey research 1972: U.S. Geol. Survey Prof. Paper 800- B, p. B67-B77. Shawe, D. R., and Bernold, Stanley, 1966, Beryllium content of vol- canic rocks: U.S. Geol. Survey Bull. 1214-C, 11 p. Shawe, F. R., Reeves, R. G., and Kral, V. E., 1962, Iron ore deposits of northern Nevada, Pt. C, in Iron ore deposits of Nevada: Nevada Bur. Mines Bull. 53, p. 79-1300. Sheldon, R. P., 1963, Physical stratigraphy and mineral resources of Permian rocks in western Wyoming: U.S. Geol. Survey Prof. Paper 313-B, p. 49-273. Shenon, P. J., and Ross, C. P., 1986, Geology and ore deposits near Edwardsburg and Thunder Mountain, Idaho: Idaho Bur. Mines and Geology Pamph. 44, 45 p. Sheppard, R. A., and Gude, A. J., 3d, 1969, Authigenic fluorite in Pliocene lacustrine rocks near Rome, Malheur County, Oregon, in Geological Survey research 1969; U.S. Geol. Survey Prof. Paper 650-D, p. D69-D74 [1970]. Sheridan, D. M., Stephens, H. G., Staatz, M. H., and Norton, J. J., 1957, Geology and beryl deposits of the Peerless pegmatite, Pennington County, South Dakota: U.S. Geol. Survey Prof. Paper 297-A, p. 1-47. Sheridan, D. M., Taylor, R. B., and Marsh, S. P., 1968, Rutile and topaz in Precambrian gneiss, Jefferson and Clear Creek Counties, Colorado: U.S. Geol. Survey Circ. 567, 7 p. Siems, H. B., 1951, Fertilizersfin Kirk, R. E., and Othmer, D. F., eds., Encyclopedia of chemical technology: New York, Inter- science Encyclopedia, v. 6, p. 376-452. Silliman, Benjamin, 1821, Virginia and Illinois fluorspar: Am. Jour. Sci., [1st ser.], v. 3, p. 248-245. Simons, F. S., 1964, Geology of the Klondyke quadrangle, Graham and Pinal Counties, Arizona: U.S. Geol. Survey Prof. Paper 461, 173 p. Snyder, F. G., and Gerdemann, P. E., 1965, Explosive igneous activity along an Illinois-Missouri-Kansas axis: Am. Jour. Sci., v. 263, no. 6, p. 465-493. Spada, Aldo, 1969, I1 giacimento di fluorite e baritina esalativo- sedimentario in "facies" lacustre, intercalato nei sedimenti piro- clastici della zona di Castel Giuliano, in prov. di Roma [Lacustrine-facies exhalative-sedimentary fluorite and barite deposits interbedded in pyroclastic sediments of the Castel Giuliano area, Rome]: Industria Mineraria, v. 20, no. 10, p. 501, 504-510, 515-518. Staatz, M. H., 1963, Geology of the beryllium deposits in the Thomas Range, Juab County, Utah: U.S. Geol. Survey Bull. 1142-M, 36 p. Staatz, M. H., and Carr, W. J., 1964, Geology and mineral deposits of the Thomas and Dugway Ranges, Juab and Tooele Counties, Utah: U.S. Geol. Survey Prof. Paper 415, 188 p. Staatz, M. H., and Griffitts, W. R., 1961, Beryllium-bearing tuff in the Thomas Range, Juab County, Utah: Econ. Geology, v. 56, no. 5, p. 941-950. Staatz, M. H., and Osterwald, F. W., 1959, Geology of the Thomas Range fluorspar district, Juab County, Utah: U.S. Geol. Survey Bull. 1069, 97 p. Stefannson, Kristinn, and Sigurjonsson, Julius, 1957, Temporary in- crease in fluorine content of water following the eruption, in Einarsson, T., Kjartansson, G., and Thorarinsson, S., eds., The eruption of Hekla 1947-1948: Vi’sindafélag Islendinga Rit, v. 3, no. 1/3, 13 p. Stenzel, H. B., 1982, Pre-Cambrian of the Llano uplift, Texas [abs.]: Geol. Soc. America Bull, v. 43, no. 1, p. 143-144. Steven, T. A., 1949, Geology and fluorspar deposits of the St. Peters Dome district, Colorado: Colorado Sci. Soc. Proc., v. 15, no. 6, p. 259-284. 1960, Geology and fluorspar deposits, Northgate district, Colo- rado: U.S. Geol. Survey Bull. 1082-F, p. 323-422 [1961]. 1968, Geologic map of the Wagon Wheel Gap fluorspar district, Colorado: U.S. Geol. Survey open-file map, 1:20,000. Steven, T. A., and Ratté. J. C., 1960, Geology and ore deposits of the Summitville district, San Juan Mountains, Colorado: U.S. Geol. Survey Prof. Paper 343, 70 p. Stuckey, J. L., 1942, Barite deposits in North Carolina, in Newhouse, W. H., ed., Ore deposits as related to structural features: Princeton, N. J., Princeton Univ. Press, p. 106-108. Sutton, A. H., 1953, Fluorspar resources of the southeast, in McGrain, Preston, ed., Techniques of mineral resources exploration and evaluation-Southeastern Mineral Symposium, 1950, Proc.: Ken- tucky Geol. Survey, Spec. Pub. 1, Ser. 9, p. 139-149. 1954, Fluorspar supplies in the United States [abs.], in Annual meeting, 1954; Am. Inst. Mining, Metall. and Petroleum Engineers, Indus. Mining Div., p. 3-5. Swanson, V. E., and Vine, J. D., 1972, Composition of coal, South- western United States: Geol. Soc. America Abs. with Programs, v. 4, no. 7, p. 688-684. Sweeney, J. W., 1971, Phosphate rock: U.S. Bur. Mines Minerals Year- book, 1969, v. 1-2, p. 907-920. Taber, J. W., 1952, Crystal Mountain fluorite deposits, Ravalli County, Montana: U.S. Bur. Mines Rept. Inv. 4916, 8 p. 1953, Montana's Crystal Mt. fluorite deposit is big and high grade: Mining World, v. 15, no. 7, p. 43-46. Temple, A. K., and Grogan, R. M., 1963, Manto deposits of fluorspar, northern Coahuila, Mexico: Econ. Geology, v. 58, no. 7, p. 1037-1053. Thompson, T. G., and Taylor, H. J., 1983, Determination and occur- rence of fluorides in sea water: Indus. Eng. Chemistry, Anal. Ed., v. 5, no. 2, p. 87-89. 98 GEOLOGY AND RESOURCES OF FLUORINE Thorarinsson, Sigurdur, 1970, Hekla, a notorious volcano: Reykjavik Almenua Bokafélagid, 62 p. Thurston, W. R., 1946, Preliminary report on the Baxter fluorspar deposit near Broken Hills, Nevada: U.S. Geol. Survey Strategic Minerals Inv. Prelim. Rept. 3-196, 3 p. Thurston, W. R., and Hardin, G. C., Jr., 1954, Moore Hill fault system, Crittenden and Livingston Counties, in Fluorspar deposits in western Kentucky, Pt. 3: U.S. Geol. Survey Bull. 1012-E, p. 81-113. Thurston, W. R., Staatz, M. H., Cox, D. C., and others, 1954, Fluor- spar deposits of Utah: U.S. Geol. Survey Bull. 1005, 53 p. Trace, R. D., 1947, The Fourth of July and Luckie No. 1 and No. 2 fluorspar veins, Greenlee County, Arizona: U.S. Geol. Survey Strategic Minerals Inv., Prelim. Rept. 3-207, 6 p. 1954, Mineral Ridge area, Livingston and Crittenden Counties, in Fluorspar deposits in western Kentucky, Pt. 2: U.S. Geol. Survey Bull. 1012-D, p. 59-79. 1960, Significance of unusual mineral occurrence at Hicks Dome, Hardin County, Illinois, in Short papers in the geological sciences: U.S. Geol. Survey Prof. Paper 400-B, p. B63-B64. 1962a, Geology of the Salem quadrangle, Kentucky: U.S. Geol. Survey Geol. Quad. Map GQ-206. 1962b, Geology and fluorspar deposits of the Levias-Keystone and Dike-Eaton areas, Crittenden County, Kentucky: U.S. Geol. Survey Bull. 1122-£, 26 p. 1966, Geologic map of the Marion quadrangle, Crittenden and Caldwell Counties, Kentucky: U.S. Geol. Survey Geol. Quad. Map GQ-547. Tweto, Ogden, Bryant, Bruce, and Williams, F. E., 1970, Mineral re- sources of the Gore Range-Eagles Nest Primitive Area and vicinity, Summit and Eagle Counties, Colorado: U.S. Geol. Survey Bull. 1319-C, 127 p. Ulrich, E. O., and Smith, W. S. T., 1905, The lead, zinc, and fluor- spar deposits of western Kentucky: U.S. Geol. Survey Prof. Paper 36, 218 p. Underwood, J. R., Jr., 1963, Geologic map of Eagle Mountains and vicinity, Hudspeth County, Texas: Texas Univ. Bur. Econ. Geol- ogy Geol. Quad. Map 26, 32 p. U.S. Bureau of Mines, 1969, Minerals yearbook, 1968: U.S. Bur. Mines, v. 1-2, 1208 p. 1973, Minerals yearbook, 1971; U.S. Bur. Mines, v. 1, 1303 p. U.S. Geological Survey, 1956, Fluorspar reserves of the United States estimated: U.S. Geol. Survey Press Release, Nov. 23, 1956, 3 p. 1968, Geological Survey research 1968: U.S. Geol. Survey Prof. Paper 600-A, 371 p. 1970, Geological Survey research 1970: U.S. Geol. Survey Prof. Paper 700-A, 426 p. [1971]. Valentine, G. M., 1949, Nonmetallic minerals, Pt. 1 of Inventory of Washington minerals: Washington Div. Mines and Geology Bull. 37, v. 1, text, 175 p., v. 2, maps, 83 p. (2d ed., 1960, revised by M. T. Huntting). Van Alstine, R. E., 1947, Fluorspar investigations, in Vanderwilt, J. W., [ed.], Mineral resources of Colorado: Denver, Colorado Min- eral Resources Board, p. 457-465. 1954, Geologic features of the fluorspar deposits of the United States [abs.], in Annual meeting 1954; Am. Inst. Mining Metall. and Petroleum Engineers, Indus. Mining Div., p. 2-3. 1964, Fluorspar, in mineral and water resources of Colorado: U.S. 88th Cong., 2d sess., Senate Comm. Interior and Insular Affairs, Comm. Print, p. 160-165. 1965, Geochemical prospecting in the Browns Canyon fluorspar district, Chaffee County, Colorado, in Geological Survey research 1965; U.S. Geol. Survey Prof. Paper 525-D, p. D59-D64. 1966, Fluorspar, in Mineral and water resources of Washington: U.S. 89th Cong., 2d sess., Senate Comm. Interior and Insular Affairs, Comm. Print, p. 199-204. 1969, Geology and mineral deposits of the Poncha Springs NE quadrangle, Chaffee County, Colorado, with a section on Fluo- rspar mines and prospects, by R. E. Van Alstine and D. C. Cox: U.S. Geol. Survey Prof. Paper 626, 52 p. [1970]. Van Alstine, R. E., Estrada, Samuel, de la Garza, Ernesto, 1962, In- vestigacion de los principales distritos de fluorita en Mexico: Mexico Consejo Recursos Nat. no Renovables Bol. 62, 62 p. Van Alstine, R. E., and Moore, R. T., 1969, Fluorspar, in Mineral and water resources of Arizona: U.S. 90th Cong., 2d sess., Senate Comm. Interior and Insular Affairs, Comm. Print, p. $48-857. Van Alstine, R. E., and Sweeney, J. W., 1968, Fluorspar, in Mineral resources of the Appalachian Region: U.S. Geol. Survey Prof. Paper 580, p. 286-289. Wallace, S. R., Muncaster, N. K., Jonson, D. C., MacKenzie, W. B., Bookstrom, A. A., and Surface, V. E., 1968, Multiple intrusion and mineralization at Climax, Colorado, in Ridge, J. D., ed., Ore deposits of the United States, 1983-1967 (Graton Sales Volume), V. 1: New York, Am. Inst. Mining Metall. and Petroleum Engin- eers, p. 605=640. Warne, J. D., 1947, Northgate fluorspar, Jackson County, Colorado: U.S. Bur. Mines Rept. Inv. 4106, 23 p. Warner, L. A., Holser, W. T., Wilmarth, V. R., and Cameron, E. N., 1959, Occurrence of nonpegmatite beryllium in the United States: U.S. Geol. Survey Prof. Paper 318, 198 p. Weaver, C. E., 1912, Geology and ore deposits of the Index mining district: Washington Geol. Survey Bull. 7, 96 p. Weis, P. L., Armstrong, F. C., and Rosenblum, Samuel, 1958, Recon- naissance for radioactive minerals in Washington, Idaho, and western Montana 1952-1955; U.S. Geol. Survey Bull. 1074-B, p. 7-48 [1959]. Weis, P. L., Schmitt, L. J., Jr., and Tuchek, E. T., 1972, Mineral resources of the Salmon River Breaks Primitive Area, Idaho, with a section on Aeromagnetic survey, by W. E. Davis: U.S. Geol. Survey Bull. 1353-C, 91 p. Weller, J. M., Grogan, R. M., and Tippie, F. E., 1952, Geology of the fluorspar deposits of Illinois: Illinois Geol. Survey Bull. 76, 147 p. Weller, Stuart, and Sutton, A. H., 1951, Geologic map of the western Kentucky fluorspar district: U.S. Geol. Survey Mineral Inv. Field Studies Map MF-2. White, D. E., Hem, J. D., and Waring, G. A., 1963, Chemical composi- tion of subsurface waters, Chap. F of Fleischer, Michael, ed., Data of geochemistry [6th ed.]: U.S. Geol. Survey Prof. Paper 440-F, 67 p. White, D. E., Thompson, G. A., and Sandberg, C. H., 1964, Rocks, structure, and geologic history of Steamboat Springs thermal area, Washoe County, Nevada: U.S. Geol. Survey Prof. Paper 458-B, 63 p. > White, D. E., and Waring, G. A., 1963, Volcanic emanations, Chap. K of Fleischer, Michael, ed., Data of geochemistry [6th ed.]: U.S. Geol. Survey Prof. Paper 440-K, 29 p. White, W. H., and MacKenzie, W. B., 1973, Hydrothermal alteration associated with the Henderson molybdenite deposit [Colorado] [abs.]: Econ. Geology, v. 68, no. 1, p. 142. Williams, F. E., 1966, Fluorspar deposits of New Mexico: U.S. Bur. Mines Inf. Circ. 8307, 148 p. Williamson, D. R., 1961, Mexican fluorite deposits: Colorado School Mines Mineral Industries Bull., v. 4, no. 2, 15 p. Wilmarth, V. R., Bauer, H. L., Jr., Staatz, M. H., and Wyant, D. G., 1952, Uranium in fluorite deposits, in Selected papers on uran- ium deposits in the United States: U.S. Geol. Survey Circ. 220, p. 13-18. Wilson, E. D., 1933, Geology and mineral deposits of southern Yuma County, Arizona: Arizona Bur. Mines Bull. 134, Geol. Ser. 7, v. 4, no. 2, 236 p. REFERENCES 99 1950, Fluorspar in Arizona: Arizona Bur. Mines Circ. 15, 13 p. 1951, Castle Dome district, Chap. 10, in Arizona lead and zinc deposits, Pt. 2: Arizona Bur. Mines Bull. 158, Geol. Ser. 19, v. 22, no. 3, p. 98-115. Woollard, G. P., 1968, The interrelationship of the crust, the upper mantle, and isostatic gravity anomalies in the United States, in The crust and upper mantle of the Pacific area-Internat. Upper Mantle Proj., Sci. Rept. 15: Am. Geophys. Union Geophys. Mon. 12 (Natl. Acad. Sci.-Natl. Research Council Pub. 1687), p. 312-341. 1969, Tectonic activity in North America as indicated by earth- quakes, in The Earth's crust and upper mantle-structure, dynamic processes, and their relation to deep-seated geological phenomena: Am. Geophys. Union Geophys. Mon. 13 (Natl. Acad. Sci.-Natl. Research Council Pub. 1708), p. 125-133. Woollard, G. P., chm., and Joesting, H. R., coordinator, 1964, Bouguer gravity anomaly map of the United States (exclusive of Alaska and Hawaii): Am. Geophy. Union and U.S. Geol. Survey. Worl, R. G., Van Alstine, R. E., and Heyl, A. V., 1974, compilers, Fluorite in the United States (exclusive of Hawaii): U.S. Geol. Survey Mineral Inv. Resource Map MR-60. Worl, R. G., Van Alstine, R. E., and Shawe, D. R., 1973, Fluorine, in Brobst, D. A., and Pratt, W. P., eds., United States mineral resources: U.S. Geol. Survey Prof. Paper 820, p. 223-235. Wright, T. L., 1971, Chemistry of Kilauea and Mauna Loa lava in space and time: U.S. Geol. Survey Prof. Paper 735, 40 p. [1972]. Wright, T. L., and Fiske, R. S., 1971, Origin of the differentiated and hybrid lavas of Kilauea volcano, Hawaii: Jour. Petrology, v. 12, no. 1, p. 1-65. Yates, R. G., and Thompson, G. A., 1959, Geology and quicksilver deposits of the Terlingua district, Texas: U.S. Geol. Survey Prof. Paper 312, 114 p. Zartman, R. E., Brock, M. R., Heyl, A. V., and Thomas, H. H., 1967, K-Ar and Rb-Sr ages of some alkalic intrusive rocks from central and eastern United States: Am. Jour. Sci., v. 265, no. 10, p. 848-870. Zies, E. G., 1929, The fumarolic incrustations and their bearing on ore deposition, Pt. 1; and The acid gases contributed to the sea during volcanic activity, Pt. 2, in The Valley of Ten Thousand Smokes; Natl. Geog. Soc., Contributed Tech. Papers, Katmai Ser., v. 1, no. 4, 79 p. Zietz; Isidore, Andreasen, G. E., and Cain, J. C., 1970, Magnetic anomalies from satellite magnetometer: Jour. Geophys. Research, v. 75, no. 20, p. 4007-4015. Zietz, Isidore, and Kirby, J. R., Jr., 1972, Aeromagnetic map of Colo- rado: U.S. Geol. Survey Geophys. Inv. Map GP-836. ++ U.S. GOVERNMENT PRINTING OFFICE 1976-777-034/12 QE 715 arm _- ? C 7 DAYS CIENCEg LIBRARY vid 9 <4 Paleozoic Origin of the Cycads GEOLOGICAL SURVEY PROFESSIONAL PAPER 934 [DOCUMENTS Derartment I JAN 19 r LIBRARY UNIVERSITY OF CALIFORNIA Paleozoic Origin of the Cycads By SERGIUS H. MAMAY GE O LOGICAL SURVEY PROFESSIONAL PAPER 934 Megasporophylls and associated plant parts from the Upper Pennsylvanian and Lower Permian rocks of the southwestern United States are interpreted as a lineage of cycads, derived from the pteridosperms. Cycadean foliage is thought to be derived from Paleozoic taeniopterids UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1976 UNITED STATES DEPARTMENT OF THE INTERIOR THOMAS S. KLEPPE, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Director Library of Congress Cataloging in Publication Data Mamay, Sergius H. 1920- Paleozoic origin of the cycads. (Geological Survey professional paper; 934) Bibliography: p. Includes index. 1. Cycadales, Fossil. 2. Paleobotany-Paleozoic. 2. Paleobotany-Southwest, New. I. Title. II. Series: United States. Geological Survey. Professional paper; 934. QE976.2.M35 561'.591 76-14384 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 Stock Number CONTENTS Page A > >> :. = . ». .. colo oo ollo o o + he cnn ne nve hr m a en Ee mme tne malar ese 1 | Systematic descriptions-Continued _ --- .- - -- -. - - cll ll o o on nne nn n pne en nnn nn na anar 1 *Cycadean Male CONOS -n. cn else csecells Acknowledgments _________ccccccccccclllclclllcllll 3 Cycadospadix yochelsoni Mamay, n. sp __________ ; ._ 4 A§soc.1ate.d plants -__- A non § Distribution of the fossils Sources of material 4 Morphologic and evolutionary considerations _______. EFEYIOHS| WOT ool lull nol nnn ener ous 5 Interrelationships of the American fossils _______ Systematic descriptions ___________________________ 6 An alternative interpretation of Archaeocycas ___ Spermopteris sp }L__________LLLLLLLLLLLLLLLL LL 7 Relationships of the American fossils to other Archaeocycas Mamay __________________________ 8 fosgll CyCAOUS - -.. 2 s 2 2 ee acl ao ue hele onle oe mace sa e marg s te Bein ua he an Archaeocycas whiter Mamay ___________________ 8 Evolution of the pinnate habit in cyeadalean Phasmatocyeas Mamay ________________________ 12 FOliAgQ | uus... uuu ine sme Phasmatocycas kansana Mamay ________________ 12 | Conclusions L ?Phasmatocyeas sp ____________________________ P1 | References cited ?Phasmatocycas spectabilis Mamay, n. sp ________ 22.1 MEK , .. . 1 o nor s nen n o eran amer ans ne an a in be mee ie ml to t PLATE FIGURE r $%» 5 m Vp fso. popa 9 10. 11. ILLUSTRATIONS [Plates follow index] Archaeocycas white Mamay. Archaeocycas white Mamay, ?Phasmatocycas sp., Spermopteris sp., Cycadospadix yochelsoni Mamay, n. and ?cycadean male cone. Phasmatocycas kansana Mamay. Seed cuticles of Phasmatocycas kansana Mamay. ?Phasmatocycas spectabilis Mamay, n. sp., and Taeniopteris sp. Reconstruction of part of seed-bearing leaf of Spermopteris coriaceq ________________________________ Diagrams showing hypothetical reduction of the megasporophyll and evolution of the ovulate cone im- plied by iving eyoentls . .. _... . ... .. .. .. 2 ul nll cull 2. soll oce s poe an eee ee me he w me e on on ae a a ne ne a fe e an jon mend e ae ae he ah in te he a an me he ae ai be a he ui as we Semidiagrammatic reconstruction of megasporophyll of Archaeocycas ______________________________ Semidiagrammatic reconstruction of megasporophyll of Phasmatocyeas ______________________________ Stratigraphic correlations showing vertical distribution of late Paleozoic cycads and related plants ____ Diagrams showing hypothetical evolutionary development of Cycas-like megasporophyll from spermop- terid pleridOspent ... - - _o o.. .. cou. a al 2 oo ul ol ol he bone oo ie nt ml al ar on n e e oe ma me p it h n an nl ie oe oe e un t An e 'as ar us ne jn js ane t on ne ht n ie n te tn iy e ae ae te to Semidiagrammatic representation of hypothetical evolution of Cycas-like megasporophyll from Archaeo- CEE oo s cs sile. as oo al he ue me oe ma me se ar sa he oe ne ne Suan e e ne Sane ue or ne he s os ne ne an ap ht he fe os he en ne s us at e n og tems mas ae sa h a te n n as ue n an an S oe ae oe ewe n hare en a d Heat ae Semidiagrammatic representation of hypothetical evolution of carpellike organ from Archaeocycas ___ Reconstructions of DioonffoG@rpidiHM .L. lll ll nll lo nln on Reconstructions of Diagrams depicting suggested evolution of cycadalean leaves from taeniopterid ancestral forms ________ III Page 28 24 26 29 29 33 36 39 43 44 47 sp., Page 2 3 11 18 29 31 33 35 37 39 43 PALEOZOIC ORIGIN OF THE CYCADS By SErcius H. Mamay ABSTRACT Upper Paleozoic rocks in the southwestern United States contain several taxa of fossil plants that appear to repre- sent a lineage from the Pennsylvanian pteridosperms to the modern cycads. The oldest genus, Spermopteris Cridland and Morris, is a megasporophyll that has surficial seeds borne in two rows along the sides of presumably abaxial faces of taeniopterid foliage; it was first found in the Virgilian Law- rence Formation of Kansas, and a poorly preserved specimen is now known from the Leonardian Wellington Formation of Oklahoma. The putatively derivative cycadean genera Phasmatocycas Mamay and Archaeocycas Mamay occur in lower Leonardian rocks of Kansas and Texas. Phasmatocycas (type-species: P. kansana Mamay) is a fertile axis that has two lateral rows of naked ovules. In P. kansana, the ovules are represented by well-preserved double cuticles similar to those of the Jurassic cyead Beania; small resinoid spherules, interpreted as remains of glands, alternate with the ovules. Taeniopterid foliage as- sociated with P. kansana bears abundant spherules identical with those on the fertile axis; on the basis of the spherules and the known seed habit of Taemiopteris, Phasmatocycas is reconstructed as a cycadlike megasporophyll with an elaminar, fertile base and a taeniopterid distal lamina. The glands may be primitive nectaries, possibly involved in an early stage in the development of entomophily. Phasmatocycas is regarded as having been derived from a spermopterid ancestor through reduction of numbers of seeds and loss of the basal foliar lamina. Although the type specimen of Phasmatocycas is from the Wellington Formation of Kansas, a poorly preserved specimen is known from the younger Vale Formation of Texas, where it is associated with structures strongly reminiscent of male cycadean cones. Archaeocycas Mamay, typified by A. white? Mamay, is ap- proximately contemporary with Phasmatocycas kansana, oc- curring at two localities in the Belle Plains Formation of Texas. It resembles Phasmatocycas in having a basal seed- bearing area and a sterile distal lamina. In Archaeocycas, however, the seeds are attached to the basal lamina, which is inrolled and at least partly encloses the seed. Archaeocycas probably represents a form intermediate between Spermopteris and Phasmatocycas. Alternatively, Archaeocycas might be a fertile appendage evolving in the direction of a carpellike or- gan through proliferation and fusion of the laminar margins; this alternative, however, is not preferred. Associated with the cycadean megasporophylls are a single specimen of Cycadospadix (C. yochelsoni Mamay, n. sp.) and various other plants. C. yochelsoni cannot be determined as cycadalean or bennettitalean, although it is possible that it is either and that the two orders may have been distinct as long ago as the Early Permian. Among other plant associ- ates, only Taeniopteris is common to all the cycad-bearing assemblages; this strengthens the hypothesis that Taenmiop- teris was the leaf of primitive cycads, as indicated by Sper- mopteris. Stratigraphic distribution of the plants supports the idea of derivation of the cyeads from the Pennsylvanian spermop- terids and of a lineage to the modern cycads through the Triassic genera Palaeocycas-Bjuvia and Dioonitocarpidium. Horizontal and vertical distribution of the Paleozoic taxa em- phasize the importance of the lower Leonardian of the south- western United States as a temporal and geographic locale for rapid and dramatic plant evolution. Modern cycadean leaves are thought to be derived from taeniopterid stock in the Paleozoic and Mesozoic through progressive incision of the laminar margin. In one line of foliar evolution, shallow marginal dentation, in which each tooth receives only one vein, becomes progressively deeper, resulting in the cyeadaceous leaf type in which each leaflet has a single vein; a parallel series of known forms involved sev- eral veins to a segment and could have produced the zamia- ceous leaf type. The Stangeria leaf type could have arisen from a compound taeniopterid with little modification, whereas marginal incision of a compound taeniopterid might have produced the Bowenia type of leaf. Cumulative evidence indicates that the cyceads extend back into the Late Pennsylvanian instead of only the Late Triassic; thus they rival the conifers as the oldest extant seed plants. INTRODUCTION The tenor of this paper is appropriately estab- lished by reference to some of T. M. Harris' percep- tive comments on the fossil history of the cycads. Harris (1961, p. 322) : wrote as follows : The fossil Cycads consist of a good many genera of isolated organs classified on more or less good evidence as Cycadales, a very few genera with two known organs and one genus with more than two organs. If we are right in our synthesis of these plants with two or more organs, and also right in our present classification of, say, half of the isolated organs (and I think it unreasonable to hope for more), then we can say that by Jurassic times the family had probably completed its evolution. * * * The known genera are no more primitive than the living ones and they do not help in linking the Cycadales with any other family, and for this link one looks first at the 1 2 PALEOZOIC ORIGIN OF THE CYCADS Pteriodosperms. If we are to find progressive evolution it must be before the Jurassic. Harris' remarks succinctly outline a particularly enigmatic deficiency in the paleobotanical record, in regard to which considerable speculatory discussion has been published. A brief supplement to Harris' statements is in order here as background to the present article, which may advance the status of our knowledge of the evolution of the cycads. The order Cycadales is a small group of gymno- sperms; most authors regard it as consisting of only nine living genera, although Johnson (1959) recog- nized a tenth. The geologic history of the Cycadales is little known and perplexing, despite the abundant Mesozoic occurrences of fossil leaves generally at- tributed to the cycads. Intermixed and often closely associated with these fossils are leaves of the extinct Bennettitales or cycadeoids-plants that reached their zenith during the Jurassic and disappeared before the end of the Mesozoic. The distinction be- tween the foliage of the two groups is based largely on features of the epidermis, and particularly on the stomatal characters. Otherwise, leaves of the cycade- oids and cyceads are so similar that they are difficult to differentiate on the basis of gross external morphology. Reproductive structures of the cycadeoids were preserved in some abundance and have been investi- gated in considerable detail (Delevoryas, 1968a, b). However, little is known of either the polleniferous or ovulate mechanisms of the fossil cycads; accord- ingly, their phylogeny is speculative. The often-cited Late Triassic genera Dioonitocarpidium von Lilien- stern and Palaeocycas Florin ("Bjuvia'" Florin) and the Jurassic Cycadospadix Schimper are the oldest fossils regarded to represent true cycadean ovulate appendages. Of much more significance is the Middle Jurassic genus Beania Carruthers, a lax conelike structure with biovulate, spirally disposed peltate appendages. Beania is consistently compared with female cones of the modern zamioid cyeads (Arnold, 1953, p. 62). The Jurassic existence of cones with advanced morphologic features such as those found in Beania indicates a considerable pre-Jurassic evo- lutionary history for the cycads ; clearly, Harris had this circumstance in mind while writing the com- ments quoted above. The Jurassic cyceads, along with the sparse but significant evidence of Triassic rela- tives, nurture a reasonable speculation that true cycads existed during Paleozoic time. Harris' designation of the pteridosperms as the probable source of a "link'" between the Cycadales and other taxa expresses a long-established line of thought, but one that has been developed on tenuous evidence. I feel that concrete evidence of a late Paleozoic cycadalean stock may now be contributed to the paleobotanical record. The cycadalean stock indeed has its beginnings in the pteridosperms, as speculated, and more specifically, in the taeniopterid pteridosperms. In 1960, Cridland and Morris described the genus Spermopteris (fig. 1), a taxonomic novelty based on fertile pteridospermous foliage from the Upper Pennsylvanian of Kansas. Spermopteris was shown to be the ovulate phase of a member of the form- genus Taeniopteris, representatives of which are common in upper Paleozoic rocks. The seed-bearing specimens were identified with Taeniopteris coriacea Goeppert, which became the basionym for the type species, Spermopteris coriacea (Goeppert) Cridland and Morris. Spermopteris is characterized by having "a row of seeds borne on the abaxial surface of each side of the midvein" (Cridland and Morris, 1960, p. 855) on what otherwise were normal-appearing ap- pendages. The authors limited their discussion of the significance of their material to a few statements regarding the infrageneric relationships of certain taeniopterid forms and refrained from speculating on the possibly broader evolutionary implications of Spermopteris. It is my opinion, however, that Sperm- opteris is one of the earliest known elements of evi- dence bearing on the paleontologic history of the cyeads from the standpoint of ovulate reproductive structures. This opinion is bolstered considerably by the characteristics of the Permian material dis- FIGURE 1.-Reconstruction of part of fertile leaf of Sper- mopteris coriacea, showing oblique view of upper (ventral) surface, and at lower right, mode of seed attachment to lamina (from Cridland and Morris, 1960, fig. 10). Approxi- mately X 3. INTRODUCTION 3 cussed here, and considered with those of Spermop- teris they form the basic theme of this paper. The Paleozoic plant collections of the U.S. Geo- logical Survey contain several Permian specimens of ovuliferous appendages that resemble Spermop- teris in the general disposition of their seeds but differ conspicuously in details of the foliar lamina- tion. Also present in the collections are other forms that may be related and are therefore included in this paper. Because of the stratigraphic and geo- graphic relationships of the two suites of specimens, it seems reasonable that some of the morphological resemblances reflect phylogenetic alliances rather than fortuities or parallelisms. Thus this paper is intended as a sequel to the Cridland and Morris work on Spermopteris. All these specimens have been con- sidered critically with the purpose of stressing cer- tain morphological features that they have in com- mon with each other, with other fossils regarded to represent true cyceads, and with the modern Cyca- dales. A certain amount of speculation is entailed in this presentation, but the ensuing pages will also describe tangible evidence for the presence, in the late Paleozoic record, of plants directly related to the living recent cycads. In the literature dealing with the modern cycads, evolutionary advancement of the various genera is usually interpreted in terms of comparative morphol- ogy of the female cones. The megasporophylls of Cyeas are held to be the most primitive ovulate or- gans of the order, because they retain leaflike fea- tures indicative of foliar derivation and are pro- duced like leaves in loose terminal crowns instead of in compact cones (fig. 2; Chamberlin, 1965, p. 145- 150). Because they are produced like leaves they are also shed like leaves, as individual units rather than as more or less complete cones. Thus one would logically look for individual detached fossil mega- sporophylls in searching for primitive fossil cycads or their progenitors. This assumption is borne out here, inasmuch as the material on which this study is based consists entirely of single detached append- ages that provide no evidence regarding modes of attachment to the parent plants. Fortunately the individual cycadalean megasporophyll, though sim- ply constructed, is sufficiently diagnostic that the in- complete nature of the fossils treated here does not detract substantially from their phylogenetic significance. ACKNOWLEDGMENTS The late David White is in no small measure re- sponsible for this paper, because he collected much FicurE 2.-Hypothetical reduction of the megasporophyll and evolution of the cone as implied by living cycads. A, Leaf; like megasporophyll of Cycas revoluta; F, loose terminal crown of megasporophylls (C. revoluta), not compacted into a cone. B, Megasporophyll of Cycas media, the leaflets re- duced to serrations; G, terminal aggregate of megaspor- phylls of same species, in a more conelike arrangement. C, Megasporophyll of Dicon edule, in which the lamina is en- tire-margined; H, loosely compacted cone of D. edule. D, Megasporophyll of Macrozamia miquelii, the lamina re- duced to a spine; I, tightly compact cone of same species. E, Highly reduced megasporophyll of Zamia floridana, hav- ing no resemblance to a leaf; J, extremely compact cone of Z. floridana. (All from Chamberlain, 1930, p. 290, with slight modifications of figure explanations.) of the critical material. Although he published noth- ing known to refer directly to this material, his curatorial markings on specimens and certain other kinds of evidence show that he was aware of the im- portance of some of the material described here and almost certainly suspected its cycadlike affinities. Thus White's considerable influence on the Paleozoic paleobotany of North America continues, even now, to exert itself. Thanks are extended to Dr. J. A. Wilson, Univer- sity of Texas, for information regarding the Patter- son locality at Lawn, Tex., and to C. O. Patterson for permission to collect on his property. Ralph 4 PALEOZOIC ORIGIN OF THE CYCADS Howe and Mart Russell of Seymour, Tex., kindly per- mitted me to collect on the "Emily Irish" land during several field seasons. T. R. Tinsley of Stamford, Tex., gave me some of the specimens described and showed me the fossil plant deposit in Haskell County. A. D. Watt, U.S. Geological Survey, assisted me in the field in 1961, 1963, 1967 and 1974. During my first collecting trip to the Lawn locality, I was accom- panied by C. B. Read and E. L. Yochelson, both of the U.S. Geological Survey. Read also gave me valua- able instruction in general procedures in locating fossiliferous deposits in the southwestern Permian red beds. TERMINOLOGY The literature dealing with cyeadophytes is some- what confusing with regard to the terminology ap- plicable to the foliar organs. Some authors prefer the terms "leaf" and "leaflet," whereas, others use "front" and "pinna," or "frond-like leaf" ; occasion- ally the alternative terms are even used interchange- ably in the same publication. Thus, this area of ter- minology seems basically to involve a matter of personal choice, especially in routinely descriptive approaches. I prefer "leaf" to "frond," particularly because this study is fundamentally concerned with entire-margined taeniopterid foliar organs of un- known origins or homologies. My use of the term "megasporophyll" for the fer- tile structures discussed here is commonly accepted usage, yet it invites a statement of qualification be- cause of an article by Meeuse (1963), in which the traditional interpretation of the ovuliferous append- age of Cyeas was challenged. Sporne (1965, p. 111-112) presented a consensus on the evolutionary status of Cycas, reiterating the substance of Chamberlain's hypothesis as presented in figure 2 of this paper. Sporne wrote "It is widely accepted by morphologists that the Cyeas sporo- phyll is the primitive type among cycads, and that, during the evolution of the group, there has been a reduction in the number of ovules to two * * *." Meeuse (1963, p. 122), however, presented the in- teresting but improbable hypothesis that the ovulate appendage of Cycas is "an organ of dual nature com- bining axial and foliar characteristics" and the "so- called 'megasporophyll' represents an abnormally developed whole cyeadalean strobilus." In attempt- ing to demonstrate that "the conventional identifica- tion of the female reproduction organ of Cycas with a 'megasporophyll' (defined as a leaf homologue bearing marginal ovules) is fallacious," Meeuse in- voked unconvincing evidence, such as structural aberrancy or resemblance of the ovuliferous organ of Cycas to a teratological case. He furthermore re- jected Palaeocycas Florin as a cycead, primarily on the basis of its entire-margined taeniopterid leaves. Because of the taeniopterid form of the lamina of Spermopteris, its clearly megasporophyllar nature, and the evolutionarily simple processes of reduction of parts and incision of foliage necessary for deriva- tion of the Cycas type of fertile appendage, Meeuse's hypothesis is herewith rejected in favor of tradition- al concepts. I regard the term "megasporophyll" as correctly applicable to Cycas and use it here accordingly. The ovulelike bodies on the basal part of the lami- na in Archaeocycas are described as attached to the lower (abaxial) surface of the megasporophyll, but the choice of this word-in preference to the con- trasting term "upper" adaxial-is an arbitrary one, based on homology with the abaxial positons of the sporangia of most ferns. The few specimens avail- able are detached, and thus it is impossible to deter- mine their orientation in relation to the apex of the parent plant. Even though proof is lacking, how- ever, the specimen shown in lateral view in plate 2, figure 7, lends the strong impression of being ori- ented in its natural attitude, that is, with its adaxial surface toward the top of the illustration. soOURCES OF MATERIAL The specimens described here came from eight localities in Lower Permian outcrop areas, one in Oklahoma, three in Kansas, and four in Texas. The localities are described below, the numbers referring to U.S. Geological Survey paleobotany localities; quotation marks indicate parts of locality descrip- tions that were taken from the U.S. Geological Sur- vey (USGS) locality register: 6233: "St. L. & S. F. Ry. in cut % mile west of station at Perry, Oklahoma. Coll. David White, Sept. 6, 7, 8, 1911. Permian." According to the geologic map of Oklahoma (Miser, 1954) this locality is within an undivided Permian rock interval indicated as "Pwa" (including the Admire Group through the Welling- ton Formation). As far as I am aware, no subse- quent collections have been recorded from this lo- cality. 8298: "Elmo limestone. Wellington formation,. Permian. In- sect Hill, 3% mi. SE of Elmo, Kansas, Barn lot on S. side of road, and adjacent localities. Coll. C. B. Read, 1931." The Elmo Limestone Member of Dun- bar (1924) of the Wellington Formation is of Early Permian (Leonardian) age. I have not visited this area. SOURCES OF MATERIAL 5 8868: "Carleton, Kansas. Wade Sterling Ranch, 4 miles south of Carleton, Dickinson Co., Kansas (Wellington Fm.). D. White, 1909." 8869: "3% miles south of Banner, Kansas (Elmo Is.). Coll. by David White, 1909." This collection was probably made very close to locality 8298, and almost certain- ly is from the same rock unit. 8877: "Wichita formation, 4% miles south-east of Fulda, Texas. Collected by D. White." This locality is within the Lower Permian Belle Plains Formation (Romer, 1958, map facing p. 178), and although the locality register and specimen labels lack the date of the collection, a brief discussion of the collecting activities by White in his paper on Gigantopteris americana (1912, p. 495) indicates that the collec- tion was made in 1910. White wrote "the exact local- ity being the bank of the stream at the crossing of the old road, one-fourth miles south of the ford of the Little Wichita River, 4 miles southeast of Fulda, a station in Baylor County." In 1955 and 1957, I at- tempted to relocate this important fossil deposit, but the vague nature of the locality description and the lack of distinctive physiographic features in this badlands terrain made it impossible to be certain of one's position. Furthermore, the plant-bearing shale is easily eroded and the original deposit prob- ably has been totally destroyed or obscured. A few nondescript fossil plant fragments were found in the supposed vicinity of White's locality, but none of these equaled White's material in quality of preser- vation. 8959: "Baylor County, Texas. East-central part of Emily Irish grant, south side of Salt Fork of Brazos River, 16 to 16% miles scaled due S.E. of Seymour. Shale in upper part of Belle Plains Formation. Coll. Read and Ervin. October, 1940." This locality is on prop- erty now owned by Roland Howe, Fort Worth, Tex., and the plant deposit is by far the richest and most diversified one known in the Permian of the south- western United States. I have collected from it on five different occasions, in 1955, 1957, 1961, 1963, 1967, and 1974. Actually, the plant deposit is at about the middle, rather than the upper part, of the Belle Plains Formation and is regarded as the ap- proximate equivalent of the Valera Shale. A detailed locality description and a measured section of the "Emily Irish" exposure have been published (Ma- may, 1968, p. 12-13). 10057: C. O. Patterson property, Taylor County, Tex. The site is reached by starting at the theater in Lawn, Tex., going 1.5 miles (2.4 km) west, 2.9 miles (4.8 km) south, 1.1 (1.8 km) miles west, and 0.5 (0.9 km) mile south to the remains of the old home of C. 0. Patterson. Fossils were found in a quarry and its vicinity, in a pasture against the east slope of a low hill as viewed west-southwest from the west side of the house. Fossil plants were especially abundant in a small lens of light gray shale about 100 yards (915 m) north of the quarry, and con- siderable material was collected in April 1955 by C. B. Read, E. L. Yochelson, and myself. According to Meyers and Morley (1929b), the Lawn deposits are in the lower part of the Vale Formation, in the Lower Permian Clear Fork Group. (See Wilson, 1953, p. 456-459 for further discussion of this lo- cality.) In 1957, A. D. Watt and I revisted this lo- cality, hoping to obtain additional material. We found that the Patterson home had recently been destroyed by fire and was then in the process of being rebuilt on another site, and our collecting plans were com- pletely thwarted because the pasture in which the fossil deposits occurred had been deep-plowed as part of a mesquite-eradication program. Small chips of the plant-bearing matrix and the under- lying caliche bed were scattered over a fairly ex- tensive area, but the plant beds may now be re- garded as having been destroyed and lost to science. 10064: Haskell County, Tex. This locality is approximately 3 miles (4.8 km) north of Stamford, Tex., and just north of the Jones-Haskell County line. To reach the locality, go north and northwest on U.S. High- way 380, starting at the city square in Stamford. After 2.25 miles (3.7 km) turn north on an un- paved crossroad; approximately 1 mile (1.6 km) north, plant fossils occur in red and gray, medium- to fine-grained sandstone which is exposed in the road shoulders and nearby pastures along both sides of the road. According to Meyers and Morley 19292), exposures in this area are within the boundaries of the Vale Formation (Lower Permian Clear Fork Group), and their geographic proximity to other Vale exposures would suggest a high posi- tion in the Vale Formation. The presence of plant fossils in these beds was told to me by T. R. Tinsley of Stamford, and in his company, A. D. Watt and I collected plants there in 1961, 1963, and 1967. PREVIOUS WORK This study is limited to Paleozoic cycadean fructi- fications and their possible affinities with the early Mesozoic cycads, and because only a few examples are known, there is no need for an extensive review of past investigations of all fossil cyead parts. Sev- eral examples of petrified stems and other vegetative remains thought to have cycadean affinities are known from the Paleozoic. These are largely from the Old World, and their occurrences were sum- marized adequately by Arnold (1953, p. 61), who speculated that "There were probably some cycads in existence during the middle or late Carbonifer- ous and by the end of the Paleozoic era they wore fairly numerous." The existence of true cyceads in North America during Late Triassic time was demonstrated by Delevoryas and Hope (1971), who described the new genus Leptocycas from the Pekin Formation of North Carolina. The material consists of leaves and stem parts, both with typical cycadean cuticular characteristics; also found was a structure resem- bling a male cone, attached to a stem. Unfortunate- ly, nothing is known of the ovulate parts. Delevoryas 6 PALEOZOIC ORIGIN OF THE CYCADS and Hope's reconstruction of Leptocyeas gracilis shows a sparsely foliate plant with pinnate leaves and a slender trunk with long internodes. Stem features of Leptocycas differ from the modern cycad trunks which have dense armors of closely spaced persistent leaf bases. Gould's (1971) anatomical study of petrified stem material from Upper Triassic beds of Arizona and New Mexico provided additional important evidence of cyeads in the American Triassic. This investiga- tion centered upon the taxon Lyssozxylon grigsbye Daughtery from the Chinle Formation. Lyssozylon was originally regarded as a member of the Bennet- titales (Daughtery, 1941), but various anatomical characters led Gould to reinterpret it as a true cycad. The first recorded hint of Paleozoic cycads in North America appeared in a list of Permian plants published by White (1912, p. 508). White listed the taxon "Cycadospadix? sp." from a locality supposed- ly near Carlton, Kans. This listing was not sub- stantiated by an illustration or description, but I found in White's collections a specimen that I be- lieve to have been the basis for that identification ; it is probably the most important specimen involved in this study. In 1916, Bassler described a specimen of Plagioza- mites planchardi (Renault) Zeiller from a coal seam in the Upper Pennsylvanian Conemaugh Group in West Virginia; this species is known from several Lower Permian localities in the Old World but had not been reported previously in North America. Originally Plagiozamites was regarded as a true cycad leaf, although conclusive proof has never been found. Bassler's material neither substantiates nor refutes a cycadean affinity for the genus, but the Pennsylvanian occurrence is noteworthy. The present paper culminates several years of study and six collecting trips. My interest was initially aroused by the presence of cycadlike speci- mens in White's Permian collections from the south- western United States and was further stimulated by Cridland and Morris' (1960) description of Spermopteris. My ideas were announced in four brief publications (U.S. Geological Survey, 1968; Mamay, 19692, 1971, 1973), in which the concept of late Paleozoic origin of cycads from Pennsylvanian spermopterids was set forth. One of my articles (Mamay, 19692) was immedi- ately preceded in the same journal by an article by Taylor (19692), in which a new type of pollenifer- ous cone was described from Upper Pennsylvanian coal ball petrifications from Illinois. Taylor (p. 294) identified his material as Pennsylvanian cyceads with the statement "The fossil record of true cycads is extended from the Upper Triassic to the Upper Pennsylvanian"; he (1969b) later reiterated this interpretation. In a subsequent publication, Taylor (1970) named the cones Lasiostrobus polysacei, at the same time modifying his interpretations so as to permit consideration of the possibility that the cones are coniferalean rather than cycadalean. Taylor's reason (1970, p. 682) for moderation of his original- ly unequivocal viewpoint was the discovery of addi- tional specimens, some showing features of the apex. Taylor's altered stance emphasizes the question- able taxonomic position of Lasiostrobus and the ne- cessity for its reevaluation. I regard it as conifero- phytic in the bulk of its characteristics and reject it as a cycad. In my preliminary announcements (Mamay 1969a, b), I said that two new genera were recog- nized in this investigation. Formal names were not proposed ; the genera were merely designated "new genus A" and "new genus B." Sporne (1971) dis- cussed the new plants briefly, drawing attention to their as yet incomplete nomenclatural status; he also introduced an incorrect finality into some of my then-tentative ideas on angiosperm origins and ac- credited me with the discovery of "The origin of the carpel" (Sporne, 1971, p. 14). So that these possi- bly controversial fossils might be discussed with legitimate names, I proposed and validated the gen- eric names Archaeocycas and Phasmatocycas (Ma- may, 19783). SYSTEMATIC DESCRIPTIONS Four genera are recognized and discussed in this paper. Two of these-Spermopteris Cridland and Morris, 1960, and Cycadospadizx Schimper, 1870- are represented by one specimen each. The genus Archaeocycas Mamay, 1973, is represented by eight specimens; both counterparts of three of these are present. Phasmatocycas Mamay, 1973, is based on one specimen, both counterparts of which were found; a number of other specimens are provision- ally assigned to this genus. Inasmuch as the formali- zation of the names Archaeocycas and Phasmato- cycas in 1973 was an expedient designed to make names available while full descriptions and discus- sions were in preparation for this paper, the diag- noses are repeated here in the interest of completeness. With the possible exception of Cycadospadix, the specimens above are clearly of megasporophyllar morphological nature. They are presented in the con- SYSTEMATIC DESCRIPTIONS 7T text of an evolutionarily interrelated complex of plants, which collectively appears to fill an impor- tant hiatus in the early history of the cycads. Two other plant structures, one foliar and one possibly a microsporangiate cone, are discussed because of their possible pertinence to the theme of this investigation. Division PTERIDOSPERMOPHYTA Order PTERIDOSPERMALES Genus SPERMOPTERIS Cridland and Morris, 1960 Spermopteris sp. Plate 2, figure 2 DESCRIPTION This identification is based on one small specimen, poorly preserved on a chip of greenish fine-grained shale from the Perry, Okla., locality. The specimen represents a taeniopterid leaf fragment 3.0 em long and 11.0 mm wide. It is preserved as an impression with only a few minor traces of adherent car- bonaceous matter, which is restricted to the midrib. The midrib is conspicuously marked with parallel longitudinal striations and is very broad in propor- tion to the breadth of the lamina, measuring about 2.5 mm in width. The leaf fragment does not appreci- ably diminish in width from one end to the other. The venation is poorly preserved, but the delicate ultimate veins may be seen by careful observation of a few small parts of the lamina. The veins are close- ly spaced, occasionally forked, and gently decurrent. Insofar as observable, the venation of this specimen is characteristically taeniopterid. The chief significance of this specimen is in the presence of two uninterrupted rows of closely con- tiguous swellings that are roughly oblong or ovoid in shape and of very minor relief. One row lies along each side of the midrib, but the swellings are best seen along the right side of the specimen (pl. 2, fig. 2). The swellings average about 2.5 mm long and 1.5 mm wide. The basal ends of the swellings are not clearly separable from each other lengthwise and substantially coalesce to form a fairly conspicuous, uninterrupted but slightly undulating trough paral- lel to the midrib and about halfway out in the lamina on each side of the midrib. The sides of the swellings form shallow parallel depressions and generally fol- low the decurrence of the lateral veins. The swellings show no evidence of extending beyond the edges of the lamina; rather, they appear to be truncated at the leaf margins. DISCUSSION I am satisfied that this fossil represents a speci- men of Spermopteris, probably preserved before the seeds had matured. Several features of the Okla- homa specimen suggest a congeneric relationship with material of Spe‘Vmoptem's that was regarded by Cridland and Morris to represent immature onto- genetic stages. In one of their specimens (Cridland and Morris, 1960, fig. 2), the positions of supposedly immature seeds are clearly shown, but the general proportions of the specimen are somewhat smaller than those of the Oklahoma specimen. Cridland and Morris illustrated a second specimen (Cridland and Morris, 1960, fig. 9, p. 859), which they cited as a fertile leaf seen from the adaxial side, with "swell- ings which indicate the positions of immature seeds." This specimen and the Oklahoma specimen compare favorably in such characters as width of the leaf; proportional widths of lamina and midrib; density and attitude of the secondary veins ; size, shape, and distribution of the seeds; and termination of the seeds apices exactly at the margins of their respec- tive parent laminae. To judge from the indistinct seed outlines in the Oklahoma specimen, it may be assumed that the surface of this specimen represents the upper or adaxial surface of the leaf and, as in the case of Cridland's and Morris' specimen, the thickness of the foliar tissue obscured the details of the seeds in the fossils. The Oklahoma specimen conforms with the gen- eric diagnosis of Spermopteris in having a row of seeds or seedlike bodies on each side of the midrib of a taeniopterid leaf, but specific alliance cannot be attempted because of the limitations of material. However, a few remarks about specific identifica- tions are added here because there is some evidence that this particular specimen was given more than passing attention by David White. In discussing the relationships of Spermopteris, Cridland and Morris (1960, p. 859) commented on Von Gutbier's (1835, p. 73) original description of Taeniopteris abnrnormis, wherein he mentioned but did not illustrate specimens with shallow cross- wrinkles ("flache Querfalten") toward the tips of the leaves. Apparently Cridland and Morris re- garded this condition as a possible point of compari- son with Spermopteris but did not pursue the idea further because of the unavailability of additional information. White (1912, p. 507) published a pro- visional list of plants from the Perry, Okla. locality. This list includes Taemniopteris multinervis Weiss and T. abnorm{is Gutbier, but no indication was given of the number of specimens of each that were present. The original collection contains just nine specimens of Taeniopteris, mostly very badly preserved. As the Oklahoma Spermopteris specimen is the only taeni- 8 PALEOZOIC ORIGIN OF THE CYCADS opterid in that suite that can be distinguished from the others on morphological grounds, it seems likely that White had studied this very specimen, equated its seedlike swellings with Gutbier's "flache Quer- falten," and identified T. abnormis in the Oklahoma collection. In the same paper, White (1912, p. 506) also listed T. abnormis among the plants collected at the Fulda, Tex. locality (USGS locality 8877), but I have not been able to ascertain the basis for that identification. This occurrence extends the known geographic range of Spermopteris from eastern Kansas to north-central Oklahoma. At the same time, the stratigraphic range of the genus is extended upward from the Upper Pennsylvanian (Virgilian) to the Lower Permian, where it is associated with Gigant- opteris americana in flora zone 14 (Read and Ma- may, 1964). Occurrence.-USGS paleobotany locality 6233. Figured specimen.-USGS 6233-1 (pl. 2, fig. 2). Division CYCADOPHYTA Order CYCADALES Genus ARCHAEOCYCAS Mamay New genus B. Mamay, 1969b. Archaeocycas Mamay, 1973. Type species.-Archaeocycas whiter Mamay. Generic diagnosis.-Bilaterally symmetrical fer- tile appendage (megasporophyll) with broad mid- rib; several pairs of apparently sessile, closely ap- pressed ovules borne in two lateral rows on surface of reduced basal part of lamina; ovules closely in- vested by lamina, each with a small circular sear of attachment to lamina near midrib. Distal part of ap- pendage expanded as flattened sterile foliar blade. Archaeocycas white Mamay Figure 3; plate 1; plate 2, figures 5-7 Archaeocycas white Mamay, 1973, p. 687, fig. 1, a-c. Specific diagnosis.-Megasporophylls cuneiform, ovate, or lanceolate, 1.7-2.3 em long, with greatest widths of 1.0 em. Midribs straight, flat, faintly marked with closely set parallel striations; midrib widths uniform or increasing distally from 1.0-1.3 mm at the bases to 3.0 mm at distal termini of fer- tile areas; midribs flaring out conspicuously or be- coming obscure in distal, laminar parts of mega- sporophylls ; sides of midribs more or less concavely scalloped through appression with ovule bases. Proximal fertile area 1.2-1.6 ecm long, 3.0-5.0 mm wide at base, nearly uniform in width or increasing distally to width of 9.0 mm. Ovules produced in op- posite pairs, four to six ovules in each row. Ovules oblong, oblanceolate, rhombic or rhomboidal, tightly appressed with each other, obliquely directed for- ward at broad angles to the rachis. Basalmost ovules proportionately shorter, broader than distal ones; ovules 1.5-3 mm long (measured along lines of ap- pression between contiguous ovules), 1.5-3.0 mm wide (measured along outer free margins and along lines of appression with the rachis), reaching 5 mm in greatest (diagonal) dimension. Ovules each with a small circular shallow depression or attachment scar, 0.7-1.0 mm in diameter, uniformly placed on the upper surfaces of ovules near the center of area of ovular appression with the midrib. Upper (adaxi- al) surfaces of ovules closely invested by foliar lamina; extent of lamina over lower (abaxial) ovu- lar surfaces unknown ; laminar investment of ovules showing fine closely spaced parallel striations per- pendicular to midrib. Megasporophylls abruptly modified into sterile laminar structures immediately distal to distal pair of ovules, and contiguous with ovules; extent, vena- tion, and margination of distal laminae obscure. Holotype.-USGS 8877-1 (pl. 1, fig. 5). Paratypes.-USGS 8877-2, 3, 4 (pl. 1, fig. 8, 4; pl. 2, fig. 7) ; USGS 8959-1, 2 (pl. 1, fig. 1, 2; pl. 2, fig. 5, 6). Occurrences.-USGS paleobotany localities 8877, 8959. DESCRIPTION This description is based on eight specimens, six of which were collected by White at the Fulda locali- ty (USGS paleobotany locality 8877) in 1910. Al- though I could not find White's locality, I was fortu- nate in discovering two additional specimens of Archaeocycas at the "Emily Irish" locality (USGS paleobotany locality 8959) in 1963; the two associ- ated floras are substantially the same and presum- ably occupy the same stratigraphic position within the Belle Plains Formation. Although White never published an account of his specimens, it is obvious that he recognized them as unique or important, be- cause I found them carefully trimmed, encircled in crayon marks, and segregated from the remainder of his Fulda collection in a separate tray. The specimens are preserved as compressions with a moderate amount of relief that aids somewhat in their interpretation. Only small amounts of car- bonaceous residue remain, but enough of this is present to provide some critical information. Three of the Fulda specimens are casts (pl. 1, figs. 4, 5; pl. 2, fig. 7), two are molds (pl. 1, figs. 1, 3), and the remaining one is available as both counterparts. SYSTEMATIC DESCRIPTIONS 9 Each of the two "Emily Irish" specimens consists of both counterparts; these are shown in plate 1, fig- ures 1 and 2, and plate 2, figures 5 and 6. Unfortu- nately no cuticularized material is preserved ; matrix samples from both localities were processed for paly- nological analyses, with negative results. The specimens are each characterized by a fertile basal area, which is neatly and symmetrically "com- partmented" (pl. 1, figs. 4, 5; pl. 2, fig. 7), and a distal, usually poorly preserved laminar extension. A superficial examination of this material might prompt the conclusion that the paired inflated ob- jects borne along the basal extent of the midrib are vegetative pinnules, but close inspection necessitates another interpretation. If these were ordinary pin- nules rather than casts and molds of expanded three- dimensional bodies, the abrupt transition from a basal area of distinct pinnulation to the rather vague, poorly defined distal lamination would con- trast markedly with the conventional type of pin- nate foliar appendage. Again, if these were ordinary pinnules, one might expect some overlapping or other irregularities of attitude such as is found ordi- narily in pinnate plant fossils. Instead, these bodies show no sign of overlapping but present the appear- ance of firmly attached solid objects of considerable depth; they abutted laterally against each other, with the result that contiguous sides were noticeably flattened (pl. 1, figs. 4, 5; pl. 2, fig. 7). In the ab- sence of compelling alternative evidence, these ob- jects are regarded as the fillings of ovules, probably buried in an immature stage of development; conse- quently, the term "ovule" is applied herewith. The casts are the most informative of the avail- able specimens. Although most of the casts are lack- ing in carbonaceous residues, most of which adhered to the molds when the specimens were split, all the casts show remnants of a substantial thickness of coalified organic material separating contiguous ovules. These films, which average about 0.15 mm in thickness, are clearly seen on plate 1, figures 4 and 5, and plate 2, figures 6 and 7. The holotype (pl. 1, fig. 5) is particularly interesting in this respect. When tilted and viewed edgewise, the ovules are seen to reach thicknesses of 1.0 mm or slightly more, and the presence of coaly films beneath the edges of the ovules indicates that the ovules are actually the sedi- mentary casts or mud fillings of hollow bodies that were of sufficient substance to resist compaction. Most of the specimens were deposited with their flat surfaces normal to the surface of the enclosing sedi- ment; the specimen shown in plate 2, figure 7, how- ever, was deposited edgewise on the bedding surface and therefore gives some idea of the thickness of the ovules, which are seen nearly end-on in a side view of the sporophyll. The coaly films separating the ovules from each other are interpreted as the organic residue of part of the investing lamina, which originally was deeply creased so as to intrude downward between the ovules. When the sporophylls were compressed, the foliar tissue was "pinched" between the flattening and laterally expanding ovules and is now preserved as a thin coalified film. The remainder of the foliar residue adhered to the counterparts, or molds; there the grooves between ovules lack any coaly residue. This distribution of organic residue between counter- parts of a given specimen is well demonstrated on plate 2, figures 5 and 6. The scalloped outlines of the midribs are also de- limited by a thin coaly film that separates the ovules from the midribs (pl. 1, figs. 4, 5; pl. 2, fig. 6). Tilt- ing and edgewise observation of these specimens show that the coaly film continues downward and toward the median axis of the midrib. Thus the mid- rib seems to have been compressed into a flangelike form, partly covering the upper surface of the foliar lamina in the fertile region of the sporophyll. Com- pression of the midrib probably caused an entrap- ment of foliar tissue between the rachis and ovules, similar to that between the ovules; the result is a gently curved longitudinal continuity of the coaly films separating contiguous ovules (pl. 2, figs. 5, 6). These features favor the interpretation of these basal objects as originally distended tightly ap- pressed bodies rather than flattened vegetative pin- nules. They are emphasized in order to justify the diagnosis of the basal objects as ovules. Unfortu- nately, none of the specimens are exposed in bottom (abaxial) view, so that details of the opposite sides of the sporophylls and ovules are not known. The ovules are arranged in opposite pairs. The specimen shown on plate 2, figures 5 and 6, has only four pairs, the smallest number of ovules observed. Some specimens have five pairs each (pl. 1, figs. 3- 5), and others have six, the largest number observed (pl. 1, figs. 1, 2; pl. 2, fig. 7). The shape of the ovules varies somewhat, those at the basal end of the sporo- phyll being proportionately the longest in the direc- tion parallel to the axis of the sporophyll ; these vari- ations in size and shape of ovules are shown on plate 1, figure 5. The overall length of the fertile areas of the sporophylls ranges from about 1.2 cm in the smallest specimen (pl. 2, fig. 5) to 1.6 em in the largest (pl. 2, fig. 7). The aggregate of ovules on a given sporophyll may be of an almost uniform width 10 PALEOZOIC ORIGIN OF THE CYCADS (pl. 1, fig. 3) or may show a wedgelike shape, the width increasing in the distal direction (pl. 1, figs. 4, 5). Many of the ovules have a single small circular depression or scar; these range from 0.7 to 1.0 mm in diameter. Each of the eight ovules shown on plate 2, figures 5 and 6, has one of these marks; these are most clearly shown on the mold (pl. 2, fig. 5). The uppermost pair of ovules on plate 1, figure 4, shows these marks as distinct, fairly deep pits. Two others are seen on plate 2, figure 7, where they are partly filled with a carbonized residue. The scars are uni- formly placed, next to the midrib at approximately the middle of the ovule. The sears on the specimen shown on plate 2, figure 5, are of additional morpho- logical interest because each one has a centrally lo- cated circular sear, approximately one-third of its own diameter. Because of their occurrence, one to an ovule, these marks are interpreted as remnants of points of at- tachment to the lamina. There is no evidence of a stalk, so the ovules are regarded to have been sessile, with a small point of attachment; the subsidiary marks on the scars shown on plate 2, figure 5, may represent points of entry of vascular strands into the bases of the ovules. The position of these marks, relative to the lamina and rachis, is essentially iden- tical with the position of seed attachment in Spermopteris, as reconstructed by Cridland and Mor- ris (1960, fig. 10). Evidence of a laminar element in the ovuliferous area of the sporophyll is variously shown in the suite of specimens but is best seen on the specimens illustrated on plate 1, figures 1 and 2. These are counterparts of the same sporophyll ; figure 1 shows the mold, figure 2, the cast. The mold has retained much of the coalified resi- due of the specimen, and many of its details are relatively obscure. However, the right half of the photograph shows not only the grooves separating the ovular impressions, but also a series of closely spaced parallel ridges and grooves, oriented at a broad angle decurrently to the midrib. These are in- dicative of a system of fine foliar venation. The cast shows an important additional detail, indicated by the arrow in plate 1, figure 2. In this specimen, the grooves between contiguous ovules are clearly shown, but the most distal groove is incom- plete; that is, it does not extend entirely to the mar- gin of the lamina, as do all the others. This is evi- dently a function of the fact that this particular groove is beyond the most distal ovule, and the ab- sence of the pressure of a contiguous ovule permitted the marginal part of the involute lamina to flatten out and merge with the distal sterile lamina. This demonstrates a continuity of lamina between fertile and sterile parts of the sporophyll. On the basis of the foregoing, it is well established that the ovules of Archaeocycas were borne on a laminar blade, which seems to have at least partly enclosed the ovules. The evidence bearing on extent of the enclosure is incomplete, however. All but one of the specimens show remnants of a distal foliar lamina immediately above the upper- most pair of ovules. The longest lamina is shown on plate 2, figure 7, where it extends about 6 mm beyond the last ovule, and constitutes nearly one-third of the total length of the specimen. The complete extent of the lamina or features of its margination are not known. As shown on plate 1, figures 3-5, and plate 2, figures 5-7, it forms as a distal continuation of the midrib, which normally flares out rather abruptly; the striations of the midrib also diverge to present a vague suggestion of a fanlike system of fine veins. The specimen shown on plate 2, figures 5 and 6, differs from the others in that the strong midrib essentially retains its identity as far as its broken distal end, and the lamina evidently decreases in width distally, instead of expanding as all the others do. The lamina shows faint striations, decurrently directed away from both the midrib and the ovules, so that the orientation of the photograph seems to be correct ; that is, the morphologically proximal end of the specimen is directed downward on the plate. This conclusion is corroborated by comparisons of the shapes and decurrent attitudes of the ovules of this specimen with all the other specimens. The acute distal tapering of the lamina seems anomalous when compared with the other specimens; this, how- ever is not an accident incurred when the matrix was split, because the laminar margin along the up- per left of the photograph is entire and intact. Thus, in this particular specimen the distal laminar exten- sion was significantly smaller than those seen in the other specimens ; and perhaps this one is nearly com- plete except for its very tip. Another variation in the form of the distal lamina is seen in the holotype (pl. 1, fig. 5). In this speci- men, the distal lamina is very broad in proportion to the widest part of the fertile area. This is indicated by a conspicuous folding of the lamina, which is seen in the photograph to begin in the middle of the mid- rib at a level approximately even with the upper- most pair of ovules. The resultant crease proceeds obliquely toward the upper right of the photograph, meanwhile deepening and continuing as far as the SYSTEMATIC DESCRIPTIONS 11 broken distal edge of the lamina. This folding or distortion had no noticeable effect on the symmetrical arrangement of the two rows of ovules below; the intactness of the symmetry may be interpreted to mean that the ovuliferous part of the appendage was quite rigid and capable of resisting distortion. The absence of the lowermost ovule on the right side of this specimen is due to breakage during splitting of the matrix, for when the specimen is tilted at a high angle and viewed from the side, remnants of the broken base of the missing ovule are evident. There is no extensive information regarding the manner of attachment of Archaeocycas. A minor bit of evidence, however, is seen at the basal end of the specimen shown on plate 2, figure 7. At the left end of the specimen, the broken end of the midrib in- creases abruptly in thickness, and remnants extend toward both the top and bottom of the figure. This indicates that the base of the sporophyll occurs im- mediately next to the proximal pair of ovules, and the sporophyll therefore had no extended stalk. It also suggests connection to a supporting axis whose orientation would parallel the left edge of the plate. In consideration of the fact that this particular specimen was laterally compressed in preservation, the basal enlargement of the rachis might be taken to indicate that the parent axis was perpendicular not only to the direction of the midrib but to the natural plane of the lamina as well. Thus the Archaeocycas specimen represents a lateral append- age equivalent to a leaf, rather than a subdivision of a compound leaf. A partly diagrammatic restoration of a sporophyll of Archaeocycas is presented in figure 3. DISCUSSION Cycadean affinity of Archaeocycas.-Morphologi- cal interpretations and taxonomic assignment of Archaeocycas revolve about consideration of the fol- lowing unique combination of structural features, which, insofar as I know, occurs in no other Paleo- zoic plant and validates the status of Archaeocycas as a new genus : 1. It is a bilaterally symmetrical flattened leaflike appendage. FIGURE 3.-Reconstruction of megasporophyll of Archaeocycas whitei, showing upper (?adaxial) surface. Fertile bagal part of megasporophyll rendered transparent to show ovules (broken lines); darkened circular areas represent points of attachment of ovules to lamina. Shape and extent of sterile distal lamina hypothetical. For view of lower sur- face of megasporophyll see figure 7A. 12 PALEOZOIC ORIGIN OF THE CYCADS 2. It is clearly differentiated into two areas, the dis- tal one which is sterile, and the basal one which is fertile. The basal or fertile part of the lam- ina is somewhat reduced, and on its lower sur- face it bears two lateral rows of large paired conspicuous bodies that are interpreted as ovules; the fertile lamina seems to have in- vested the ovules tightly and was involute so as to at least partly enclose the lower surfaces of the ovules. A simple small circular sear occu- pies the same relative position on each ovule. The small circular sears on the fertile laminae are difficult to interpret in terms other than sears of at- tachment; likewise, the large size of the bodies attached thereto strongly suggests ovules. Addition- al evidence in support of their ovular nature is found in their resemblance to the seeds of Spermopteris and the manner in which the bodies were borne. The crowded closely appressed arrangement and shapes of the ovules of Archaeocycas are probably a func- tion of immaturity at the time of preservation, and a function of the way in which the ovules were en- closed by the foliar lamina. When the plant matured, the fertile base of the sporophyll probably elongated rapidly to accommodate the increasing size of the ovules. The fertile lamina may have ceased to grow as the rachis and ovules enlarged, with the result that only inconspicuous distorted remnants of the fertile lamina remained to maintain the attachment of ovules to sporophyll. Thus, mature sporophylls, if found fossilized, could be difficult to recognize as later ontogenetic stages of the material upon which this description of Archaeocycas is based. Phyllospermy-the condition in which seeds were borne directly upon more or less unmodified leaves (Sahni, 1920, p. 300) -is a phenomenon that became well established during late Paleozoic time. Several examples appear repeatedly in the literature as the result of important work by Halle (1929). Spermop- teris is a clear example of phyllospermy ; the Texas Permian has recently produced two distinctive but probably unrelated forms (Padgettia Mamay, 1960; Tinsleya Mamay, 1966), and now Archaeocycas may be added to the growing list. Archaeocycas stands apart from all other fossils of this general nature in having its ovules restricted to the basal part of the sporophyll. From the standpoint of symmetry of ovular insertion, striking comparisons may be drawn between Archaeocycas and Spermopteris, but the latter genus shows no evidence of sharp segregation of the reproductive function to the basal part of the sporophyll. The trend toward reduction of the basal part of the lamina, as demonstrated in Archaeocycas, is another important point of distinction between Archaeocycas and Spermopteris. A continuation of that trend obviously could result in a megasporophyll much like that of Cyeas, and for this reason I regard Archaeocycas as an extremely significant stage in the evolution of the cycads. In recognition of these Permian fossils as representatives of a precursive cycadean stock, I have therefore applied the generic name Archaeocycas. The species A. white? is named for David White, who found most of the specimens on which the description is based. Genus PHASMATOCYCAS Mamay New genus R. Mamay, 19692. Phasmatocycas Mamay, 1973. Type species.-Phasmatocycas kansana Mamay. Generic diagnosis.-Fertile axis with two lateral rows of sessile, broadly attached gymnospermous cvules; ovules simple, with blunt, funnel-shaped micropyles, two distinct cuticular layers, and a thick megaspore membrane; laminar tissue unknown, in- terseminal appendages lacking. Phasmatocycas kansana Mamay Figure 4; plates 3, 4 Phasmatocycas kansana Mamay, 1973, p. 689, fig. 1, d-g. Specific diagnosis.-Axial fragment 2.5 em long, relatively stout, 2.5-4.0 mm wide, rather abruptly broader at (?) basal end of specimen ; surface of axis with obscure narrow longitudinal striations, lacking other characteristic markings. Bilateral arrangement of ovules symmetrical; no evidence of ovules between lateral rows. Ovules in- serted alternately to suboppositely, extending essen- tially perpendicularly to axis, closely arranged as to juxtapose against but not overlap each other. Ovules 2.8-4.0 mm long, 2.0-2.6 mm wide, ovoid to oblong, broadly attached with point of attachment to axis nearly as long as greatest width of ovule; ovular apices blunt, some specimens with a shallow apical cleft or notch. Cuticular envelopes slightly smaller than impres- sions of ovules, lacking ornamentation; cuticles ap- parently complete except for micropylar openings; apices of cuticles gradually decreasing in diameter, without marked constriction, terminating in micro- pylar openings 0.2-0.5 mm wide ; specialized chalazal features absent. Outer cuticle very thin, diaphanous, with elongate cells reaching 20 by 80 um (microme- tres) in dimensions, parallel to long axis of ovule. Inner cuticle apparently thicker and separate from outer; cells thick walled, walls abruptly increasing SYSTEMATIC DESCRIPTIONS 13 in thickness toward the micropylar area; cells in micropylar area essentially isodiametric, averaging 10 um in diameter. Megaspore membrane essentially filling cavity within cuticles; membrane dense, lack- ing ornamentation. Small (0.25-0.33 mm in diameter) spherical ses- sile bodies with resinlike luster regularly alternating with ovules in each row; one such body partly im- bedded in axial tissue at point between and approxi- mately level with bases of two adjacent ovules. Holotype.-USGS 8869-1 (pls. 3, 4). Occurrence.-USGS paleobotany locality 8869. DESCRIPTION Phasmatocycas kansana is based on one specimen from the Banner, Kans., locality. Fortunately, both counterparts of the specimen are available, and al- though only a short fragment of the axis was found, it is sufficiently well preserved that some important details are observable. The specimen is preserved in a matrix of very hard dark-gray limestone of poor fissibility and apparently not very rich in plant material. Both counterparts of the specimen are shown on plate 3 (figs. 1, 2). As the axis is broken at both ends, it is impossible to ascertain what the total length of the fertile structure was, but it may have been considerably more than the 2.5 ecm preserved here. The axis itself lacks any conspicuous morphologi- cal features, and only the fact that one end of the fragment is somewhat broader than the other gives one any basis for differentiating between the distal and proximal ends; I assume that the wider end (4.0 mm) is proximal and the narrow end (2.5 mm) is distal. Parts of each of the axial counterparts carry flecks of a fairly thick carbonaceous residue, but no traces of a cuticle are preserved. Faint, closely spaced longitudinal striations ap- pear in a few small areas on the axial surface; these are distributed over nearly the entire width of the axis and probably are best interpreted as striations caused by wrinkling of the original surface or as impressions of the vascular system of the axis. There are no traces of a foliar lamina between or projecting beyond the apices of the attached seeds, and, in con- sideration of the generally good preservation of the specimens, it may be concluded that this fragment is indeed lacking in any sort of foliar lamina. Indirect evidence regarding foliation of Phasmatocycas is presented on subsequent pages. The most impressive feature of Phasmatocycas is its two lateral rows of seeds, the near-perfect aline- ment of which attests to the bilateral morphology of the appendage and precludes any interpretation of the seed insertion as a spiral system. Further- more, the axial surface between the two rows of seeds shows no evidence whatsoever of sears of seed attachment that might provide a basis for challeng- ing a bilateral interpretation for the seed arrange- ment of Phasmatocycas. The seeds are mostly preserved as flattened molds, each of which contains a double seed cuticle and megaspore membrane on one or the other of its counterparts. The outlines of the molds are usually slightly wider and longer than the contained cuticles. This preservational feature is well shown in the de- tached seed at the left of figure 1, plate 3; here the mold is about 2.5 mm wide and the enclosed cuticle is slightly less than 2.0 mm wide. The consistent presence of these flattened, slightly hollow seed cavi- ties in the matrix surrounding the completely flat- tened cuticles suggests that the outermost layers of the seeds were fleshy integumental layers that were not adapted to fossilization. These integuments did resist compaction to a certain extent, however, and resulted in production of the slight cavities in the limestone, left there as molds when the integumental tissues disintegrated. Flakes of fairly thick coalified residue on the surfaces of some of the cuticles attest to the original presence of such an integumental layer. In general proportions, most of the seed molds are less than twice as long as they are broad ; their broad bases and blunt apices thus lend them a nearly ob- long shape in most instances. Most of the seed bases are only slightly narrower than the widest parts of the seeds, and the actual point of attachment may even slightly exceed the width of the seed. The broadly sessile nature of the ovules is best seen on plate 3, figure 4. The same figure shows another in- teresting, although perhaps not especially significant feature of the seed molds-a shallowly bifid or notched apex that appears in several of the speci- mens. These notches range form 0.2 to 0.3 mm in depth and tend to give one the initial impression that these are the molds of platyspermic seeds, because several previously described genera of supposedly platyspermic seeds have apical notches. Cridland and Morris (1960) described and illustrated similar apical notches in the seeds of Spermopteris. They speculated (Cridland and Morris, 1960, p. 857) that the apical cleft "may have been caused by fracture of the micropyle during compression or may be a natural feature of the seeds." In the case of Phasma- tocycas, I prefer to interpret the apical clefts as a 14 PALEOZOIC ORIGIN OF THE CYCADS mechanical result of compression rather than a natural feature of the seeds. This seems to be a pref- erable interpretation because no consistent feature of the cuticles indicates a cleft micropyle in Phasmatocycas. Associated with the axis are 21 seeds; one is de- tached but the others are preserved in their original positions of attachment. It is not possible to ascer- tain whether or not they were arranged in morpho- logically opposite pairs, because some are subop- posite or slightly alternate to each other, perhaps as a result of slight preservational distortion of posi- tion. Nearly all the seeds, however, extend outward from the axis at right angles (pl. 3, figs. 1, 2) and give the impression that they were firmly attached. They are so close to each other that their sides are in physical contact with each other, but there is little overlapping. The seeds are essentially the same size and in the same condition of preservation from one end of the specimen to the other. Thus it would seem that all the ovules matured simultaneously. All the seed cuticles are, in part, only loosely ad- herent to the limestone matrix. Consequently it was possible to remove three nearly complete cuticles and parts of two others. These were cleared of coalified remnants of integumental tissues and other humic components by the standard palynological clearing technique, using a mild Schultze's solution followed by washing and treatment with potassium hydrox- ide. This method cleared the specimens sufficiently for microscopic examination. The cuticles were then mounted on glass slides in a synthetic resin mount- ing medium. Some of the specimens are shown on plate 4. The evidence for an ovulate nature of the inflated bodies in Archaeocycas is partly inferential, but Phasmatocycas has genuine gymnospermous seeds. The description of the cuticles on the follow- ing pages leaves no room for doubt. The best cuticular specimen, shown on plate 4, figure 3, is nearly complete except for the absence of a large patch of material within the almost-intact margin of the lower half of the seed. This specimen is 3.2 mm long and 1.8 mm wide. The broken speci- men shown on plate 4, figure 4, is 3.3 mm long, and in its entirety it must have been at least 2.2 mm wide. Another nearly complete specimen, not illus- trated, is 2.0 mm wide, but its length is not known. This one is the broadest of the cuticular specimens removed from the matrix; other larger specimens are still attached to the axis. As shown on plate 4, figure 3, the megaspore mem- brane is a heavily cutinized thick envelope. It is vari- ously folded and wrinkled, as indicated by the dark stripes where two or more thicknesses of the mem- brane are involved. Thus, the light-colored patch below the middle of the seed shown on plate 4, figure 3, represents a single layer of megaspore membrane; the darker area above it consists of a double layer, and the still darker stripes represent folded areas of four layers. Although the chalazal outlines of the megaspore membranes are quite plain, their apical contours are difficult to differentiate because of op- tical interference by the thickened surrounding cuti- cles in the micropylar areas of the seeds. On plate 4, figures 4 and 5, however, the dim outline of the apex of the megaspore is seen passing over the upper part of the seed approximately 0.5 mm below the seed apex, or the area that presumably constituted the bottom of the pollen chamber. Although the mega- spores are heavily cutinized membranes, they do not have any conspicuous or characteristic ornamentation. It is clear from the apical areas of the seeds that two distinct cuticles are present. This is shown in all of the photographs on plate 4. The cuticles are difficult to differentiate through the entire length of the cleared specimens, however, and they may not have been sufficiently developed in the basal parts of the seeds to be preserved. In the specimen shown on plate 4, figure 3, the faint outlines of cuticular cells are observable down to about the midsection of the seed ; in that area, only one layer is visible, however, and the differentiation into two layers is not ap- parent below approximately the apical 0.75 mm of the seed. The specimen shown on plate 4, figure 4, seems to be the best preserved from this standpoint, for cellular outlines of the isodiametric type that characterize the inner cuticular layer are visible proximally about as far as the median part of the seed. At about the level of the "shoulders" of the seeds, or 0.7 to 0.8 mm below the seed apices, the inner cuticle thickens abruptly, whereas the outer cuticle remains very thin and barely visible in many places. Apparently the outer cuticle is separate from the -\ inner; the distinction between the two layers is shown on plate 4, figures 1 and 5. Most cells of the outer cuticle are elongate, reaching lengths of 50-80 pm, and have widths of one-fourth or less of the length ; these cells are oriented with their long axes parallel to the length of the seed. The isodiametric cells of the inner cuticle average about 10-15 um in any direction ; their walls are very thick, particularly in the area immediately above the distal surface of the megaspore membrane. As shown on plate 4, fig- ures 1 and 5, this wall thickening decreases abruptly in the terminal 0.2-0.3 mm of the micropyle. SYSTEMATIC DESCRIPTIONS 15 The micropylar narrowing of the cuticular enve- lope is shown on plate 4. All the illustrated specimens show a gentle tapering of the apical "funnel." The actual openings of the micropylar canals range from 0.2 to 0.5 mm in width. The tips of the micropyles shown on plate 4, figures 1, 2, and 3, are slightly damaged, but the one shown on figures 4 and 5 is essentially intact and shows no evidence of a bifid tip. If any specialized pollen-receiving structures were within the seed apices, no evidence of such is preserved in the fossils. The present material per- mits only the assumption that the seeds of Phasma- tocycas had a simple pollen chamber, accessible through a short, evidently unspecialized micropylar canal, which was delimited by the thick inner cuti- cle. The cleared specimens show no evidence of inter- nal vasculature or pollen grains within the micro- pylar areas. Interpretation of the morphological nature of the two cuticles of Phasmatocycas is aided by reference again to two of the preservational features of these seeds. First, the presence of flat cavities or seed molds in the matrix, somewhat larger than the cutic- ular outlines, indicates that a thick integument originally surrounded the seeds. This deduction is supported by the second pertinent feature, the pres- ence of thick flecks of coalified organic residue on the outer surfaces of the cuticles. Thus, the absence of any trace of an outer integumental cuticle, which should have directly lined the seed molds, indicates that the thin outer cuticles of the cleared specimens represent the inner cuticular lining of the integu- ment. Accordingly, the thickened inner cuticles are interpreted as the surface of the nucellus. These interpretations are supported to some ex- tent by Harris' (1941) study of the seeds of the Jurassic cycead Beania. Attention is directed to a comparison of the cleared specimens of Phasmato- cyeas (pl. 4) with Harris' (1941, pl. 5, fig. 14) preparation of the inner cuticles of a seed of Beania gracilis. In both instances, very similar sets of thick- ened membranes are preserved. The Beania contains a thick megaspore membrane surrounded by a rela- tively thick cuticle (nucellar), which in turn is en- closed by a thin inner integumental cuticle. The "apical spike" of Harris' Beania is not present in Phasmatocycas, nor does the nucellar cuticle of the Beania flare upward into the conelike projection present in Phasmatocycas. These differences in de- tails of apical morphology of the two genera of seeds reduce the exactness of comparison, but the similarities point up the fact that the Phasmatocycas seeds are similar to those of the Jurassic cycads. The excellent preservation of seed cuticles sug- gested the possibility that a palynological examina- tion of the matrix might corroborate a cycadalean interpretation of Phasmatocycas. Chips of the lime- stone were submitted to R. A. Scott who reported (written commun., 1968) as follows: The search for palynological evidence for the presence of cycadophytes was unsuccessful. Only a very few apparently monosulcate grains were seen; these were considered as un- reliable because they 1) could be folded fern spores or 2) resembled the pollen of gingkophytes more closely than that of cyeadophytes. Because of poor preservation, these results do not in any way preclude the presence of cyeadophyte pollen; they sim- ply fail to establish it. One other feature of Phasmatocycas invites close attention. Both counterparts of the holotype bear either the impressions or the actual substance of a number of tiny, almost perfectly spherical bodies, which measure 0.25-0.33 mm in diameter. Where the actual spherule is present on one counterpart, the impression is usually visible on the opposite counter- part. These bodies do not seem to have been com- pressed or distorted from a spherical shape during the process of preservation. The spherules are dark brown to black, and although most have lustrous sur- trous surfaces, faint rugosity is evident on some specimens. Some are translucent and reddish brown suggestive of a resinoid composition. One might first be tempted to interpret these spherules as a fortuitous concentration of foreign objects not related to Phasmatocycas. However, their distribution indicates that they are not all randomly placed, which necessitates the conclusion that the spherules are a natural part of the fertile axis. A few of the spherules are present on the axis in random positions, whereas a few others are in the matrix a short distance away from the specimen. The most significant aspect of their distribution, however, is the fact that many of the spherules occur on the axis at the bases of the seeds, and most of these occupy positions exactly between two adjacent seeds ; none occur directly on the seeds. This regular arrangement is shown on plate 3, figures 3, 4, and 5. Plate 3, figure 3, shows this feature particularly well. Such a positional relationship between the seeds and spherules can hardly be accidental. Rather, it indicates that the spherules are a natural feature of Phasmatocycas and, inasmuch as they are inter- calated between adjacent seeds, they should be ap- proximately equal in number to the 21 seeds found on the type specimen. By actual count, the specimen contains the remains or impressions of 24 spherules. 16 PALEOZOIC ORIGIN OF THE CYCADS The slight disparity between numbers of spherules and seeds may reflect an originally weak attachment of spherules to the axis, some having been displaced from their natural positions and shifted to other locations during burial of the specimen. If one accepts the distributional pattern of the spherules as evidence that they were originally at- tached to the axis of Phasmatocycas in two rows like the rows of seeds, one spherule placed between each two successive seeds, one then faces the problem of interpretation of these objects. Because of their uniformity in size, shape, and distribution, the spherules are best interpreted as the remnants of regularly placed glands; the spher- ules represent the coagulated acellular contents. These hardened before burial and retained their naturally spherical shapes during induration of the matrix. Regardless of their functions, the glands constitute a point of comparison with associated leaf remains; this comparison warrants further elabora- tion because of its inferential bearing on the morphological identity and systematic placement of Phasmatocycas. DISCUSSION Evidence for a megasporophyllar interpretation of Phasmatocycas.-Phasmatocycas kansana shows no direct evidence of attachment to a foliar lamina. Because of the strikingly cycadlike nature of this fertile organ, however, it is well to examine closely whatever evidence is available in the direction of es- tablishing or refuting a megasporophyllar character for Phasmatocycas. Certain compelling evidence is indeed present in the form of associated taeniopterid leaves bearing resinoid spherules that seem identical with those on the ovuliferous axis of Phasmatocycas. The small assemblage of plants associated with the holotype of Phasmatocycas kansana contains seven fragments of narrow taeniopterid leaves, most of which are poorly preserved ; the largest specimen is about 10 em long. All these specimens bear either the impressions or the actual substance of small spher- ules, which compare closely in size with those of the fruiting axis. Their preservation is mostly inferior to preservation of the glandular remains on Phas- matocycas, but in two of the leaves, the spherules are identical with those of the fertile specimen in color and surface features. Distribution of the spherules or glands varies; in some of the leaves, the glands are more or less evenly scattered over the surfaces, but in one specimen they are concentrated near the margins of the lamina. Only a few of the glands ap- pear directly on the midribs. Another small collection from a nearby site also has taeniopterid material containing resinoid spher- ules. This collection (USGS paleobotany locality 8298) was made "314 mi. SE of Elmo" ; the sources of this material and of the holotype of Phasmatocy- cas (USGS paleobotany locality 8869; "814 mi. 8. of Banner") are probably the same outcrop, because Dunbar (1924, p. 174) pointed out that Elmo was "at one time called Banner City." Lithologic compari- sons suggest that the two collections were made from the same rock unit. This collection contains nine fragments of taeniop- terid foliage, which is not distinguishable from the leaves from locality 8869. Most of these leaves are poorly preserved, but some show evidence of at least a few of the spherical bodies. One of these specimens, by contrast, is quite well preserved and is remark- able because of the large number of resinoid glandu- lar remains or hemispherical impressions borne on its surface. This specimen is shown on plate 5 (figs. 5-7). Plate 5, figure 5, shows the entire specimen, which is 6.5 em long and 1.4 ecm wide at its broadest point. The basal part of the specimen is missing, but the fragment clearly shows a stout midrib and simple typically taeniopterid venation. Plate 5, figure 6, a higher magnification, shows the many small spherical bodies, more or less evenly distributed over the entire lamina. A small part of the lamina is shown at 20 magnifications on plate 5, figure 7. The glands average about 0.25 mm in diame- ter and are so abundant that most of them are with- in 0.5 mm of another. They are restricted to the nar- row laminar areas between the closely set veins, and the midrib itself is devoid of glands. The spherules are easily removed from the matrix, and many of them were detached, presumably when the specimen was collected. The positions of missing spherules, however, are clearly indicated by hemispherical de- pressions in the intercoastal areas. Most of the glandular bodies are virtually uncompressed, and in their general physical characteristics they are re- markably similar to those attached to Phasmatocy- cas kansana. Excellent examples of the glandular remains also appear on taeniopterid foliage contained in USGS collection 8868, from a locality near Carleton, Kans. Attention was originally directed to spherical bod- ies on taeniopterid foliage by Sellards (1901, 1908). In 1901, Sellards described some Permian taeniop- terid material from the Banner City locality; this included specimens identified as Taenmiopteris cori- acea Goeppert, T. coriacea var. linearis, Sellards SYSTEMATIC DESCRIPTIONS 17 var., T. new Fontaine and I. G. White, and T. sp. Much of that paper was devoted to description of '"interneural bodies" and "scars on or near the rachis" on the foliage. The scars were compared with insect damage and fungal remains, but no definite interpretations were adopted. The "interneural bodies'" were described (Sellards, 1901, p. 5) as being "small, oval, resistant bodies, situated between the veins, half immersed in the epi- dermis of the frond, nearly globular in shape, some smooth on top, but more often showing a slit across the top." To judge from Sellards' illustrations, those bodies are identical with the spherules described here, with the exception of the distal slits ; I have not observed such slits in the material at my disposal. Sellards briefly compared these bodies with dots on other genera from other localities. In attempting to analyze them in terms of morphology and function, he appeared to favor interpretation of the bodies as sporangia or a new type of fructification ; in his sub- sequent paper (Sellards, 1908, p. 447), however, he regarded the question of function of the objects as an equivocal matter. Cridland and Morris (1960, p. 859) commented that the bodies "have never been shown to be anything but globules of iron oxide." Much of Sellards' paper of 1901 was incorporated without alteration in his subsequent summary article on the late Paleozoic plants of Kansas (Sellards, 1908). Although no new or supplemental informa- tion regarding the taeniopterid complex or the "in- terneural bodies" was presented, that paper provided important information regarding the biological sig- nificance of the spheres. Sellards described 26 species (among nine genera) of plants represented by foliar organs, from the Banner City locality. "Interneural bodies" were discussed only in connection with the taeniopterids in the Banner City flora; comparable objects were not noted on any of the associated taxa. This consistency of association is borne out by my observations. The genera Callipteris, Glenopteris, and Odontopteris are present in the material I have examined from the vicinity of Banner City, but these show no evidence of spherical bodies. From the foregoing discussion, one can scarcely avoid the deduction that the spherical objects are a natural biological attribute of the taeniopterid leaves from the Elmo Limestone Member of Dunbar (1924) at or near the Banner City plant locality. Their con- sistency of size, shape, and placement on the lamina between veins militates against their origin as acci- dents of mineralization or organic degradation. They are most likely glandular in origin, and their re- stricted occurrences on just one foliar form genus in a rather diverse generic assemblage demand serutiny of possible relationships between these leaves and any associated plant remains on which similar glandular bodies occur. The resemblances between the glands of Phas- matocycas kansana and those of the associated taeni- opterid leaves lead to examination of the obvious inference, that is, that the fruiting axis and the leaves are parts of the same species of plant, if not the same morphological unit. Several circumstances add support to this inference; these and previously noted evidence are summarized here: 1. The glands are not found on any of several as- sociated genera, although many of the associ- ated taeniopterid leaf specimens show evidence of their presence. 2. Neither the occurrence nor distribution of the glands on Phasmatocycas and the associated taeniopterid leaves is reasonably ascribable to accident; these objects are doubtless a bio- logically controlled phenomenon. 3. The cycadlike fundamental morphology of Phas- matocycas requires only a terminal foliar part to complete the qualitative morphological re- quirements of a cycadean megasporophyll. 4. A seed-bearing taeniopterid leaf (Spermopteris Cridland and Morris) is already known; its geographic and stratigraphic occurrences and morphologic details make it a logical predeces- sor for a Phasmatocycas-like appendage with terminal taeniopterid foliation. 5. The putative Triassic cyceads Dioonitocarpidium Lilienstern and Bjwvia Florin both have been supposed to have had taeniopterid foliage. Viewed as an aggregate, these points of circum- stantial evidence provide the basis for a tentative reconstruction, in which Phasmatocycas is presented as a cycadalean megasporophyll with a taeniopterid foliar lamina distal to the basal seeds. This is shown in figure 4. Cycadean affinity of Phasmatocycas.-Phasmato- cycas and Archaeocycas resemble each other funda- mentally in their bilateral symmetry, each bearing two lateral rows of paired seedlike objects. In the case of Archaeocycas, substantial indirect evidence indicates that these objects are the remains of seeds, partly enclosed by a reduced lamina and attached to it a short distance from the rachis. In Phasmatocy- cas, on the other hand, the bilaterally paired objects are unequivocably the remains of gymnospermous seeds, as established by their well-preserved mega- spore membranes and double cuticles. 18 8 a PALEOZOIC ORIGIN OF THE CYCADS e '.- e-4 6 cs sy. ,.XOI-l %*% W” gg fii‘a‘.‘ Olé & .A.A‘ 6 A O l O p b\ bo Phasmatocycas shows no development of a lamina in the ovuliferous region, so that if it is considered in the light of a reduction series, Spermopteris being the hypothetical ancestor, it would represent the ulti- mate stage in elimination of a foliar element in the fertile area of the appendage; the result would en- tail direct attachment of the seeds to a naked midrib. Only one specimen of Phasmatocycas is available, and absolute proof of attachment of a terminal foliar lamina is lacking. However, if my appraisal of in- direct evidence and my resultant reconstruction of Phasmatocycas are correct, then Phasmatocycas may be regarded as the perfect fulfillment of a theoretical reduction series, beginning with the completely lam- inar Spermopteris and possibly involving Archaeo- cycas as an intermediate form. (See fig. 6). Detailed comparisons of Phasmatocycas with oth- er fossil gymnosperms in order to derive taxonomic conclusions would be pointless because no close simi- larities exist among the fossils. The seed-bearing groups known to exist in the Permian include the pteridosperms, the early conifers (cordaiteans and walchians), and the ginkgophytes, none of which constitutes a logical taxonomic receptacle for Phas- matocycas. This fossil shows no evidence of the cupu- lar seed investment, complex pollen-receiving mech- anism, or other critical features of the pterido- sperms. A pteridospermous alliance for Phasmato- cycas may be dispensed with, except in the sense that it was probably derived from the pteridospermous Spermopteris. It is equally obvious that Phasmatocycas is no cordaitean or conifer. It lacks any evidence of the complex investiture of spirally arranged bracteate appendages that appear in cordaitean cones. It has no resemblance to the short shoots involved in the evolution of coniferous cone seales or of Ginkgo-like fructifications. There likewise appears to be no rea- son to postulate a Bennettitalean alliance for Phas- matocycas; Harris' (1954, p. 290) observations on the absence of a cutinized megaspore membrane in the cycadeoids, considered against the well-developed membrane in Phasmatocycas, weigh heavily against an already remote point of possible speculation. The remaining approach lies in comparisons of qualitative features of Phasmatocycas with other gymnosperms, irrespective of relative ages. Such an FIGURE 4.-Reconstruction of megasporophyll of Phasmato- cycas, showing bilaterally arranged basal ovules, alternat- ing glands, and taeniopterid distal lamina. Relative length of distal lamina reduced; venation and glands on lamina shown less dense than in actual specimens. From Mamay, 19692, figure 1A. SYSTEMATIC DESCRIPTIONS examination reveals that, in points of fundamental | morphology, Phasmatocycas is more like a mega- sporophyll of the modern genus Cycas than any oth- er gymnospermous fructification, living or fossil. In all the living cyeads, the ovules (usually only two) are bilaterally arranged, but Cycas usually has two or more pairs of ovules, indicating a multiovulate ancestral megasporophyll. Phasmatocycas at least hypothetically performs that role. The seeds of Phasmatocycas were apparently of simple construction and are more like those of the cyeads than of the pteridosperms or cordaites. The cuticular details of Phasmatocycas are sufficiently similar to those of the Jurassic cyead Beania to warrant increased emphasis of the other cycadlike features of Phasmatocycas. In its overall qualitative aspects, then, the one ex- tant specimen of Phasmatocycas is remarkably like a megasporophyll of Cycas. Only the terminal leaf- like proliferation of the Cycas megasporophyll is missing from Phasmatocycas, but evidence for its presence, already discussed, is substantial. Converse- ly, such a terminal proliferation is indeed present in Archaeocycas, which I believe to represent an evo- lutionary form just one step below Phasmatocycas in this primitive cycadalean lineage. The name Phasmatocycas derives from the modern generic name Cycas and the Greek "phasma," mean- ing apparition or specter. In connection with a cycadean interpretation of Phasmatocycas, it is of interest that White's (1912 p. 508) list of provisional plant identifications from the Permian of Kansas includes reference to "Cy- cadospadix? sp., C," the "C" referring to Carlton, Kans. (according to the USGS paleobotany locality register, this is locality 8868, "4 miles south of Carl- ton"). The Carlton collection contains nothing that might suggest a cycadean structure; conversely, Phasmatocycas is the only cycadlike specimen in the Banner collection, but White's list indicated nothing cycadlike from Banner. In attempting to determine the exact location of the Banner plant beds, certain facts have been found that may explain the above inconsistency and that consequently may identify White's "Cycadospadix? sp." as the specimen de- scribed here as the holotype of Phasmatocycas. The town of Banner no longer exists under that name; its site is now occupied by the town of Elmo, which formerly was less than a mile west of Banner on the Missouri Pacific Railroad. Carlton still exists and is about 3 miles west of the former town of Banner. Collection 8868 was made "4 miles south of Carlton," and collection 8869 was made "314 miles 19 south of Banner." As Banner was just 3 miles east of Carlton, the two fossil localities were less than 4 miles apart. The collections could have been, and probably were, made the same day (collecting dates for the two localities are given only as "1909" in the USGS locality register). Parts of the two collections are lithologically indistinguishable, and it thus is possible that the two collections were mixed and mis- labeled, and that White consequently identified a cycadlike fossil from Carlton instead of Banner. The two collections are very similar, and I am reasonably certain that White had in hand the holotype of Phas- matocycas when he made the identification of "Cy- cadospadix ? sp.". If this analysis is not correct, then an important part of the Carlton collection was lost and White failed to notice this important specimen in the Banner collection. Either alternative seems unlikely, and the combination of circumstances is even more difficult to envisage. Function of the glands.-The glandular bodies are so regularly placed between the seeds of Phasmato- cycas and so abundant on the associated taeniopterid leaves-putatively the distal laminae of megasporo- phylls-that speculative teleological comments are in order. The function of the glands will likely re- main conjectural at best, but the important biologic process of entomophily is indeed suggested by cer- tain indirect evidence. The presence of glandlike bodies on taeniopterid leaves is not a new idea, for Halle's (1927) memoir on Paleozoic plants from Shansi, China, included descriptions of two species of Taemiopteris that had "black dots" between the veins. Halle (1927, pl. 37, fig. 1) also showed similar bodies on the type speci- men of 7. multinervis from the Permian of Saxony. Halle regarded the dots as probably glandular but made no comments on their possible functions. In a later paper, however, Halle (1929) presented a lengthy and pertinent discussion of that possibility of entomophily in the Paleozoic; his argument was based largely on the presence of glandular bodies on other kinds of foliage, but it may be applied to Tae- niopteris as well. Halle's discussion (1929, p. 18-23) aimed primari- ly at trends in seed position in the pteridosperms, but modes of pollination were also considered. He wrote as follows : There remains the possibility of insect pollination. Oliver and Scott, in their famous memoir 'On the structure of the Palaeozoic seed Lagenostoma Lomazxi etc.' draw attention in this connection to the capitate glands found on the cupule of Lagenostoma Lomaxzi and on various parts of Lyginopteris old- hamia. Following a suggestion of the late Sir Joseph Hooker, 20 PALEOZOIC ORIGIN OF THE CYCADS the authors briefly discuss the possibility of insect pollina- tion in the Lagenostoma seeds * * * Insect pollination is not peculiar to the Angiosperms, since it has been found to oc- cur in several genera of living Cycads (cf. Seward, Fossil Plants, vol. 3, 1917, p. 28). Several groups of insects being known to have existed in Carboniferous times, the possibility of entomophily in Paleozoic plants cannot be denied. The fre- quent occurrence of structures which seem to represent ex- ternal glands is, indeed, a striking feature in the seed-bear- ing leaves of Pteridosperms known at present. . . . But the relative frequency of these gland-like structures among plants which were certainly or probably Pteridosperms, coupled with the appearance at about the same time of a rich insect-life is suggestive. Additional knowledge of their distribution in dif- ferent groups of Paleozoic plants is therefore desirable. Thus there is evidence of a marked trend toward production of glandiferous foliage and accessory parts among late Paleozoic seed plants, and Phas- matocycas provides an excellent additional example. The variety of insect life in the Paleozoic was pointed out by Carpenter (1952, p. 16) who listed nine orders of insects known from Permian strata and two orders from older rocks. Carpenter com- mented "Altogether, the lower Permian insect fauna was very diverse-more so, in fact, than any other insect fauna known." Not only was the Paleozoic insect fauna taxo- nomically diverse, but according to Carpenter (1971), it showed significant adaptive modification of mouth parts. Carpenter commented as follows (1971, p. 1238, 1240, 1241) : A very interesting feature of the Palaeodictyoptera, Mega- secoptera and Diaphanopterodea that has been discovered in recent years is the presence of piercing-sucking mouthparts. There can be no question that beaks in the Paleodictyoptera and the other two orders mentioned were used for piercing and for sucking liquifi nutriment. What the types of food were, we have no way of knowing; presumably they were of plant origin. The most likely sources were probably some of the succulent lycopods that were extensive and abundant during the Carboniferous and Permian. Carpenter (1971, p. 1241) continued thusly : The significance of this large array of hemipterous insects in the Paleozoic is obvious: there must have been abundant sources of liquid plant food available. With the Palaeodictyop- tera, Megasecoptera and Diaphanopterodea also present, it is perhaps no exaggeration to state that nearly half the species of Paleozoic insects had piercing-sucking mouthparts. 1 It is doubtful that the lycopods, although abundant in the Paleo- zoie, constituted an important source of food for herbivorous insects be- cause of certain anatomical features of the lycopods. Their structural strength derived mainly from a thick outer periderm layer which was nonconductive and therefore not very succulent. Likewise, the thick hypodermal layers of the leaves probably discouraged foraging insects. Possibly some soft-skinned amwhibians were plagued by these sharp- beaked insects, just as man suffers mosquito bites today. If one adheres to the idea of an herbivorous habit for these Paleozoic insects, however, alternative sources of food other than the lycopods deserve consideration. There seems to have been a compatibility of size between the sucking beaks of some of the Paleozoic insects and the micropylar canals of some of the pteridospermous seeds. Perhans the contents of the pollen chambers were attractive to the insects, who used their beaks to obtain these substances, but it is very doubtfnl that enough seeds were pro- duced to sustain a sizeable insect population. Thus, external glands on the plants seem to constitute a logical source of insect nutriment. Considering the sizable number of Paleozoic in- sects that had specialized-feeding mouthparts and their coexistence with many kinds of seed plants, it seems reasonable to assume that the feeding habits of at least some of these insects were involved in the pollination of seed ferns and precursive cyeads. The contents of the external glands were possibly attrac- tive to insects and thus served the function of nec- taries. It is intriguing to imagine pollen-dusted insects, having previously been attracted to the pol- leniferous organs by odor or other stimuli, visiting the putative taeniopterid distal lamina of Phasmato- cycas for the food contained within the abundant glands. Thence, they would descend along the seed« bearing axis, feeding on the interseminal glands and meanwhile depositing pollen on the micropyles or pollen drops of the ovules. The regular placement of the glands of Phasmatocycas would render each ovule equally attractive to the pollen-bearing insects, thus assuring pollination of all. A mechanism for initial visual attraction of in- sects to a female infructescence of Phasmatocycas could conceivably have been one similar to that pos- tulated by Leppik (1971, p. 172) for the Triassic Palaeocycas integer. This plant supposedly had a central crown of megasporophylls surrounded by a closely set group of leaves. According to Leppik, Such patterns must have been clearly distinctive to the food- searching phytophagous insects * * * the adaptive form and arrangement of these clusters of sporophylls must have made them characteristic attractants to food searching insects, al- lowing them to distinguish these from the less palatable ferns and horse tails. The question of insect pollination among the cy- cads is variously argued. Seward (1917, p. 28) wrote: Recent observations point to the probability that insects play a part in the pollination of cyceadean ovules. Kraus drew attention to the strong smell emitted by the microstrobili of Dioon edule and noticed that small bees were attracted to the ripe strobili of Macrozamia, while odourless cones of a neigh- boring Ceratozamia received no attention. Pearson and Rattray have obtained evidence that beetles and weevils act as pollina- tors to species of Encephalartos. The reports of insect pollination in Encephalartos are interesting, but Coulter and Chamberlain (1910, p. 141) observed that "It is generally accepted that the cycads are wind-pollinated." Both Chamberlain (1935, p. 127-128) and Gaussen (1944, p. 56) pointed out that the supposedly pollinating beetle actually bores into the female gametophyte of En- cephalartos, and the gametophyte is destroyed either in the process or by the hatching larvae. In this in- efficient procedure, the ovules, although possibly SYSTEMATIC DESCRIPTIONS pal pollinated, can mature only if the boring ovipositing process is not extensively damaging or if the beetle eggs fail to hatch. Crepet (1972, p. 1055) further discussed this interaction in presenting conjectures on the possibility of animal pollination of the Meso- zoic cycadeoids. Even though there is significant skepticism toward entomophily among the cycads, an affinity of sorts exists between certain modern cycads and certain insects. Perhaps this is the distant reflection of a dubiously successful experiment that began with Phasmatocycas, other Paleozoic plants, and insect pollinators unknown. ?Phasmatocycas sp. Plate 2, figure 1 DESCRIPTION This account is based on a single specimen from the Haskell County, Tex., locality; it was found by Mr. T. R. Tinsley, who presented it to me for study. The specimen consists only of an impression in red siltstone ; it lacks any organic matter but shows suf- ficient character to warrant discussion relative to Phasmatocycas and perhaps Spermopteris. The specimen consists of a fragment of a stout straight axis, which bears the impressions of two lateral rows of oval objects that probably represent seeds. The axial fragment is 5 em long, about 5.0 mm wide at its broadest point, and shows little tapering throughout its length. Along either side of the axis is a row of shallow oval impressions closely resembling the symmetrical system of placement of the seeds in Phasmatocycas kansana. Toward what I presume to be the top of the axis, the impressions are evenly paired, but the positional relationships between the units in the two rows become somewhat obscure toward the opposite end of the axis where preservation is less complete. There are 24 impressions in all, 14 on the right and 10 on the left side, part of which is missing. It is evident that one row contained approximately the same number of impressions as the other. The impressions are those of small round to oval bodies, 3.0-4.0 mm long and 2.0-3.0 mm wide. They show no tapering or beaklike apices, but evidently reflect broad attachments to the central axis by bases nearly as wide as their greatest widths, like seeds of Phasmatocycas kansana. The impressions are close- ly arranged but neither abut upon nor overlap each other; there are intervals of 1.0-1.5 mm between adjacent impressions. No evidence of accessory structures or sears of attachment of additional seed- like bodies is present on the surface of the axis be- tween the lateral rows. One additional feature of this specimen remains for comment. A slightly elevated ridge almost com- pletely surrounds the specimen except for the basal area where preservation is extremely obscure. This ridge more or less outlines the axis at a distance of approximately 3 mm beyond the apices of the seed- like impressions. It is best seen at the presumed apex of the specimen, where it passes over the tip of the axis and lends a tonguelike outline to the organ (pl. 2, fig. 1). The rounded apical outline resembles the tip of a leaf or pinna, but if the ridge does in fact represent the margin of a laminar appendage, there are no impressions of veins to confirm this. Figured specimen. 10064-1 (pl. 2, fig. 1). Occurrence.-USGS paleobotany locality 10064. DISCUSSION This specimen is only provisionally assigned to Phasmatocycas because of its preservational defici- encies. Nonetheless it merits attention because its symmetrical bilateral construction conforms with the cycadlike pattern of the other fossils under con- sideration. Its central axis is more robust than that of Phasmatocycas kansana in proportion to the size of the seedlike impressions; although the seedlike impressions are not as closely arranged as the seeds of P. kansana, this specimen is, in most respects, al- most exactly what one might expect to see if the holotype of Phasmatocycas kansana had been pre- served in the same manner as the specimen from the Vale Formation. I thus have no reservations in re- garding the Vale specimen as a seed-bearing axis, and further, as a representative of some taxon in the Permian complex of ancestral cycads. If we could establish beyond doubt that the Vale specimen is correctly assignable to Phasmatocycas, then the question of specific status would pose diffi- culties because only one specimen is available from each of the two localities concerned. The size and other differences between Phasmatocycas kansana sp. may fall within a reasonable range of specific variation,. but until more material becomes available this problem must remain unanswered. Of more moment is the matter of the generic as- signment of the Vale specimen. Were it not for the vague suggestion of the outline of a foliar lamina surrounding the axis and its complement of pre- sumed seeds, there would be a rather stronger case for assigning the specimen to an unnamed species of Phasmatocycas without the provisional question mark. This outline must be considered, however. If 22 PALEOZOIC ORIGIN OF THE CYCADS it actually represents the margin of a leaf blade rather than an anomalous sedimentary feature of the rock, then the Vale specimen cannot be regarded as a Phasmatocycas but as something more akin to Spermopteris. The seedlike bodies of the Vale speci- men appear to have been attached directly to the axis; in this respect, the specimen differs from Spermopteris, whose seeds were attached to the lamina. Conversely, the mode of seed attachment re- lates the specimen to Phasmatocycas. If we assume that a lamina was present and the seeds were at- tached directly to the axis, then we may speculative- ly regard the Vale specimen as a derivative of Spermopteris, in which the foliar lamina retained its identity but in which the seeds had undergone a "phyletic slide" in the direction of the axis, to the extent of actually being attached to the latter. This possibility is interesting in the light of relative ages of material considered here. The Vale specimen is the youngest of the entire complex, but if it indeed retains the Spermopteris-like foliar lamina, it may represent a conservative line of cyead evolution, as opposed to the older Phasmatocycas, which has lost the foliar blade in the fertile area. ?Phasmatocycas spectabilis Mamay, new species Plate 5, figures 1-4 Specific diagnosis.-Fertile axes large, robust, as long as 20 ecm or more; widths as much as 1.0 em or slightly more, distally tapering very gradually. Bi- laterally inserted rows of seeds conspicuous, the seeds as long as 7.0 mm, as wide as 5.0 mm; seeds oval or ovoid with blunt or slightly pointed apices, broadly attached by bases 2.0-3.0 mm wide, oppo- sitely paired or slightly alternating in position, per- pendicular to the axis or slightly ascending in atti- tude, crowded so as to abut against one another or more laxly arranged with intervening spaces 1.0- 2.0 mm wide. Holotype.-USGS 8959-3 (pl. 5, fig. 1). Paratypes.-USGS 8877-5, 6, 7 (pl. 5, fig. 2, 8, 4). Occurrences.-USGS paleobotany localities 8877, 8959. DESCRIPTION As was the case with Archaeocycas, my attention was originally directed toward the existence of this large fertile structure by a suite of specimens col- lected by David White at the locality south of Fulda, Tex. (USGS paleobotany locality 8877), where he also found the first specimens of Archaeocycas. Like Archaeocycas, these 10 specimens were found segre- gated in a separate tray, and some had been marked with crayon ; thus it is obvious that White had noted these as fossils of special interest and probably had intended ultimately to describe them. In retrospect of these events, it is rather strange that he made no reference to them in his list of plants from the Fulda beds, published in his paper on Gigantopteris ameri- cana (1912). In 1963, I found one small specimen of ?P. spec- tabilis, at the "Emily Irish" locality (USGS paleo- botany locality 8959), and in 1967, I found several more specimens at the same place. The specimens from White's locality are better preserved and have more of a carbonaceous residue in a somewhat more durable matrix. There is no doubt, however, that the two suites of material represent the same taxon. None of the axial surfaces between the lateral rows of seeds show any features that would aid in characterizing this species. Some of the specimens, particularly those from the Fulda locality, retain thin films of coalified plant residue and show vaguely defined longitudinal striations (pl. 5, fig. 4), but these may be shrinkage phenomena and probably are meaningless. The longitudinal striations are also seen in a few small areas on the holotype (pl. 5, fig. 1). Attention is directed to the "Emily Irish" speci- men shown on plate 5, figure 4, because it has a rather thick film of coalified material adherent to the interseminal surface of the axis. This film is fragmented into many small, nearly cubical pieces but is particularly significant in that it shows no con- spicuous points of interruption or discontinuity that would indicate places of attachment of seeds or ac- cessory structures. This and all other observable fea- tures of this suite of specimens show that in ?P. spectabilis we are again dealing with a bilaterally symmetrical organ. Unfortunately the seeds are poorly preserved and contain no cutinized parts ; pol- len analyses were attempted with samples from both localities, but results were negative. Nonetheless, the arrangement of the seeds conforms closely with the patterns seen in the other materials discussed in this paper. The outstanding morphological feature of ?Phas- matocycas spectabilis is its large size. Some indica- tion of this is provided by the holotype, shown on plate 5, figure 1. This incomplete specimen is broken at both ends, measures 20 cm in length, and has the attached remains of about 60 seeds. As there is scarcely any diminution in its width from one end to the other, one has little basis for estimating the com- plete length of this specimen ; it could have been con- siderably longer than 20 em, however. Approximate- ly 20 smaller fragments were found closely associ- SYSTEMATIC DESCRIPTIONS 28 ated with the holotype, but these cannot be estab- lished as parts of the same specimen. None of the shorter fragments exceeds the width of the axis of the holotype (approximately 1.0 ecm) ; probably this was the maximum width attained by this species. White's specimens from Fulda, one of which is shown on plate 5, figure 2, substantially confirm a maximum axial width of approximately 1.0 ecm. The best Fulda specimen, shown on plate 5, figure 2, is 10.8 em long and has a maximum width of 2.3 cm ; the latter measurement embraces the two later- al rows of seed. The longest specimen in this suite (pl. 5, fig. 4) is 11.5 em long, and the remaining 10 are considerably shorter. Among the "Emily Irish" specimens, none approaches the holotype in total length. None of the other "Emily Irish" specimens show any unusual dimensions. The specimen shown on plate 5, figure 3, however, shows a definite de- crease (from 3.5 to 2.0 mm) in width of the central axis and may represent the terminal part of the fruiting region. The seeds are oval or ovoid, and those of the Fulda specimens generally appear to be somewhat more pointed than those from the "Emily Irish" locality. Exact dimensions are difficult to ascertain because the specimens are generally poorly preserved, but their size ranges from 5.0 to 9.0 mm in length and 4.0 to 7.0 mm in width. The largest seeds present in the collection are attached to the specimen shown on plate 5, figure 4; the fifth seed from the bottom in the right-hand row, for example, is 9.0 mm long and 7.5 mm wide, and others on this specimen are of com- parable sizes. Some of the seeds on the holotype are considerably smaller and seem disproportionately small in relationship to the size of the axis. Some of the seeds show vague suggestions of bifid apices, but such structures are not a consistent fea- ture and may merely be the result of splitting of the integuments at the apices during compression. The presence of a thick integument is indicated by the "Emily Irish" specimens shown on plate 5, figures 2 and 3. A few of the seeds are represented only by flattened coalified films. Others, however, are more or less completely delimited by deep cavities in the matrix; these are evidently cavities left behind by the decay of the seed integuments, and the extent of the cavities is indicative of some considerable bulk to the original seed coats. The seeds are uniformly sessile, being attached by very broad bases. In other features there is a certain degree of variation, all of which could be attributa- ble to different relative positions on the parent axes. In most specimens the seeds are so closely inserted as to abut against each other (pl. 5, fig. 2), but in others they are considerably more widely spaced (pl. 5, fig. 1). Most of the seeds extend away from the axis per- pendicularly, but in a few specimens they are di- rected forward at broad angles; the latter condition is most prevalent in specimens where the seeds are loosely arranged, as in the holotype (pl. 5, fig. 1). In some specimens, the seeds seem to be arranged in distinctly opposite pairs, but most commonly they appear to alternate with each other. In none of the seeds is there any indication of cupular or bracteate investment; some seeds show irregular outlines and have erratic projections from their overall oval shapes, but I attribute these to preservational distortions. Furthermore, this spe- cies shows no remnants of the interseminal resinoid spheroids that appear on Phasmatocycas kansana. A final important feature of ?P. spectabilis is the absence of evidence of a foliar lamina. Because a complete specimen is not available, the presence or absence of a distal foliar lamina cannot be established. DISCUSSION ?Phasmatocycas spectabilis is of considerable morphological interest because of its large size, the limits of which are yet unknown. More importantly, it establishes the presence in the Permian of still an- other representative of the cyceadalean complex with which this paper is concerned. ?P. spectabilis repeats with remarkable clarity the bilateral symmetry and other qualitative features of the fruiting organs under study and thereby emphasizes the probability of phylogenetic derivation from Spermopteris or a similar ancestral form. The absence of cuticularization is unfortunate; be- cause of this absence, I refer this species to Phas- matocycas with the provisory question mark. Were cuticles present, I am reasonably confident that an unqualified generic assignment to Phasmatocycas could be justified. The specific name spectabilis refers to the overall showiness of these large fructifications. ?Cycadean male cones Plate 2, figure 4 DESCRIPTION An interesting element in the Haskell County flora is represented by 20-0dd impressions or molds of axial fragments which were collected and given to me by Mr. Tinsley. The specimens are stout and unbranched, with no evidence of the apical or basal 24 PALEOZOIC ORIGIN OF THE CYCADS organization. The specimen shown on plate 2, figure 4, is the most informative of the lot. The axes have surface patterns of closely imbri- cated, spirally arranged rhomboidal impressions with considerable depth ; in most specimens the im- pressions are elongate transversly to the axis. The impressions range from 7.0 to 10.0 mm in width and average about 5.0 mm in height. Extending away from most of the axial specimens and ascending up- ward in gently curved attitudes are molds of spine- like appendages, 2.0-3.0 ecm long ; these extend across bedding planes in a manner suggesting that the appendages were stiff and rigid at the time of burial and resisted compaction. Several are seen clearly in the illustrated specimen. Figured specimen.-USGS 10064-2 (pl. 2, fig. 4). Occurrence.-USGS Paleobotany locality 10064. DISCUSSION These robust axes resemble to a certain extent the impressions of the characteristic bark patterns of the arborescent lycopods. However, the depth of the prominent surface impressions is much greater than that of fossil lyecopod leaf cushions, which ordinarily show only a moderate amount of relief; these are more like the impressions that would be made by the distal faces of cyceadean microsporophylls. If the Texas specimens are to be regarded as lycopod re- mains, this isolated appearance is anomalous with the known history of arborescent lycopods in North America. I am aware of no representatives of that complex in American rocks younger than those of the Belle Plains Formation, where lycopod remains occur rarely in an otherwise rich plant assemblage. It is therefore highly probable that the giant lyco- pods became extinct on this continent before the end of Belle Plains time, and identification of the Texas specimens with the lycopods seems unwarranted. The preferred alternative of tentatively treating these specimens as cycadean male cones is encour- aged by their geographic and stratigraphic proxim- ity to Phasmatocycas and its putative foliage, Taeni- opteris. That indirect evidence finds support in the resemblance of the Texas material to the specimen of the Jurassic Androstrobus zamioides figured by Schimper (1870-1872, pl. 72, fig. 1), especially in size, general proportions of the specimens, and the overall appearance of their surface markings. The cycadlike aspects of the Texas specimens is further accentuated by the molds of spinelike, apparently rigid extensions of the individual members. These may well represent the attenuated termini of micro- sporophylls, like those produced in the male cones of Cycas cireinalis (Schuster, 1982, fig. 10A) or of Macrozamia (Johnson, 1959). All factors taken into consideration then, I prefer at least provisionally to regard the Texas specimens as the essentially mod- ern-appearing male cones of an early cycad rather than as representatives of a lingering remnant of the giant lycopod lineage. Only the absence of evi- dence of microsporangia on the individual units of these strobiloid structures restrains me from apply- ing a formal binomial indicative of cyceadean affinity. Several kinds of presumably male strobiloid fruc- tifications are present in other Permian collections under study ; some are apparently new, but none can be identified as cycadean. Among these, those most like fossil cycadean male cones are the strobiloid structures from Fulda, Tex., described by White (1912, p. 500-501) as the "supposed polleniferous strobili" of Gigantopteris americana. These were composed of reniform bracts inserted in distichous rows, with "oval, pendant sacs" on the lower sur- faces of the bracts. A few reniform bractlike organs are indeed present in the Fulda collection, but my examination of the original specimens does not sub- stantiate either White's interpretation of these bracts as the actual appendicular components of the strobili, or his observation that the cone scales are distichously arranged. Furthermore, the material is too poorly preserved to allow observation of any details of sporangial structure or disposition. None- theless, the Fulda specimens present the overall ap- pearance of rather loosely constructed cones, proba- bly with spirally arranged appendages. The speci- mens illustrated by White (1912, pl. 48, figs. 3, 4) are remarkably similar to some of the Jurassic male cycadean cones, Androstrobus wonnacotti Harris, illustrated by Thomas and Harris (1960, pl. 2, figs. 8-13) ; the similarity is most marked if one turns White's illustrations upside down and compares his figure 4 with Thomas and Harris' figure 10 or 12. The absence of epidermal details or in situ pollen in the Fulda specimens, however, renders further com- ment on this similarity pointless. Order uncertain (CYCADALES or BENNETTITALES) Genus CYCADOSPADIX Schimper, 1870 Cycadospadix yochelsoni Mamay, new species Plate 2, figure 3 Specific diagnosis.-Appendage with short, broad peduncle and lanceolate lamina. Margin of lamina entire from broadest point to base, deeply pinnatifid distally; lamina with heavy, unbranched ribs, sev- eral to the undissected part, one each to the distal lobes. SYSTEMATIC DESCRIPTIONS 25 Appendage 21 mm long, 12 mm broad at widest part of lamina. Peduncle 5 mm wide, 4 mm or more long. Incised part of lamina with eight teeth along each margin; teeth as much as 3 mm long, the long- est about halfway towards the sharply pointed tip. Teeth directed obliquely forward at broad angles. Holotype.-USGS 10057-1. Occurrence.-USGS paleobotany locality 10057. DESCRIPTION Only one specimen of this species has been found ; the better of its counterparts is shown on plate 2, figure 3. The incompletely preserved peduncle is a very broad and flat structure, the distal end of which extends as a broad wedge up into the base of the lamina and appears to have formed a slight ridge or suture at the line of juncture with the lamina. There is no indication of the nature of the vascular system of the peduncle other than a series of fine, closely spaced parallel surface striations that continue dis- tally as far as the tip of the lamina. The lamina flares out abruptly and at a broad angle from its point of confluence with the peduncle, soon reaching its maximum width of 12 mm. Be- tween the base of the lamina and its widest point the margin remains entire. The broad, parallel, distally directed ribs probably represent the original vena- tion of this appendage. Each side of the undissected part of the lamina contains six such ribs, which de- part from the distal end of the peduncle in a pseudo- palmate fashion. Those that enter the margin lobes of the remaining two-thirds of the lamina, however, are arranged in a more typically pinnate manner. A distinct midrib is not evident in the lamina ; the only indication of structural differentiation in the median part of the appendage is the weak parallel striations that extend into the lamina from the peduncle and continue to the tip. The lobes of the dissected part of the lamina vary in angles of departure, the more proximal ones being directed forward at about 45° and the distal ones extending out at right angles from the median line of the appendage. The lobes number eight or nine to each margin. The longest ones are about 3 mm long and less than 1 mm wide; they decrease in length so that the last one or two directly beneath the spicate tip of the lamina are slightly less than 1 mm long. DISCUSSION Cycadospadix Schimper is an artificial genus of isolated bractlike or megasporophyll-like organs, known from only a few occurrences, mostly of Meso- zoic age. As stated by Harris (1932b, p. 96), it is characterized by "a stalk expanded distally and bearing several fingerlike branches; the fossils re- semble in miniature the megasporophylls of Cycas." The resemblance to Cycas led Schimper (1870-1872, p. 207) to express the opinion that there could be no doubt as to a cycadean affinity for his material. Schimper referred two species to Cycadospadix : C. Hennocquei (Pomel) Schimper and C. Moraeanus (Pomel) Schimper. In his diagnosis of the genus, Schimper drew attention to sears on the sides of the petiole which he interpreted as indications of points of seed attachment ; this was purely speculative, and Schimper's illustration of C. Hennocequei shows no seeds attached. Emberger (1944, fig. 346) published a drawing of a specimen of that species that had a seed attached to the petiole, but Harris (1961, p. 316) stated that "no seed is known attached." Harris (1932b, p. 97-98) described Cycadospadizx dactylota on the basis of three specimens from the Rhaetic of Greenland. The cuticles were preserved in his material, and on the strength of stomata of the bennettitalean type, Harris emended the generic diagnosis of the genus, thus inferentially assigning all species of Cycadospadix to the Bennettitales; Harris interpreted his specimens as the basal scales of a bennettitalean "flower" rather than megasporo- phylls. Florin (1933, p. 34) rejected Harris' emenda- tion of Cycadospadix and proposed a new genus, Bennettitolepis, for the Greenland material, typified by the recombination B. dactylote (Harris) Florin. Harris evidently prefers his original designation, because in 1961 (p. 316-317) he referred to the Greenland material only as Cycadospadix and made no mention of Bennettitolepis. In his extensive article on the Mesozoic cycads, Florin (1933) restudied Nathorst's original material of Cycadospadix integer from the Rhaetic of Sweden. Florin was able to determine the cuticular details, and on the basis of epidermal structure, he concluded that the organs were in no way related to the Ben- nettitales but instead showed close affinity with the modern Cyceadales. He proposed a new genus, Palaeo- cycas, typified by the species of P. integer (Nat- horst) Florin, and reconstructed an entire plant. Florin's (1933, p. 91) often-reproduced reconstruc- tion shows a plant with a dense aggregate of termi- nal megasporophylls of the Palaeocycas integer type and taeniopteroid leaves called Bjuvia simplex Flor- in. A single megasporophyll is shown as having a broad entire-margined lamina and two pairs of seeds attached to the sides of a thick stalk. Harris (1961, p. 315) remarked that none of the figured specimens has seeds attached, although bulges on the stalks of 26 PALEOZOIC ORIGIN OF THE CYCADS Palaeocycas are suggestive of points of seed attach ment, The Texas specimen conforms well with the gener- al concept of Cycadospadix but, as is the case with all other taxa heretofore assigned to Schimper's genus, there is no evidence of seed attachment. The broad stalk of C. yochelsoni suggests a stout seed-bearing axis, but no evidence of sears of attachment is visi- ble; less than 5 mm of the stalk is preserved, how- ever, and the possibility that more complete speci- mens would demonstrate evidence of seed attach- ment cannot be discounted. The distal laminar part of the appendage is remarkably similar to the top of a Cycas megasporophyll. This resemblance is such, in fact, that if C. yochelsoni had been found attached to any of the seed-bearing organs described here, there would scarcely be room for doubt as to a cy- cadalean affinity. The cuticle of C. yochelsoni is not preserved ; thus, no evidence of stomatal or other epidermal features is available to aid in interpreting relationships of the specimen. Furthermore, laminar form is of little assistance. With its distally laciniate and basally en- tire lamina, C. yochelsoni is intermediate between Bennettitolepis, a presumed cycadeoid, and Palaeo- cycas, a presumed cyead. A short while before Florin reassigned Cycadospadix integer to the new genus Palaeocycas because of epidermal characteristics, Harris (1932b, p. 96) suggested that the undissected margin of the lamina of C. integer probably was sufficient basis for its reference to another genus. Similarly, additional material from the Vale Forma- tion may necessitate a reassessment of the generic status of C. yochelsoni. Regardless of the taxonomic histories of organs originally assigned to Cycadospadix but later reas- signed as the result of discovery of anatomical char- acteristics more definitive than gross external form, the original name Cycadospadizx retains a consider- able degree of utility as an ordinally noncommittal receptacle for structures that could be either cyca- deocidean bracts or cycadalean megasporophylls. Con- sequently, the Texas specimen is appropriately re- tained in Cycadospadix. The question of its true sys- tematic position-bennettitalean or cycadalean- may, for the time being, be examined principally in the light of indirect evidence, some of it negative but all of it significant. The stratigraphic and geograph- ic relationships between Cycadospadizx yochelsoni and the other materials considered here are over- whelmingly indicative of a cycadalean alliance for C. yochelsoni. Conversely, the absence of credible evi- dence of bennettitalean plants in the Paleozoic of North America militates against a bennettitalean in- terpretation of the Texas specimen. If C. yochelsoni is indeed a bennettitalean appendage, it stands strangely isolated-biotically, chronologically, and geographically. If its resemblance to the distal lami- na of a Cycas megasporophyll is taken at face value, however, it assumes a harmonious niche in an as- semblage of fossils whose collective impact is to extend the known history of the cycads downward through the greater part of two geologic systems. Cycadospadix yochelsoni is named after E. L. Yochelson, who found the specimen. ASSOCIATED PLANTS The following summary enumerates the plants as- sociated with the fossils described here. It is pre- sented to emphasize that the cycadlike fossils share little in the way of floral associates that could be considered pertinent to questions of taxonomic affin- ity. The overall effect of this résumé is to strengthen the circumstantial evidence that the sporophyll of Phasmatocycas kansana had a taeniopterid lamina. The Perry, Okla., flora (USGS paleobotany locali- ty 6233), of which the specimen of Spermopteris is part, was discussed briefly by White (1912, p. 506- 507) in connection with occurrences of Gigantop- teris americana. He listed the following seven gen- era: Pecopteris, Gigantopteris, Taeniopteris, Annu- laria, Sphenophyllum, Walchia, and Araucarites. I have confirmed the presence of all of those genera except the two arthrophytes ; if present, they are too poorly preserved for positive identification. The identification of Araucarites is doubtful, but the col- lection contains a few small seeds that may have been the basis for White's identification; two are partly cutinized and resemble somewhat the seeds of Phasmatocycas kansana. Two specimens of the asymmetrically winged seeds attributed by White to Gigantopteris are present, and I have identified Odontopteris and Callipteris. This is a typical lower Permian assemblage, referable to Zone 14 of Read and Mamay (1964). *The Banner, Kans., flora, supposedly the source of the holotype of Phasmatocycas kansana (USGS paleobotany locality 8869) was indicated by White (1912, p. 507-509) to contain the following six gen- era: Callipteris, Glenopteris, Taemopteris, Spheno- phyllum, Noeggerathia?, and Carpolithes. The col- lection also contains specimens of cordaitean leaves and several small undetermined sporangia, but I am hesitant to verify a noeggerathialean element in the collection. This is a sparse assemblage, but by virtue ASSOCIATED PLANTS 27 of the presence of the genus Glenopteris it is also assignable to Zone 14. The small collection from Insect Hill (USGS paleobotany locality 8298) con- tains no elements that are not present in the Banner assemblage. Sellards (1908), however, described the following additional genera from the vicinity of "Banner City" in his review of the Permian flora of Kansas: Pecopteris, Aphlebia, Neuropteris, Odon- topteris, Sigillaria, Cordaites, Cordaianthus?, Rhab- docarpos, Cardiocarpon, and Aspidiopsis. White (1912, p. 505-506) recognized 14 genera associated with Gigantopteris americana at the Ful- da, Tex., plant beds (USGS paleobotany locality 8877). These are Diplothmema, Pecopteris, Odontop- teris, Newropteris, Taemniopteris, Annularia?, Sphe- nophyllum, Sigillaria, Cordaites, Poacordaites, Wal- chia, Gomphostrobus ?, Aspidiopsis, and Araucarites. Names of plants from a nearby locality, 2% miles south of Fulda ("Castle Hollow," locality unnum- bered) were incorporated in the same list because of the presumable identical stratigraphic positions of the two deposits. The "Castle Hollow" beds apparent- ly contained most of the genera present in the Fulda beds, as well as the two additional genera Aphlebia and Sigillariostrobus. Thus the total number of gen- era from the two localities, as noted by White, is 17. My examination of White's Fulda collection sub- stantiates his identifications with the exceptions of Sigillaria, Poacordaites, and Diplothmema. Con- versely, this collection contains Stigmaria, Aphlebia, Annularia, Discinites (Mamay, 1954), and a curi- ous apparently new peltate structure that in some features resembles the genus Dawbreeia Zeiller. I have not located White's original "Castle Hollow" collection and must assume that it has been lost. Its absence, however, only reduces the total list of gen- era known from that area by one, Sigillariostrobus. The overall aspect of this flora is unusual only from the standpoint of the presence of the Daubrecia-like structure, Discinites, Archaeocyceas, and ?Phasmato- cycas. The flora is recognized as an assemblage typi- cal of Zone 14. The most diverse flora involved in this study is that of the "Emily Irish" deposit (USGS paleobota- ny locality 8959). These beds are regarded as the approximate stratigraphic equivalent of the Fulda beds; they contain all the genera recognized in White's collection with the exception of the Daw- breeia-like organ. The "Emily Irish" flora is richer than the Fulda flora, however, in numbers of species as well as genera. It contains an unusually diverse spectrum of callipterids in addition to Lobatannu- laria, Sphenopteris, and the curious genus Russell- ites (Mamay, 1968). Also present are several types of undescribed microsporangiate fructifications. Gi- gantopteris americana is a conspicuous element in this flora, which is referable to Zone 14. The Taylor County, Tex., flora (USGS paleobota- ny locality 10057) in which the specimen of Cyca- dospadix yochelsoni was found, is somewhat younger than the Fulda and "Emily Irish" assemblages and has been assigned to the youngest of the recogniz- able upper Paleozoic floral zones, Zone 15. It con- tains essentially the same generic association as the older flora but is distinguishable from it on the basis of several notable specific differences ; the most strik- ing of these differences is the appearance in the Lawn flora of a new, as-yet-undescribed species of Gigantopteris (G. n. sp. B; Read and Mamay, 1964, pl. 19, fig. 2). In addition to Gigantopteris and Cyca- dospadix, the flora contains Amnmularia, Aphlebia, Taeniopteris, Sphenophyllum, Pecopteris, Gomphos- trobus, Walchia, Callipteris, Odontopteris, and sev- eral taxa of unidentified seeds. Some of the odontop- terids are of an extremely large-pinnulled type remi- niscent of certain elements in the Siberian Angara flora, also of Permian age. The flora also contains some curious hooked appendages of unknown alli- ance, two specimens that resemble liverworts, and what appears to be a new genus of conifers. A de- seription of this flora and related assemblages from other parts of the Vale Formation is being prepared. The Haskell County, Tex., exposures (USGS pale- obotany locality 10064) are higher in the Vale For- mation than the Lawn deposit, and, accordingly, they contain the youngest flora involved in this study. This flora is also the simplest in terms of gen- eric composition, although its taxonomically limited aspect is not regarded as a function of limited ma- terial. The known plant association consists only of the single specimen of ?Phasmatocycas sp., Taemniop- teris, walchian conifers, an undescribed new species of Gigantopteris, and the axial specimens described as ?eycadean male cones. This sparse list of plant types is the result of not only my personal examina- tion of literally hundreds of specimens scattered over an area of several acres, but it also reflects the care- ful collecting activities, over a period of many years, of Mr. T. R. Tinsley, who gave me the specimen of ?Phasmatocycas sp. as well as my choice of any oth- er specimens in his collection. Mr. Tinsley had searched the site and immediate environs frequently, in the hope of finding "new" fossils. On my first visit there, I was astonished by not only the great num- bers of specimens lying loose on the surface, but also by the predominance of specimens of Gigantopteris 28 PALEOZOIC ORIGIN OF THE CYCADS n. sp. This taxon constituted possibly all but 1 per- cent of the total number of specimens examined; walchian conifers were next in frequency. Only a few specimens of Taenmiopteris have been found at this site. By virtue of the numbers of specimens ex- amined, I believe that a reasonable representation of the original flora is preserved at the fossil site. After elimination of the obviously pteridophytic or coniferous elements in the preceding assemblages, only seven pteridophyllous genera (Callipteris, Gi- gantopteris, Glenopteris, Neuropteris, Odontopteris, Russellites, and Taeniopteris) remain for considera- tion as possible foliar components of the plants that bore the cycadlike structures under investigation. From the standpoint of its pinnately compound mor- phology and parallel-veined foliar segments, Russell- ites is the most cycadlike of these genera. The possi- bility of cyceadean affinity for Russellites has been considered briefly (Mamay, 1968, p. 110), but as it is known from only one site (USGS paleobotany locality 8959), it is ineligible for consideration as a common ancestral form for both Phasmatocycas and Archaeocycas. Odontopteris, Callipteris, and Newropteris are all regarded as pteridosperms ; although they are strati- graphically and geographically widespread, none ap- pears in the Haskell County flora. Although G/- gantopteris is abundant at the Haskell County site and in other parts of Texas and Oklahoma, the genus is not known in Kansas; likewise Glenopteris, abun- dant in a small area of Kansas, is not known outside of that State. The distributions of Glenopteris and Gigantopteris, viewed in terms of paleophysiography and other physical factors, led Read and Mamay (1964) to present some conclusions regarding lim- ited geographic provinciality of certain Permian plants in North America. The conclusions from the foregoing remarks are that Taeniopteris is the least common botanical de- nominator; it is the only foliar element present at all the localities where the cycadlike remains have been found. Although inconclusive in itself, the evi- dence of consistent association must always be given careful consideration in drawing general conclu- sions; in this instance it is particularly germane because of the convincing evidence of the known seed-bearing habit of Spermopteris and the acces- sory evidence provided by the resinoid spherules of Phasmatocycas and associated taeniopterids. DISTRIBUTION OF THE FOSSILS Stratigraphic positions of the fossils are shown in figure 5. The limited vertical distribution of the plants adds substantially to the distinctiveness of the guide assemblages that characterize floral zones 11 to 15 of Read and Mamay (1964). The searcity of taxa in the Wolfcampian interval is clearly a func- tion of limited collections. Wolfcampian beds have not yet yielded large amounts of good fossil plant material, and the few collections on hand have cast no light on the cycadean complex under investigation. Geographically, the plants are distributed over a small part of the southwestern United States, extend- ing only from north-central Texas to east-central Kansas. Although a very small area is involved, in terms of general geographic distribution of Paleo- zoic floras, the distribution of these plants may be significant from the standpoints of ecologic toler- ances and evolutionary plasticity. Read and Mamay (1964, p. K16-K18) recognized three Permian floral provinces in Leonardian rocks of the southwestern United States; these provinces were based on lateral distributions of the genera G- gantopteris, Glenopteris, and Supaia, as related to paleophysiography, paleoedaphics, floristic charac- teristics, and other factors. The Glenopteris flora is restricted to a small province in eastern Kansas, and because of its consistently close association with saline rocks, the flora was presumed to have evolved in a physiologically xeric environment. Immediately south and southwest, the Glenopteris flora is dis- placed, without evidence of overlap, by the Gigantop- teris flora, which occupied a rather less limited geo- graphic province having presumably less rigorous environmental conditions. A third province, that of the Supaia flora, existed west of the ancestral Rocky Mountains; this province seems to have been char- acterized by harsh ecologic conditions which pro- duced an impoverished flora. The plants described in this paper were not re- stricted to any one of the three Permian floral prov- inces, but collectively they occupied all three. Al- though Spermopteris probably evolved before the beginning of the differentiation of upper Paleozoic floral provinces, it first appeared on the eastern mar- gin of the area later to become the province of Glen- opteris, then migrated southwest, and eventually was found in north-central Oklahoma in the G@- gantopteris province. Phasmatocycas appears in the Glenopteris province and, assuming that my identifi- cations are correct, in the Gigantopteris province as well. Only Archaeocycas and Cycadospadix are pres- ently limited to one province-that of Gigantopteris. Taeniopteris, however, the one floral element in com- mon association with all the others, is found in all DISTRIBUTION OF THE FOSSILS 29 w 3 i 0 © o 5 6 3 North Texas Oklahoma Kansas t L- N C om Choza Formation ?Cycadean a g male cones C : 3 & Vale Formation _7PhaS/sngfocycas 0 .s > * 5 §l8i | ame. g o C Formation » © C g G 0 /= €. | N S Leuders S a Formation S Archaeocycas E ~d re rchaeocy 2 . whiter i oC F ® * 5 £. Clyde Formation Wellington f Wellington Phasmatocycas ¥ o 8 20 - < v r t €Spermopteris sp . ; e- . % 6 Sp $6| Belle Plains spectabilis ormation Formation ansana ® = f f s Formation & -3 & | n ~ fa boe $ 3 a S a ¢ a $ Fa § | 2 f 8 o c ® had o gq o | ~ =< jWJ/MMQEVWA’J x mamma m I $ a .o & g HC _ G 6 - £ ‘—1 Spermo- < | < - § a Lawrance 1 Hf conacea 8 a | ® | o § 2 Formation RIS (ee S g S S S | o N ~ a orion nees v W’MMNWAW FIGURE 5.-Vertical distribution of late Paleozoic cyceads and related plants. Modified from Dunbar and others (1960), Moore and others (1944), and Read and Mamay (1964). three provinces; it is a conspicuous element in the Supaia flora, three species having been described from the Hermit Shale of the Grand Canyon, Ariz. (White, 1929). Thus Taeniopteris was chronological- ly persistent, geographically adventuresome, and ecologically resilient. If those qualities reflect a gen- etically robust constitution for the taeniopterids, the phylogenetic acquisition of such traits would equip any Paleozoic derivatives of the taeniopterids ad- vantageously for worldwide dispersal during the Mesozoic. MORPHOLOGIC AND EVOLUTIONARY CONSIDERATIONS INTERRELATIONSHIPS OF THE AMERICAN FOSSILS Appraisal of the evolutionary significance of these fossils is best begun by an attempt to weigh their relationships with each other. If their stratigraphic positions were such that a vertically linear sequence of forms could be used as a framework for phylo- genetic analyses, those interpretations would be a relatively simple matter. As shown in figure 5, how- ever, a linear phylogeny is not apparent among these plants because two of the principal forms, Phas- matocyeas and Archaeocycas, are clearly representa- tive of different evolutionary stages ; in their earliest appearances, however, they occur in essentially syn- chronous deposits. Thus they probably represent sep- arate evolutionary tangents, derived from a common spermopterid ancestor. The simplest phyletic facet of this association is in the relationship between Spermopteris and Phas- matocycas. The role of Spermopteris as the morpho- logic equivalent of, if not the actual archetypic plant from which Phasmatocycas kansana arose, becomes apparent upon examination of all the pertinent facts. The stratigraphic relations fit the demands for such a phylogenetic affinity, for Spermopteris, the ances- tor, is considerably older than Phasmatocycas, the descendant. The geographic proximity of known oc- currences of the two genera adds substance to a hypothetical genetic relationship. Seeds of the two, 30 PALEOZOIC ORIGIN OF THE CYCADS as far as they are known, are similar. The intimate association of the genera with sterile taeniopterid foliage emphasizes the probability of physical identi- ty. This probability is virtually converted into reality by the presence of the regularly distributed resinoid spheroids on the axis of Phasmatocycas kansana and of similar bodies on only taeniopterid foliage; only the actual organic continuity between Taemiopteris and Phasmatocycas remains to be found. The total effect of the foregoing lines of evidence is to promote acceptance of Phasmatocycas as a cycadean mega- sporophyl) with basal ovules and a taeniopterid dis- tal lamina, as reconstructed in figure 4. Phyletic derivation of Phasmatocycas from Sper- mopteris would involve the morphogenetic processes of migration of seed positions, partial sterilization of the megasporophyll, and reduction of the lamina. According to Cridland and Morris (1960, p. 855), the seeds were confined to the terminal regions of the megasporophyll of Spermopteris but in all the cy- cads and in the reconstruction of Phasmatocycas the seeds were basal. As seed position varies considera- bly among the pteridosperms that have simple, non- cupular phyllospermy, there is no compelling reason to believe that terminal seed positions need have been phyletically immutable among the spermopter- ids. If some spermopterid megasporophylls original- ly were only basally fertile, or became basally fertile through distal sterilization of completely fertile megasporophylls or through basal migration of orig- inally terminal seeds, one has found a phyletically immediate antecedent for Phasmatocycas. Eventual organogenetic elimination of the basal fertile part of the lamina would then effect a migration of the points of seed attachment from laminally superficial positions to positions along the sides of the foliar midrib, which would thus become the fertile peduncle of the megasporophyll. This process would preserve the original bilateral disposition of the laminally borne seeds of Spermopteris and would be reflected in the bilateral ovuliferous habit of all cycadean megasporophylls, living and fossil. This hypothetical series of modifications is shown diagrammatically in figure 6. The ultimate derivative form (fig. E), representing Phasmatocycas kan- sana, is essentially a modern Cycas-type of mega- sporophyll, except for certain minor morphologic divergences such as its relatively large number of ovules and its entire-margined lamina. Inasmuch as it epitomizes a cycadean megasporophyllar form that had substantially completed the important qualita- tive phases of its evolution by the onset of Leonardi- an time, closely related taxa might reasonably be expected to be found ultimately in any post-Paleo- zoic rock unit. Even though a certain amount of doubt attends the identity of the Haskell County specimen from the Vale Formation, the presence there of Phasmatocycas may well be expected. Con- sidered in relation to other pertinent material, the Vale specimen would more logically be a true Phas- matocycas than an intermediate form with Spermop- teris-like foliation; nonetheless, the possibility of affinity with Spermopteris cannot be discounted without more satisfactorily preserved material. With the exception of Archaeocycas, which proba- bly represents an evolutionary stage similar to that shown in figure 6D, the age relationships of the known elements in this evolutionary series suggest that forms intermediate between Spermopteris and Phasmatocycas must be sought in Wolfecampian rocks, where evidence of cycadean fossils is yet absent (fig. 5). The specimens of ?Phasmatocycas spectabilis are impressive because of the large size of the sporo- phylls and the great numbers of seeds. These fea- tures make it difficult to assess possible affinities be- tween ?P. spectabilis and its contemporary, P. kansana, which is known from only one small speci- men. Although the bilateral distribution of the seeds of the Texas specimens strongly suggests their as- signment to Phasmatocycas, the absence of informa- tion regarding ovular details precludes the positive generic identification. Organic evidence for correla- tion of the fertile axes with a given foliar type, such as presented by the resinoid spherules in the Kansas material, is not available in the Texas specimens. However, some of the associated Texas taeniopterid leaves are so large (laminae as much as 10.0 cm wide and midribs as broad as 15.0 mm) that their establishment as distal laminae of the large fructifi- cations of ?P. spectabilis would not seem dispro- portionate. Accepting the probability that ?°P. spectabilis and P. kansana are congenerically related, the impressive quantitative differences and essentially contempo- raneous stratigraphic relationships of the two taxa would seem to discourage the concept of a linear phylogenetic alliance. More likely they are coordi- nate descendents of a common spermopterid ances- tor, ?P. spectabilis having undergone reduction of a considerable length of foliar lamina while retaining a large complement of seeds. In the latter respect, P. spectabilis is to be regarded as the most prolific megasporophyllar type yet discovered among the cycads. It is also to be regarded as primitive, if the concept of reduction of numbers of parts is to be MORPHOLOGIC AND EVOLUTIONARY CONSIDERATIONS 31 E" | ~ o 3 e 2 a C CA O CA U _ ] FIGURE 6.-Hypothetical evolutionary development of primitive Cycas-like megasporophyll from Spermopteris-like pteri- dosperm. A, Dorsal view of ancestral spermopteroid megasporophyll with taeniopterid lamina and superficial seeds (small solid-black areas at seed bases indicate points of attachment to lamina). B, Derivative form with reduced num- bers of seeds. C, Further reduction, with seeds restricted to basal area of lamina. D, Form with seed production essen- tially constant, but basal part of lamina reduced, with basalmost seeds attached directly to stalk of megasporophyll. E, Ultimate Cycas-like derivative (=Phasmatocycas), with basal part of lamina completely reduced and all seeds at- tached to megasporophyll stalk. retained as indicative of a trend in the direction of evolutionary advancement. Because of its morphologic intermediacy, Archaeo- cyecas whitei stands out as the most intriguing and phylogenetically important taxon found in this study. It is a curious mixture of both primitive and ad- vanced features; thus the nature of its morphologic origins and the direction of its phyletic destiny seem apparent. The critical characteristics of this plant are consistently observable in such a large suite of specimens that they cannot be coincidental ; thus, an obvious set of interpretations may be set forth with confidence, The spacing of the small circular scars of seed attachment is remarkably regular in Archaeocycas, showing that the seeds were attached to the lamina 32 PALEOZOIC ORIGIN OF THE CYCADS only a short distance away from the midrib. This method of seed attachment is very much like that in Spermopteris, as diagrammatically shown at the lower right of Cridland and Morris' (1960) figure 10 (fig. 1 of this paper). Thus Archaeocycas retains the primitive feature of laminally borne seeds; by virtue of this characteristic, Archaeocycas may be regarded as a primitive megasporophyll, derived from Spermopteris or a Spermopteris-like ancestor. Judged on its laminar morphology, Archaeocycas is distinctly intermediate between Spermopteris and the Cycas-type of megasporophyll. Although its lam- ina was still essentially complete, the basal part had become sharply modified and partly reduced, form- ing a closely fitting indusiumlike covering for the ovules. The sterile distal part of the lamina is known to have been abruptly expanded immediately above the fertile area. Its distal details are incompletely known, but there is no reason to doubt its foliar na- ture; the specimens shown on plate 1, figures 1, 2, and plate 2, figures 5, 6, show hints of a taeniopterid type of venation. The extent to which the ovules were enclosed by the laminar '"indusium," is unknown, but, in keeping with the cycadean theme of this pa- per, I would venture to guess that the lower surfaces of the ovules were largely exposed. Judged next on the basis of positions and numbers of seeds, Archaeocycas contrarily demonstrates a distinctly advanced morphological level, particularly in comparison with Spermopteris. Although the seeds of Spermopteris are said to be restricted to terminal parts of the sporophyll, those of Archaeo- cycas are definitely basal and therefore of substan- tially modern Cycas-like appearance. Furthermore, the number of ovules of Archaeocycas is not known to range beyond the extremes of 8 and 12 on a given sporophyll, with an average of 10. This modest level of ovule production is again an essentially modern feature; inasmuch as it is consistent among a hand- ful of specimens from two different localities, this seems to have been a stable genetic feature of at least the type species, Archaeocycas whitei. Com- parison of this characteristic would seem to indicate a higher level of advancement for Archaeocycas than either Spermopteris or Phasmatocycas. In Spermop- teris, the number of seeds on a given sporophyll is not evident from Cridland and Morris' (1960) ac- count, but evidently the seeds were produced in large numbers; the type specimen of Phasmatocy- cas kansana had 21 or more seeds, ?P. sp. had 24 or more, and ?P. spectabilis produced 60 or more. All morphologic features considered, then, Arch- aeocycas must be the derivative of some plant with the salient attributes of Spermopteris. It needed only to undergo morphogenetic reduction of the basal indusiumlike covering of the ovules and migration of the points of seed attachment to the midrib to attain the morphologic equivalence of a modern Cycas megasporophyll. Archaeocycas is, in fact, a classic example of evolutionary intermediacy; it is the type of fossil that would eventually have been predicted on a paper diagram, had it not been dis- covered in actual substance. In the vernacular, it is a critical "missing link." In terms of the hypothetical evolutionary sequence diagrammed in figure 6, Archaeocycas would most closely approximate the stage depicted in 6C or 6D ; the chief qualitative difference is that the laminar "indusium" extending around the ovules is not shown in the diagram. A more complete representa- tion of my visual concept of Archaeocycas and its evolutionary development into a Cycas-like mega- sporophyll is given in figure 7. The progressively deeper marginal incision of the distal lamina, as shown in forms C and D, represents a necessary step in the derivation of a pinnate appendage from an entire-margined one; certain fossil evidence for such a process is discussed below. The kinds or degrees of natural affinity that ex- isted between Archaeocycas and Phasmatocycas are difficult to weigh, primarily because both genera appeared at approximately the same time. Were it chronologically intermediate, Archaeocycas might be regarded as phylogenetically intermediate be- tween Spermopteris and Phasmatocycas, in much the same manner that it seems intermediate between Spermopteris and Cycas; it might even be construed as a common ancestor of both Phasmatocycas and Cycas. The essentially comparable ages of the two Permian fossils, however, introduces the need for caution in reading their relationships in terms of a simple linear phylogeny. This picture is further be- clouded by the fact that ?Phasmatocycas spectabilis occurs at the same two localities and in the same beds in which Archaeocycas has been found. A clear- er understanding of the affinities between P. kansana and ?P. spectabilis would be helpful because, as pre- viously stated, their phylogeny probably is not a linear one. If one is not derived from the other, it is possible that a common ancestor, intermediate be- tween Spermopteris and those two Permian species, exists in older Permian rocks, and that the same taxon was ancestral to Archaeocycas as well. In the light of age relationships of the fosils, then, it seems preferable to regard Archaeocycas as the result of a MORPHOLOGIC AND EVOLUTIONARY CONSIDERATIONS 33 -> ¢ " "3 FIGURE 7.-Hypothetical evolution of Cycas-like megasporophyll from Archaeocycas (modified from Mamay, 19692, fig. 1). A, Presumed dorsal view of megasporophyll, showing partial enclosure of paired basal seeds by sporophyllar lamina (extent of enclosure conjectural). B, C, Hypothetical derivatives, showing progressive reduction of basal lamina and resultant attachment of basalmost seeds to midrib, or stalk, of sporophyll; margin of distal lamina of form depicted in C has become slightly incised. D, Cycas-like ultimate form, with basal lamina completely reduced and all seeds at- tached directly to stalk of sporophyll; distal lamina deeply pinnatifid, substantially as in Cycas revoluta. line of descent independent from that which pro- duced Phasmatocycas. Cycadospadix yochelsoni is difficult to interpret from the standpoint of phyletic position. Almost any of the necessarily indirect bases for assaying its natural relationships to the other fossils discussed here are obscured by the possibility that this species is actually a cycadeoidean bract rather than part of a cycadean megasporophyll. Nonetheless, be it cyca- dean or cycadeoidean, its presence in Lower Permian rocks is chronologically precocious, like that of its cycadean associates. Cycadospadix yochelsoni is fur- ther intriguing in that it would not appear biologi- cally incongruous if it were established as the distal extension of one of the terminally incomplete fructi- fications described in preceding pages; rather, it would seem harmonious in such a position and would complete the portrayal of a Cycas-like megasporo- phyll in any of those instances. AN ALTERNATIVE INTERPRETATION OF ARCHAEOCYCAS The sketches presented in figure 7 were made on the premise that the ovules of Archaeocycas were incompletely enclosed by the fertile lamina. Place- ment of the margins of the lamina was rendered speculative because the few specimens available for this study were preserved only in lateral or ventral view, and delicacy of matrix and preservation lim- ited mechanical preparation to a point just short of destruction of the specimens. Thus the lower or dor- sal aspects of the megasporophyll of Archaeocycas 34 PALEOZOIC ORIGIN OF THE CYCADS are not known. The lack of evidence of constriction of foliar tissue immediately distal to the most distal pairs of ovules, however, suggests that at least that part of the lamina was incompletely inrolled and that the plant was therefore gymnospermous. With the morphology of Spermopteris and other cycadalean entities in mind, then, I applied the most conservative interpretation of Archaeocycas and de- picted it as a fertile leaf with superficial ovules and incompletely involute margins. As such it represents as uncomplicated as possible a theoretical derivative of a Spermopteris-like ancestor, yet with significant attributes of a primitive Cycas-like megasporophyl Only a modicum of morphogenetic reduction of the lamina was necessary to produce a Phasmatocycas or Cycas type of megasporophyll; thus Archaeocy- cas, if interpreted correctly, stands as an important Permian intermediate between the Pennsylvanian Spermopteris and the modern Cycas. The uncertain extent of involution of the mega- sporophyll of Archaeocycas introduces the necessity for considering an alternate interpretation. It seems at least remotely possible that, even if not coalescent in the known specimens of Archaeocycas, the oppos- ing margins of the fertile base of the lamina were in an initial stage of proliferation rather than reduc- tion, so that they subsequently met and began phylo- genetic fusion. Continuation of the process of fusion would eventually have resulted in an organ in which the ovules were completely enclosed. This structure would thus show basic characteristics of an angio- spermous carpel; as illustrated in my preliminary article (Mamay, 19692, fig. 1H), this hypothetical organ was said by Uhl and Moore (1971, p. 990) to resemble the carpel of the palm Nypa. This theoretical process is illustrated in figure 8, adapted from the preliminary article (Mamay, 19692, fig. 1). Its significance lies in the inference that plants having at least one important angio- spermlike reproductive feature could have been de- rived on a more or less direct line from the late Paleozoic pteridosperms, and in the corollary evi- dence from the fossil record, that the carpel is in- deed a modified seed-bearing leaf. Thus the sugges- tion emerges that possibly the angiosperms and cy- cads are closely related through a common ancestry -the spermopterid pteridosperms-or even that the angiosperms arose directly from the cyeads during the late Paleozoic. The latter idea could be particu- larly attractive to those who espouse the concept of Paleozoic angiospermy. (See Eames, 1959.) At the same time it would shed doubt on my arbitrarily delineated distinction between dorsal and ventral sides of Archaeocycas, because it is accepted that angiosperm ovules are produced on the adaxial (ven- tral) side of the megasporophyll (Eames, 1961, p. 186) ; thus, either the ovules of Archaeocycas were attached to the wrong side of the lamina or my ter- minology has been incorrectly applied. The possibility that the specimen of Cycadospadizx in the Texas flora is a Bennettitalean appendage resuscitates the question of relationships between the cycads and cycadeoids. Chamberlain (19835, p. 158, figs. 171, 172) proposed that both groups arose from a common ancestor among the Paleozoic pteri- dosperms and presented a series of drawings depict- ing two reduction series leading to the cycads and cycadeoids; the hypothetical ancestral megasporo- phyll had a terminal ovule and several lateral ones intercalated between foliar segments. In one series the terminal ovules aborted, and continued reduc- tion of parts resulted in several of the modern types of cycadean megasporophylls; in the other series, some of the terminal ovules were retained, all others were aborted, and with the reduction of laminar tissue, the ovule-bearing stalks and sterile "inter- seminal scales" of Bennettites were produced. Neith- er the ancestral nor the intermediate forms have been discovered, so that Chamberlain's hypothesis remains unsubstantiated. Delevoryas - (1968b) agreed that the pteridosperms are the most likely source of the cycadeoids, but he also pointed out that the cycads and cycadeoids have always been distinct orders. If a common ancestry is involved among the pteridosperms, one probably must look for it in pre- Pennsylvanian sediments. Certainly the present study sheds no light on the question, and if the Texas specimen of Cycadospadix is indeed bennettitalean, it is more than ever apparent that the two orders were distinct entities for as long as either can be recognized. It would be impractical to attempt to review here all the thoughts that have been published in regard to the ancestry of the angiosperms. Suffice it to say that all the various groups of gymnosperms-par- ticularly the pteridosperms-have at one time or other been regarded as probable progenitors of the flowering plants. The cycads were especially favored as angiosperm ancestors by Chadefaud (1986) who interpreted the fertile megasporangiate organs as homologous and wrote (Chadefaud, 1986, p. 590- 591) : Or, la feuille carpellaire des Angiospermes est manifeste- ment un feuille fertile a ovules lateraux * * * On en devra évidenment conclure que les Angiospermes sont plus étroitement apparentées aux Cycadales qu'a aucun autre MORPHOLOGIC AND EVOLUTIONARY CONSIDERATIONS 35 FIGURE 8.-Hypothetical evolution of carpellike organ from Archaeocycas (modified from Mamay, 19692, fig. 1). A, Pre- sumed dorsal view of megasporophyll, showing partial enclosure of ovules by sporophyllar lamina. B-D, Hypothetical derivatives, showing progressive proliferation and fusion of margins of lamina, resulting in complete enclosure of ovules (D). groupe de Gymnospermes. En particulier, cest aux Cycadales et non aux Bennettitales qu'il faut les rattacher. Chadefaud's ideas have not been widely accepted by subsequent writers. Eames (1961, p. 464-467) re- viewed the possible origins of angiospermy and omitted the cyceads from consideration. Cronquist (1968, p. 40-46) again considered the possibilities and eliminated all the gymnosperms except the pter- idosperms (Lyginopteridales) as likely angiosperm forerunners, meanwhile urging caution toward ac- ceptance of the lyginopterids in that capacity. Cron- quist conceded, however, that the cycads probably arose from the pteridosperms; this view implicitly indicates a distant but common ancestry for the cyeads and angiosperms. Carlquist (1969, p. 353), in commenting on fossil evidence on floral theories, discounted any evidence that "Cordaitales, Ginkoales, Coniferales, Cycade- oidales or even Cycadales ever had leaf-like mega- sporophylls." Carlquist also commented that "The chance of finding a fossil that shows ancestral car- pel form is vanishingly small-and would we recog- nize it if we saw it, or would we misidentify it be- cause of preconceptions ?" Carlquist's comment on the "vanishingly small" chance seems incongruous with the steadily increas- ing ranks of paleobotanists, the attendant increase in collecting activities, and the fluid overall state of fossil botany. And perhaps an early "ancestral car- pel form" may be represented by an Archaeocycas- like plant, and only better material is needed to es- tablish its identity as such. The foregoing discussion is not intended as my endorsement of the idea that Archaeocycas is a car- 36 PALEOZOIC ORIGIN OF THE CYCADS pel or even an early form in the evolution of the carpel; I regard the cycadalean interpretation as more consistent with the attendant facts and cir- cumstances, and particularly with the morphology of Spermopteris. The discussion is rather intended to point up the increasingly evident fact that Permian time produced many botanical innovations (see Ma- may, 1962, 1966, 1968; White, 1912), some of which may have involved carpellike seed enclosure and the beginnings of angiospermy. RELATIONSHIPS OF THE AMERICAN FOSSILS TO OTHER FOSSIL CYCADS Appraisal of the natural affinities between the American taxa and other fossil cyead fructifications ish not substantially more complicated than that of the interrelationships of the American fossils them- selves. The problem is that so few points of com- parison exist; only a few post-Paleozoic ovuliferous cycadean fructifications are recorded, and recently Harris (1961) has reviewed reasons for doubting the validity of some of those. The most thoroughly understood of the Mesozoic ovuliferous cycad fructifications is the genus Beania Carruthers from the Jurassic of Yorkshire, Eng- land; our knowledge of Beania is the result prin- cipally of the many contributions of T. M. Harris and H. H. Thomas. Beania is commonly regarded as the most advanced of the known Mesozoic cycads, for its seeds are of essentially modern aspect and the female infructescence, although long, lax and not very conelike, is composed of simple biovulate sporo- phylls of zamioid construction. Thus Beania does not bear any significant morphological resemblance to any of the American fossils and is probably not as closely related to Phasmatocyceas or Archaeocycas as it is to the modern genera; if related within a simple infrafamilial phylesis, Beania is certainly closer to Phasmatocycas than to Archaeocycas, be- cause of the more modern aspects of Phasmatocycas. If Beania arose from a member of the spermopterid stock of which the American fossils are clearly the progeny, the evidence for such derivation has not begun to be recovered; when found, it may be diffi- cult to recognize as such. Beania, then, is of little aid in interrelating the American fossils with fossils of the Old World or of the post-Paleozoic. Further- more, the Jurassic Period was sufficiently younger than the Permian, and cycad evolution apparently so rapid, that even if a more complete spectrum of female fructifications were known from the Jurassic, probably none of those fossils would be very en- lightening on the phylogenetic implications of the American material. We must therefore consider the Triassic cycads. Only two Triassic taxa need to be considered here: Dioonitocarpidium Lilienstern 1928 and Palaeocycas Florin 1933. Both genera have been widely cited as credible examples of Triassic cycads, but in each instance challenges have been directed at cycadean interpretations. Riibhle von Lilienstern described Dioonitocarpidi- um on the basis of Upper Triassic (Keuper) ma- terial from Bavaria. It is a composite taxon, consisting of taeniopterid sterile foliage (Danaecop- sis angustifolia Schenk) and pinnate sporophylls (Dioonites pennaeformis Schenk), each sporophyll supposedly bearing a basal pair of ovules. The type species is D. pennaeforme (Schenk) Lilienstern. Lili- enstern's original reconstruction of the apical part of a plant is presented here in figure 94. Although this reconstruction has been duplicated by various authors (Gothan and Weyland, 1954; Krausel, 1950; Magdefrau, 1956), it has not been universally ac- cepted. It was criticized by Florin (1983, p. 122), who expressed doubt that a plant with simple taeni- opterioid leaves would produce pinnate sporophylls. Florin furthermore pointed out that the foliage (Danaeopsis angustifolia) may not even be gymno- spermous and that the name Dioonitocarpidium pos- sibly embraces two unrelated plants, one a fern, the other a cyeadophyte. An element of doubt as to morphological detail at- tends Lilienstern's reconstruction of the fertile base of a megasporophyll (fig. 9B). The drawing depicts a broad stalk with slender hairy pinnae and a pair of basal seeds. Within bounds of artistic license the reconstruction is reasonably acceptable, except for what appear to be anomalous subsidiary flanges of tissue extending from the sporophyll stalk below the attached ovules and continuing distally to the upper limits of the sketch. This improbable feature was not discussed in Lilienstern's text and is possibly at- tributable to a drafting error. The probability of an error gains support from the fact that the detailed reconstruction of the sporophyll base (see fig. 9C) was modified in two textbooks by Gothan and Wey- land (1954, p. 288; 1964, p. 314), so as to eliminate the superfluous flange, although Lilienstern's overall reconstruction of the plant is faithfully reproduced in both the cited instances. Inasmuch as the Gothan and Weyland references included no explanation of the discrepancies between their reconstruction and the original, I inquired of Professor Weyland as to a possible reason. Weyland (written comm., 1968) advised me that "The illus- MORPHOLOGIC AND EVOLUTIONARY CONSIDERATIONS 37 z 22 ermm sas FFT FZ tes K & DF 5 2) N8 F. FIGURE 9.-Reconstructions of Diconitocarpidium Lilienstern. A, Apical part of plant of D. pennaeforme, showing taeniop- teroid sterile foliage and pinnate sporophylls (from Lilienstern, 1928, fig. 1). B, Lilienstern's reconstruction of basal part of megasporophyll, showing pair of seeds and peculiar subsidiary flangelike structure beneath the pinnae (from Lilienstern, 1928, pl. 6, fig. 4). C, Modified reconstruction of basal part of megasporophyll of D. pernaeforme (from Gothan and Weyland, 1964, fig. 208 a). D, Reconstruction of megasporophyll of D. keuperianum (Krasser) KrAusel, showing two pairs of basal seeds and pinnate lamina (from Krausel, 1949, fig. 4). tration of the sporophyll originated with Gothan and, so far as I know, it is not published elsewhere. Lilienstern's reconstruction probably seemed incor- rect to him. However, I do not know where the speci- men forming the basis of Gothan's illustration is located today." Whatever the exact basis of Gothan's illustration was, it is apparent that he disagreed with Lilienstern's depiction of the fertile base of a sporophyll, meanwhile accepting the overall recon- struction of Dioonitocarpidium as a plant with pin- nate sporophylls and taeniopterid sterile leaves. The modifications introduced by Gothan into the recon- struction of the individual megasporophyll have re- sulted in a morphologically more reasonable form than the original of Lilienstern. Two additional species of Dioonitocarpidium-D. keuperianum (Krasser) Kriusel, and D. liliensterni Kriausel-were subsequently described; these also are from the Upper Triassic Keuper beds of Bavaria. D. keuperianum (Krausel, 1949) was reported as a megasporophyll with a pinnate distal part and a slightly pinnatifid basal flange of laminar tissue to the surface of which "several" seeds were allegedly attached ; Krausel published a reconstruction of the sporophyll, which appears here as fig. 9D. In a later publication, Kriusel (1953, p. 106) discussed addi- tional material of D. keuperianum, which was said to "partly confirm, partly broaden and correct" his earlier description of that species. Krausel (1953, p. 106) stated that seeds were plainly visible on one of 38 PALEOZOIC ORIGIN OF THE CYCADS the specimens, and that at least five were attached to one side of the sporophyll. Unfortunately, the photo- graphic illustrations in both publications are of poor quality and do not permit verification of the critical details of Kriausel's reconstruction or of his diagnosis (Krausel, 1949, p. 48) of D. keuperianum. A similar circumstance concerns the third species, D. liliensterni. In 1953, Krausel described D. lilien- sterni on the basis of a single sporophyll. This spe- cies differed from Krausel's concept of D. keuperi- anum in that its basal, supposedly fertile part was much longer and broader than that of D. keuperianum and was entire-margined. Krausel (1953, p. 107) stated that three seeds were attached to one side of this organ. Here again the photographic illustrations fall short of convincing demonstration of attached seeds. It may well be that the broad base of the leaf in question truly represents a reproductive speciali- zation, but the requisite seeds remain in question. Thus, in consideration of the deficiencies of illustra- tion, a certain amount of reservation appears neces- sary in drawing comparisons of D. keuperianum and D. liliensterni. If one assumes that the descriptions and recon- structions of the three species of Dioonitocarpidium are accurate, then one must weigh the question of their taxonomic relationships, because their modes of seed attachment show as much difference, funda- mentally, as the difference between the seed attach- ment of Archaeocycas and Phasmatocycas. In D. pennaeforme the seeds are sessile and supposedly are attached directly to a naked sporophyll stalk, as is the case with Phasmatocycas; conversely the seeds of D. liliensterni and D. keuperianum are supposedly attached to a basal laminar extension of the sporo- phyll, similar to the manner of seed attachment in Archaeocycas. Thus the Dioonitocarpidium concept may in fact consist of two genera, differentiable on the basis of characteristics of the fertile bases of the sporophylls. From this speculative standpoint, one might postulate two Permian-Triassic evolutionary lines of cycads, one leading from Phasmatocycas to Dioonitocarpidium pennaeforme and an attendant reduction in number of seeds, the other leading from Archaeocycas to D. liliensterni and D. keuperianum. Admittedly this is a tenuous approach to phyletic considerations because of the many deficiencies in the physical evidence on hand. The American and European taxa may very well not be related by way of a linear phylesis. However, they show approxi- mately comparable morphological qualities and seem to express similar evolutionary trends. Both groups contain members with laminally produced seeds, a fact which, through conservative interpretation, sug- gests that both groups were derived from a sper- mopterid stock. Thus it is not entirely unrealistic to suppose that the two groups had a common ancestry, if not a lineal relationship. The evidence of common association with taeniopterid foliage seems to bear out the foregoing lines of thought. The one Triassic taxon remaining for considera- tion is Palaeocycas-Bjuvia, a cycadean plant from the Rhaetian (Upper Triassic) of Sweden, which was described and reconstructed by Florin in 1933; his reconstructions appear here as figure 10. Like Dioonitocarpidium, Palaeocycas (for the sake of brevity of reference, Florin arbitrarily selected this as the name for the entire plant) is a composite genus. It consists of large taeniopterid leaves (Tae- niopteris gigantea Nathorst), to which Florin ap- plied the binomial Bjuvia simplex, and presumed megasporophylls with ovate entire-margined distal laminae; Florin named the latter Palaeocycas inte- ger. No seeds were found attached to the megasporo- phylls, but the stalks of these organs each bore two pairs of prominent lateral protrusions. Florin inter- preted the latter as points of seed attachment and reconstructed the sporophyll as a leaflike appendage with four basal seeds attached to the naked stalk. Because of their association in the same rock strata and similar epidermal anatomy in the leaves and sporophylls, Florin deduced that the two organs be- longed to the same plants. The reconstruction of this plant shows a stout armored trunk and loose termi- nal crown of leaflike megasporophylls, which strik- ingly resemble those of the modern Cycas. The out- standing discrepancy between Palaeocycas and Cy- cas is in the architecture of their respective fronds; those of Palaeocycas are simple, whereas those of Cycas are pinnate. The circumstantial nature of the evidence for Florin's synthesis of Palaeocycas and Bjuvia into one plant was discussed fully by Harris (1961). Har- ris' comments were mildly skeptical; his emphasis was on the absence of seeds as an important negative factor. Notwithstanding this type of criticism, how- ever, Florin's reconstruction continues to appear in recent textbooks (Sporne, 1965, p. 106; Emberger, 1968, p. 457) as an acceptable Mesozoic cycad. In- deed, this is possibly the most widely reproduced ex- ample of paleobotanical synthesis; Florin's recon- struction seems to have enjoyed classic status from the time of its original publication. The relevance of Palaeocycas to the present study is in the important resemblances that are seen be- tween this Triassic plant and my concept of Phas- MORPHOLOGIC AND EVOLUTIONARY CONSIDERATIONS 39 FIGURE 10.-Reconstructions of Palaeocycas-Bjuvia. A, Hy- pothetical reconstruction of entire plant. B, Single mega- sporophyll, reconstructed with two pairs of basal seeds. From Gothan and Weyland, 1964, fig. 209. matocycas; a much more convincing set of argu- ments can be presented in support of direct phyletic relationships between Phasmatocycas and Palaeo- cycas than is possible in comparing Dioonitocarpidi- um with Archaeocyceas and Phasmatocycas. Al- though Palaeocycas and Phasmatocycas are both in- completely known, the gaps in our knowledge, which appear in different areas of their diagnostic images, are considerably minimized by taxonomically sig- nificant similarities. Indeed, their reconstructions are mutually corroborative on the grounds of com- plementarily indirect evidences. In both genera, sterile taeniopterid leaves and megasporophylls are regarded as attributable to a common parent on similar bases: (1) consistent as- sociation in the same strata, and (2) morphological features shared by both pairs of organs-epidermal characteristics in the case of Palaecocycas and glands in the case of Phasmatocycas. The seeds are absent from the sporophyll of Palaeocycas but its distal lamina is known, whereas the reverse is true for Phasmatocycas; its seeds are known but the distal lamina is not. Nonetheless, the evidence for correla- tion of parts in each instance points toward two plants with very similar ovulate sporophylls and both with taeniopterid foliage. Although my recon- struction of the megasporophyll of Phasmatocycas entails a longer, more leaflike distal lamina and more seeds than in Palaeocycas, I believe that the overall aspects of the two plants were probably very similar and that, with only minor modifications, Florin's habit reconstruction of Palaeocycas might represent Phasmatocycas as well. The sporophyllar differences between the two genera are of minor quantitative nature and may be ascribed to reductive evolution in the direction of Palaeocycas, the younger of the two genera. From the standpoint of all morphologic evi- dence as well as stratigraphic relationships, a Palaeo- cycas-like plant would be a credible derivative of Phasmatocycas. Therein lies the most apparent rela- tionship between American and other fossil cycads. Archaeocycas must be related to Palaeocycas in much the same way as it is related to Phasmatocycas. EVOLUTION OF THE PINNATE HABIT IN CYCADALEAN FOLIAGE Considerable indirect evidence favors the assump- tion that all the known pre-Jurassic cyeads, with the exception of the Upper Triassic Leptocycas, bore entire-margined taeniopterid foliage. One thus is prompted to seek fossil evidence of an evolutionary process leading from simple undivided laminae to pinnate leaves of the type that characterizes most of the modern cycads. Gradual incision or laciniation of an originally Taenmiopteris-like foliar lamina seems the most logical theme. As hypothetically proposed on preceding pages, the taeniopterid distal lamina of the sporophyll of Archaeocycas or Phasmatocycas became incised to produce a Cycas-like megasporo- phyll, and there is no good reason to discount a co- ordinate modification of the sterile leaves as well. The essence of such an evolutionary sequence among the cycads was briefly touched upon by Gaussen (1944, p. 20), who wrote: "Pour passer de la feuille de Palaeocycas & celle des Cycadites ou du type Cy- cas, il faut que le limbe se dechere en pinnules et que dans chacune la nervure fine se transforme en une épaisse." Taeniopteris or closely allied forms are abundant in the late Paleozoic fossil record, and among these are some whose margination suggests, and would easily fit into, an evolutionary series that produced a Cycas-type of leaf in which each laminar division contains a single vein. Taemiopteris serrulata, from the Lower Shihhotse Series (Lower Permian) of 40 PALEOZOIC ORIGIN OF THE CYCADS Shansi, China (Halle, 1927, p. 160-161, pl. 42, figs. 13-18, pl. 64, fig. 13), is a species that well might represent an early evolutionary stage in such a series. T. serrulatg has a simple leaf and is quite ordinary in most morphological aspects; however, it has a finely serrate margin. The marginal teeth are very small, but each one is occupied by a single vein ending (cf. fig. 11B). A further step in this evolutionary process is ex- emplified by the lamina of Taeniopteris? koreanensis Cheong and Lee, 1970, from the upper part of the Sadong Formation in Korea. The associated plants indicate an Early Permian age for T.? koreanensis and place it in a harmonious stratigraphic light within this hypothetical evolutionary series. Previ- ously given the new generic name Serratopteris (Cheong, 1969), this material consists of a "proba- bly unipinnate" leaf, in which all but the basal part of the lamina is deeply divided into a series of es- sentially equal incipient "pinnules" that extend about halfway from the margin of the lamina to the midrib (cf. fig. 11C). As in T. serrulata, each divi- sion received a single vein. The illustrations suggest that T.? koreanensis is a compound leaf and that the serrations represent incipient pinnules rather than pinnae, but this point is not entirely clear. It is also not known whether the deep serrations extend to the tips of the laminae, because the specimens are incom- plete. Notwithstanding, this species is a near-perfect intermediate between a normal, entire-margined Taeniopteris and a Cycas leaf with its single-veined leaflets; T. serrulata would be the expectable start- ing point. Interpretation of these plants as an in- fallibly credible series is beclouded by the uncertain- ty whether T.? koreanensis is a simple or compound leaf, but certainly the two examples demonstrate the possibility of a regular increasing process of laminar laciniation in the Paleozoic, each resultant ultimate foliar division supplied by a single vein, as in the modern monotypic family Cyceadaceae (classification of Johnson, 1959). Culmination of the process pro- ducing T. serrulata might be envisioned in the Juras- sic genus Paracycas Harris. A natural segregate from the form-genus Cycadites, Paracycas has Cycas-like pinnation and venation and is grouped among the cycads by Harris (1964, p. 65). With the exception of Stangeria, all other genera of modern cycads are characterized by pinnae or pinnules that lack midribs and have several to many longitudinal dichotomous veins; those genera were placed by Johnson (1959) in the family Zamiaceae. Stangeria, monotypifying the family Stangeriaceae, differs markedly in that its pinnae each have a pro- nounced midvein with a dense system of transverse, parallel, dichotomously divided secondary veins. Just as putatively early stages in the development of the Cycas type of leaf are present among the Paleozoic taeniopterids, there are kno wn fossil forms that may be regarded as possibly early representa- tives of the zamioid type of foliage. Halle's Taeniop- teris? serrata (Halle, 1927, p. 161, pl. 42, figs. 9-12), a pinnate form associated with T. serrulata in the Permian Shansi flora, has a lamina with an irregu- larly serrate margin. In this species, however, the marginal areas between adjacent teeth receive three or four veins, rather than one as in T. serrulata (cf. fig. 116). Taeniopteris pseudobrevis is a species described by Barnard (1967, p. 725, pl. 61, figs. 1-4; 1970, p. 37-38, pl. 5, figs. 1, 4, text-figs. 3B, D) from Upper Triassic rocks of Afghanistan. In this plant, the foliar lamina is regularly corrugated at intervals of 2-3 mm and is incised to as much as one-sixth the total width of the blade (cf. fig. 11F). Each partial segment contains the ends of several simple veins. This leaf thus gives the impression of a simple taeni- opterid leaf well on the way toward becoming a pin- nate one; the multiple-veined segments would each correspond to a single leaflet of a Zam{ia-like leaf. Zamioid foliage appears more fully blown in the Upper Triasic of North Carolina, in the genus Lep- tocycas Delevoryas and Hope. The leaves of Lepto- cycas were compared with the common Jurassic genus Pseudoctenis, which, according to Harris (1964, p. 71), "would include the leaves of more than one genus of Recent Cycads." The lamina in this case is completely divided into essentially equal linear segments, each containing several parallel veins. This leaf is interpreted here as a remarkably precocious end product of laminar dissection, and consonantly Leptocycas, if truly a zamioid cycad, represents the earliest known appearance of the zamioid lineage. Jurassic time, during which explosive develop- ment of the cycadophytes took place, produced a broad variety of foliar types. Some forms are regu- larly pinnate and quite cycadlike in most aspects; others, however, show significant intermediate steps in the laciniation, or pinnation, of originally entire- margined taeniopterid laminae. The specimen illus- trated by Seward (1917, fig. 615) as Pterophyllum (Anomozamites) Nilssoni from the Middle Jurassic beds of Yorkshire, England, is a particularly inter- esting form. It has a very irregular but basically taeniopterid lamina, some parts of which are di- vided as far as the midrib into unequal truncated ¥ Y MORPHOLOGIC AND EVOLUTIONARY CONSIDERATIONS 41 segments; the apical one-third of the lamina is nearly entire, however. This species was later placed in Anomozamites by Harris (1969, p. 80) and is re- garded as bennettitalean. The other species of Ano- mozomites, A. thomasi Harris, has nearly uniform segmentation of the lamina. The genus Nilssonia is especially pertinent to the rationale of this theory of cycadean foliar evolution, because Nilssonia species show extremes of variable laminar dissection that occur in no other contempo- rary plant. Harris' (1932b, p. 40-52) descriptions of the Lower Jurassic species of Nilssonia from East Greenland revealed interesting examples. In some, the lamina may be consistently entire (N. undulata Harris), but in others (N. polymorpha Schenk; cf. fig. 11G) the lamina is entire or deeply incised into irregular segments. Usually the segments on one side of the leaf have no regular relationship to those on the other. In N. incisoserrata Harris, (1932a; p. 49- 52) the leaf is mostly "incisoserrate," uniform seg- mentation reaching only part of the distance to the midrib (cf. fig. 11F) ; this form resembles Taeniop- teris pseudobrevis Barnard in the partial segmenta- tion of the lamnia. N. obtusa (Nathorst) Harris usually has entire leaves, but some have deep inci- sions at long intervals. Asama (1968) pointed out a trend observed in some Japanese forms of Nils- sonia in which the entire-margined Early Jurassic N. orientalis supposedly gave rise to the smaller Late Jurassic N. orientalis var. minor. Asama assumed the series to culminate in the still smaller, Early Cretaceous N. schaumburgensis, in which the lamina is dissected into irregular segments as in N. tenuin- ervis. Asama's article, however, was intended to demonstrate a decrease in laminar size with the pas- sage of time, and did not express any ideas regard- ing cycadean leaf derivation from an entire- margined ancestral form. Other examples of irregular segmentation in Nils- sonia are widely documented, but only one more, N. tenuinervis Seward, need be discussed here because of its established cycadean affinity. Harris (1941) presented compelling evidence that Androstrobus, a cycadean male cone; Beania, a female cone; and the leaf Nilssonia are parts of the same plant and that the plant is certainly a cyead. Thomas and Harris (1960) defined specific relationships with the con- clusion that Beania mamayi, Androstrobus wonna- cotti and Nilssonia tenuinervis, all of Middle Juras- sic age, belonged to the same plant. They (p. 147) stated that the Nilssonia plant is to be classified as a member of the Cycadales, but though nearest to the Zamioideae in its female cones, it should not be included in the Zamioideae because the female cones and probably the male cones differ from the cones of all Zamioideae in their lax and open construction. In the same article, Thomas and Harris (1960, p. 150) concluded that Androstrobus prisma Thomas and Harris and Pseudoctemis lanei Thomas emend. Harris, were male cone and leaf of the same plant. Unlike Nilssonia tenuinervis, Pseudoctenis lanei is in most aspects remarkably similar to leaves of Zami@ or Encephalartos, and Thomas and Harris (1960, p. 153) remarked on the zamioid aspects of this combination of organs : Taking the two fossils together the agreement with the Zamioideae is much enhanced, and without speculating about the unknown female cone, we can say that in Pseudoctenis lanei with Androstrobus prisma we have a plant which is so like them that it may prove to belong to one of these living genera. Harris' conclusions regarding the relationships between Nilssonia tenuinervis, Androstrobus won- nmacotti and Beania mamayi led him to publish a restoration of a cycad bearing those three organs (Harris, 1961, text-fig. 2). It shows the irregular construction of the taeniopterid leaves, with long laminar segments separated by clefts extending as far as the midribs. The morphologic variability discussed here clearly expresses the manifestation of a trend involving evolutionary modification of leaves constructed on the simple taeniopterid pattern. The oldest taeniop- terids are of late Paleozoic age and are mostly entire- margined ; in the Permian, however, the early begin- nings of marginal incision are seen. The trend pro- gresses to the extent that, by Late Triassic time, laminar dissection was complete in at least one ex- ample; by Middle Jurassic time it was complete in several taxa, although resultant foliar segments were not always of uniform size. To be sure, certain anomalies appear in this trend: the apparent pre- cocity in Leptocycas and the conservatism in Nils- sonia. These express the vagaries of morphogenetic rates within a group of rapidly evolving organisms. Nonetheless, the degree of laminar dissection seems generally consistent with decreasing geologic age of materials. Furthermore, the evolutionary process involved stages leading to two of the basic types of cycadean foliage-the Cycas type having single- veined leaflets, and the Zami@ type with multiple- veined divisions. It is even feasible to postulate de- rivation of the Stangeria type of leaf from Paleo- zoic stock, for pinnately compound taeniopterids are known, and each leaflet of a Stangeria leaf is basical- ly equivalent to a simple Taeniopteris leaf in overall | 53 I ~ V/vv/x//////fi///////// W Utz wlll TTT MORPHOLOGIC AND EVOLUTIONARY CONSIDERATIONS 43 shape, venation, and margination. Thus, a pinnately compound taeniopterid leaf could evolve into a Stan- geria leaf with only relatively minor modifications. Further, the bipinnate Bowenia type of leaf might have evolved from a pinnate taeniopterid ancestor through incision of the leaflets into multiveined segments. Intimate association of leaves and fructifications has been of utmost importance in tracing the course of early cyeadean evolution; cuticular comparisons, although not always possible, have also constituted critical evidence. In all instances where fossil leaves and reproductive parts have been suspected or dem- onstrated to represent the same plant and that plant has been interpreted as a cycad, the leaves have been taeniopterid or almost certainly of taeniopterid de- rivation. This circumstance strengthens the proposi- tion that the cyeads arose from taeniopterid pterido- sperms and that modern cycad leaves have been derived from either simple or pinnate taeniopterid leaves through relatively little evolutionary modifi- cation, or through an evolutionary process no more complicated than laminar dissection. Substantially the same statement may be made in regard to cycadeoid foliage. Taemiopteris has long been accepted as an artificial (form) genus repre- senting Cycadales, Bennettitales, and Pentoxylales; nomenclatural and taxonomic adjustments have been made to accommodate the situation (Harris, 1969, p. 68; Sahni, 1948, p. 50). Although conclusions cannot yet be drawn relative to a possibly common ancestry between the Cycadales and Bennettitales or the age of the ancestral complex, it seems reasonable, on the basis of present knowledge of Bennettitalean leaves, to suppose that they too arose from a simple taeniop- FIGURE 11.-Suggested evolution of cycadalean leaves begin- ning with (A), an entire-margined ancestral taeniopterid leaf and, through progressively deeper incision of margins, resulting in two basic types of cycadalean leaves (D, cyead- aceous type; I, zamiaceous type). All diagrams are based on known fossil or modern forms. The cycadaceous lineage, indicated by heavy arrows, consists of forms in which marginal teeth and ultimate foliar segments involve a sin- gle vein each (B, cf. Taemiopteris serrulata; C, cf. T.? koreanensis; D, cf. Cycas sp.). The zamiaceous lineage, in- dicated by light arrows, consists of forms in which marginal incisions and ultimate foliar segments involve several veins each (E, cf. Taemiopteris? serrata; F, cf. T. pseudobrevis or Nilssonia incisoserrata; G, cf. Nilssonia tenuinervis or N. polymorpha; H, cf. N. tenuicaulis; I, cf. Zamia sp.) ; forms such as E and G might have been derived independ- ently from A, whereas H may have arisen from either F or G. Both cycadaceous and zamiaceous lineages evolved substantially concurrently. Ages of forms: A, Pennsyl- vanian-Permian; B, E, Permian; C, Triassic; F, G, Triassic- Jurassic; H, Jurassic; D, I, Mesozoic-Cenozoic. terid leaf form through a process parallel to that proposed here for the cycads. The sum of presently available evidence, then, suggests that Taemiopteris constitutes a logical source of evolutional raw ma- terial for derivation of the open-veined cycadophytic leaf, be it bennettitalean or cycadalean. CONCLUSIONS The morphology, age, and geographic occurrences of the fossils discussed here combine to present a significant set of conclusions, bearing on various as- pects of the history of some of the gymnosperms. Archaeocycas and Phasmatocycas are sufficiently cycadlike to be classified among the cycads; Phas- matocycas is particularly convincing because of its twin-ranked, well-preserved seed cuticles. At the same time, these genera are so logically derivable from Spermopteris that the cycadean lineage is clearly extended to the Pennsylvanian pterido- sperms; this conclusion corroborates the opinions of Harris (1961) and others and injects additional credence into the few known occurrences in the Pale- ozoic of petrified wood with cycadlike anatomical features (Cycadoxylon; Andrews, 1940). These fos- sils eminently represent the pre-Jurassic "progres- sive evolution" postulated by Harris (1961, p. 322). Thus, the stratigraphic continuity of the cycadalean lineage is extended from the Upper Triassic to the Upper Pennsylvanian, and the cycads join the coni- fers as one of the two oldest groups of extant seed plants. On the basis of present information, we cannot know when the spermopterid ancestors of the cycads ceased to be pteridosperms and became cycads, even though the relationship between the two taxa seems clear. Concordantly, the possible impact of the Paleo- zoic fossils on the broad classification of the cycads cannot yet be assayed, and a revised system will not be attempted here; such a system will depend ulti- mately on the acquisition of knowledge of other parts of the Permian and Triassic cycads, such as the male organs and vascular anatomy. Nonetheless, Spermopteris, Phasmatocycas and Archaeocycas ap- pear to uphold the concept that Cycas, with its leaf- like megasporophylls, is the primitive genus among extant cycadales; this appraisal has been the popu- larly accepted one, although Thomas (1947, p. 250; 1953, p. 582) held the contrasting view that Zamia, rather than Cycas, is the primitive member of the Cycadales. The evidence for taeniopterid foliation among the early cycads and the indications that gradual lami- 44 PALEOZOIC ORIGIN OF THE CYCADS nar dissection of entire leaves produced the pinnate types of modern cycad foliage are an important facet of this study, but they also pose a perplexing problem-the origin of taeniopterid foliage-whose solution will necessitate extensive searches of pre- Pennsylvanian rocks. Taemiopteris appears "sudden- ly" in Pennsylvanian rocks and gives no clues as to its ancestry ; it is a conspicuous element in the Per- mian and Triassic floras. In keeping with his theories of leaf evolution, Asama (1960, p. 258) be- lieves that the simple leaf of Taemiopteris is an ex- ample of an "enlarged leaf," formed from an origi- nally pinnate form. Kon'no, Asama, and Rajah (1970, p. 559), however, stated that no substantial evidence yet exists in the Cathaysian flora for this opinion, and I know of none in the Angara. Euro- pean, or American floras. It is conceivable, of course, that Asama's notions are correct and that intermedi- ate forms ancestral to Taeniopteris simply have not been found. Geologic age and geologic occurrence of the fossils discussed here are significant considerations, for they contribute to the ever-growing number of indi- cations that the Lower Permian deposits of the southwestern United States contain proof that this area was an important center of botanical evolution and radiation. Late Pennsylvanian plants such as Spermopteris emphasize the importance of that time in respect to the beginnings of major innovations in plant form. Wolfcampian time produced the Permi- an guide fossil, Callipteris, but this genus has no more morphological significance than that of a younger type of seed-fern with a "marker" type of foliage. Leonardian time, however, saw the introduction into the American botanical community of radically different forms, some of which leave no grounds for interpretation as to antecedents or progeny; Gi- gantopteris americana White and Russellites taemi- ata (Darrah) Mamay are prime examples. Repro- ductive methods of these two genera are yet specu- lative. They are probably seed plants, but this spec- ulation is based only on their age and the otherwise demonstrated progress of plant evolution as shown by fossils. Thus far, the advent of cycads is the most im- portant botanical event attributable to the Leon- ardian or even to the Permian System. It emphasizes the importance of Leonardian rocks as the physical representation of a time interval worthy of close examination in the search for progressive evolution among higher plants. REFERENCES CITED Andrews, H. N., Jr., 1940, On the stelar anatomy of the pteridosperms, with particular reference to the second- ary wood: Missouri Bot. Garden Annals, v. 27, no. 1, p. 51-118. Arnold, C. A., 1953, Origin and relationships of the cyeads: Phytomorphology, v. 3, nos. 1-2, p. 51-65. Asama, Kazuo, 1960, Evolution of the leaf forms through the ages explained by the successive retardation and neoteny: Tohoku Univ. Sci. Repts., 2d ser. (Geology), spec. v., no. 4 (Shoshiro Hanzawa Memorial Volume), p. 252-280. 1968, The evolution of Nilssonia orientalis: Tokyo Nat. Sci. Mus. Mem., no. 1, p. 33-36. (In Japanese with Eng- lish abstract and title.) Barnard, P. D. W., 1967, Two new plants from the Upper Triassic of North East Afghanistan: Riv. Italiana Pale- ontologia Stratigrafia, v. 73, no. 3, p. 723-728. 1970, Upper Triassic plants from the Kalawch River, , North-East Afghanistan: Italian Exped. Karakorum (K°") and Hindu Kush, Sci. Repts. 4, Paleontology- Zoology-Botany, v. 2, Fossils of north-east Afghanistan, p. 25-40. Bassler, Harvey, 1916, A cyeadophyte from the North Ameri- can coal measures: Am. Jour. Sci., 4th ser., v. 42, no. 247, p. 21-26. Carlquist, Sherwin, 1969, Toward acceptable evolutionary in- terpretations of floral anatomy: Phytomorphology, v. 19, no. 4, p. 332-362. Carpenter, F. M., 1952, Fossil insects, in U.S. Dept. Agricul- ture, Yearbook of Agriculture, 1952: Washington, D.C., U.S. Govt. Printing Office, p. 14-19. 1971, Adaptations among Paleozoic insects: North Am. Paleont. Convention Proc., pt. 1, 1236-1251. Chadefaud, M., 1936, Lorigine cycadéene des Angiospermes et la classification générale des Phanerogams: Soc. Bot- anique France Bull., v. 83, p. 586-592. Chamberlain, C. J., 1930, Elements of plant science: New York, McGraw-Hill Book Co., 394 p. 1935, Gymnosperms: structure and evolution: Chicago, Univ. Chicago Press, 483 p. 1965, The living cycads: New York, Hafner Publishing Co., 172 p. (facsimile of 1919 ed.). Cheong, C. H., 1969, Stratigraphy and paleontology of the Samcheog coalfield, Gangweondo, Korea (1): Geol. Soc. Korea, Jour., v. 5, no. 1, p. 13-56. Cheong, Chang Hi, and Lee, Ha Young, 1970, Some fossil plants from the Jangseong Formation of the Gangweon coal mine, Korea: Geol. Soc. Korea Jour., v. 6, no. 1, p. 13-27, pl. 1-4. Coulter, J. M., and Chamberlain, C. J., 1910, Morphology of gymnosperms: Chicago, Univ. Chicago Press, 458 p. Crepet, W. L., 1972, Investigations of North American cyca- deoids: pollination mechanisms in Cycadeoidea: Am. Jour. Botany, v. 59, no. 10, p. 1048-1056. Cridland, A. A., and Morris, J. E., 1960, Spermopteris, a new genus of pteridosperms from the Upper Pennsylvanian Series of Kansas: Am. Jour. Botany, v. 47, no. 10, p. 855-859. Cronquist, Arthur, 1968: The evolution and classification of flowering plants: Boston and New York, Houghton Miffiin Co., 396 p. REFERENCES CITED 45 Daugherty, L. H., 1941, The Upper Triassic flora of Arizona, with a discussion of its geologic occurrence by H. R. Stagner: Carnegie Inst. Washington Pub. 526, 108 p. 34 pl. Delevoryas, Theodore, 1968a, Investigations of North Ameri- can cycadeoids-structure, ontogeny and phylogenetic considerations of cones of Cycadevidea: Palacontogra- phica, v. 121, Abt. B, Lief. 4-6, p. 122-133, pl. 34-38. 1968b, Some aspects of cycadeoid evolution: Linnean Soc. London Jour., Botany, v. 61, no. 384, p. 137-146; reprinted in K. L. Alvin and others, eds., Studies on fossil plants, London and New York, Academic Press. Delevoryas, Theodore, and Hope, R. C., 1971, A new Triassic cyead and its phyletic implications: Postilla, no. 150, p. 1-21. Dunbar, C. O., 1923, Kansas Permian insects, pt. 1. The geologic occurrence and the environment of the insects; with a description of a new paleodictyopterid, by R. J. Tillyard: Am. Jour. Sci., 5th ser., v. 7, no. 39, p. 171- 209, 2 pls. Dunbar, C. O., chm., and others, 1960, Correlation of the Permian formations of North America: Geol. Soc. Ameri- ca Bull., v. 71, no. 12, pt. 1, p. 1763-1805, chart 7. Eames, A. J., 1959, The morphological basis for Paleozoic origin of the angiosperms: Internat. Bot. Cong., 9th, Montreal, 1959, Proc., v. 2, Abstracts, p. 100-101. 1961, Morphology of the angiosperms: New York, Mc- Graw-Hill Rook Co., 518 p. Emberger, Louis, 1944, Les plantes fossiles dans leurs rap- ports avec les végétaux vivants (éléments de paléobotani- que et de morphologie comparée): Paris, Masson et Cie, 492 p. (2d ed., 1968, 758 p.). Florin, Rudolf, 1933, Studien iiber die Cycadales des Meso- zoikums nebst Erorterungen liber die Spaltoffnungsap- parate der Bennettitales: Kgl. Svenska Vetenskapsakad. Handl., ser. 3, v. 12, no. 5, 134 p., 16 pl. Gaussen, Henri, 1944, Les Cycadales, Chap. 3 of Les gymno- spermes actuelles et fossiles, fase. 1-2: Toulouse, Fac. Sci., [106] p. (First part of a monographic study of fos- sil and living gymnosperms.) Gothan, Walther, and Weyland, Hermann, 1954, Lehrbuch der Paliobotanik: Berlin, Akad.-Verlag, 535 p. (2d. ed., 1964, 594 p.). Gould, R. E., 1971, Lyssoxylon grigsbyi, a cyead trunk from © the Upper Triassic of Arizona and New Mexico: Am. Jour. Botany, v. 58, no. 3, p. 239-248. Halle, T. G., 1927, Paleozoic plants from Central Shansi: Palaeontologia Sinica, ser. A, v. 2, pt. 1, 316 p., 64 pl. 1929, Some seed-bearing pteridosperms from the Per- mian of China: Kgl. Svenska vetenskapsakad. Handl., ser. 3, v. 6, no. 8, 24 p. Harris, T. M., 1932a, The fossil flora of Scoresby Sound, East Greenland-Pt. 2, Description of seed plants in- certae sedis, together with a discussion of certain cyeado- phyte cuticles: Medd. Grgnland, v. 85, no. 3, 114 p. 9, pls. 1932b, The fossil flora of Scoresby Sound, East Green- land-Pt. 3, Caytoniales and Bennettitales: Medd. Grgn- land, v. 85, no. 5, 133 p., 19 pls. 1941, Cones of extinct Cycadales from the Jurassic rocks of Yorkshire: Royal Soc. London Philos. Trans., ser. B, v. 231, no. 577, p. 75-98, 2 pl. 1954, Mesozoic seed cuticles: Svensk Bot. Tidskr., v. 48, no. 2, p. 281-291. 1961, The fossil cyeads: Palaeontology, v. 4, pt. 3, p. 313-323. 1964, The Yorkshire Jurassic flora-Part 2, Caytoni- ales, Cycadales and pteridosperms: London, British Mus. (Nat. History), 191 p., 7 pl. 1969, The Yorkshire Jurassic flora-Part 3, Bennet- titales: London, British Mus. (Nat. History), 186 p., 7 pl. Johnson, L. A. S., 1959, The families of cycads and the Zamiaceae of Australia: Linn. Soc. New South Wales Proc., v. 84, pt. 1, no. 389, p. 64-117. Kon'no, Enzo, Asama, Kazuo, and Rajah, S. S., 1970, The Late Permian Linggiu flora from the Gunong Blumut area, Johore, Malaysia: Nat. Sci. Mus. [Tokyo] Bull., v. 13, no. 3, p. 491-580. Kriusel, Richard, 1949, Koniferen und andere Gymnospermen aus der Trias von Lunz, Nieder-Osterreich: Palaeonto- graphica, v. 89, Abt. B, If. 1-8, p. 35-78, pl. 6-18. 1950, Versunkene Floren-eine Einfiihrung in die Paliobotanik: Frankfurt am Main, Waldemar Kramer, 152 p., 63 pls. 1953, Ein neuses Dicoonitocarpidium aus der Trias von Lunz: Senckenbergiana, v. 34, no. 1-8, p. 105-108, 1 pl. Leppik, E. E., 1971, Paleontological evidence on the morpho- genetic development of flower types: Phytomorphology, v. 21, no. 2-3, p. 164-174. Lilienstern, H. R., 1928, Dioonites pennaeformis Schenk, eine fertile Cycadee aus der Lettenkohle: Paléont. Zeitschr., v. 10, p. 91-107, pl. 5-6. Magdefrau, Karl, 1956, Paliobiologie der Pflanzen: 3d ed., Jena, Gustav Fischer 443 p. Mamay, S. H., 1954, A Permian Discinites cone: Washington Acad. Sci. Jour., v. 44, no. 1, p. 7-11. 1960 (1962), Padgettia, a new genus based on fertile neuropteroid foliage from the Permian of Texas: Paleo- botanist, v. 9, no. 1-2, p. 53-57. 1966, Tinsleya, a new genus of seed-bearing callipterid plants from the Permian of north-central Texas: U.S. Geol. Survey Prof. Paper 523-E, p. E1-E15, pl. 1-3. 1968, Russellites, new genus, a problematical plant from the Lower Permian of Texas: U.S. Geol. Survey Prof. Paper 5983-I, p. I1-I15, pl. 1-3. 1969a, Cyeads-fossil evidence of late Paleozoic origin: Science, v. 164, no. 3877, p. 295-296. 1969b, Fossil evidence of Late Paleozoic origin of the cycads: Internat. Bot. Cong., l1th, Seattle, 1969, p. 139. 1971, Cyceadales, in McGraw-Hill yearbook of science and technology: New York, McGraw-Hill Book Co., p. 158-160. 1973, Archaeocycas and Phasmatocycas-new genera of Permian cycads: U.S. Geol. Survey Jour. Research, v. 1, no. 6, p. 687-689. Meeuse, A. D. J, 1963, The so-called "megasporophyll" of Cyeas-a morphological misconception. Its bearing on the phylogeny and the classification of the Cycadophyta: Acta Botanica Neerlandica, v. 12, p. 119-128. Meyers, P. A., and Morley, H. T., 1929a, Geologic map of Jones County, Texas: Texas Univ. Bur. Econ. Geology map, scale 1: 48,000; revised, 1937. 46 PALEOZOIC ORIGIN OF THE CYCADS 1929b, Geologic map of Taylor County, Texas: Texas Univ. Bur. Econ. Geology map, scale 1:48,000; revised 1987. Miser, H. D., 1954, Geologic map of Oklahoma: U.S. Geol. Survey in cooperation with Oklahoma Geol. Survey, scale 1:500,000. Moore, R. C., chm., and others, 1944, Correlation of Penn- sylvanian formations of North America: Geol. Soc. America Bull., v. 55, no. 6, p. 657-706, pl. 1 (chart 6 of series). Read, C. B., and Mamay, S. H., 1964, Upper Paleozoic floral zones and floral provinces of the United States: U.S. Geol. Survey Prof. Paper 454-K, p. K1-K35. Romer, A. S., 1958, The Texas Permian redbeds and their vertebrate fauna, in Westoll, T .S., ed., Studies on fossil vertebrates: London Univ., Athlone Press, p. 157-179. Sahni, Birbal, 1920, On the structure and affinities of Acmo- pyle Pancheri, Pilger: Royal Soc. London Philos. Trans., ser. B, v. 210, p. 253-310, pl. 9-11. 1948, The Pentoxyleae: a new group of Jurassic gymnosperms from the Rajmahal Hills of India: Bot. Gaz., v. 110, no. 1, p. 47-80. Schimper, W. P., 1870-1872, Traité de paléontologie végétale ou la flore du monde primitif: Paris, J. B. Bailliére et Fils, 3 v., and atlas containing 110 pls. Schuster, J., 1932, Cycadaceae, in Engler, Adolph, Das Pflanzenreich: Leipzig and Berlin, W. Engelmann, v. 4, pt. 1, p. 1-168. Sellards, E. H., 1901, Permian plants-Taeniopteris of the Permian of Kansas: Kansas Univ. Quart., v. 10, no. 1, p. 1-12, pl. 1-4. 1908, Fossil plants of the Upper Paleozoic of Kansas: Kansas Univ. Geol. Survey, v. 9, p. 386-480, pls. 44-69. Seward, A. C., 1917, Fossil plants: London, Cambridge Univ. Press, v. 3, 656 p. Sporne, K. R., 1965, The morphology of gymnosperms: Lon- don, Hutchinson Univ. Library, 216 p. 1971, The mysterious origin of flowering plants: Lon- don, Oxford Univ. Press, 16 p. Taylor, T. N., 1969a, Cyeads-evidence from the Upper Penn- sylvanian: Science, v. 164, no. 3877, p. 294-295. 1969b, Evidence for the presence of cycads in the Upper Pennsylvanian: Internat. Bot. Cong., 11th, Seattle, 1969, Abstracts, p. 216. 1970, Lasiostrobus gen. n., a staminate strobilus of gymnospermous affinity from the Pennsylvanian of North America: Am. Jour. Botany, v. 57, no. 6, pt. 1, p. 670- 690. Thomas, H. H., 1947, The history of plant form: British Assoc. Adv. Sci. Reports, v. 4, no. 15, p. 248-254. 1953, The relationships between the Bennettitales and Cycadales: Internat. Bot. Cong., 7th, Stockholm, 1950, Proc., p. 581-582. Thomas, H. H., and Harris, T. M., 1960, Cycadean cones of the Yorkshire Jurassic: Senckenbergiana Lethaea, v. 41, no. 1-6, p. 139-161, pl. 1-4. Uhl, N. W., and Moore, H. E., 1971, The palm gynoecium: Am. Jour. Botany, v. 58, no. 10, p. 945-992. U.S. Geological Survey, 1968, Geological Survey Research, 1968: U.S. Geol. Survey Prof. Paper 600-A, 371 p. Von Gutbier, C. A., 885, Abdriicke und Versteinerungen des Zwickauer Schwarzkohlengebirges und seiner Umgebun- gen: Zwickau, G. Richter, 80 p., 11 pl. White, C. D., 1912, The characters of the fossil plant Gigan- topteris Schenk and its occurrence in North America: U.S. Natl. Mus. Proc., v. 41, no. 1878, p. 498-516. 1929, Flora of the Hermit shale, Grand Canyon, Ari- zona: Carnegie Inst. Washington Pub. 405, 221 p., 51 pl. Wilson, J. A., 1953, Permian vertebrates from Taylor County, Texas: Jour. Paleontology, v. 27, no. 3, p. 456-470. INDEX [Italic page numbers indicate descriptions and major references] Page A abnormis, Taeniopteris ________________ 7, 8 americana, Gigantopteris ___ 5, 8, 24, 26, 27, 44 Androstrobus 41 PTEMG: 1 41 wonmacotti 24, 41 zamioides angustifolia, Danacopsis Anmmularia ________________ Anomozamites --..... Nilssoni, Pterophyllum ____________ (Anomozamites) Nilssoni, Ptcrophyllum 40 ADMEDTE ! 122.004 240000 Araucarites Archaeocycas __ 4, 6, 8, 10, 11, 12, 14, 17, 18, 19, 22, 27, 28, 29, 30, 31, 32, 33, 84, 35, 36, 38, 39, 43 whiter _______ 8, 11, 12, 31, 32,; text-fig. 3; pls. 1, 2 Aspidiopsis 27 B (2 uuu 2, 15, 19, 36, 41 15 mamayi = __ Bennetbibes Benmettitolepis ____________________ AACEYIOLA ce BJUUNE "22200 ieee 00 enne he cu nees 2, 17, 38, 39 simplex __ ____ 25, 38 BOENIG CLSS I2 CLIA m RL s dne Plc wss 43 C Callipteris | 17, 26, 27, 28, 44 Cardiocarpon 27 Corpolithes | 26 Ceratozamia | 20 circinalis, Cycas -= 24 Cordatanthus | 27 COFAMLER ! 2 cc.... 27 coriacea, Spermopteris ________________ 2 Taentopterig ____..___...._........ 2, 16 linearis, Taemiopteris _. 16 ! Lucu nnn nn doin nee iu Cycadospadiz dactylota Hennocqueti _______________________ 25 integer _.... -_ 25, 26 Moraeanus .________________. 25 yochelsoni __ __ 26, 33; pl. 2 SPME ne aA er canne ch ennis neue awe 6, 19 CUCRUORYION 43 CyCRE 3, 4, 12, 19, 25, 26, 80, 31, 32, 33, 34, 38, 40, 43 24 Medi 22 22 eon nle n nah on becca ines 3 Fegoluth 3, 33 $D rone essas s 43 D dactylota, Benmettitolepis _____________ 25 Cycadospadiz ___. isk e Me 25 Page Danacopsis angustifolia _______________ 36 Dambrecia | 27 Dioom edule ___________________________ 3, 20 Dioomites pennaeformis _______________ 36 Dioonitocarpidium _ __ - 2, 17, 36, 37, 38, 39 Ieuperianum | _____________________ 37, 38 liliensterni | 37, 38 pennaeforme - __________________ 36, 37, 38 Diplothmema = _________________________ 27 27 E edule, Dioom __________________________ 3, 20 Emeephalartos | ________________________ 20, 41 F floridama, Zamia ______________________ 3 G gigantea, Taeniopteris ________________ 38 Gigantopteris ______________________ 26, 27, 28 americana | _____________ 5, 8, 24, 26, 27, 44 M. 27 - Ansemes 18 Glemopteris | _______________ -- 17, 26, 27, 28 Gomphostrobus | _______________________ 27 gracilis, Beamia _______________________ 15 Leptocycas 6 grigsbye, Lyssoxylon 6 H Hennocquei, Cycadospadiz ____________ 25 I incisoserrata, Nilssonia _______________ 41, 43 integer, Cycadospadiz ___ -__- 25, 26 Palaeocycas _ ________ -__. 20, 25, 38 kansama, Phasmatocycas _ 12, 21, 23, 26, 29, 30, 32; text-fig. 4; pls. 3, 4 keuperianum, Diconitocarpidium ______ 37, koreanensis, Taemiopteris _____________ 40, Lagenostoma - _________________________ Lomazi ______ lamei, Pseudoctenis . Lasiostrobus __________________________ \ Leptocycas GFAOIHI® liliensterni, Diconitocarpidium __ linearis, Taemiopteris coriacea _________ Lobatannularia Lomazi, Lagenostoma _________________ Lyginopteris oldhamia ________________ Lyssoxylon grigsbye __ ___________________ 5, 6, 29, 40, 38 43 41 Page M ! 1.2.0 0000. usind paced 20, 24 MIGHOUWE -.. . ...- 2 uous eee nade aos nalang 8 BEWMG 41 media, Cycas _____________ 2 8 minor, Nilssonia orientalis __ P 41 miquelii, Macrozamia _____ M 3 Moreanmus, Cycadospadiz ______________ 25 multinervis, Taeniopteris _____________ 7,19 N - 222. cece. 27, 28 newberriana, Taeniopteris . 17 Nilssomia ________. 41 incisoserrata __ 41, 43 ODHIGG - 202000 uns nen tas wud s chew 41 OrIENIINNS . ... 212000000000 nas bu tine 41 HMMON: 22000000000, ves SE 41 41, 43 schaumburgensis 41 fenitignitlée ol 43 ~ 12 0s elt 41, 438 - .s. 41 : .. coll o rda ve anis wai 26 | .. ue 34 0 obtusd, Nilesonts ._...___..____...__.. 41 Odontopteris ______ & __ 17, 26, 27, 28 oldhamia, Lyginopteris _______________ 19 orientalis, - Nilssonta | _______________.. 41 minor, . NHegonkt 41 P . onces duties eu 12 _ 2, 4, 25, 26, 36, 38, 39 -_ 20, 25, 38 Palaeocycas _. integer __ Paracycas ___ £s 40 Pebopleris 26, 27 pennaeforme, Dioonitocarpidium ____ 36, 37, 38 pennaeformis, Dioonites _______________ 36 Phasmatocycas __ 6,13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 28, 24, 27, 28, 29, 30, 31, 32, 33, 34, 36, 38, 39, 43 kansama - ________- 12, 21, 23, 26, 29, 30, 82; text-fig. 4; pls. 3, 4 spectabilis = _____ -22, 30, 32; pl. 5 §D . Mel cee 27; pl. 2 - 1... -_. locos cess 6 planchurdi. 6 planchardi, Plagiozamites _____________ 6 . polymorpha, Nilssonia __. polysacci, Lasiostrobus ________________ 6 prisma, Androstrobus _________________ 41 pseudobrovis, Taemiopteris _________ 40, 41, 43 PSEHIOGEOHS .L. usta ee. 40 lamei 41 Pterophyllum 40 48 Page R CYUCGB a- cece os 3, 83 Rhabdocarpos 27 RHSSEHIEE 22s a anns 27, 28 | Cee cel lc een ne owen cose 44 S schaumburgensis, Nilssonia _. -% 41 serrata, Taeniopteris ______ __ 40, 48 Berratopteris 40 serrulata, Taemiopteris _____________ 39, 40, 43 N Al- onn 27 Sigillartostrobuig 27 simplex BJAOMG: 25, 38 spectabilis, Phasmatocycas ___ 22, 30, 32; pl. 5 Spermopteris ______ 2, 8, 4, 6, 7, 8. 10, 12, 18, 1% 18, 21, 22, 23, 26, 28, 29, 30, 31, 32, 34, 36, 43, 44 * 221 ILI cnc nee 2 Sp INDEX Page Sphenophyllum ________________________ 26 Sphenopteris | _________________________ 27 Stangeria Stigmaria Supaia ___________ T taeniata, Russellites __________________ 44 Taeniopteris _____ 2, 7, 19, 24, 26, 28, 29, 30, 39, 41, 43, 44 abmormig | 7, 8 coriacea 2, 16 HEATS 16 GIGU@IMENE koreanensis __ multinervis ___ newberriana ___ pseudobrevis ___________________ 40, 41, 43 SETYAEL | 40, 43 serrulata _. 39, 40, 43 §D 17 Page tenutonults, Nilssonia ._.........._..._ 43 tenuinervis, Nilssonia _________________ 41, 43 thomasi, Anomozamites _______________ 41 TSREIEYG 1...1... .u 200000000002 12 U undulata, Nilssonia ___________________ 41 w ",. . c culo unless oe 26, 27 whitei, Archacocycas ________ 8, 11, 12, 31, 32; text-fig. 3; pls. 1, 2 wonnacotti, Androstrobus _____________ 24, 41 Y yochelsoni, Cycadospadiz ________ 26, 33; pl. 2 Z ONN S cel, r 40, 41, 43 ROHAURE . . 02 202 Leb eac 3 SP LOL I= ables bacs ae a me al nne uh 43 zamioides, Androstrobus ______________ 24 ¢ U.S. GOVERNMENT PRINTING OFFICE: 1976 O-211-317/1 44 PLATES 1-5 Contact photographs of the plates in this report are available, at cost, from U.S. Geological Survey Library, Federal Center, Denver, Colorado 80225 PLATE 1 FIGURES - 1-5. Archaeocycas whiter Mamay (p. 8). 1. Megasporophyll exposed in oblique-lateral position, showing carbonaceous remnants of original tis- sues, indusiumlike laminar covering of ovules and faint striations suggestive of fine, parallel vena- tion. Paratype, USGS 8959-1b. x 8.5. 2. Counterpart of specimen illustrated in figure 1, showing adaxial view of six closely appressed ovules. Note incomplete fold in lamina at upper (right) edge of most distal ovule (arrow), indicating flat- tening out of inrolled lamina covering the ovules. Paratype, USGS 8959-1la. x 8.5. 3. Mold of adaxial side of megasporophyll, showing stout midrib, five pairs of oppositely attached ovules, and distal expansion of sterile lamina. Paratype, USGS 8877-3. x 6. 4. Cast of adaxial side of megasporophyll, showing stout midrib, five pairs of opposite ovules, and coali- fied residue of laminar tissue that originally intruded between contiguous ovules. Several ovules show circular scars of attachment; these are best shown in the distal (uppermost) pairs of ovules. Para- type, USGS 8877-2. X 6. 5. Cast of adaxial side of megasporophyll, with nine ovules (lower right ovule broken off). Note fairly strong development of sterile distal lamina, with fold extending distally and diagonally from midrib. Holotype, USGS 8877-1. x 6. GEOLOGICAL SURVEY PROFESSIONAL PAPER 934 PLATE 1 i 494 # ARCHAEOCYCAS FIGURE 1. 5-7. PLATE 2 ?Phasmatocycas sp. (p. 21). Fragmentary impression of stout axis with evidence of approximately 12 pairs of opposite seedlike bodies and suggestion of outline of surrounding foliar lamina. USGS 10064-1. x 2. Spermopteris sp. (p. 7). Fragmentary impression of taeniopterid leaf, bearing two marginal rows of ovoid swellings suggestive of immature ovules. USGS 6233-1. x 2. Cycadospadix yochelsoni n. sp. (p. 24). Small appendage with broad peduncle and deeply pinnatifid lamina. Holotype, USGS 10057-1. x 3. ?Cyeadean male cone (p. 23). Fragmentary mold of conelike organ, with deep rhomboidal impressions, probably representing distal faces of microsporophylls. Note distally curved spinelike molds, apparently impressions of rigid attenuated termini of microsporophylls. USGS 10064-2. x 1. Archaeocycas whiter Mamay (p. 8). 5. Mold of adaxial side of megasporophyll with four pairs of ovules. Each ovule shows a circular scar of attachment near the rachis; three lower left ovules in turn have tiny central markings, prob- ably representing vascular strands from lamina to base of ovules. USGS 8959-2a. x 6. 6. Counterpart of specimen shown in figure 5. USGS 8959-2b. x 6. 7. Cast of megasporophyll, exposed in lateral view. Most proximal ovule partly broken away; contiguous ovule showing attachment scar near midrib. Note expanded base of midrib, indicative of perpendic- ular attachment to an axis; also note fragmentary but apparently well-developed sterile lamina at right. Paratype, USGS 8877-4. x 6. GEOLOGICAL SURVEY PROFESSIONAL PAPER 934 PLATE 2 ?PHASMATOCYCAS, SPERMOPTERIS, CYCADOSPADIX, AND ARCHAEOCYCAS PLATE 3 FIGURES 1-5, Phasmatocycas kansana Mamay (p. 12). 1,2. Counterparts of holotype, showing stout axis with two lateral rows of seeds, one displaced from orig- inal position. Figure 1, USGS 8869-1a; figure 2, USGS 8869-1b. x 4. 3. Seed on USGS 8869-1b, oriented with apex upward. Spherical glandlike bodies at base of seed (indi- cated by arrows) one at either side. X 21. 4. Part of USGS 8869-1b, enlarged. Second complete seed from top at left side appears in figure 3. X 9. 5. Line tracing of figure 4, showing distribution of glandlike bodies relative to seeds. Solid black cir- cles represent actual glands; unshaded circles represent impressions of glands, X 9. PROFESSIONAL PAPER 934 PLATE 3 GEOLOGICAL SURVEY PHASMATOCYCAS PLATE 4 FigUrES - 1-5. Phasmatocycas kansana Mamay (p. 12). Chemically cleared cuticular systems of three seeds removed from holotype (USGS 8869-1b). 1,2. Apical region of seed fragment, showing short, relatively broad micropyle and two cuticles. Outer cuticle (inner integument) is thin and delicate; inner layer (nucellar integument) is thicker, with isodiametric, heavy-walled cells. Note absence of any pollen-chamber mechanisms. Fig 1, X 100; fig. 2, X 30. 3. Nearly complete specimen, showing double nature of cuticle to approximate "shoulders," and heavy megaspore membrane. Light area below center represents damage in which part of megaspore membrane is missing; darkest streaks represent folds in megaspore membrane, in which at least three thicknesses of membrane are present. X 30. 4,5. Two views of third specimen, fig. 5 showing details of micropylar area, double cuticle, and top of megaspore membrane seen as a dark are across bottom of figure. Fig. 4, X 30; fig. 5, X 100. GEOLOGICAL SURVEY PROFESSIONAL PAPER 934 PLATE 4 PHASMATOCYCAS PLATE 5 FIGURES - 1-4. ?Phasmatocycas spectabilis Mamay, n. sp. (p. 22). 1. Unusually long, stout axis with about 60 seeds, borne in two lateral rows. Holotype, USGS 8959-3. Xx 1. 2-4. Paratype specimens, clearly demonstrating lack of evidence of seed attachments between lateral rows. Fig. 2, USGS 8877-5, x 1; fig. 3, USGS 8877-6, x 2; fig. 4, USGS 8877-7, X 1. 5-7. Specimen of Taemiopteris sp., associated with holotype of Phasmatocycas kansana, and showing abundant spherical, intercostal, resinoid glandlike bodies. Compare those in fig. 7 with similar bodies shown in pl. 3, fig. 3. Figure 5, X 2; fig. 6, X 5; figure 7, X 20. USGS 8869-2. GEOLOGICAL SURVEY PROFESSIONAL PAPER 934 PLATE 5 ?PHASMATOCYCAS AND TAENIOPTERIS SOLIDIFICATION OF ALAE W EVEL LVLU mam 1g --- The Eruption of August 1963 and the Formation of Alae Lava Lake, Hawaii By DALLAS L. PECK and W. T. KINOSHITA SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII CEOLOGECAL SURVEY PROFESSIONAL PAPER 035 A A detailed description of the formation and surface features of a thin, ponded basalt flow UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1976 OCT 14 1976 semtoi de % Wh UNITED STATES DEPARTMENT OF THE INTERIOR THOMAS S. KLEPPE, Secretary GEOLOGICAL SURVEY L V. E. McKelvey, Director Library of Congress Cataloging in Publication Data Peck, Dallas Lynn, 1929- The eruption of August 1963 and the formation of Alae Lava Lake, Hawaii. (Solidification of Alae Lava Lake, Hawaii) (Geological Survey professional paper ; 985-A) Bibliography : p. Supt. of Does. no.: I 19.16 1. Kilauea. 2. Alae Lava Lake, Hawaii. I. Kinoshita, W. T., joint author. II. Title. III. Series. IV. Series: United States. Geological Survey. Professional paper ; 985-A. QE523.K5P4 557.3'08s [551.2'1] 76-608005 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 Stock No. 024-001-02848-4/Catalog No. I 19.16:985 A CONTENTS Metric-Emglish equivalents ________________________ M N2 - 2. -o co ol- a ans abl aan eos cana nan aenaanes al "= 2 2. cou. 222 nce co ee eles aa ot, _ l. c CGeolppicuisSetting Pre-eruption uplift of the summit region and seismic CLV ILY S- - 2 2 o PHL cleo o ole nual al ae a au mn a an daw ies Description of the August 21-23, 1963, eruption _____ FIGURE 10. 11. 12. 13. 14. 15. 16. 27. Page IV A1 1 2 8 5 Temperature measurements L_._______LL____________ Deflation of the summit region and seismic activity during and after the eruption ......--.-.._.______.. Physical features of the lake Surface features other than joint cracks ________ Inferred circulation of lava during the eruption __ Development of joint cracks ._.___________._ __. References Clled . ILLUSTRATIONS Map showing the five volcanoes consti- tuting the island of Hawaii ________ Index map of the summit and upper east rift zone of Kilauea Volcano as it ap- peared in 1968 Map of Alae Crater showing distribution of volcanic rocks from the August 1963 eruption East-west ground tilting of the Kilauea summit area from Dec. 1962 to Sept. 1963 Chronology of events during the August 1963 eruption of Kilauea ___________ Eruption in Alae Crater at 19"20", Au- gust 21, 1968 Volume of lava in Alae lava lake during the August 1963 eruption __________ Eruption in Alae Crater at 11"10", Au- TED 2 s on e o o ee o a o ar oe aree she we aha iaa ale i h Eruption in Alae Crater at 15", August Po .e rer nees Fountain in Alae Crater at 17"45", Au- PUSH 22 2 o n d el 2 ue ao o alee nene ee wane tase haes te in ee Eruption in Alae Crater at about 18", AUupHist 22 2... ...see cs Eruption in Alae Crater at about 21", Anvevest 22; Jin u unl Eruption in Alae Crater at 22"45", Au- ~. ene Eruption in Alae Crater at 23"16", Au- PULTE 12 o een ule o lle hone ann he e ae a's n in ie ing sa Northwest part of Alae lava lake on 81 2. _ Lull U ece en asana s Changes in altitude of benchmarks along the Chain of Craters Highway be- tween mid-July and early September 1963 Map and cross sections of Alae lava lake, showing contours at the base and margin of the lake Page A3 10 11 12 12 12 13 14 16 FIGURE 18. 19. 20. 21. 22. 28. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. Aerial photograph of August 1963 lake__ Pressure dome in moat at south edge Of Nake Lull n Panorama of Alae lava lake __________ Map of surface features of Alae lava lake exclusive of joint cracks ______ Pressure dome of jumbled slabs near northwest edge of the lake ________ Twisted-ribbon bomb on the surface of Alne 194v8 I8KC ooo eae eee an ae ean nas Cross sections of the surface of the lake Pressure ridge formed from the crust of a linear Surface features exclusive of joint cracks of a central part of Alae lava lake Filamented surface Outer margins of active lava circulation, shear zones, and inferred directions of lava flowage during the evening of August 22, 1908 Cross section along the major axis of Alae lava lake during the evening of August 22 showing inferred lava cir- culation Open joint cracks at the surface of the lava "lake '=: ...ll inn ree tue Map of open joint cracks on the lava lake that formed during drainback near the close of the eruption _____. Map showing cracks on part of the Alae lava lake 1 IO.. Urea en Troughs and crests on part of Alae lava 1@KG ... . clos ol aol oa una alona ae baad Alae lava lake on October 21, 1963 ___ Alae lava lake on August 25, 1964 ___. Areas bared of filamented surface layer near joint cracks in the lake _______ Alae lava lake on August 18, 1965 __. III Page A13 14 15 15 22 24 32 Page A17 18 19 20 21 21 21 22 28 24 25 26 26 27 28 29 30 81 31 32 IV CONTENTS METRIC-ENGLISH EQUIVALENTS Metric unit English equivalent Metric unit English equivalent Length Specific combinations-Continued millimetre (mm) = 0.03937 inch (in) litre per second (1/s) sa 0353 cubic foot per second metre (m) =;:8.28 feet (ft) cubic metre per second kilometre (km) == .62 mile (mi) per square kilometre [ (m#/s) /km*] = O1L.47 cubic feet per second per Area square mile [ (ft?/s) metre per day (m/d) = 8.28 feet per day (hydraulic square metre (m) = 10.76 square feet (ft") conductivity) (ft/d) square kilometre (km) = - .386 - square mile (mi?) metre per kilometre y hectare (ha) = 247 acres (m/km :: ~5.28 feet per mile (ft/mi) kil?me;11'le per hour $112 ' £ oal km/h) s 4 oot per secon t/s Volume metre per seciiond (an/s) = -8.28 feet per second 7 as .0 (oj 3 metre squared per day $1330 (Clinumetre o # 69.031 53812 £31195 ay (m*/d) = 10.764 _ feet squared per day (ft"/d) cubic metre (m?) = 85.31 __ cubic feet (ft?) (transmissivity) cubic metre z 00081 acre-foot (acre-ft) cubicsmetre per second f cubic hectometre (hm) =810.7 acre-feet (m*/s) = 22.826 - million gallons per day litre = 2.118 pints (pt) (Mgal/d) litre =i: 1.06 quarts (qt) cubic metre per minute litre = © 26 gallon (gal) (m/min) =264.2 gallons per minute (gal/min) cubic metre = .00026 million gallons (Mgal or litre per second (1/s) = 15.85 gallons per minute 10° gal) litre per second per $ cubic metre = 6,200 barrels (bbl) (1 bbl=42 gal) metre [(1/s)/m] -= gallons per minute per foot I 3 [ (gal/min) /ft] j ometre per hour Weight (km/h) = - .62_ - mile per hour (mi/h) gram (g) = 0.035 ounce, avoirdupois (oz avdp) gig: lak segfcnd (m/s) =.. miles per hour s 9 { R g per cu tine to S C H res :~. centimetre (g/cm?) = 62.48 _ pounds per cubic foot (Ib/ft?) = 98 t 1 2.240 1 gram per square al pn Pons (2.25000) centimetre (g/cm) - 2.048 pounds per square foot (Ib/ft") f R s gram per square Specific combinations centimetre .0142 pound per square inch (lb/in?) kilogram per square centimetre (kg/cm?) =+ :0.06 atmosphere (atm) Temperature kilogram per square e m centimetre = .98 bar (0.9869 atm) degree Celsius (°C) =< 1.8 degrees Fahrenheit (°F) cubic metre per second degrees Celsius (m/s) = $5.8 cubic feet per second (ft/s) (temperature) =[(1.8x°C) +32] degrees Fahrenheit SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII THE ERUPTION OF AUGUST 1963 AND THE FORMATION OF ALAE LAVA LAKE, HAWAII By Darras L. Prox and W. T. KinosHrra ABSTRACT After the seismic episodes and summit collapses of May, July, and early August 1963, Kilauea Volcano, Hawaii, erupted along the east rift zone in and near Alae Crater from August 21 to 23, only 8% months after the December 1962 eruption in nearby Aloi Crater. The eruption started at 18"10", August 21, after nearly 4% hours of summit defla- tion, low-amplitude tremor, and many small earthquakes. Fountains on the floor and north wall of Alae Crater fed lava, at rates as high as 2.3% per hour and at maximum temperatures of about 1,160°C, into a growing lava lake in the southeast pit of the crater. During the last 12 hours of the eruption, a decrease in the rate of extrusion to near zero, accompanied by a decrease in temperature to 1,140°C, led to stagnation of the lava lake at its maximum volume of 8.4%10'm'. Near the end of the eruption at 08"10", August 23, 1.8%10'm" of lava drained back into the vents leaving a stagnant lake 305 m long, 245 m wide, and as much as 15 m deep, composed of homogeneous tholeiitic basalt containing 3.5 percent olivine. A low spatter ridge partially covered and bordered the northwest side of the lake and continued as coalescing spatter cones on the north wall of the crater. The flat central floor of the lake was 14.0 m deep, but the northern part thinned to 11.3 m over a buried spatter ridge. The surface of the lake was complex in detail and showed a variety of flow features, which are a record of the many events that took place during and after the eruption. A 20- m wide levee of discontinuous slabs and pressure ridges bor- dered all but the northwest margin of the lake and was in turn bounded for most of its length by a moat as deep as 5 m. The main part of the lake was crossed by pressure ridges and linear oozeups, which marked the position of glowing cracks during the eruption. Shear zones marked the bound- aries between crust formed on relatively static lava and crust formed on lava flowing out from the vent area. Observation during the eruption and later mapping of the surface fea- tures indicated that during the later part of the eruption, the lava lake consisted of an outer stagnant part and an inner part marked by active lava circulation; lava flowing outward from the vent rafted crust against the stagnant margin, shov- ing up piles of broken slabs and buckling the erust along lines of weakness. The surface of the lake was broken by cracks formed large- ly as the result of stresses induced by thermal contraction of the cooling lava. Cracks that developed near the end of the J eruption formed a random, orthogonal network outlining 0.3- to 0.7-m-high hummocks, which were divided by ortho- gonal cracks into 3- to 6-sided polygons averaging 1% m in diameter. Many new cracks, mostly short ones near the crests of hummocks, opened during a period of heavy rain- fall 5 to 9 months after the eruption. Most of these cracks were sites of deposition of sublimates of sulfur and calcium sulfate. Between 1 and 2 years after the eruption, when the lake had completely solidified, the surface was broken by long swarms of short cracks that originated in the zone of maxi- mum cooling at depths of 5 to 10 m in the lake and propa- gated upward to the surface. Field studies of the August 1963 lake ended in February 1969, when the lake was covered by 72 m of new lava. INTRODUCTION Kilauea Volcano erupted in and near Alae Crater from August 21 to 23, 1963, 84 months after the eruption in nearby Aloi Crater. The eruption left in Alae Crater a lava lake about 15 m deep of tho- leiitic basalt, which solidified during the following 10 months and cooled to less than 90°C by August 1967, 4 years after the eruption. This report de- scribes the eruption and surface features of the lava lake. Three eruptions of Kilauea Volcano during the last 2 decades ponded flows in accessible pit craters, forming stagnant lakes of molten lava that have been intensively studied-the 1959 eruption in Kil- auea Iki Crater, the August 1963 eruption in Alae Crater, and the March 1965 eruption in Makaopuhi Crater. These lava lakes are natural laboratories that have provided unique opportunities to learn more about the properties of basaltic lava and the cooling processes of ponded flows. They have been the target of a major continuing effort by the U.S. Geological Survey's Hawaiian Volcano Observatory. Studies begun in the 111-m-deep lava lake in Kilauea Iki Crater included precise-level measure- ments, core drilling, temperature measurements, chemical and petrographic study of drill core, A1 A2 ground magnetic studies, and field electromagnetic surveys. These have been described by Ault and oth- ers (1961, 1962), Macdonald and Katsura (1961), Decker (1963), Rawson and Bennett (1964), and Richter and Moore (1966). When the August 1963 eruption produced a shal- low ponded flow in Alae Crater, a variety of similar studies were begun. Drilling equipment was lowered to the lake surface by an aerial tram, and the first of many drill holes through the erust was started on August 29, 6 days after the end of the eruption. Temperature measurements in the drill hole were begun on the following day. During the following months many different studies were started, some of which were continued until August 1967, when the lake had approached ambient temperature. These in- cluded the following : repeated core drilling through the crust and into melt; sampling of gases and melt, and measurement of temperatures in the drill holes; study of drill core, including examination by binocu- lar microscope, measurement of density, petrograph- ic study, chemical analysis, mineral separation, and measurement of magnetic susceptibility ; installation of a grid of stations on the lake surface and repeated measurement of the altitude and intensity of the magnetic field at the stations; recording of rainfall at the rim of the crater ; and mapping of surface fea- tures and joint cracks. Previous reports on Alae lava lake include: a pre- liminary description of the eruption and of tem- peratures in the lake (Peck and others, 1964) ; analy- sis of volcanic tremor associated with the eruption (Shimozuru and others; 1966;) descriptions of crys- tallization of basalt in the lake (Peck and others, 1966) ; magnetic properties and oxidation of iron- titanium oxide minerals (Grommé and others, 1969) ; the formation of joint cracks (Peck and Min- akami, 1968) ; sulfide-rich blebs in an ooze sample from the lake (Skinner and Peck, 1969) ; lava coils on the surface (Peck, 1966) ; and a report on infra- red radiation from the lake (Decker and Peck, 1967). In March 1965, an eruption of Kilauea Volcano produced a new lava lake 83 m deep in Makaopuhi Crater (Wright and others, 1968). Studies of this lake, which have been carried out under the direc- tion of T. L. Wright, include microprobe study of core samples (Hakli and Wright, 1967; Wright and Weiblen, 1968), measurement of oxygen fugacity in drill holes (Sato and Wright, 1966), collection of molten lava samples from beneath the erust (Wright and others, 1968), collection of gas samples from drill holes (Finlayson and others, 1968), and the measurement of viscosity in molten lava beneath the SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII crust (Shaw and others, 1968). The prehistoric lava lake that was exposed in section on the mezzanine cliff face of Makaopuhi Crater has also been studied in detail (Moore and Evans, 1967; Evans and Moore, 1968). Some of the interpretations of the present series of papers on Alae lava lake have benefited from these later studies. The studies of all three recent lava lakes have been summarized by Peck (1974) and Wright, Peck, and Shaw (1976). Later eruptions of Kilauea Volcano, starting in February 1969, have buried the lakes in Alae and Makaopuhi Craters with new lava (Swanson and others, 1971), bringing field studies of these lakes to a close. Alae Crater was later completely filled with lava during a series of episodic and complex eruptive events (Swanson and others, 1973; Swan- son and Peterson, 1972), and by 1972 the site was covered by a broad, 100-m-high, basaltic dome, satel- litic to the new shield, Mauna Ulu. In this report, accordingly, we refer to the crater and its 1963 lava lake in the past tense. As of 1974, studies including level measurements and core drilling of the still partly molten lake in Kilauea Iki Crater were continuing. ACKNOWLEDGMENTS We are indebted to many who helped observe the eruption in Alae Crater, and carry out the program of studies of the lava lake. These included a visiting group of Japanese scientists under the direction of Prof. T. Minakami of the Earthquake Research In- stitute, Tokyo University. Other members of the group were Profs. D. Shimozuru, S. Aramaki, and K. Kamo and Messrs. T. Miyasaki and S. Hiraga. The work reported here was a team effort of the staff of the observatory under the general guidance of the successive Scientists-in-Charge, J. G. Moore and H. A. Powers. Directly concerned with the stud- ies are the following: T. L. Wright, B. J. Loucks, W. H. Francis, J. C. Forbes, E. T. Endo, R. T. Okamura, R. Y. Koyanagi, and G. Kojima. The Hawaiian Vol- cano Observatory and Alae Crater are in Hawaii Volcanoes National Park; the Park Superintendent at the time of the study, F. T. Johnston, and other personnel of the Park Service helped facilitate the studies. The U.S. Weather Bureau supplied the con- tinuously recording rain gauge for the study. Major P. K. Nakamura, Commander of the Air National Guard unit in Hilo, made aircraft available for photographing the eruption area. Many of the prob- lems encountered in the study have been clarified as the result of discussions with our colleagues of the Geological Survey, particularly H. R. Shaw, D. ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE B. Stewart, R. S. Fiske, M. Sato, and A. H. Lachenbruch. GEOLOGIC SETTING The basaltic shield volcano Kilauea is one of five volcanoes on the island of Hawaii (fig. 1), at the southeast end of the Hawaiian Archipelago. Kilauea “Q f KOHALA®O e em MAUNA LOA VOLCANO e 12 900 10 900 o> o m MAUNA KEA VOLCANO A8 has erupted repeatedly in historic time in its summit caldera and along the two rift zones that stretch from the summit to the east and southwest. After nearly 18 years of quiescence, Kilauea began the current series of eruptions in June 1952 (Mac- donald, 1955), with activity in Halemaumau Crater, in the summit caldera (fig. 2). In each of the next ISLAND OF HAWAII +20°N Site of Alae Crater 20 MILES 30 KILOMETRES +19°N 155° W FIGURE 1.-Map showing the five volcanoes constituting the island of Hawaii. SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII A4 are eJop|e> eont[IY *sour[ faeoy se umoys pue sime1o yo sourepunog 'g96I UI poreadde 41 se ourpjoA tont[Iy Jo auoz j;I1 jsta doddn pus jtunums ay; yo dew xopuy-*z SUODIAX (y'. O % | K S3HIJ3WOTX ¢ F f A ors | | I 1 | it fs T T I s 0 S37IW € ¢ I o 02.61 k eunyeyamp N Hd 159M dM ndoey2ep W _ enuy y & jajawmowsiag Y G NOILVNY14X3 o® A x g ain8y & N40 AvyOX jo eaiy § e ANd Q4; - esp - His Ac8 Has (C! (t -o MXVW ran YHALVHD S YALVHYD FXVIIH 3NOZ { 318 jeoiy| siimag Jol G I y & > 18v3 AnnHw/iw|\\\\\\\||//r;:f11‘|;///////// $" & y y 0 Yy" uarvxo0 y> & nyWIHNAd § y o & /.\)/ \w4/o N nyWKAyWNATVH - gm - .Sz.61 UYHALVHD vVUH0TVD oueat0 A IXI VYADVIIM -{ \ wewemey vanyIIx ¢ f_ > ¥ coo /\}\ | | 1OTeSST 1STSST OceSST ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE three pairs of eruptions, a summit phase was fol- lowed by a flank eruption along the east rift-in 1954-55 (Macdonald and Eaton, 1957, 1964), 1959- 60 (Richter and Eaton, 1960), and in 1961 (Richter and others, 1964). During the next 5 years, all erup- tions were from fissures along the east rift zone with no intervening summit phases: in Aloi Crater in December 1962 (Moore and Krivoy, 1964), in Alae Crater in August 1963 (this report), in Napau Crater and along the middle part of the east rift in October 1963 (Moore and Koyanagi, 1969), in Maka- opuhi Crater and along the middle part of the east rift in March 1965 (Wright and others, 1968), and in Aloi Crater again in December 1965 (Fiske and Koyanagi, 1968). Lava also moved into the east rift zone from the summit reservoir, but without attend- ant eruptions during three seismic episodes in 1963 (Kinoshita, 1967). Since 1965, Kilauea has erupted repeatedly, both at the summit and along the two rift zones. The week-long eruption in February 1969 fed flows into Alae Crater (Swanson and others, 1973) covering the August 1963 lava lake. Alae Crater was on the east rift zone of Kilauea Volcano, about 8 km southeast of the summit caldera (fig. 2). The filled crater now underlies the eastern flank of Mauna Ulu. Alae was an elliptical pit crater elongate in a northwesterly direction, 640 m long and 460 m wide (fig. 3). Most of the observations and measurements of the 1963 eruption were made from a turnout and observation point on the Chain of Craters Highway at the south edge of the crater. Alae Crater was a double crater formed by two overlapping collapse pit craters. The first and larger pit was filled by a prehistoric lava lake and by an overlying 6-m-thick lava flow to within 100 m of the crater rim. At the time of the 1963 eruption, this old crater fill was still preserved as a bench on the northwest side of the crater. A second prehistoric collapse of the southeast part of the first crater drop- ped the original crater fill at least 60 m, producing an approximately circular pit 160 m deep and 300 m in diameter. In 1840, lava erupted from fissures in the crater (Dana, 1849, p. 188), formed a spatter rampart on the mezzanine and floor of the southeast pit, spread as thin shelly flows over parts of the floor and cliff face of the mezzanine, and ponded in a shallow flow on the floor of the southeast pit. PRE-ERUPTION UPLIFT OF THE SUMMIT REGION AND SEISMIC ACTIVITY Uplift of the summit region of Kilauea Volcano resumed almost immediately after the collapse as- A5 155°11'30" I T 0 0 ff 3068 x 19°22" |- o 0 200 400 FEET * 0 100 METRES CONTOUR INTERVAL 100 FEET EXPLANATION Lava lake Spatter, thin flows, and pumice w** Eruptive vents Letters refer to vents described in text and shown in figure 6 FigurE 3.-Map of Alae Crater showing distribution of vol- canic rocks from the August 1963 eruption. Location of crater is shown in figure 2. sociated with the small flank eruption of December 1962 (fig. 4) and continued until May 9, 1963, when about 8 million m* of lava moved from the summit reservoir into the southwest rift zone, accompanied by harmonic tremor and many shallow earthquakes (Kinoshita, 1967). The intrusion of lava produced extensive new cracking along a zone extending 5 km from Puu Koae to a point 3 km south of Halemau- mau. Collapse of the summit ceased on May 12, and rapid reinflation began. This continued until July 1, when about 8 million m* of magma moved from the < S & [~ # a a a on § C E C Cf 5 VOLCANIC ACTIVITY % 6} E, eruption => j C, collapse & a 5/~ , West (uplift) Z & E P" m & 3}- / Gf 3L $8; ; o ~1}- "East (subsidence) 5 0 NL .) 1 yr i MO. -A L MeO. gcd | A 1962 f 1963 FIGURE 4.-East-west ground tilting of the Kilauea summit area as indicated by daily readings of the short base water- level tiltmeter at Uwekahuna from December 1962 to Sep- tember 1963. Periods of eruption (E) and collapse (C) are shown at top. A6 summit reservoir into the eastern part of the Koae fault zone. The intrusion produced extensive crack- ing and deformation along the Koae fault near its junction with the east rift zone. Uplift began again on July 2 and continued until August 3, when nearly continuous tremor and earth- quakes were recorded at Kilauea as about 2 million m* of magma moved from the summit reservoir into the upper east rift zone near the junction with the Koae fault zone. If the volume calculations are even approximately valid, we can infer that most, perhaps all, of the lava erupted into Alae Crater from August 21 to 23 was derived from lava stored in the rift zone after the collapses of May, July, and August 1963. After this episode, uplift began again, slowly, and continued until August 21. During this period, local caldera quakes averaged about 70 per day. DESCRIPTION OF THE AUGUST 21-23, 1963, ERUPTION The first indication of the impending eruption was the onset, at *13"46" August 21, 1963, of nearly con- tinuous small-amplitude tremor that recorded more strongly on a seismograph near the Makaopuhi Cra- ter than on seismographs near Kilauea Caldera (fig. 5). Beginning at the same time, many small earth- quakes, one to four per minute, were recorded. Many of these were from epicenters near Makaopuhi, which yielded S minus P intervals of 0.5 and 0.6 seconds at Makaopuhi. More than 800 distinct earth- quakes were recorded from the upper east rift zone during the remainder of August 21. At 13"50"=, dis- placement of the traces of the Press-Ewing seismo- graphs in the Uwekahuna vault near the Hawaiian Volcano Observatory indicated the beginning of a distinct tilt to the east and a less distinct tilt to the south. Tremor died down to a low level after 15°00", but at 18"00"= it began to increase again. A sharp in- crease in tremor amplitude at 18"10"= probably marks the beginning of the eruption. A swarm of small earthquakes began at about 13"50= after three earth- quakes of magnitude about 2 on the Richter scale. The frequency of earthquakes increased to about 5 per minute at 14°30" and then decreased to 0 at 18"15=. When the eruption was first observed, shortly after 18"15=, vigorous lava fountains on the floor of southeast pit and on the north wall had already formed a pool of lava covering most of the floor of the southeast pit. At 19"20", the fountains on the floor formed an almost continuous curtain 5 to 10 m high extending 80 m N. 65 E. from near the east * All times given are in hours and minutes, Hawaiian standard time. SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII £ w d T T T T T T T T § Ys xm" I RATE OF EXTRUSION AND DRAINBACK Y e ( - ke | got == < €. g 16F al € 14l- TREMOR AT MAKAOPUHI € I x12} - W g 10}- > 2. °C 4 F 6}- £ 4 |- P 21 < 2 A w 0 c 8 6 TREMOR AT WEST PIT fa 0 at 8 2}- a 0 5 in 9 a < 26 EAST-WEST GROUND TILT UWEKAHUNA - a 1. s 25° ~ { elm 3 ys rane | ue fand I open a zZ" &L Kg "k 6: NORTH-SOUTH GROUND TILT AT UWEKAHUNA x oi \ Z*, \ el 2.91: \_/ i f 4 ___ m #5 resists 0 0 C a 0 & 3 2300, EARTHQUAKES c (Short-period seismograph at Makaopuhi) 7 & {3 200 |- ic < S ? £100} { & < u I 1 1 1 1 1 J 1 HOUR: 12 18 0 6 12 18 o 6 12 18 AUGUST 21 AUGUST 22 AUGUST 23 FIGURE 5.-Chronology of events during the August 1963 eruption of Kilauea. edge of the southeast pit (long eruptive vent, fig. 3). Lava was erupting from 11 vents on the north wall along a fissure parallel to the fissure on the floor of the crater but displaced 30 m to the northwest. This lava cascaded down the cliff and plunged beneath the surface of the growing lake of lava (fig. 6). The most rapid discharge was obliquely upward and westward from one of the vents in the middle of the line, about 100 m below the north rim (vent H of fig. 6). The dark crust of the lake was broken by glow- ing zigzag cracks that splayed outward from the curtain of fire and from the base of the lava cascades. At this time the lake had an estimated depth of al- most 10 m and had completely covered the floor of the southeast pit, an area of about 65,000 m*, with almost 400,000 m* of lava (fig. 7). The level was ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE FIGURE 6.-Eruption in Alae Crater at 19"20" August 21, 1963, as viewed from overlook at south rim of crater. Let- ters refer to vents described in text. Fountains extend over a vertical distance of 100 m on the north wall of the crater. n o w co T ‘ T ~ / 1 p VOLUME,IN CUBIC METRES x10 ° w no v‘ | , cas. 0 L- AM - 1 HOUR:18 0 6 12 18 0 6 12 18 0 6 12 AUGUST 21 AUGUST 22 AUGUST 23 AUGUST 24 -y FIGURE 7.-Volume of lava in Alae lava lake during the Au- gust 1963 eruption, as determined by transit sightings (dots) from the rim during the eruption and by triangula- tion and level measurements after the eruption. rising at an estimated rate of about 2 m per hour, indicating a rate of extrusion of approximately 230,000 m* of lava an hour (fig. 5). Spatter from a 150-m-long en echelon line of fountains in the jungle, 60 m beyond the north rim, was visible above the trees. These fountains died by 19"45= after send- AT ing thin flows of shelly pahoehoe over an area of 7,000 m*. By the uppermost vents on the north wall (A, B, and C, fig. 6) had died, and the thin lava streams that had issued from them were dark. Dur- ing the remaining 4 hours of August 21, many of the other vents on the north wall of the crater died -vent D at 21" and E and F at 22"; vent F emitted bluish gases for several hours after it had stopped spattering. By midnight only three vents were still erupting vigorously. The line of fountains on the floor of the crater had become discontinuous; the fountains were playing to a height of 8 m over a length of 40 m, and an isolated fountain at the southwest end of the original line was bubbling weakly. As determined by transit sightings from the overlook, the surface was rising at a decreasing rate-1.5 m per hour at 21", 1 m per hour at 22", and 0.6 m per hour at 23". The rate of extrusion had dropped from about 200,000 m* per hour at 20" to 45,000 m* per hour at 23". As the rate of extrusion from the vents on the north wall decreased, the apparent viscosity of the lava increased (probably because of a decrease in temperature), and lava cascading down from the fountains formed a notice- able delta at the edge of the lake. At 01"40" on August 22, surges in the fountain on the floor of the crater produced waves in the lake with a frequency of 22 cycles per minute; these broke like surf on the shore northwest of the fountain and traveled 70 to 100 m to the southwest in the lake before dying out. Through the night, the lake had been broken by glowing zigzag cracks that radiated out from the vents. Photographs taken at 10-minute intervals between 01"50" and 02°10" show correspondence in the crack pattern between successive photographs, indicating that the crust of the lake was being rafted outwards on lava flowing from the vents at a rate of about 200 m per hour, to be piled in a narrow slabby levee at the edge of the southeast end of the lake. By 03"00%, the line of fountains on the floor of the crater had become erratic, momentarily dying back to a length of 15 m, and then lengthening to 35 m. The fountains on the north rim continued to dimin- ish in vigor. The lake was about 17 m thick and prob- ably was rising at an hourly rate of about 0.3 m (15,000 to 20,000 m* of lava per hour). Transit sightings to the foot of the lava cascade on the north shore of the lake continued to show a decrease in the angle of depression with time, suggesting con- tinued extrusion of 30,000 m* of lava per hour until 05°30=. However, the later increase in the angle of depression between 05°30" and 07", indicating an A8 apparent subsidence of the lake surface, and obser- vations of the north shore the following morning, suggest that the transit sightings were reflecting the slow buildup and lowering of the surface of a broad lava delta at the foot of the cascades, rather than changes in altitude of the surface of the lake. By 06"00=, the lava lake had reached almost its maximum depth (18 m) and volume (8.4% 10° m'). The increasing volume of lava in the lake for the rest of the eruption was so small that it was not measur- able. The vent on the floor of the crater had begun to fountain irregularly; instead of throwing up a curtain of flowing lava 7 to 10 m as before, it er- ratically spewed bursts of spatter to heights as great as 50 m, accompanied by great booming noises. The change in activity of the fountains at this time, and the lack thereafter of any substantial increase in volume of the lake, probably recorded a change in the type of material being supplied to the fountains from the rift zone. Before 06", this included a sig- nificant proportion of lava, the quantity of which steadily decreased with time; after 06", it was most- ly gas. The lava thrown up by the fountains may largely have been lava from the lake, blown out by expanding gases rising upward through the lake from the underlying vent. The three remaining foun- tains on the north wall (H, J, and K of fig. 6) were spattering weakly and sending two sluggish rivers of aa down to a small low delta on the lake. In the morning light, we could see that the vents on the north wall had built a row of coalescing spatter cones and that spatter and cinders from the foun- tains on the floor of the crater had accumulated to form a low spatter and cinder cone along the north- west side of the lake. The eruption continued at a diminishing rate dur- ing the morning. At 11", a party circled around to the north side of the crater, where they could look down directly into the two remaining active vents on the north wall. One, vent K (fig. 6) was spattering weakly within its cone and sending a red ribbon of lava down to the lake. The other, vent H, was emit- ting much fume but only a little lava. This lava formed a sluggish stream which at times had a rough dark aa crust that first formed at the lower end and appeared to grow upstream, only to be re- peatedly dislodged and carried downstream by the red lava. Two major fountains and several minor ones played on the floor of the crater (fig. 8). The lava lake was impounded behind a slab levee, but occa- sionally it overflowed or undermined the levee, pro- ducing festooned flows or tumuli on the lake margin. SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII FIGURE 8. -Eruption in Alae Crater at 11"10", August 22, 1963, as viewed from the north rim. Incandescent bubbling areas (light colored) are about 10 m in diameter. The crust of the lake continued to founder locally, mostly along cracks near the edge of the lake or above the former southwest end of the curtain of fire. Pumice that had been erupted during the night was collected from the ground near the north rim, and samples were gathered from the spatter ram- part and shelly pahoehoe flows in the jungle beyond the north rim of the crater. By 15", the line of fountains on the floor of the crater had shortened to 30 m; they bubbled to an average height of 3 m but sent occasional bursts of spatter to 50 m. The vents on the north wall were almost dead, and lava rivers from them were crusted over. The slabby rim of the lake increased in width from the new slabs of crust that were continually added to it from the lava lake. The crust of the lake had a distinctive mosaic appearance (fig. 9) caused by patterned differences in surface texture that ap- peared as alternating light and dark bands parallel to the glowing cracks. These textural differences are discussed on page 22. At 17", the fountains on the floor of the crater were seen from the mezzanine to bubble vigorously with whooshing and thundering noises from two large and several small intermittent centers along a length of 35 m (fig. 10). At each center, the activity ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE A9 FIGURE 9.-Eruption in Alae Crater at 15", August 22, 1963, as viewed from overlook at the south rim of the crater. Behind and to the left of the vent is a 50-m cliff at the end of the forested mezzanine. The degassing vents on the far wall are almost dead, but the inconspicuous line of foun-. tains on the far floor are bubbling to an average height of 3 m. The lake surface looks like a mosaic because of patterned differences in surface texture parallel to glowing cracks. resembled the bubbling of a mud pot; a bubble of lava 1 m or so in diameter would appear, grow rapid- ly larger as the gas inside expanded, and then break into long ribbons of lava that were cast as high as 50 m. Gases rising from below the lake apparently provided the driving force for the fountaining. Spat- ter was not thrown evenly in all directions; more than half fell to the west on the lakeshore behind the fountains, forming a spatter and cinder cone that was about 2 m high at this stage. Possibly this un- evenness reflected only the prevailing wind direc- tion, which is typically a counter-trade wind direc- tion on the floor of the crater; from the mezzanine, however, the fountaining appeared to be inclined in that direction, perhaps because lava in the lake im- pinged on the breaking lava bubbles. The dark surface of the lake was broken by glow- ing cracks; these consisted of two sets, one of con- centric cracks near the fountains and the other of thin cracks that radiated out from the fountains. The latter cracks diverged farther away and divided the lake surface into distinctive lobate areas (fig. 11). The outer margins of the lobate areas were bor- dered by wider glowing cracks, which were the sites of foundering at many localities. We scrambled down the talus to the southwest edge of the lake and found that the lake was brim full behind a 15-m-wide levee of folded and faulted pahoehoe slabs. The surface of the lake was not flat; it rose as much as 0.5 m from the edge of the lake to the site of the fountains, reflecting the hydraulic gradient of the viscous fluid. Small outflows and foundering took place at the distal end of the lobate areas, which were commonly about 0.3 m higher than the lake surface at its margin. Foundering was related to outflows of fresh lava from these cracks. The weight of the outflowing lava would cause a slab of crust to bend down and break off. This slab would glide inward and tilt down toward the center of the lake until the leading edge would sink and the trailing edge rise up in the air-looking for all the world like a sinking surfboard. Crustal foundering on a much larger scale was observed during the for- mation of Makaopuhi lava lake (Wright and others, 1968, fig. 10). The levee was separated from talus mantling the crater walls by a moat 15 m wide and about 3 m deep. Inside the moat were several small tumuli about 8 m long and 2 m high; the medial crack of each was filled by a snakelike mass of lava, still glowing red in marginal cracks. At many places the levee had been breached and overflowed by small festooned flows of pahoehoe. We marked the level of the lake surface on the talus mantling the walls of the crater by sighting with a compass level. Later surveying with a transit from the floor of the lake fixed this level at 785.5 m (2,577 feet), 3.0 m above the present floor of the lake and 18.0 m above the base of the crater before the eruption. Beginning at about 21", the glowing cracks at the south and east edges of the lake began to darken (fig. 12) as the actively circulating part of the lake shrank because of diminished vigor of fountaining. By 22"45= (fig. 13), the southeast half of the lake was dark ; by 23"16" (fig. 14), only two small lobate areas were left; and by 23"55", the lake was dark except for the fountain. The position in plan of the glowing cracks was later determined by examination of photographs taken during and after the eruption and by mapping the surface features of the lake. These studies provide clues to the pattern of lava circulation in the lake during the evening of August 22, as discussed on pages 22-24. The activity of the line of fountains on the floor of the crater decreased through the evening of the SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII F, 4 FIGURE 10.-Fountain in Alae Crater at 17"45", August 22, 1963, as viewed from the mezzanine. Height of foun- tain about 5 m. Ribbon spatter on top was produced by bursting of a bubble of lava an instant before the pho- tograph was taken (compare with fig. 23). Photograph by S. Aramaki. ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE publ FIGURE 11.-Eruption in Alae Crater at about 18", August 22, 1963, as viewed from overlook. Two sets of glowing cracks break the dark crust, one near to and concentric with the fountain, which was about 35 m long, and another radial to the fountain which outlines lobate areas of crust. Crustal foundering (light spots) occurs at the distal ends of the lobes. A marginal levee of slabs and small festooned flows can be dimly seen in the left foreground. Photograph by T. Miyasaki. 22d and the early morning hours of the 28d. By 23"30= of the 22d, fountaining was taking place at 5 centers along a length of 25 m. By midnight, the fountains became more erratic and died completely for a second or two before resuming again. By 02"10= of the 23d, only three areas were bubbling, and the periods of quiet had increased to 5 seconds. By daylight, at 05°30, only one bubbling area was left; it burst slowly about every 5 seconds. At this time the lake surface was at the same level as at 19" the previous evening, when the level of the lake sur- face had been marked with a sample bag on the talus mantling the lower crater walls after sighting on the lake surface with a brunton compass level. By the vent was still bubbling every 5 to 7 seconds; drainback had lowered a semicircular area about 30 m (100 ft) in diameter around the vent. By 08"10%, the bubbling at the vent had stopped. The eruption was over. Slow drainback into the orifice beneath the lake at the vent, however, apparently continued for several hours. Vertical-angle transit sights from the rim of the crater on August 24 showed that the surface of the flow had been lowered about 3 m by drainback since the end of the eruption. Later, precise level measure- ments from the flow surface to the marker on the talus and more precise triangulation from the rim A12 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII FIGURE 12.-Eruption in Alae Crater at about 21", August 22, as viewed from the overlook. Stagnation of the lava lake in the foreground has resulted in the nearly complete darkening of glowing cracks visible earlier there (fig. 11). FIGURE 13.-Eruption in Alae Crater at 22"45", August 22, as viewed from the overlook. Glowing cracks in the entire outer part of the lake have darkened because of continued stagnation of the lake. FIGURE 14.-Eruption in Alae Crater at 23"16", August 22, as viewed from the overlook. The actively circulating part of the lake has become restricted to an area less than 100 m long adjacent to the fountains. ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE showed that the average elevation of the surface had fallen 3.0 m below the marker-a loss of 180,000 m* of lava. The lake was left with a volume of 660,000 m*. An additional 30,000 m* of flows, spatter, and pumice had accumulated in the jungle beyond the north rim of the crater and in the low spatter ridge that covered and bordered the northwest side of the lake and continued as coalescing spatter cones up the north wall of the crater (fig. 15). TEMPERATURE MEASUREMENTS Temperatures were estimated during the eruption with a glowing-filament optical pyrometer. These estimates have been supplemented by temperatures deduced from a study of pumice from the eruption and by temperatures measured with thermocouples later in the Alae lava lakes. Heavy fume obscured the view of the fountains during much of the eruption, making pyrometer measurements of little value. When the fume was moderate to light, repeated pyrometer measurements of the fountains were made from the overlook at the south rim of the crater-a distance of 350 to 400 m from the fountains. During the evening of August 21, from 20" to 23", maximum readings were re- peatedly found to be 1,090° to 1,100°C The actual maximum temperatures of the fountains were proba- bly substantially higher. Pumice from this fountain- ing, which was collected August 22 from the north rim of the crater, contains only 7 percent crystals- 2.0 percent olivine, 4.1 percent augite, and 0.9 per- cent plagioclase (sample H 186, T. L. Wright and D. L. Peck, unpub. data). Laboratory melting stud- ies of drill core from Alae lava lake (Tilley and oth- ers, 1967) and data from the 1965 eruption in Maka- opuhi Crater (Wright and others, 1968) indicate that the temperature of first appearance of plagio- clase in these lavas is 1,160+5°C. The scarcity of plagioclase (and other crystals) in pumice from the early stage of the Alae eruption, accordingly, sug- gests a maximum eruption temperature of about 1,160°C. During the evening of August 22d, maximum tem- perature measurements showed a steady increase, from 1,085°C at 19°30" to 1,115°C at 20°15 and finally to 1,140°C at 21"20=. This increase very likely resulted from a decrease in radiation-absorbing fume between the fountain and the observers (as a result of the decrease in activity), allowing more accurate temperature measurements, rather than an actual increase in temperature of the dying fountains. Pumice from this late fountaining, which was col- A183 FIGURE 15.-Northwest part of Alae lava lake as viewed from overlook on August 31, 1963, showing low spatter ridge near former vent on floor of crater and coalescing spatter cones marking positions of former vents on north wall. Forested mezzanine atop a 50-m-high cliff can be seen to the left of the spatter cones on the north wall of the crater. lected near the vent after the eruption (sample DPH-77, T. L. Wright and D. L. Peck, unpub. data), contains 13 percent crystals-2 percent olivine, 8.5 percent augite, and 2.5 percent plagioclase. Compari- son of this abundance of crystals with the abun- dance in samples of known temperature from the 1965 Makaopuhi lava lake (Wright and others, 1968) indicates a temperature of about 1,140°C ; this is in excellent agreement with the maximum pyrometer readings. The maximum initial temperature of the lava lake at the end of the eruption can be inferred from tem- perature data collected with thermocouples in drill holes in the lake after the eruption (Peck and others, 1964). The highest temperature measured in the lake was 1,136°C. This value was obtained on Novem- ber 8, 1963, 24 months after the eruption, at a depth of 5.46 m, 2.2 m below the base of the crust. Extra- polation of the temperature gradient downward 2 m to the estimated depth of the maximum temperature in the lake suggests a temperature of about 1,140°C. Maximum temperatures in the central part of the lake were measured starting July 2, 1964, and con- tact temperatures at the base of the central part of *On the basis of the MgO/MgO-|-FeO-|-Fe:Os: content of analyzed glass from sample DPH-77 and the calibration of Tilley and others (1964, fig. 23), the temperature of the later stage of the eruption has been esti- mated to be 1,065°C (Wright and others, 1968, table 5). We favor a lower temperature that is more in accord with the temperature and crystallinity measurements in Makaopuhi lava lake. A14 the lake were obtained beginning on December 8 from drill holes that pierced the lake. These values were, respectively, 1,050° and 700°C. Analysis of these and later values using the heat-conduction theory (J. C. Jaeger, written commun., 1967) indi- cates initial maximum temperatures in the lake of 1,140°C, Thus we conclude that during the early vigorous part of the August 1963 eruption, the maximum tem- perature of the fountains was 1,160°C. During the waning stages of the eruption, the maximum tem- perature fell to 1,140°C. By the end of the eruption, the maximum temperature of the resulting lava lake was 1,140°C. DEFLATION OF THE SUMMIT REGION AND SEISMIC ACTIVITY DURING AND AFTER THE ERUPTION The sharp collapse of the Kilauea summit area that began at 13"50= on August 21 was completed by about 19", August 21 (fig. 5); thereafter the summit subsided slightly until 15", August 22, when renewed inflation took place. SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII Precise level measurements of benchmarks along the Chain of Craters Highway after the August 21 to 23 eruption (fig. 16) showed inflation of the rift zone between Aloi and Makaopuhi Craters since mid- July. The maximum uplift measured was 0.131 m at the overlook at Alae Crater, relative to the YY22 benchmark at the junction of the Kilauea Crater rim and the Chain of Craters Highway. The level meas- urements also showed deflation of the rift zone near Hiiaka Crater, at the intersection of the Koae fault zone and the east rift. Several lines of evidence indi- cate, however, that this subsidence took place during the August 3 collapse and not during the August 21 to 23 eruption: (1) measurements of the altitude of about one-half the benchmarks after the August 3 collapse but before the eruption in Alae crater; (2) the location, during the August collapse, of epicen- ters of shallow earthquakes that cluster near Hiiaka crater; and (3) the distribution of cracks formed during the collapse. After the outbreak of lava at about 18"10" on August 21, earthquake activity at Alae Crater and beneath the summit and upper east rift zone of Kilauea decreased to a very low level. Later, at x o 0 < o > O u a R £ £ o V s § § $ £ C C o o 6 f o = C 82 3 3 3 (s) o © < a (9) ) < 6 s o x E x © c 3 w+ .:.2r £ 6 = "g o L &C 3 ag © 9 o © - o o b a o s u u d ip A Q 3 & & A ® o ® m m (2 m m m it X- 2+1 9 Ara INS Z MEz = E Lu x/ \x\\ me uf m tenn f" 4 o red o ¢ > D ue" -o Lie - - -B Fust C' O+ =f eres fic» -0 t,“ - ~ x - 4 X vert Z 3 2: y & , xs. _"" t wh TAL" uff u *C" o 5000 10,000 FEET (29 O < <2: ) 1000 2000 3000 METRES -| -.5I a L o ~: 2 EXPLANATION +------ Altitudes of stations in mid-July 1963 set equal to zero o----o--- Changes in altitude mid-July to mid-August 1963, relative to station 25 Changes in altitude mid-July to early September 1963, relative to station 22 x -- ---x FIGURE 16.-Changes in altitude of benchmarks along the Chain of Craters Highway between mid-July and early Sep- tember 1963 on the basis of second- and third-order leveling. Location of craters is shown in figure 2. ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE about 02"00= on August 22, earthquake activity in- creased, centered mostly beneath Kilauea caldera. From August 24, the day after the eruption was over, to September 1, only 10 earthquakes occurred along the upper east rift. The number of Kilauea caldera earthquakes during the rest of August averaged about 65 per day. Harmonic tremor recorded during the August 1963 Alae eruption was unusual in its attenuation pattern and maximum amplitude. The highest ampli- tudes were recorded by the seismometer at Makao- puhi Crater, only 2% km from the site of the erup- tion. The tremor attenuated rapidly as distance in- creased, however, and was never recorded at the Desert and Mauna Loa seismometers, which were 21 km west and 24 km northwest, respectively, of the eruption. This attenuation pattern is similar to that of shallow-focus earthquakes beneath the summit of Kilauea. The maximum amplitude of tremor re- corded by the Makaopuhi seismometer was one-third to one-fourth of that recorded during the October 1963 and March 1965 upper east rift eruptions, even though the eruptive vents were closer to the Makao- puhi seismometer during the Alae eruption than were some of the vents of the October 1963 eruption (Shimozuru and others, 1966). A possible explana- tion for this observation is that the amplitude of tremor recorded at seismometers within 3 km of a vent is an indication of the rate of extrusion at that vent. Thus, the amplitude of tremor recorded by the Makaopuhi seismometer was much higher during the October 1963 eruption (Moore and Koyanagi, 1969) and the March 1965 eruption (Wright and others, 1968), when the rates of extrusion at the vents near Makaopuhi Crater were much higher. PHYSICAL FEATURES OF THE LAKE SURFACE FEATURES OTHER THAN JOINT CRACKS The lava lake in Alae Crater, formed from August 21 to 23, 1963, was an elliptical lens 305 m long, 245 m wide, and as much as 15 m thick. The lake was composed of homogeneous tholeiitic basalt contain- ing 3.5 percent olivine (Peck and others, 1966). The outline of the lake and the location of surveying sta- tions and a study area are shown in figure 17, to- gether with cross sections along the major and minor axes of the lake. The pre-1963-eruption topography of the bottom of the crater (fig. 17) was prepared by photogrammetric methods from aerial photographs taken February 1963 by the U.S. Geological Survey and supplied to us by William A. Fischer. The lake deepened inward from the edge to an average maxi- A15 mum thickness of 14.0 m. The floor of the lake was nearly flat but sloped at a small angle to a low area beneath the northcentral part of the lake. The north- ern part of the lake thinned to 11.3 m over a buried spatter ridge formed during the eruption of 1840. The surface of the lake was surprisingly complex and showed a great variety of features recording events that took place during and after the eruption. Many of the features which were observed forming, as described earlier in this paper, were studied in detail because they are potentially important in understanding the physical regimes of flowing lava and cooling basalt flows. Similar features on well- preserved flows formed elsewhere during eruptions that were not observed, may provide clues to the temperatures, viscosities, and flow patterns of the erupting lava, the relative chronology of eruption, and the conditions under which the flows cooled. Many of the surface characteristics of Alae lava lake are shown in an aerial photograph (fig. 18) taken 1 month after the eruption. The southeast half of the lake was separated from talus flanking the crater walls by a moat (cross-section A-A'; fig. 17) of August 1963 lava. The moat was 8 to 15 m wide and about 2 m deep for most of its length, and the base sloped irregularly southeast to a maximum depth of 5 m below the levee at the southeast end of the lake. In many places it was partially filled by small festooned flows that broke through or over the levee. Judging from the altitude of the base of the moat at the southeast end of the lake [778.8 m (2,555 ft)], the moat was formed within the first few hours of the eruption. Two pressure domes, or tumuli, rose from the floor of the moat (fig. 19) ; each was about 10 m long, 5 m wide, and 2 to 3 m high and had an axial crack partly filled with bulbous lava. These domes were formed by lava forced along channels beneath the levee by hydrostatic pressure from the lake inside the levee. After the eruption, the crests of the domes were at or above the present surface of the main body of the lake; when the domes formed, however, the crests were below the surface of the lake. The main body of the lake was separated from the moat and from the talus flanking the crater walls by a levee about 30 m wide that borders all but the northwest end of the lake. Figure 20 shows the levee and moat as viewed from the talus at the south edge of the lake, and they are indicated on the map (fig. 21), which shows all major features of the lake sur- face except for joint cracks. The map was prepared using an aerial photograph (fig. 18) and pace and compass methods from a grid of stations on the lake A16 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII __ Area of figures 26, 32 and 33 -_ o 0 Sse... a" f EXPLANATION B. ® Surveying station on lake surface 100 200 300 FEET | I 25 50 75 METRES A! METRES 790 FEET 2600 780 770 2590 - 2500 B' METRES B 2600 7. 90 2550 780 770 2500 DATUM IS MEAN SEA LEVEL FIGURE 17.-Map of Alae lava lake showing contours (in feet above sea level) at the base and margin of the lake. Shaded area was studied in detail. Cross sections along the major axis (A-A4') and the minor axis (B-B') of lake have the same horizontal and vertical scales. Topography prepared by photogrammetric methods from aerial photographs taken February 15, 1963, by the U.S. Geological Survey. neg ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE A17 FIGURE 18.-Aerial photograph of Alae lava lake taken September 25, 1963, from an altitude about 900 m above the east rim of the crater. Letters designate features as follows: a, moat; b, levee; c, spatter cone formed by fountains on floor of crater; d, spatter ridge formed by vents (A-K) on north wall of crater; e, shear zones; f, landslide block thrown down by an earthquake on September 21, 1963. surface, which were laid out by means of a transit and steel tape. The panorama of figure 20 was photo- graphed from near point A of figure 21. The levee was made up of discontinuous pressure ridges- mostly symmetrical anticlinal folds, but also folds overturned away from the center of the lake, faults overthrust toward the edge of the lake, and jumbled piles of broken slabs of crust. The average height of the levee rose and the width narrowed as the levee was traced from near point A around the west side of the lake toward the vent. In the same distance, the pressure ridge between the levee and the main A18 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII FIGURE 19.-Pressure dome in the moat at the south edge of Alae lava lake. body of the lake changed northwestward from an anticlinal fold to a fault with a searp as high as 2 m facing inward toward the center of the lake. This change probably resulted from a greater and more abrupt withdrawal of lava from the lake near the vent during drainback at the end of the eruption. Lava in the lake farther from the vent had stagnated earlier (p. 8) and hence was colder and more vis- cous. The levee was formed from slabs rafted across the lake by lava flowing from the vent beneath the crust during the eruption. When first noted at 02"30#, August 22, the levee was 6 to 10 m wide. The pressure ridges and thrust faults were formed by the push of lava-rafted crust against the static crust of the levee, which yielded by folding and faulting along lines of weakness. The central part of the lake had a hummocky fila- mented surface that sloped almost imperceptibly to- ward the vent. Along the north side of the lake were several domes which consisted of jumbled piles of slabs of crust and were as high as 2 m (fig. 22). These domes are similar in origin to the levee rather than to the pressure domes in the moat formed by hydrostatic pressure. They were formed at the inter- section of convergent streams of flowing lava, most- ly where lava flowing from vents on the floor of the crater converged with lava from vents on the north wall of the crater. Near the former position of the vent on the floor of the crater was an asymmetrical spatter ridge (cross section A-A', fig. 17) that mantled the north- west end of the lake and the adjacent talus and con- tinued as coalescing spatter cones up the north wall of the crater. Below the vents, the wall of the crater was mantled by drapery, which led downward to rough aa flows of oxidized cindery blocks overlying the talus. Beyond the spatter cone (which on fig. 21 is arbitrarily bounded where the spatter cover is less than 30 cm thick), the lake was mantled by a thin layer of spatter, which consisted mostly of cowdung bombs and twisted-ribbon bombs (fig. 23). Formation of ribbon bombs such as that shown in the figure is captured in the photograph of the foun- tains on the evening of August 22 (fig. 10). Still farther from the vent the spatter cover was discon- tinuous. The greater the distance from the vent, the smaller the proportion of surface covered. The quar- ter of the lake farthest from the vent had almost no spatter. Because all the crust formed near the vent was covered by spatter, the decreasing proportion of spatter on the lake at greater distances from the vent recorded the decreasing proportion of original crust. By the time the crust was rafted three-fourths of the way across the lake, all of it had either foun- dered or been covered by thin flows of lava that spread through cracks from beneath the crust. The surface of the central part of the lake was crossed by pressure ridges and linear oozeups that marked the position of glowing cracks during the eruption. These are shown on the map-of.-the-lake (fig. 21) and also on the cross sections of figure 24. The undisturbed oozeups had a thin shell of crust, usually less than 2 cm in thickness, enclosing a cen- tral gas cavity. When they formed, gas exsolved from the underlying molten lava was escaping through cracks in the crust; when the incandescent lava filling a crack crusted over, some of the gas was trapped, forming the central cavity. The pressure ridges rose as much as 1 m above the surface of the surrounding lava and ranged from faulted anticlinal folds to piles of broken slabs of crust formed from the thin crust of oozeups (fig. 25). During the erup- tion, many overflows and squeezeups could be seen forming along glowing cracks at the distal ends of the lobate areas of dark crust (fig. 11). When the cracks darkened as a result of stagnation of the un- derlying molten lava, the oozeups formed lines of weakness because of the thinness of their crust. Most of them were soon folded and faulted as a result of the pressure of lava-rafted crust nearer acon 0 . + clits C ai z A19 ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE 'JoyeIo ;o uo sute3uno; £q pouuro; ouoo 'p 'qsnao yo sqefs Suisojppus 4jerred smog soysoyed 'a foomal 'q po oop 'e :smo[jO; se ore somnjea; 'I7 omsy ;o y qurod xeau wou; 'g96t 'Ig ;sn@ny usoye; ydeaSojoug 'oye| oy}; Jo aspo y;nos oy; ge snjej; oy} woIJ; otly Jo ewreJoueq- 0s A20 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII 0 100 I f 2(I)0 3?0 FEET Mapped October and November, 1963 | I I I 0 25 50 75 METRES EXPLANATION Spatter cone; spatter more than 30 cm thick £59 Pressure dome E Pressure dome; crest broken by crack filled with linear Thin spatter, spatter less than 30 cm thick ooze-up of lava Pressure ridge ————— Linear ooze-up of lava ~=mx>> Shear zone Lake levee; surface marked by abundant pressure ridges fi and faulted slabs of crust Central part of lava lake Scarp; hachures on downthrown side @ _ Lava coil; clockwise coil Moat (part of rim below level of central part of lake) 9 Lava coil; counterclockwise coil FIGURE 21.-Map of the surface features exclusive of joint cracks of Alae lava lake. the vent against the crust of the stagnant part of the | tween relatively static crust and crust moving above lake. a flowing stream of lava. Voluted strips of lava The lake surface was also broken by shears (Peck, | (lava coils, Peck 1966, fig. 2), as high as 30 cem, 1966, fig. 3), which were formed at the contact be- | were found along the shears together with irregular ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE A21 FIGURE 23.-Twisted ribbon bomb (1 m long) on the surface FIGURE 22.-Pressure dome of jumbled slabs near the north- of Alae lava lake. west edge of the lava lake. A a' FEET METRES 2590 - | 2580 -- -785 \ 2570 - a 2560 - -780 I I I 2550 h DATUM IS MEAN SEA LEVEL B FEET M ETRES 2580 785 2570 3 ‘\ 2560 200 FEET 1 0 25 50 METRES FIGURE 24.-Cross sections of the surface of Alae lava lake. Vertical scale exaggerated 10 times. Section A-A' is along the major axis of the lake, and section B-B' is along the minor axis of the lake, as shown in figure 17. Short cracks shown as light lines; long cracks as heavy lines. FIGURE 25.-Pressure ridge formed from the crust of a linear oozeup on Alae lava lake. clots and strips of lava and indicated the relative direction of movement of the lava. Coils that spiral inward in a clockwise direction record right-lateral shear ; coils that spiral inward in a counterclockwise direction indicate left-lateral shear. Other coils, mostly broad coplanar forms as large as 1 m in di- ameter, do not occur along recognizable shear zones. The surface features of the lake were studied in more detail in an area approximately 30 m (100 ft) square near the center of the lake (fig. 17). The fea- tures were mapped at a scale of 10 feet to the inch (1:120). Nails were driven into the lake surface at 3-m (10 ft) intervals along two opposite sides; steel tapes were stretched between the nails, and the fea- tures mapped by inspection. Figure 26 shows the distribution of the surface features except for joint cracks. The area was bisected by a northwest-trend- ing shear zone consisting of two offset segments. This zone marked the position of a glowing crack that, during the eruption, was the source of many small flows of slightly different age on each side of the zone. The northwest side of the area intercepted a major zone of pressure ridges and linear oozeups formed from lava in the glowing crack at the margin of the active part of the lake at 23"16" on August 22. The northeast edge cut across the flank of a low pressure dome. Several different ages of crust were present; the oldest crust, which had a thin spatter SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII cover, was at the north corner. The area was crossed by many narrow linear oozeups, which occupied cracks formed near the end of the eruption. Several minor features are shown in figure 26. Fine lines over most of the map indicate the trend of filaments of the surface crust. The surface of the lake consisted of l-cm-thick layer of highly vesicu- lar froth made up of intertwined filaments or threads of glassy lava (fig. 27). The filaments on each small flow were generally alined parallel to the direction of flowage or stretching of the crust; this was par- ticularly clear on the crust of linear oozeups. The filaments appear to have formed by the drawing out during flowage of vesicles in the still-plastic glassy crust until only threads of glass were left between vesicles. A filamented surface is characteristic of many Kilauea flows extruded at moderate tempera- tures (1,150°C +); at higher temperatures (1,175° to 1,200°C), a smooth or sharkskin surface is formed, such as that on the lava lakes in Kilauea Iki and Makaopuhi craters. The filamented surface bent near the medial shear zone in the area of detailed study, particularly at the south end of the north seg- ment, indicating a left-lateral displacement along the shear. The sense of lava coiling along the shear suggested the same relative displacement. The fila- mented surface was also offset by fractions of 1 em or sharply bent along obscure lines ( marked by long dashes on the map), which indicated the same sense of movement. Apparently these lines were minor shears marking interruptions in the extrusion of the small flows. These discontinuities, together with shears and the fronts of small flows, outlined bands parallel to the cracks from which the flows issued. The bands were obscure after the eruption, except when viewed under oblique illumination early and late in the day, but were conspicuous during the eruption (fig. 9). INFERRED CIRCULATION OF LAVA DURING THE ERUPTION Clues to the pattern of lava circulation in the Alae lava lake during the latter part of the August 1963 eruption are provided by observations made and photographs taken, particularly during the evening of August 22 (figs. 11 to 14), and by the distribution of pressure ridges, linear oozeups, shears, and lava coils on the lake (fig. 21). Early in the evening of the 22d, lava was circulating beneath most of the surface of the lake. The lava flowed from the foun- tains to the inner margin of the levee along a surface that sloped gently outward at a gradient of 1 or 2 to 1,000. The rafted crust above the circulating lava ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE A23 \le peHri ad \ ¥ _ (d \ AN 3 Bloke f $ a i a O- 0 10 20 FEET ! 0 5 METRES EXPLANATION + Margin of flow; crosstic shows relative age (straight seg- _ Clockwise lava coil, indicating right-lateral shear ment on side of older flow, curved on younger) C 9 Counterclockwise lava coil, indicating left-lateral shear ' -- -_ Trend of filaments of flow surface W | Discontinuities in filamented surface showing relative ® Pressure ridge direction of offset of filaments _-_im m =o A- A : Q-_3 Shear zone, showing sense of movement in 3 ~ mm Linear oozeup (I3 0) Spatter-covered areas FIGURE 26.-Surface features exclusive of joint cracks on a part of the central part of Alae lava lake. Area of map and location of numbered surveying stations are shown in figure 17. was so newly formed and thin that cracks in it ex- | overflows in its passage across the lake to be re- posed hot incandescent lava. Crust formed near the placed by still newer thinner crust. Lava from the vent was destroyed by foundering and covered by | fountains moved outward in diverging streams, A24 FIGURE 27.-Filamented surface of Alae lava lake. Knife in center of photograph is 10 ecm long. which were bordered at the surface by glowing shear fractures. The outer margin of the active part of the lake was bounded by breaks where the lake surface step- ped down to the lower level of the outer stagnant part of the lake, exposing red-hot lava. These breaks formed conspicuous glowing lines, along which many small flows were extruded, and much foundering of the crust took place. The crust of the outer stagnant part of the lake was dark. Because the surface had not been renewed by overflows and foundering, the crust had become thicker with time, and cracks no longer penetrated hot, brightly glowing lava. The outward push of lava-rafted crust at the margins of the inner part of the lake caused the crust of the outer part to buckle along lines of weakness, mostly the thin-crusted lin- ear oozeups that occupied cracks along former mar- gins of the inner lake. The inferred pattern of lava circulation during the evening of August 22 and the margin of the inner active part of the lake at three different times are shown in figure 28, adapted from figure 7 of Peck (1966). Early in the evening, the strongest flow of lava from the fountains was in a stream under the crust northeast of the long axis of the lake. Between 17"45=® and 22"45", the stream decreased in length and shifted to a position southwest of the major SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII axis. The stream continued to decrease in length, and by 23"16" it extended only 100 m from the vent. The flow of lava was deflected each time at the southeast end of the lake, probably by a buttress of static viscous lava. Subsidiary currents split off from the stream towards the margin of the lake, diminishing in vigor with greater distance from the vents. Because the surface of the lake rose only imper- ceptibly during the last 26 hours of the eruption, we conclude that the outward flow of lava near the sur- face was compensated by slow flowage of relatively dense degassed lava downward and backward to- ward the vents in the main body of the lake, as shown in the cross section in figure 29. Presumedly, this lava was either recirculated in the lake or seeped back down inactive parts of the feeding fissure. DEVELOPMENT OF JOINT CRACKS When first traversed 4 days after the eruption, the lake surface was found to be broken by a network of cracks outlining irregular polygons of crust. Some of these cracks had been widened, forming gaping crevasses as much as 30 ecm wide and 2 m deep (fig. 30). Many were still glowing red at the bottom on August 30 and yielded temperatures as high as 960° C. The cracks more than 2.5 em wide at the surface formed an open network in the central part of the lake that radiated outward towards the edge (fig. 31). The pattern is reminiscent of the network of faults over a piercement dome. Presumedly, cracks formed as a result of thermal contraction in the thin crust of the lake during the latter part of the erup- tion were widened at the end of the eruption because the surface crust was extended by downsagging dur- ing drainback of the underlying molten lava. The formation of joint cracks by thermal contrac- tion of the cooling crust of the lake was studied in detail for 2 years by repeated mapping (at 10 ft to the inch, 1:120) of new cracks and deposits of sub- limates in an area approximately 30 m (100 ft) on a side near the center of the lake (fig. 17). Observa- tions and conclusions derived from this study and from studies of Makaopuhi and Kilauea Iki lava lakes have been summarized by Peck and Minakami (1968). Surface features in the area, other than cracks and sublimates, are shown in figure 26. The cracks were first mapped during late October, November, and early December 1963, when the crust was 3 to 4 m thick. The area was found to be divided by a random orthogonal network (Lachenbruch, ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE A25 300 FEET | 25 50 75 METRES 0 | | 0 EXPLANATION @ - Clockwise coil | Spatter rampart 9 - Counterclockwise coil f 17h45r - Outer margin of active lava circulation, dashed where ap- proximately located, at 17h45m, 22h45m, and 23h16 m on August 22, 1963, during a late stage of the eruption. ; Levee and moat Interior part of lake Margins are loci of pressure ridges and linear squeeze ups <== - Shear zone <--- Inferred direction of subcrustal lava flowage FIGURE 28.-Outer margins of active lava circulation, shear zones, and inferred directions of lava flowage on August 22, 1963. 1962, p. 48) of straight to gently curving contrac- | most of which had 3 to 6 sides. Less abundant were tion cracks, which ranged from hairline fractures to | short cracks near the centers of crack polygons and gaping cracks as wide as 20 em (fig. 32). The cracks | branching off from longer cracks. Many of the outlined irregular polygons of crust 1.5 to 6 m wide, | cracks follow preexisting flaws in the lava lake, such A26 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII FIGURE 30.-Joint cracks at the surface of Alae lava lake formed during drainback near the close of the eruption. Photograph taken August 11, 1965. Map case for scale. A A' Outward flow progressively diminished by lateral flow 2580 - Froth of accumulated gas and I@aV@ - outflows and foundering ‘‘‘‘‘ (- 785 2570 - | Post-drainback surface o $ B mean f "A is a i ° .Spatter' and pumice 2560 -| Slow flowage downward and back toward vent of relatively dense, degassed lava J - 780 ye. / lava / \ 2550 - #4 / *. ark #. / |- 775 *~. 2540 - Phisct a & andl 1’/ m__. 2530 - a- mL -.. i- 770 2520 - 0 100 200 FEET 0 25 50 METRES 2510 765 DATUM IS MEAN SEA LEVEL FIGURE 29.-Cross section along the major axis of Alae lava lake (see fig. 17) during the evening of August 22, showing inferred lava circulation. Vertical exaggeration, X 10. as the fronts of small flows and shear cracks. Where irregularities in the walls of the cracks or features such as shear zones on the adjacent lake surface permitted correlation across the cracks, no horizontal offset along the crack could be identified. In most places, only displacement normal to the plane of the crack surface was present, although, in a very few areas, one side was displaced vertically as much as 5 em with respect to the other. More than 90 percent of the crack intersections for which the angle of intersection could be determined were orthogonal (marked by solid dots in fig. 32). In many places, orthogonal intersection was achieved by curvature of one or both of the cracks close to the point of intersection. Many of the cracks in the area were very incon- spicuous, particularly during this early stage of the cooling history of the lake, when few cracks were marked by incrustations of sublimates. Some could be found only by scraping away the filamented sur- face layer and following the course of the crack on hands and knees. In parts of the area, the cracks were concealed by thin surface flows, in the cracks at the fronts of small flows, or in shallow surface ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE A27 200 300 FEET | | ] 75 METRES 0 100 | | o Mapped October and November, 1963 EXPLANATION Cracks <7.5 cm in width Cracks >7.5 cm and <10 cm in width Cracks >10 cm in width Shallow cracks <2 m in depth FIGURE 31.-Map of joint cracks on Alae lava lake that formed during drainback near the close of the eruption. cracks formed during the eruption. No doubt some cracks were missed during the mapping; radiation temperatures in December 1964 along a line near the northwest edge of the area (Decker and Peck, 1967) indicate the presence of several that had been overlooked. The surface of the area was not flat but was made up of broad hummocks separated by sharp troughs. The surface of each hummock sloped outward with increasing steepness toward the marginal trough, somewhat like an inverted saucer, as shown in the cross section of figure 32. The relief, excluding pres- sure ridges and oozeups, had a maximum of about 60 cm but averaged less than 30 cm. Each hummock was made up of one or more crack polygons and, with few exceptions, bounded by cracks on all sides. A28 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII 20 FEET FEET 0 5 METRES 2570 - | 2560 - | ! METRES 1-778 DATUM IS MEAN SEA LEVEL EXPLANATION Cracks and sublimates, December 1963 Crack >5 cm in width Crack >2.5 cm and <5 cm in width; dashed where covered -- Crack<2.5 cm in width; dashed where covered Deposit of sublimates New cracks and sublimates, May 15, 1964 New cracks and sublimates, November 16, 1964 New cracks and sublimates, August 11, 1965 Crack intersections 90° 60° or 120° FIGURE 32.-Cracks on part of the August 1963 Alae lava lake. Location of area is shown in figure 17; surface features other than cracks and sublimates are shown in figure 26. Cross section is along the southeast edge of the area. Numbers at corners of the map area refer to tagged nails marking surveyed stations. ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE A29 Where boundary cracks were not evident, they prob- | ranged from 3 to 6 m in diameter, averaging about ably were present at shallow depth but were ob- | 4% m. Most were roughly equant, but a few were secured at the surface by thin overlying flowlets. The | elongate parallel to the long dimension of the lake. troughs delimiting hummocks and the crests of the | Hollow cavities occurred at shallow depths beneath hummocks are mapped in figure 33. The hummocks | the crests of many of the hummocks (fig. 33) and & 4 A5 o 10 20 FEET Fery-- T o 5 METRES EXPLANATION Major trough -- _ Minor trough -|- Crest of dome 0 Cavity at shallow depth 33.-Map showing troughs and crests on part of Alae lava lake. Location of area is shown in figure 17. Crests of domes are marked by crosses. Cavities at shallow depth (located by stamping on the surface and listening for a hol- low sound) are marked by circles. Numbers at corners of the map area refer to tagged nails marking surveyed sta- tions. A8O were encountered in drilling at depths of 0.3 to 1.2 m. The appearance of similar features in flows ex- posed on the walls of craters and in roadcuts sug- gests that the cavities in the Alae lake were lensoid openings having convex-upward tops and flat floors. Observations during and after the Makaopuhi erup- tion indicate that hummocky topography starts forming within a few hours after formation of the crust and is completed within 1 to 4 days (Peck and Minakami, 1968, p. 1154). The uplift of polygon centers and the formation there of lensoid cavities and vesicular zones apparently is caused by the trap- ping of gas exsolved from the underlying molten lava. Escape of the gas through cracks at polygon boundaries, in contrast, leads to downsagging of the margins. When detailed mapping was first begun in late October, very few of the cracks had incrustations of sublimates along them. No sublimates are evident on a photograph from the overlook taken October 21, 1963 (fig. 34), and none were noticed during the work on the lake until late October. By the time the mapping was completed on December 17, 1963, the lake surface had cooled sufficiently for sublimates to be deposited. Many of the cracks had incrustations, FIGURE 34.-Alae lava lake as viewed from the overlook at the south rim of crater on October 21, 1963. Few sublimates have been deposited on the lake surface as yet because of the relatively high temperatures at shallow depth in the lake. Compare with figure 35. SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII yellow deposits of sulfur and white deposits of gyp- sum and anhydrite, along them. The sublimates were particularly noticeable along the short cracks within the polygons, many of which had opened during the course of the mapping. During this period, yellow sublimate-laden gases could be seen rising along some of the cracks. Most of the sublimates were washed away by heavy rains during the next several months, and the areas of incrustation became inconspicuous. On almost every day of work on the lake during the next several months, cracks could be heard open- ing in the lake with a sound like that of a sharp dis- tant explosion. New cracks and strips incrusted with sublimates were therefore remapped in the study area on May 15, 1964 (fig. 32), when the crust was 7 m thick. The mapping revealed many new cracks- mostly short cracks near the crests of hummocks but also a few cracks that extended across crack polygons and subdivided them into nearly equal parts. The short cracks at the crests of hummocks probably extended to depths of only a few feet, be- cause of the presence of underlying cavities. Almost all of the new and many of the preexist- ing cracks were by now bordered at the surface by incrustations of sublimates. Deposition of sublimates at the surface was accelerated during the first few months of 1964, as shown by the photograph (fig. 35) taken on August 25, 1964, because of drastic cooling of the uppermost part of the crust by heavy rains, which brought 167 em of rainfall between January 2 and May 28. In late December, only the upper 10 cm was below 100°C; by May, this thick- ness had increased to 1 m. As a result, gases rich in sulfur and calcium sulfate that were rising along the cracks were cooled sufficiently to deposit the subli- mates instead of carrying them into the atmosphere to be blown away by the wind. Heavy rainfall is not a necessary requirement for such deposits; subli- mates were deposited on the cooler margins of Alae lava lake and more sparingly on the center before the onset of the heavy rain. During the period from December to May, sublimates were deposited not only along new cracks but also along some preexist- ing cracks, which suggested renewed downward propagation of at least some of the cracks in the thickening crust of the lake. The thin filamented surface layer of the lake had now loosened from the underlying lava over much of the lake. This apparently resulted from the wide- spread formation of small subhorizontal fractures at the base of the layer, presumedly the result of de- vitrification of the glassy filaments and differential ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE FIGURE 35.-Alae lava lake as viewed from the head of the aerial tram at the south end of the crater on August 25, 1964. Incrustations of sublimates are apparent along some of the joint cracks as a result of more than 170 cm of rainfall during the preceding 8 months. Compare with fig- ure 34. thermal contraction of the frothy layer and the rela- tively dense underlying lava during cooling. Over narrow areas adjacent to some cracks, the loose pieces had been washed by rainwater and blown by wind into the open cracks, leaving a smooth surface of nonfilamented lava exposed (fig. 36). The rains had chilled the crust of the lake so drastically by the end of May that temperatures in the upper 2 m of the crust did not fall during the following 7 months, but instead stayed approximate- ly constant. When the study area was remapped on November 16, 1964, 15 months after the eruption and 2 months after solidification of the last interstitial melt in the lake, few new cracks or areas incrusted by sublimates were discovered (fig. 32). When the area was mapped again on August 11, 1965, almost 2 years after the eruption, many new cracks were found (fig. 32). The maximum tempera- ture in the lake by this date had fallen to about 680° C, 300°C below the solidus temperature. The new cracks, which were short and sublimate incrusted, occurred in long swarms that crossed major pre- existing cracks (Peck and Minakami, 1968, pl. 1, fig. 2). Newly deposited sublimates were also found along preexisting cracks crossed by the swarms. The AS1 FIGURE 36.-Areas bared of the filamented surface layer near joint cracks in Alae lava lake. Photograph taken March 1965. alinement and continuity of the swarms of short cracks indicate that they were the surface expres- sion of major continuous cracks at depth in the lake. A few swarms could be seen outside the study area in February 1964. More can be seen in the August 1964 photograph (fig. 35) and even more in the August 1965 photograph (fig. 37). Some of these appear to outline large polygons 30 m or so across. The mapping shows that these did not correspond to preexisting crack polygons but instead cut across them. The swarms probably represent new cracks that originated at depths of 5 to 15 m in the lake and propagated upward to the surface, as suggested by the discontinuous nature of the individual cracks in each swarm, the length of each swarm and its continuity cross major preexisting cracks, and the large dimensions of polygons outlined by some of the swarms. Between May 1964 and August 1965, temperatures in the upper metre of the crust de- creased very slightly; thus little stress induced by thermal contraction developed near the surface of the lake. As a result, the upward-propogating cracks feathered out near the surface. In contrast, tempera- tures deep within the crust continued to fall after May 1964. The maximum change in temperature between May 1964 and August 1965, and hence the maximum accumulated stress, was at a depth of about 10 m, which was slightly below the middle of A32 FIGURE 37.-Alae lava lake on August 18, 1965, as viewed from the head of the aerial tram at the south rim of the crater. Swarms of sublimate-incrusted cracks can be seen, some of which outline large crack polygons in the lake. the lava lake. Apparently the downward growth of preexisting cracks did not adequately release the ac- cumulated stress in this zone so that new cracks opened at points distant from preexisting cracks and propagated upward into the zone of small accumu- lated stress. REFERENCES CITED Ault, W. U., Eaton, J. P., and Richter, D. H., 1961, Lava temperatures in the 1959 Kilauea eruption and coooling lake: Geol. Soc. America Bull., v. 72, no. 5, p. 791-794. Ault, W. U., Richter, D. H., and Stewart, D. B., 1962, A temperature measurement probe into the melt of the Kilauea Iki lava lake in Hawaii: Jour Geophys. Re- search, v. 67, no. 7, p. 2809-2812. Dana, J. D., 1849, Geology. United States exploring expedi- tion, during the years 1838, 1839, 1840, 1841, 1842, * * *: Philadelphia, v. 10, 756 p. Decker, R. W., 1963, Magnetic studies on Kilauea Iki lava lake, Hawaii: Bull. Volceanol., v. 26, p. 23-35. Decker, R. W., and Peck, D. L., 1967, Infrared radiation from Alae lava lake, Hawaii: U.S. Geol. Survey Prof. Paper 575-D, p. D169-D175. Evans, B. W., and Moore, J. G., 1968, Mineralogy as a func- tion of depth in the prehistoric Makaopuhi tholeiitic lava lake, Hawaii: Contr. Mineralogy and Petrology, v. 17, no. 2, p. 85-115. Finlayson, J. B., Barnes, I. L., and Naughten, J. J., 1968, Developments in volcanic gas research in Hawaii, in Knopoff, Leon, Drake, C. L., and Hart, P. J., eds., The crust and upper mantle of the Pacific area: Am. Geophys. SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII Union, Geophys. Mon. 12 (Natl. Acad. Scii-Natl. Re- search Council Pub. 1687), p. 428-438. Fiske, R. S., and Koyanagi, R. Y., 1968, The December 1965 eruption of Kilauea Volcano, Hawaii: U.S. Geol. Survey Prof. Paper 607, 21 p. Grommé, C. S., Wright, T. L., and Peck, D. L., 1969, Mag- netic properties and oxidation of iron-titanium oxide minerals in Alae and Makaopuhi lava lakes, Hawaii: Jour. Geophys. Research, v. 74, no. 22, p. 5277-5293. Hikli, T. A., and Wright, T. L., 1967, The fractionation of nickel between olivine and augite as a geothermometer: Geochim. et Cosmochim. Acta, v. 31, no. 5, p. 877-884. Kinoshita, W. T., 1967, May 1963 earthquakes and deforma- tion in the Koae fault zone, Kilauea Volcano, Hawaii: U.S. Geol. Survey Prof. Paper 575-C, p. C173-C176. Lachenbruch, A. H., 1962, Mechanics of thermal contraction cracks and ice-wedge polygons in permafrost: Geol. Soc. America Spec. Paper 70, 69 p. Macdonald, G. A., 1955, Hawaiian volcanoes during 1952: U.S. Geol. Survey Bull. 1021-B, p. 15-108. Macdonald, G. A., and Eaton, J. P., 1957, Hawaiian volcanoes during 1954: U.S. Geol. Survey Bull. 1061-B, p. 17-72. 1964, Hawaiian volcanoes during 1955: U.S. Geol. Survey Bull. 1171, 170 p. Macdonald, G .A., and Katsura, Takashi, 1961, Variations in the lava of the 1959 eruption in Kilauea Iki: Pacific Sci., v. 15, no. 8, p. 358-369. Moore, J. G., and Evans, B. W., 1967, The role of olivine in the crystallization of the prehistoric Makaopuhi Tholeiitic lava lake, Hawaii: Contr. Mineralogy and Petrology, v. 15, no. 3, p. 202-223. Moore, J. G., and Koyanagi, R. Y., 1969, The October 1963 eruption of Kilauea Volcano, Hawaii: U.S. Geol. Sur- vey Prof. Paper 614-C, p. C1-C13. Moore, J. G., and Krivoy, H. L., 1964, The 1962 flank erup- tion of Kilauea Volcano and structure of the east rift zone: Jour. Geophys. Research, v. 69, no. 10 p. 2033- 2045. Peck, D. L., 1966, Lava coils of some recent historic flows, Hawaii: U.S. Geol. Survey Prof. Paper 550-B, p. B148- B151. 1974, Thermal properties of basaltic magma: results and practical experience from study of Hawaiian lava lakes-summary, in Colp, J. L., and Furumoto, A. S., eds., The utilization of volcano energy. The Proceedings of a United States-Japan Cooperative Science Seminar * * * held at Hilo, Hawaii, February 4-8. 1974: Albuquerque, N. Mex., Sandia Labs., p. 287-295. Peck, D. L., and Minakami, Takeshi, 1968, The formation of columnar joints in the upper part of Kilauean lava lakes, Hawaii: Geol. Soc. America Bull., v. 79, no. 9 p. 1151- 1165. Peck, D. L., Moore, J. G., and Kojima, George, 1964, Tem- peratures in the crust and melt of Alae lava lake, Hawaii, after the August 1963 eruption of Kilauea Volcano-A premilinary report: US. Geol. Survey Prof. Paper 501- D, p. D1i-D7. Peck, D. L., Wright, T. L., and Moore, J. G., 1966, Crystalliza- tion of tholeiitic basalt in Alae lava lake, Hawaii: Bull. Volcanol., v. 29, p. 629-655. Rawson, D. E., and Bennett, W. P., 1964, Results and power generation implications from drilling into Kilauea Iki lava lake, Hawaii: United Nations Conference on New ERUPTION OF AUGUST 1963 AND FORMATION OF ALAE LAVA LAKE Sources of Energy, Rome 1961, Proc., v. 2,-Geothermal energy, [pt.] 1: New York, United Nations, p. 347-360. Richter, D. H., Ault, W. U., Eaton, J. P., and Moore, J. G., 1964, The 1961 eruption of Kilauea Volcano, Hawaii: U.S. Geol. Survey Prof. Paper 474-D, p. D1-D34. Richter, D. H., and Eaton, J. P., 1960, The 1959-60 eruption of Kilauea Volcano: Smithsonian Inst. Ann. Rept. 1960, p. 349-855. Richter, D. H., and Moore, J. G., 1966, Petrology of the Kilauea Iki lava lake, Hawaii: U.S. Geol. Survey Prof. Paper 537-B, p. B1-B26. Sato, Motoaki, and Wright, T. L., 1966, Oxygen fugacities directly measured in magmatic gases: Science, v. 153, no. 3740, p. 1103-1105. Shaw, H. R., Wright, T. L., Peck, D. L., and Okamura, Reginald, 1968, The viscosity of basaltic magma-An analysis of field measurements in Makaopuhi lava lake, Hawaii: Am. Jour. Sci., v. 266, no. 4, p. 225-264. Shimozuru, Daisuke, Kamo, Kosuku, and Kinoshita, W. T., 1966, Volcanic tremor of Kilauea Volcano, Hawaii, dur- ing July-December, 1963: Tokyo Univ. Earthquake Re- search Inst. Bull., v. 44, pt. 3, p.1093-1133. Skinner, B. J., and Peck, D. L., 1969, An immiscible sulfide melt from Hawaii: Econ. Geology Mon. 4, p. 310-322. Swanson, D. A., Jackson, D. B., Duffield, W. A., and Peterson, D. W., 1971, Mauna Ulu eruption, Kilauea Volcano: Geotimes, v. 16, no. 5, p. 12-16. A838 Swanson, D. A., Duffield, W. A., Jackson, D. B., and Peter- son, D. W., 1973, The complex filling of Alae crater, Kilauea Volcano, Hawaii: Bull. Volcanol., v. 36 (1972) no. 1, p. 105-126. Swanson, D. A., and Peterson, D. W., 1972, Partial draining and crustal subsidence of Alae lava lake, Kilauea Vol- cano, Hawaii: U.S. Geol. Survey Prof. Paper 800-C, p. C1-C14. Tilley, C. E., Yoder, H. S., Jr., and Schairer, J. F., 1964, New relations on melting of basalts: Carnegie Inst. Wash- ington Yearbook 63, 1963-1964, p. 92-97. 1967, Melting relations of volcanic rock series: Car- negie Inst. Washington Yearbook 65, 1965-1966, p. 260- 269. Wright, T. L., Kinoshita, W. T., and Peck, D. L., 1968, March 1965 eruption of Kilauea volcano and the formation of Makaopuhi lava lake: Jour. Geophys. Research, v. 73, no. 10 p. 3181-3205. Wright, T. L., Peck, D. L., and Shaw, H. R., 1976, Kilauea lava lakes: natural laboratories for study of cooling, crystallization, and differentiation of basaltic magma: Am. Geophys. Union Geophys. Mon. [In press.] Wright, T. L., and Weiblen, P. W., 1968, Mineral composi- tion and paragenesis in tholeiitic basalt from Makaopuhi lava lake, Hawaii [abs.]: Geol. Soc. America spec. paper 115, p. 242-243. # U. S GOVERNMENT PRINTING OFFICE : 1976 O - 209-884 VOL. 81, NO. 29 JOURNAL OF GEOPHYSICAL RESEARCH OCTOBER 10, 1976 Reply EverEtT D. Jackson aAnp HErBERT R. Saw U.S. Geological Survey, Menlo Park, California 94025 We thank Seeger for amplifying the ideas presented in his earlier abstract [Seeger, 1973]. In essence, Seeger pro- poses that the stress field in the Pacific plate is dominated by divergence at its margins such that compressional stresses are parallel to the translation direction of the plate and tensional stresses are parallel to a great circle through the plate pole passing just southeast of Hawaii. These stress directions are roughly coincident with those that we deduced from the distribution of volcanic frac- tures in the central Pacific. We did not select a single mechanism to account for the stress distributions implied by our data but attempted to evaluate the possible forces acting to produce both transmitted and inherent stresses in the plate. We concluded that the resultant effective stresses in the plate over the last 40-50 m.y. had been such that they could be considered to have been caused by the net effect of dynamic forces resulting in a dominantly right-lateral rotational couple acting within the plane of the Pacific plate. We concluded that the dominant stress orientations prior to 40-50 m.y. ago were caused by a tend- ency for left-lateral stress rotations. Seeger's divergence model indeed has the proper sense to account for the more recent right-lateral stress rotations. So do several other mechanisms capable of producing transmitted stresses which we also explored, i.e., convergent forces at Pacific trenches and drag at the lithosphere-asthenosphere boundary. It is plainly simplistic to plead for the domi- nance of one stress-producing mechanism until quantita- tive modeling of all transmitted and inherent stresses sets some limits on relative stress magnitudes. Beyond proposing that the divergent plate boundary model accounts for present stress distributions in the Pacific plate, Seeger implies that his model accounts for the localization, geometry, and geochronology of the Hawaiian-Emperor chain. It is not clear, however, how the model accounts for the propagation direction of the chain {Jackson, 1976], for the episodic nature of the volcanism in time and space [Shaw, 1973; Shaw and Jackson, 1973], or for the systematics of the curvature of the loci [Jackson et al., 1972, 1975]. Seeger's [1973] speculation was not, moreover, a pre- decessor to our hypothesis that curved volcanic loci map resultant intraplate stresses caused by rotational couples acting on the Pacific plate. Indeed, our thoughts grew from the observations that Hawaiian-Emperor volcanic Copyright © by the American Geophysical Union. 5347 loci were systematically sigmoidal [Jackson et al., 1972] and from the observation that the rift zones of isolated Hawaiian shields were parallel to loci [Fiske and Jackson, 1972]. In fact, as we noted in the paper, we considered our work to be an extension of the explanation of Jackson et al. [1972] of volcanic loci as extensional fractures in the oceanic crust and upper mantle. Further, we wish to emphasize that the data set of shield centers of Jackson et al. [1972] does not necessarily represent 'topographic' highs but rather the intersections of rift zones deduced from bathymetric maps by use of the principles of Fiske and Jackson [1972]. The locations of these rift zones are given in Bargar and Jackson (19741. Whereas the development of our ideas on stress fields in the Pacific plate had quite different origins than those of Seeger, we are gratified that his independent deductions are consistent with our conclusions on the general orienta- tions of maximum and least principal stresses. We are pleased to see his follow-up of the ideas he presented in 1973 and hope that more will be done along these lines to provide further documentation of conditions affecting the dynamic states of the lithosphere. REFERENCES Bargar, K. E., and E. D. Jackson, Volumes of individual shield volcanoes in the Hawaiian-Emperor chain, U.S. Geol. Surv. J. Res., 2, 545-550, 1974. Fiske, R. S., and E. D. Jackson, Orientation and growth of Hawaiian rifts: The effects of regional structure and gravita- tional stresses, Proc. Roy. Soc. London, 329, 299-326, 1972. Jackson, E. D., Linear volcanic chains on the Pacific plate, in The Geophysics of the Pacific Ocean Basin and Its Margin, Geophys. Monogr. Ser., vol. 19, AGU, Washington, D. C., in press, 1976. Jackson, E. D., E. A. Silver, and G. B. Dalrymple, The Hawaiian- Emperor chain and its relation to Cenozoic cireumpacific tec- tonics, Geol. Soc. Amer. Bull., 83, 601-618, 1972. Jackson, E. D., H. R. Shaw, and K. E. Bargar, Calculated geochronology and stress field orientations along the Hawaiian chain, Earth Planet. Sci. Lett., 26, 154-155, 1975. Seeger, C. R., Plate tectonics, mantle plumes, and the origin of the Hawaiian Islands (abstract), Fos Trans. AGU, 54(4), 240, 1978. Shaw, H. R., Mantle convection and volcanic periodicity in the Pacific, Evidence from Hawaii, Geol. Soc. Amer. Bull., 84, 1505-1526, 1973. Shaw, H. R., and E. D. Jackson, Linear island chains in the Pacific: Result of thermal plumes or gravitational anchors?, J. Geophys. Res., 78, 8634-8652, 1973. (Received March 9, 1976; accepted March 30, 1976.) October 10, 1976 Vol. 81, No. 29 JOURNAL OF GEOPHYSICAL RESEARCH a CCMKEXT ON 'STRESS FIELDS IN CENTRAL PORTIONS OF THE PACIFIC PLATE: | 6 DELINEATED IN TIME BY LINEAR VOLCANIC CHAINS! BY EVERETT D. JACKSON AND HERBERT R. SHAW C. Ronald Seeger Department of Geography and Geology,/Western Kentucky University Nuss me vinny ur" > Bowling Green, Kentucky 42101 Jackson and Shaw [1975] have made a very in- teresting contribution to the understanding of the geology of the Pacific plate. They have re- lated the occurences of linear island chains to the stress field within the plate in a new way. Since they kindly reference my unpublished sug, gestions [Seeger, 1973] with regard to the con- vergence or divergence of plate boundaries as important mechanisms for producing compressional or tensional stresses within the plate, I should put those suggestions on record. Indeed, seme of my speculations may have been predecessors to the present 'rational stress' hypothesis of Jackson and Shaw, although I fully realize that others, referenced by Jackson and Shaw, have also specu- lated about stresses in the Pacific lithosphere. My model, like that of Jackson and Shaw, also varies from these others in that the stress dis- tributions lie at different attitudes, that they vary with time, and that they are controlled by the net dynamic effects imparted by the global kinematics of plate motion. The idea that Jackson and Shaw reference is simply that the topology of motions of plates on the surface of a sphere (the earth is this shape within the limitations of this discussion) re- quires that most edges parallel to the direction of plate motion be either divergent or convergent from one plate to an adjacent plate. These boundaries are in addition to the more obvious boundaries of divergence (the spreading plate boundaries) and convergence (the subduction zones). Needless to say, the stress conditions in the two adjacent plates are controlled by which of these (divergence or convergence) is the case at the plate boundary. -For an example of the possible effect of this, suppose we construct a great circle passing just southeast of the Hawaiian Islands from the Samoa Islands past Vancouver Island and through the suggested pole of rotation of the Pacific plate at 67°K, 90°W [Hey and Morgan, 1971] (other sug, gested poles do not deviate much). This circle probably represents the line across the Pacific plate perpendicular to the direction of the pre- sent plate motion toward the northwest. If we then examine the tectonic trends toward the northwest from Samoa and from Vancouver islands, we find that for distances greater than 1000 km the tectonic elements indicate divergences (open- ing) at plate boundaries from the required small circle paths of about 10° and from 10° up to 40° or even 60°. Figure 1 shows the tectonic elements of Johnson and Molnar [1972] at the southern end of the great circle near Samoa. The heavier line is the great circle, and the lighter line is per- Copyright 1976 by the American Geophysical Union pendicular to it. Notice the divergence between the latter line and the tectonic elements as one goes from the great circle toward the northwest. This implies that the deviatoric stress within the Pacific plate as it moves in this north-, westerly direction should be tensional in the direction of the great circle. At the northern edge of the Pacific plate, where the great circle passes just south of the Queen Charlotte Islands, a similar situation exists. Figure 2 shows the same great circle line and a perpendicular to it in a northwesterly direction. The active plate bounding fault in this region is the Fairweather fault (there have been recent earthquakes on it). Again, note that the fault trace diverges from the perpendicular to the great circle in a northwesterly direction. This again indicates that, as the Pacific plate moves northwest relative to the North America plate, the deviatoric stress should be tensional l I I I T T T T T Pacific - Plate ~120" Australian Plate Pacific Plate 140°€ 160°E 180° Fig. 1. Map of the tectonics between the Pacific plate and the Australian plate near the Samoa Islands to the west [from Johnson _ and Molnar, 1972]. The heavy line is the great circle perpendicular to the direction of plate motion, and the lighter line is per- pendicular to it in the direction of plate motion. Note the divergence between the latter line and the tectonic lines west of Samoa. 5344 Seeger: Commentary : 5345 y $92 EF. ~E;%' .C. 3 Fig. 2. A part of the tectonic map of North America [King, 1969] at the boundary of the Pacific plate and the North America plate near Queen Charlotte and Vancouver islands. The heavy lines represent the great circle perpendicular to the present plate motion and a perpendicular in the direction of present plate motion. Note the divergence between this perpendicular and the Fairweather fault, the plate boundary. + Slax! \Nefe) A H Now Menges oe th £002 26.005 1 09s daves parallel to the great circle direction. Note both boundaries of the Pacific plate along this also the depression to the west of Queen great circle direction. I further suggested that Charlotte Islands. these tensional stresses could be transmitted in Thus tensional deviatroic stresses parallel to the manner proposed by Elsasser [1971] and there- the great circle direction are to be expected at ' fore the Pacific plate presently fails by ex- 5346 LEGEND HEAVY LINES THDICATE LOCI CF SHIELD VOLCANOS - --- X TOPOGRAPHIC KIGH (AFTER JACKSON AND OTHERS. 1972) T SHEAR STRESS ~ NORMAL STRESS O3 LEAST PRINCIPAL STRESS HAWAIIAN DEEP Fig. 3. Map by Jackson et al. [1972] of the loci of shield volcanoes of the Hawaiian, Emperor seamount chain to which has been added the great circle perpendicular to the dir- ection of Pacific plate motion, here desig- nated 03 (see text for discussion), and a sug- gested Mohr failure envelope for the Pacific plate (see text). tension fracture or low conjugate shear angle faulting as it passes the location of the great circle near Hawaii. The sigmoidal character of the volcanic loci is consistent with this inter- pretation. This is illustrated by Figure 3, which uses the map of the loci of shield vol- canoes and topographic highs of Jackson et al. [1972] as a base; note that these features are nearly perpendicular to the previously establish- ed great circle direction which is now labeled 03, the least principal deviatoric stress. Figure 3 also shows how this pattern can be un- derstood in theory by the use of a simple Mohr diagram. The Mohr circles are drawn on the basis cf the experimental evidence of Handin [1966] for a pyroxenite which failed in both compression and extension. Notice that the failure envelope passes very close to the origin. This means that if the Mohr failure envelope in Figure 3 is a good approximation for that of the Pacific plate, then it could be failing by extensional fractur- ing, or at low conjugate shear angles, and at low tensional stress, as the Elsasser mechanism would seem to allow (e.g., the plate acting as sort of a 'stress guide' and the deviatoric stress values never needing to be very high). I further speculated [Seeger, 1973] that the Seeger: Commentary Emperor seamount chain (whatever the age turns out to be) could possibly have formed when the plate boundaries were essentially convergent and therefore compressional deviatoric stresses ex- isted within the Pacific plate. The angle 0 in Figure 3 might be an appropriate fault angle for this condition. Of course the details of the plate boundary history have not been worked out; more details need tc be incorporated into the model, For ex- ample, it may be more than accidental that the Mendocino fracture zone (extended) passes just north of the 'bend' between the Emperor chain and the Hawaiian chain and that a line from the southeast or recent end of the latter extended eastward, normal to the magnetic anomalies, would nearly intercept the southern tip of Baja California. I congratulate Jackson and Shaw for continuing to develop the model to which I subscribe; my simpler version, though, still has a certain appeal in accordance with 'Occam's Razor.' References Elsasser, W. M., Two-layer model of upper- martle circulation, J. Geophys. Res., 76, 4744-4753, 1971. Handin, J., Strength and ductility, in Handbook of Physical Constants, Mem. 97, edited by §. P. Clark, Jr., pp. 223-289, Geological Society of America, Boulder, Colo., 1966. Hey, R. N., and W. J. Morgan, Parallel seamount chains in the northeast Pacific (Abstract), Eos Trans. ACU, 52, 236, 1971. Jackson, E. D., and H. R. Shaw, Stress fields in central portions of the Pacific plate: Delin- eated in time by linear volcanic chains, J. Geophys. Res., 80, 1861-1874, 1975. Jackson, E. D., E. A. Silver, and G. B. Dalrymple, Hawaiian-Emperor chain and its re- lation to Cenozoic circumpacific tectonics, Geol. Soc. Amer. Bull., 83, 601-618, 1972. Johnson, T., and P. Molnar, Focal mechanisms and plate tectonics of the southwest Pacific, J. Geophys. Res. 77, 5000-5032, 1972. King, P. B., Tectonic map of North America, U. S. Geol. Survi, Washington, D. C., 1969. Seeger, C. R., Plate tectonics, mantle plumes, and the origin of the Hawaiian Islands (Abstract), Eos Trans. AGU, 54, 240, 1973. (Received December 8, 1975; accepted March 30, 1976.) Cooling and Vesiculation of Alae Lava Lake, Hawan By D. L. Prox ALIMON OF AEAXE_LAVA LEARKE, H AW A 1d Repeated measurements of temperatures in drill holes and the altitude of the surface document the cooling of a thin ponded basalt flow UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON: 1978 UNITED STATES DEPARTMENT OF THE INTERIOR CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY W.A. Radlinski, Acting Director Library of Congress Cataloging in Publication Data Peck, Dallas Lynn, 1929- Cooling and vesiculation of Alae Lava Lake, Hawaii. (Solidification of Alae Lava Lake, Hawaii) (Geological Survey Professional Paper 935-B) Bibliography: p. 49-50. Supt. of Docs. No.: 119.16:935-B 1. Basalt-Hawaii--Alae Lava Lake. I. Title. II. Series. III. Series: United States Geological Survey Professional Paper 935-B. QE462.B3P4 $82'.2 77-608 302 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 Stock Number 024-001-0305 3-5 CONTENTS Page Page ADSETACL 00 SEI ALL UIN ered al oa anus neend 1 | Density of Alae lava and variation with time-Continued introduction "->. 2222 tase so ea e neden en - dea ae wake 2 Density and vesicularity 19 Acknowledgmente s.:... cern 3 Altitude of the lake surface 23 Methods of study :...... :. ser neil deus ages 3 Changes in altitude during solidification ____________ 24 Drilling. es:: nob. 3 Changes in horizontal distance during solidification __ 30 Temperature measurements 5 Changes in altitude during further cooling __________ 31 Measurements of the altitude of the lake surface _____ 6 Interpretation of changes in altitude ____________________ 82 Precipitation and air temperatures ._...._________________.__ 7 Subsidence of the lake surface and calculation of the in the lava lake 7 coefficient of thermal expansion __________________ 32 Definition of crust and fAuid lava 7 Changes in altitude of the surface during solidification Effect of coolant water on temperatures__________________ 11 ofthe lake kene cay 33 Temperatures in fluid lava below the crust ______________ 12 Calculation of the density of fluid lava in the lake ________ 35 Thermal history of the Jake. _. 13 | The pattern of cooling in three dimensions __________________ 37 Maximum temperatures and temperatures at the basal Thermal properties. of Alac basalt 40 contact 1-2. bec. ere Ie ll cane aaa ed inl. ased 17 | Thermal modeling 2 ea 45 Density of Alae lava and variation with time of the altitude of Conclusions:. s.. cll el all Ia LIL 48 theilake SUMACE -_. "22.2 Lenn clon n enn nece 19 A References cited.. .... Gece 49 ILLUSTRATIONS Page FicurE 1. Map of Alae lava lake showing contours at the base and margin and the location and identification number of surveying stations and drill holes -. _s - :~ 2a: .no oll e itn ee LOLI GL anl ane anat dane nac 4 2 Photograph showing drilling in Alac ava lake LL rotel. 5 3-17. Graphs showing: 64 Representative temperature co cunt c_ ._ DLOs MOLL LiL [ILL. ua. 9 4. Temperature and depth of the tip of a hollow mullite probe as a function of time during its insertion into fluid lava -__- 10 5. Thickness of the upper crust near the main drilling site as a function of time 10 6. Temperature profile across the zone of crystallization at the base of the crust on December 30, 1963 ______________ 14 7. Temperatures near the bottom of drill hole 3 after drilling through the base of the crust _______________________-- 12 8. Depth of isotherms in drill hole 4 before and after drilling on November 27, 1963 12 9. Temperature profile across partly molten crust and into underlying fluid lava on November 8, 1963 ______________ 13 10. Depth of isotherms and rainfall from August 1963 to August 14 11. Depth of isotherms from August 22 to September 17; 19069 15 12. Temperature profiles in the upper part of drill hole 5 showing the effects of abundant rainfall ____________________ 16 13. Temperature profiles in five drill holes on January 21, 1965, and a computed profile ____________________________ 17 14. Maximum temperatures at three sites as a function Of time LQ 19 15. Proportionate depth of the maximum temperature -.2... _... _ n Neral lyses 20 16. Average core density and porosity (vesicularity) as a function of depth 20 17. Maximum size of vesicles as a function of depth: en tLe d 21 18.. Photomicrograph iof vesicular basalt "-.. LL nba nn ners 22 19: Graph showing vesicularity as a function of time l.. LE LO 23 20. Graph showing estimated rate of vesiculation and cumulative abundance of vesicles as a function of temperature. 23 21. Map showing rate of change in altitude of the lake surface during solidification of the lake from August 30 to September 24 10 20, 1063 : :c 200.00. s. cu Jal l ce ce en aar n on's age anand ae d an bee ae ae ad ela e alea eed 24 22. Map showing rate of change in altitude of the lake surface during solidification of the lake from September 24 to 26, 1903, to September 24. 1064. "__... 00 nl Ou DOLL lot eben ln Pete ay ot __ 29 28.. Graph showing cumulative changes in altitude of five stations LiccLLLG.LLLI.L.L.L 30 II IV FicurEs 24-27. 24. 25. 26. 27: TABLE 28-30. 28. 29. 30. 31. 32. 33. 34. 35-39. 35. 36. 37. 38. 39. 40. $ 19 sip to bo 1. CONTENTS Page Maps showing: Total uplift of the lake surface an ane ne seen enne me nen gand an 30 Changes in horizontal distance between stations during solidification of the lake _____________________________--- 30 Rate of subsidence of the lake surface after complete solidification in late September 1964 _____________________-- 31 Estimated daily average decrease in temperature in the lake from January to August 1965, as determined from measured rates of subsidence _L >: l.,... elec EL em careens oen ble aun bea ame alue e 33 Graphs showing: Rates of uplift and subsidence of stations on the lake compared with maximum temperatures in the lake _______- 34 Observed and calculated rates of uplift and subsidence at the main drilling site 34 Temperature profiles on September 25, 1963, and August 19, 1965, and a section through the lake on September 25, 1968; L. ter lune eae i nen nece ee oen end bee daren aaa E nene s war co on Be ns me al ue ae beige fees 36 Map showing approximate position of the outer margins of the 1,080°C isothermal surface in the lake as a function of Hime. . cr sL . AA pans eb enaine cs - aol ces aik be mals ol a ante ee be oe au a areal i 38 Map showing outer margins of the 1,000°C isothermal surface (approximately the solidus) as a function of time _- 39 Longitudinal cross section showing the positions of the 1,000°C isothermal surface as a function of time _______--- 40 Longitudinal cross section showing the positions of the 700°C isothermal surface as a function of time _________--- 41 Graphs showing: Thermal conductivity at 35°C of alr-saturated Alae basalt cl 42 The ratio of conductivity at high temperature to conductivity at room temperature of four samples of basalt from Alae Crater on rse ear LULU. inne nica ban are aas Se Pe areal a nn all an ame a au nn aie bos > alue Bd aie oe 43 Calculated heat capacity of Alae basalt as a function of 44 Calculated diffusivity of solidified Alae basalt as a function of depth and time 45 Computed and observed temperature profiles in Alae lava lake 46 Generalized radial section showing computed position of the 1,000°C isotherm after 120, 240, and 360 days -__- 48 TABLES Page ARecord of drill holes:in Alac lava lake ._. ...-. lio iero .D Coot s one lass ae 5 Rainfall-at the south rim of Alae Crater 000. races - es en te asm 8 Temperatures in Alae lava - 2. 2220. __ 1 oce ep coons cer cei oe oa oan ain able mn nied amie a ll a aln and an a a ale miele a mpm a bee a 52 Depths and values of maximum temperatures and of basal contact temperatures in Alae lava lake ____________-- 18 Bulk density and porosity of:drill core cedc 20 Altitude of stations on the lake surface .. llc ner ee Ot nage ce assole 25 Thermal conductivity of four samples of basalt from Alae crater as a function of temperature _________________--- 42 Specific heats of minerals and calculated heat capacity of Alae basalt _______________________________QQ_QQQQQ_Q_ 44 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII By D.L. PECK ABSTRACT The August 1963 eruption of Kilauea Volcano, Hawaii, left a stag- nant lake of molten lava as much as 15 m deep in Alae pit crater. Temperatures in the lava lake, which initially were as much as 1,140°C, were monitored by means of thermocouples in drill holes in the crust of the lake starting 7 days after the eruption and continu- ing until the entire lake had cooled to less than 100°C 4 years later. Solidification of the lake proceeded by the slow increase in thick- ness of crusts at the top and bottom of the lake. Crystalline basalt in the upper crust at any given time graded downward into fluid lava through a zone of partly molten crust in which temperatures and melt content increased smoothly downward. The base of the crust, defined as the horizon where yield strength abruptly decreased, was at a temperature of 1,065°C+5°C; at this temperature crystals and melt were in approximately equal abundance. The lava was com- pletely crystallized except for a sparse silicic residuum at 980+10°C. The thickness of upper crust and the depth of isotherms in the crust and melt increased linearly with the square root of time during the first three months. Throughout this period, the thickness of upper crust in meters was 0.00132 V?-0.18, where t equals time in seconds after the eruption. The lake solidified (that is, cooled to less than 1,065°C) 10% months after it had formed, and it cooled below the solidus temperature of 980°C 2% months later. Two years after the eruption the maximum temperature in the lake was 672°C. Two years later, it had fallen to 86°C. As the lake cooled, the maximum temperature was found at steadily increasing depths; the propor- tionate depth was 0.6 at solidification, 0.8, 2 years after the eruption, and greater than 1.0 (that is, below the lake) 2 years later. Rainfall, which ranged from 2.0 to 3.35 m/year, apparently did not greatly decrease the time of solidification, but it did strongly affect tempera- tures in the upper crust during solidification and greatly increased the rate of postsolidification cooling. Temperatures at the basal con- tact of the lake were measured repeatedly in drill holes that com- pletely pierced the lake after it solidified. Uncertainties as to the precise location of the contact in each hole limit the accuracy of the determined values, but the data indicate a maximum contact tem- perature of 700+15°C, 60 percent of the initial temperature of the lake. The molten lava in Alae vesiculated because of the exsolution of gases as it cooled and solidified. The abundance of vesicles, as deter- mined by point counts and density measurements on drill core, de- creased from 40 percent near the surface to a nearly constant 11 percent below 3 m. Large spheroidal vesicles, which were abundant near the surface but decreased sharply in size and abundance with depth, apparently recorded degassing of the super-saturated lava during the first 100 days after the eruption. Below 6 m the vesicles are almost entirely minute, angular pores formed at lower tempera- tures from gas exsolved from the lava during crystallization. Because of vesiculation, the molten lava in the lake expanded as it solidified, causing the surface of the lake to extend horizontally and to rise vertically. The solidified margin, in contrast, subsided because of thermal contraction. During the year of solidification, the surface above the thick central part of the lake rose as much as 0.25 m whereas the margin subsided as much as 0.06 m. Repeated surveys of the altitude of stations on the lake surface, accordingly, outlined the lens of fluid lava and partly molten crust at temperatures of more than 1,000°C and recorded its shrinkage with time. They show the location of the last lava to solidify and demonstrate that the lake solidified more rapidly where it thinned over a buried 1840 spatter cone. The maximum uplift during any given time interval was above the lateral edge of the 1,080°C isothermal surface, that is, near the crust-fluid lava interface in the lake. After the last interstitial melt in the lake had solidified in late September, 1964, the entire lake surface subsided at a rate that decreased from 0.006 cm/day in late 1964 to 0.003 cm/day between May, 1966, and September, 1967. In the first year after solidification, subsidence was greatest between the center and edge of the lake in a belt that migrated inward with time and presumedly defined the part of the lake undergoing the most rapid cooling-possibly because of the cooling effect of circulat- ing rain water in the cooler margin of the lake. Comparison of subsidence rates with changes in temperature in the lake indicates that the effective bulk coefficient of linear thermal expansion was 3+1x10 °/° C. This value is approximately one-half that of the values reported from measurements on similar rocks in the laboratory. The differences probably can be attributed to the formation of microfractures during cooling of the basalt as the result of differential thermal contraction of the constituent mineral grains and glass. Where one uses this value for the coefficient of thermal expansion and relevant temperature data, the density of molten lava in Alae in September 1963 can be calculated from the average den- sity of drill core and the change in volume during solidification and further cooling indicated by the level surveys. The calculated value of 2.79+0.03 g/cm? is the bulk density at 1,130°C of fluid basaltic lava containing an estimated 15 percent crystals. The calculated density of the melt fraction itself, assuming that exsolved gases are negligi- ble, is 2.74 g/cm. Several thermal properties of Alae basalt were measured in the laboratory or were calculated from laboratory data on the con- stituent minerals. The conductivity at room temperature of basalt from Alae lava lake and of similar Hawaiian basalts decreases with increasing porosity and increases with increasing olivine content (Robertson and Peck, 1974). Conductivity at room temperature, KRT, for the relatively olivine-poor Alae basalt with air in the pores is > SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII related to porosity, ¢, by the expression KRT = (1.929-1.554q). Previously unpublished experimental measurements on basalt from Alae crater by Kaoru Kawada (1966) and by H. R. Shaw (1966) indicate that the conductivity of vesicular basalt increases linearly with temperature at a rate of 0.03-0.09 percent/°C, presumedly be- cause of radiative heat transfer across the vesicles, convective heat transfer by air in the vesicles, or progressive closing of microfrac- tures as the result of thermal expansion of the sample during heat- ing. The heat capacity (c) was calculated by R. A. Robie from modal and laboratory data and was fitted to the following function of tem- perature (T, in Kelvins): c = 0.2356 + 4.3635 x 10~°T- 6.3440 x Thermal diffusivity varied in the cooling lava lake with time and depth. Calculated values typically increased from about 0.005 cm*/ sec near the surface to 0.006 cm*®/sec at depths below 2-5 m. The latent heat of completely crystalline Alae basalt, as calculated from experimentally determined heats of fusion of the constituent miner- als (and the normative minerals in the glass) is 94 to 85 cal/g. The calculated latent heat that was released during crystallization in the lake, however, would have been almost 20 percent less than that value, 77-67 cal/g, as the result of the crystallization prior to erup- tion of 13 percent olivine, pyroxene, and plagioclase and the reten- tion of nearly 7 percent silicic residuum at the end of crystallization. The cooling history of the lake was analyzed by comparing mea- sured temperature profiles with profiles computed on the basis of numerical analysis of one- and two-dimensional lava lake cooling models. Measured temperature profiles in the central part of the lake can be approximated with one-dimensional models that use a con- stant density for the basalt, a constant diffusivity of 0.006 cm*/sec, and the calculated heat loss of 620 cal/cc from heating and vaporiza- tion of the measured rainfall. A better correspondence is obtained using more complex models incorporating variable thermal and physical properties for the basalt. In the computations, diffusivity was allowed to vary with the calculated increase of heat capacity with temperature, with the measured increase in density with in- creasing depth in the lake, and with variations of conductivity with temperature and basalt vesicularity. Heat loss by heating and vapor- ization of rain water was computed using the measured density structure of the lake. Low values for density and conductivity were given to a shallow cell in the lake to simulate the highly vesicular and cavernous zone found at depths between 0.3-1.2 m. The closest match of computed and observed temperatures was obtained using a latent heat of 80+10 cal/g (from 80 percent crystallization) and con- ductivity (KT) that was related to temperature, T°C, and conductiv- ity at room temperature, KRT, by the expression Kr = KRr(1+BT), where B equals 0.0006+0.0001. In the computations, 10 representa- tive profiles measured during the first year of cooling were matched at 1-foot (0.3 m) intervals with an average deviation of 2°C. Computations using two-dimensional models confirmed the lateral variations with time of temperatures in the lake inferred from tem- perature measurements in drill holes and precise level studies, and indicate that temperatures in the lake were decreased more than 1°C by marginal cooling during the first year only in the outer 15 m of the lake, and in the fourth year only in the outer 45 m. Computations using "rain" and "no rain" models indicate that the lake had cooled to 100°C by mid-April 1967, 4% months before the final temperature measurements. Without rain the lake would have taken 19.5 years to cool to 100°C. INTRODUCTION Six days after the August 1963 eruption of Kilauea Volcano, Hawaii, had formed a shallow lake of basaltic lava in Alae crater, a hole was drilled 0.86 m into the crust of the lake to within 0.01 m of the molten lava. Temperatures were measured in the drill hole the fol- lowing day, and a grid of stations was established over the surface of the lake. During the ensuing year of solidification of the lava lake and the following 3 years of cooling to less than 100°C, several additional holes were drilled in the lake, temperatures were measured repeatedly, and the relative altitudes of the stations were periodically redetermined. These measurements provide a unique body of data on the rate and pattern of cooling of a ponded basaltic lava flow. Data from addi- tional measurements on drill core from the lake, in- cluding density, vesicularity, and thermal conductiv- ity, together with thermal coefficients calculated from the modal mineral composition of the core, were used in detailed comparison of the observed cooling history with numerical cooling models. This report is the second of three chapters describing the recorded history of Alae lava lake. Chapter A de- scribes the eruption of August 1963, and the surface features of the lava lake including joint cracks. Chap- ter C (Wright and Peck, 1978) presents the results of chemical and petrographic studies of pumice from the eruption and drill core from the lake. Aspects of the cooling history of the Alae lava lake have been de- scribed in the following preliminary reports: Peck, Moore and Kojima (1964) reported temperature mea- surements in the lake through February, 1964; Peck, Wright, and Moore (1966) described the crystallization of basalt in the lake; Decker and Peck (1967) analyzed infrared temperatures at the lake surface. Shaw, Hamilton, and Peck (1977) formulated numerical mod- els for the cooling history of lava lakes, which were applied to Alae lava lake by Peck, Hamilton, and Shaw (1977). Temperature measurements in other recent Kilauean lava lakes are reported by Ault, Eaton, and Richter (1961), Ault, Richter, and Stewart (1962), Rawson (1960), Richter and Moore (1966), Wright, Kinoshita, and Peck (1968), and Wright and Okamura (1977). Wright, Peck, and Shaw (1976) summarize studies made on all three Kilauean lava lakes. Alae crater was one of several pit craters on the upper east rift zone of Kilauea Volcano, Island of Hawaii. I refer to the crater and the August 1963 lava lake in the past tense because the lava lake was buried by eruptions starting in February 1969 (Swanson and others, 1971) and the crater was later completely filled with lava and the site covered by a broad, 100-m-high basaltic dome satellitic to the new shield Mauna Ulu (Swanson and others, 1973; Swanson and Peterson, 1972). The crater was about 8 km southeast of the summit caldera, and in 1963 was adjacent to the hard- surface Chain-of-Craters Road (Peck and Kinoshita, 1976, fig. 2). The surface of the lava lake was 150 m below the rim of the crater and could be reached by a COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 3 rude trail assisted by ropes over the steeper parts. Be- cause of the steepness of the trail, an aerial tram into the crater was constructed for lowering equipment and water soon after the August 1963 eruption, before a program of drilling and other studies was started. The eruption that formed the August 1963 lava lake lasted from 18h10m' August 21 to 08h10m August 23. The bulk of the lava in the lake was erupted during the first few hours, when fountains on the floor and north wall of the crater fed lava into the lake at rates as great as 2.3 x per hour and maximum temperatures of about 1,160°C. During the last 12 hours of the erup- tion, the rate of extrusion dropped to near zero and maximum temperatures fell to 1,140°C. Near the end of the eruption, 1.8 x 105m? of lava drained back into the vents, leaving a stagnant lake containing 6.6 x of homogeneous tholeiitic basalt. The lake was an elliptical lens 305 m long, 245 m wide, and as much as 15 m deep. The outline of the lake, contours at the base and margin, and the location of surveying stations and drill holes are shown in figure 1, together with cross sections along the major and minor axes. All but the northwest end of the lake was bordered by a levee of strongly jointed, discontinuous pressure ridges, and the levee was bounded for most of its length by a moat 8-15 m wide and about 2 m deep. The lake within the levee had a hummocky surface that sloped almost imperceptibly toward the vent at the north end of the lake. The surface was nearly flat having an av- erage relief of less than 0.3 m, but it was crossed by sharp pressure ridges and linear ooze ups that stood as high as 1 m above the lake surface. The lake deepened sharply inward from the edge to an average maximum depth of 14.0 m. The base of most of the lake was also nearly flat, but it sloped at a small angle to a low area beneath the north-central part of the lake, and the lake thinned to 11.3 m over a buried spatter ridge formed during the eruption of 1840. The lake surface was bro- ken by a network of joint cracks, some of which were opened during the later part of the eruption to form gaping crevasses as much as 30 cm wide and 2 m deep (Peck and Kinoshita, 1976, fig. 31). Other joint cracks formed as the result of thermal contraction during later cooling of the lake (Peck and Minakami, 1968). Solidified lava from depths below 1.5 m in the lake is homogeneous gray basalt containing sparse olivine phenocrysts in a fine-grained intergranular groundmass. The average modal composition in weight percent (Wright and Peck, 1978) is olivine, 3.5; clinopyroxene, 42.1; plagioclase, 41.6; ilmenite, 4.4; magnetite, 1.0; apatite, 0.65; and clear residual glass, 6.75. The Alae basalt is a tholeiite having a very uni- ! All times given are in hours and minutes, Hawaii Standard Time. form bulk chemical composition as follows (in weight percent): SiO, 50.47; AlaOs, 13.67; FesOs, 1.30; FeO, 9:60, MgO. 7.55;,.Cad, 11.11; NaO, 2.38; KaQ, 0.54; 2.74; POs, 0.27; MnO. 0.17. ACKNOWLEDGMENTS The study of Alae lava lake was a team effort by the staff of the Geological Survey's Hawaiian Volcano Ob- servatory. It was started under the direction of Scientist-in-Charge, James G. Moore, and was com- pleted under his successor, Howard A. Powers. Thomas L. Wright helped in all aspects of the field studies. In the gathering of the temperature data, I am particu- larly indebted to the following staff members: George Kojima, who constructed the thermocouples; William Francis, Reginald Okamura, Eliot Endo, and Willie Kinoshita, who helped in the drilling and in the tem- perature measurements; Burton Loucks, who made much of the special equipment used in the study; and John Forbes, who made the rainfall measurements. The continuous recording of temperatures in drill holes in the lake was a joint effort with a visiting group of scientists and technicians from Japan that included Profs. T. Minakami, D. Shimozuru, S. Aramaki, and K. Kamo, and Messrs. T. Miyasaki and S. Hiraga. Dr. Aramaki also helped in the drilling. Reginald Oka- mura and Eliot Endo helped with the level surveys and made most of the measurements of bulk and grain den- sity. Analysis of the temperature data was made possi- ble by the laboratory measurements of thermal proper- ties by E. C. Robertson, Prof. Kaoru Kawada, and Joel H. Swartz and by the numerical analysis by Prof. J. C. Jaeger, H. R. Shaw, and M. S. Hamilton. R. A. Robie, D. R. Wones, and H. S. Yoder supplied essential data on other thermal properties. METHODS OF STUDY DRILLING Holes were drilled in the solidified crust of the lava using tungsten carbide bits in a portable core drill powered by a 9-horsepower gasoline engine (fig. 2). Most of the holes were 2.86 cm in diameter, yielding 1.43 ecm diameter core, but the upper part of one hole (drill hole 9) was drilled to 3.8 cm diameter with EX bits. Twelve holes were drilled during the 15 months following the eruption; the holes had a total depth of 100 m. Dates and depths of drilling are given in table 1, and the location of the holes are plotted in figure 1. A total of about 18 kg of core was recovered from the lake during the drilling. All except the first two holes were drilled by mounting the drill on a portable mast an- chored in the crust and cooling the bits by means of water pumped through the drill pipe by a portable 38% horsepower gasoline engine. The water was fed SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII EXPLANATION Surveying station on lake surface B * _ Drill hole --- Edge of lake 0 50 100 METERS 122.20 e E GAA ITL bccn cnd Contours in feet above sea level A $ 2600 /Spatter ridge Moat 2500 FEET ABOVE SEA LEVEL 0 25 METERS L_] B 2600 2500 FEET ABOVE SEA LEVEL FicurE 1.- Map of Alae lava lake showing contours at the base and margin and the location and identification number of surveying stations and drill holes. Cross sections along the major axis (A-A') and the minor axis (B-B') of the lake have the same horizontal and vertical scale. The structure beneath the northern part of the lake is a buried spatter ridge formed in the 1840 eruption. Topography prepared by photogrammetric methods from aerial photographs taken February 15, 1963, by the U.S. Geological Survey. COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAI 5 FiaurE 2.-Drilling in Alae lava lake, November 1963. The mast-mounted, portable drill is forced downward by apply- ing pressure to a lever arm. Coolant water pumped through the drill stem turns to steam in the hot crust of the lake and rises along joint cracks in the drilling area. through hoses from the rim of the crater and stored in a 50-gallon (190 liter) steel drum near the drilling site. About 100 liters of water were used for each linear meter of drilling. The first two holes were drilled by holding the drill and pouring water into the hole by hand. Seven of the holes were drilled through the crust and into the underlying molten lava. In each hole, the drill fell under its own weight when the drill bit reached the base of the upper crust. When the drill stem was with- drawn, in most instances, superheated steam blasted up through the drill hole, sounding like a jet engine and continuing for as long as 12 minutes. In one hole, temperatures of 675°C were measured in the water vapor, and the collar of the drill hole was heated to glowing red. Four holes were drilled through the center of the lake and into the lower crust to near the basal contact, and core of the underlying 1840 lava was reco- vered from one of these (DH 9). Drilling rates averaged 1 m an hour when all equip- ment was working properly. Most of the drilling, how- ever, was slowed by mechanical failure of the water pump or drill. All equipment used on the lake corroded at an accelerated rate because of the high tempera- tures and rainfall and the abundance of sulfur in gases from the lake. TEMPERATURE MEASUREMENTS Temperatures in drill holes were measured re- peatedly during the 4-year study of the cooling lake. TaBLs 1.-Record of drill holes in Alae lava lake Drill hole No. Depths of interval Date drilled (meters) Comments 1963: Aug: 20 ._....._ 1 0-0.863 Boss Septs :_______ 1 0.814-1.14 1 1.10-1.40 To base of crust. 2 0-.655 Sept. 17 .-..... 1 1.28-1.70 alel Oct: I and Oct. 2: ' 3 0-2.84 To base of crust. Oct 24 ......__ 3 2.38-3.475 Do. Novel ...._:.. 3 3.02-4.08 Do. Nov. 20 .._... 4 0-4.66 Do. Nov. 27 ......_. 4 3.44-4.66 Do. Dec. 16 ...... 5 0-8.90 Through 0.7 m of melt; base of lake estimated to be at 9+1 m. 1964: Jan. 209 _...... 5 5.33-6.40 Plug of coze. Apr. 1 and Apr. 2 6 0-9.83 Through 2 m of melt into lower crust. Apr: 15.....:.. 6 5.41-13.05 Through plug of ooze and lower crust to near base of lake. Apri 22 ._..._.. 6 5.91-10.7 Through plug of ooze. Apr. 32. ...... 7 0-0.25 Mast hole. 7 0-6.1 Ara June4t..:__:s...: 7. 6.00-6.64 socked June 10 7 6.51-7.13 sess June d7 ._.... 8 0-6.325 egs June 18 _._.__. _. 8 6.325-8.73 Through 0.3 m of melt into lower crust. Jaly 28 9 0-8.57 July 20 ._:....; 9 8.57-8.87 west Sept. 11... 9 3.87-6.92 Through partly molten center of lake into lower crust. Oct 20: 9 10.52-16.12 Through base of lake at approxi- mately 14.5 m and into 1840 lava (cored between 14.6 and 15.2 m). Oct. 22 ._. ;.... 10 0-4.9 Tas. 11 02.7 Tese Nov:17 ...:... 11 2.17-4.9 ere 12 0-4.79 eros Nov.: :19 --...... 12 4.179-14.3 Through base of lake at approxi- mately 13.7 m (inferred from presence of caver- nous layer between 13.1 and 14.3 m). NotE.-Location of drill holes: 1. Near the major axis of the lava lake on the opposite side from the vents, 50.3 m S 44°E of the center of the lake (fig. 1); altitude of collar 2,567.2 feet (782.48 m) on Feb. 10, 1964. 2. 4.88 m N 75°E of drill hole 1; altitude of collar 2,567.3 feet (782.51 m) on Feb. 10, 1964. 3. 6.1 m S 15°E of drill hole 1; altitude of collar 2,567.3 feet (782.51 m) on Feb. 10, 1964. 4. 0.3 m N 75°E of drill hole 3; altitude of collar 2,567.3 feet (782.51 m) on Feb. 10, 1964. 5. 93.0 m S 49°E of drill hole 1; altitude of collar 2,565.8 feet (782.06 m) on Feb. 10, 1964. 6. 73.2 m S 17°E of drill hole 1; altitude of collar 2,567.4 feet (782.54 m).on Sept. 24, 1964. 7. 11.0 m S 85°W of drill hole 1; altitude of collar 2,566.9 feet (782.39 m) on May 12, 1964. 8. 0.3 m SW of drill hole 7: altitude of collar 2,566.9 feet (782.39 m) on May 12, 1964. 9. 79.2 m N 42°W of drill hole 1; altitude of collar 2,567.5 feet (782.57 m) on July 31, 1964. 10. 94.5 m N 16°E of drill hole 1; altitude of collar 2,568.0 feet (782.73 m) on Nov. 3, 1964. 11. 100.6 m S 87°W of drill hole 1; altitude of collar 2,566.9 feet (782.39 m) on Nov. 3, 1964. 12. 0.7 m due east of drill hole 2; altitude of collar 2,567.5 feet (782.57 m) on Jan. 21, 1965. 6 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII Most of the measurements were made with stainless- steel-sheathed thermocouples of heavy-gage chromel- alumel, using a portable millivolt potentiometer and a 0°C reference junction in an ice-filled vacuum bottle. Temperatures of the lake surface were measured by means of a mercury thermometer. During the first few months, repeated attempts were made to measure temperatures in the holes with platinum-platinum plus 10 percent rhodium thermocouples made with fine-gage wire. Most of these attempts were unsuccess- ful because the wires broke near the junction im- mediately after inserting the thermocouple, pre- sumedly as a result of the stress caused by thermal expansion of the steel sheathing. The thermocouples were constructed in the Ha- waiian Volcano Observatory by stringing 1-inch (2.54 cm)-long ceramic beads onto thermocouple wire, and inserting the wire and beads in %-inch (0.95 em) O.D. stainless-steel tubing welded shut at one end. Hastalloy-C tubing proved to be most resistant to cor- rosion. Most of the thermocouples had five junctions, spaced 1 or 3 feet (0.305 or 0.914 m) apart. Many of the thermocouples were more than 15 m long and required considerable care in handling. The potentiometer used in the measurements has a reproducible reading accu- racy of 0.01 my. Temperature readings of several but not all of the thermocouples were compared with those of a National Bureau of Standards platinum-platinum plus 10 percent rhodium thermocouple at the Obser- vatory. Most of the temperature measurements proba- bly have an accuracy of +0.5 percent at 500°C and +1 percent at 1,000°C, but uncertainties in the depth of measurement decrease the accuracy of the recorded temperatures. The depth of the thermocouple junctions in the drill holes was determined by marking each thermocouple sheathing, using a steel tape, and were recorded to an accuracy of +0.05 feet (1.5) em). Thermal expansion of the sheathing of thermocouples made ac- curate depth measurements in the deeper holes dif- ficult; thermocouples inserted rapidly to the bottom of the 15 m holes expanded as much as 6 cm within a few minutes. Each thermocouple could be used only a few times because of corrosion of the sheathing which allowed sulfur- and water-rich gases given off by the cooling basalt to come in contact with the thermo-couple wires. These gases reacted rapidly with the chromel wire, de- stroying continuity in the thermocouple. During times of heavy rains, water collected within the sheathing of the thermocouples left on the lava lake. When these were subsequently used in the drill holes, steam caused shorting between the wires. Consistent readings could not be obtained until the entire length of the thermo- couple was baked dry in a deep drill hole in the hot lake. In several drill holes, temperatures were measured both in the crust and in the underlying melt. Mea- surements in the melt were achieved by attaching a ceramic probe to the drill string and pushing it into the melt through the bottom of a drill hole after drilling through the base of the crust. Temperatures were then measured with a thermocouple inserted into the drill string and probe. Probes as long as 3.0 m were con- structed from 0.75 m lengths of 2.5-em O.D. mullite tubes (the lower one closed at one end), were joined with 10-cm sleeves cemented with silica cement, and were attached to the drill steel with a machined fitting. The whole unit was placed in the drill hole after drill- ing was completed, and the temperature in the bottom of the drill hole was measured by means of a ther- mocouple inserted in the probe. After the temperature had reached 1,065°C-1,070°C, the unit could be pushed slowly into the melt. Sometimes, but not always, it could also be pulled out. Use of excessive pressure caused cracking of the probes, allowing melt to enter. An unbroken probe was not emplaced into the melt successfully until December 2, 1963, after six unsuc- cessful attempts over the preceding 3 months. Later work in the 1965 lava lake in Makaopuhi crater has shown that stainless-steel probes can be used success- fully for long periods in the melt, and they can be emplaced with far less hazard of breakage. MEASUREMENTS OF THE ALTITUDE OF THE LAKE SURFACE On August 30, 1963, 17 stations were laid out at 100-foot (30.48 m) intervals along the major and minor axes of the lake by means of a transit and steel tape (fig. 1), and 2 additional stations (12 and 19) were marked by chiseling squares on the highest parts of 2 talus blocks at the ends of the minor axis. This grid was augmented in mid-September with 12 stations along two cross lines parallel to and 200 feet (60.96 m) from the minor axis. A few additional stations were added during the following year. The stations were marked by driving 5 cm concrete nails into the crust through stamped copper and aluminum tags. Distances be- tween stations were measured on August 30 to +1.5 cm; during the following year they were remeasured several times with more care to +0.3 ecm. Relative al- titudes of the stations with respect to station 19 on the talus were determined with a fiberglass stadia rod and a surveyor's transit (using the transit as a level by means of the striding bubble); later, a self-leveling surveyor's level was used. Relative altitudes of the sta- tions were measured 21 times during the 4 years of study-at 2-week intervals during the first month, 1-month intervals during the following year, approxi- mately 3-month intervals during the second year, and finally, on days near the end of the third and fourth COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 7 years. With few exceptions, all level lines were closed and the errors distributed. The measurements of altitudes on August 30, 1963, were taken with a damaged transit and may be in error by as much as 1.5 cm. Later altitudes relative to the stations on the talus are accurate to +0.2 cm, and the use of the surveyor's level after September, 1964, re- duced the error to about +0.1 em. The relative altitudes of stations on the surface of the lake were measured relative to the rim of the crater to +0.3 m by resection and vertical angle with a transit from two stations on the rim joined by a taped base line. PRECIPITATION AND AIR TEMPERATURES Alae crater is within the rain forest that blankets the northeast slope of Kilauea Volcano. Some rain falls during all but a few weeks each year, but most of the precipitation comes during "Kona" (southerly) storms in the fall and winter. Precipitation was monitored by means of a continuously recording rain gage at the south rim of the crater from September 12, 1963, to March 18, 1965, and also by means of a standard gage at the same location from November 16, 1964, through September 1967. Rainfall during the 4 years covered by this report varied from 200 cm to 335 cm per year and totaled 1,012 ecm. The recorded precipitation for weekly inter- vals is given in table 2. Air temperatures were recorded at Alae crater only during September, 1963. From September 12 to Oc- tober 1, 1963, the temperature at the south rim ranged fro 17.1° to 25.5°C and averaged 19.8°C. An average air temperature during the years that the Alae lava lake cooled can be estimated from record- ings at Hawaii Volcanoes National Park Headquar- ters. From October 1963 through September 1966, temperatures there ranged from 5°C to 26°C and aver- aged 15.6°C. As the mean temperature on the island of Hawaii falls at the rate of about 1°C per 150 m rise in altitude (Stearns and Macdonald, 1946, p. 209), the mean air temperature during that period at the rim of Alae, which is 300 m lower than Park Headquarters, was probably about 18°C. TEMPERATURES IN THE LAVA LAKE Temperatures in Alae lava lake were monitored by means of thermocouples in drill holes on 47 separate days over a 4 year period beginning 7 days after the eruption. Measured temperatures ranged from 16°C at the surface of the lake margin during the early morn- ing of December 18, 1964, to 1,135°C in molten lava 2.3 m below the crust of the lake on November 8, 1963. The lake solidified (cooled to less than 1,065°C) by mid- July 1964, 10% months after the eruption, and the last interstitial melt solidified by late September 1964, 2% months later. The entire lake cooled to less than 100°C in less than 4 years, as shown by the last measured temperature profile on August 31, 1967. Five repre- sentative temperature profiles, including the first and last ones, measured, are shown in figure 3, and selected temperature data are given in table 3. These are mostly from temperature profiles least affected by coolant water used during drilling and not affected by failing thermocouples, but include some data showing the chilling effect of coolant water and later thermal recovery of the crust. DEFINITION OF CRUST AND FLUID LAVA While it solidified, Alae lava lake, like other recent Kilauean lava lakes, consisted of a fluid interior and upper and lower crusts. Above the base of the upper crust (and below the top of the lower) was a broad zone of partly molten crust in which temperatures and abundance of the melt fraction increased smoothly downward from completely solidified crust to fluid lava containing subordinate crystals (Richter and Moore, 1966; Wright and others, 1968). The zone extended from a temperature of 980°C downward to the base of the crust at 1,065°C, at which temperature the yield strength of the lava abruptly decreased. At progres- sively greater depths below the upper crust, the abun- dance of crystals in the fluid lava steadily decreased to the depth of the maximum temperature in the lake. When the base of the crust was penetrated in drill- ing, the lava no longer offered strong resistance to the drill, and the drill fell under its own weight. In the six complete penetrations of the base of the crust in Alae, the horizon at which failure took place was between 0.1 and 1 m below the depth of the 1,065°C isotherm and at a temperature (before drilling) of between 1,070°C and 1,100°C. During drilling, the environment at the base of the crust was complex. As the drill penetrated the zone of crystallization, coolant water pumped through the drill stem quenched the melt fraction of the enclos- ing lava to glass until a level was reached where lava temperatures were so great that little or no selvage of glass was formed. The exact temperature of this hori- zon as reconstructed from temperature profiles mea- sured before and after drilling varied from one drilling operation to another, depending on the rate of drilling and amount of coolant water used. After the base of the crust was penetrated by drilling in this way, a hollow probe could be pushed through the drill hole and into the underlying fluid lava. After each penetration we had to wait some time until the quenched lava at the base of the hole had become suffi- ciently warmed, a period ranging from a few hours 8 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII Taste 2.-Rainfall at the south rim of Alae crater TABLE 2.-Rainfall at the south rim of Alae crater-Continued Date Between dates August. 22,1063 29 haas + onn, to pa 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 a November 1k :}. 2 13 souls sal bo pa 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 i - Co i. .' _ 'na." January 2 10964 ._: __ bo i- mM I 1 I 1 I 1 I 1 I 1 I I I 1 I 1 1 1 I 1 1 I 1 1 I 1 I I 1 I 1 I I I 1 1 1 I I I 1 I b o- bo co I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 pa June Ain nul dane nes ©. on. Coogmpnmiuvmarnclsimma- Gs O1 Or ho ho ho ho ho ho BQ b b bet - g geted (4 $o mim is m fo _i p so =a on on on its its. bo So \p so ~1 ~a -A is bo F- oren ant t do Or ho ~1 ho to to O 1 -1 -1 t i- Oo Oo On ho <1 A 1A Go +- C> SA lant i- c © 5 to w co or- mam oue coum i co . co on co w fob bo I cD > bo bo m - co Total 276.1 283.2 283.2 285.2 286.0 286.3 286.6 290.7 290.7 297.1 300.9 320.0 334.2 366.0 367.8 378.7 421.9 426.0 429.6 432.9 435.9 437.4 438.7 440.2 441.0 442.5 445.3 447.6 449.1 450.1 451.1 452.9 453.2 453.7 463.6 464.9 465.4 466.2 469.2 474.5 483.9 484.2 493.6 540.1 544.9 548.5 550.0 565.7 570.0 571.3 572.3 576.1 579.7 589.1 598.5 602.1 602.6 603.4 603.7 612.1 613.6 614.1 618.4 620.7 622.7 627.0 627.8 637.2 638.5 641.3 643.1 644.4 645.9 650.0 COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII o 2.-Rainfall at the south rim of Alae crater-Continued Rainfall (cm) Between dates Total 655.1 655.9 661.5 669.1 670.6 671.6 672.1 676.7 680.0 682.0 685.0 689.6 699.3 704.6 705.6 720.6 724.4 725.2 760.8 762.8 768.6 786.6 788.1 793.2 795.7 800.0 822.1 824.6 829.2 830.5 831.5 832.5 846.5 849.5 850.0 854.6 861.0 874.7 884.9 893.0 900.4 908.0 915.6 $19.9 923.7 944.0 945.0 945.0 947.0 948.8 950.6 952.1 954.9 960.0 964.0 970.4 976.0 978.5 986.6 1,012.3 1,012.8 1,013.1 fare mith, .C" September 1 October - ~ 1 1 1 1 i 1 I 1 1 1 1 I I 1 1 I 1 1 1 I 1 i to November bo % 1 1 1 1 1 1 1 1 1 1 1 1 i 1 1 1 1 1 1 1 1 a December go (an w J Me 1 I 1 i 1 1 1 I I I I 1 1 I 1 1 bo January February or i i 1 1 1 1 1 1 1 1 1 1 i 1 1 1 1 1 1 i 1 i a to tn \1i- tn an in oo i- oo &n oo do & & & to bo to i in im i- ho in n in 0 0 D J to in on i= to in i= in o oo 0 n bo do 0 o io Su n 0 0 to n in 0 in s os bo i- March abed April May - £" 1 I 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 bo June July August mmpmmAmNHHQ—amwr—Aopopfilflgpo La 'g 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 bo 'Estimated rainfall. when the crust was thin and little coolant water was used to several days when much coolant water was used in repeated drilling at one site through the thicker crust. Each probe, which consisted of a pointed mullite tube 1-inch (2.54 cm) in diameter with walls 0.25 cm thick, was pushed slowly under a load of about 0 o T DEPTH, IN METERS 3 T ts anny APPROXIMATE BASE OF LAKE 15 ¥ ¥ F E7 Ira 0 200 400 600 800 1000 TEMPERATURE, IN DEGREES CENTIGRADE 1200 FiGurE 3-Representative temperature profiles in Alae lava lake. 200 pounds. The temperature and depth of the tip of the probe during its insertion on October 2, 1963 is shown as a function of time in figure 4. On this date, as in each of the six attempts to emplace a probe in the melt, rapid penetration took place when the tempera- ture at the tip of the probe was between 1,065°C and 1,070°C. The best value for the temperature of inser- tion was 1,067%>2°C, as reported earlier (Peck and others, 1964), but this value was later rounded to 1,065°C (Peck and others, 1966). The value, which rep- resents the softening temperature of basaltic glass and crystals under the conditions of the experiment, is somewhat arbitrarily assigned as the temperature of the base of the crust. It is identical with the value determined in Kilauea Iki lava lake (Ault and others, 1962; Richter and Moore, 1966) and within the limits of error of the value of 1,070°C determined in Makaopuhi lava lake (Wright and others, 1968). The rate at which the upper crust of Alae lava lake grew in thickness decreased rapidly from the begin- ning of solidification at the end of the eruption until a short time before complete solidification of the lake. The thickness of crust increased at the rate of about 0.14 m per day at the end of 1 day, 0.07 m per day at 10 days, 0.02 m per day at 100 days, and 0.015 m per day at 250 days. After 270 days the rate increased again at the main drilling site, reaching 0.06 m per day shortly before solidification was complete there at 300 days. The thickness of crust in the upper part of the lake at the main drilling site (DH1, 3, 4, 7, and 8), as deter- mined from temperature profiles in the drill holes, is plotted against time in days after the eruption in figure 5. The depths at which fluid lava was reached in drill- ing are also shown. As discussed above, these depths 10 2.88 T T T T T T T T T T DEPTH, IN METERS to D n 0 to to [=] & N T T T 1 1 1 to O B T 1 to O Co D Cs & o 0 T 1 1060 l=] ~ o T 1 1080 T 1 TEMPERATURE IN DEGREES CENTIGRADE -s o el l=) T 1 1100 1 1 1 1 1 1 1 1 1 1 1 0. 10 20. 30. 40 50 60 70: 80 90 100 110120 TIME, IN MINUTES, AFTER 10" 40" OCTOBER 2, 1963 Ficur®E 4.-Temperature and depth of the tip of a hollow mullite probe as a function of time during its insertion into fluid lava through the base of drill hole 3 on October 2, 1963, after drilling was completed at 10h40m. were consistently greater than the depth of the 1,065°C isotherm. As shown in figure 5, the thickness of the crust as a function of time plots as a curved line having progressively decreasing slope during the first few months after the eruption. With respect to the square root of time, however, the thickness of the crust and the depth of isotherms in the crust and melt plot as straight line segments during the first 3 months, as discussed on page 13. Petrographic study of drill core from the lake shows that lava at a temperature of 1,065°C at the base of the crust contained 50 percent liquid (quenched to glass in the core) in a meshwork of crystals. At the inferred solidus temperature of 980°C and at lesser tempera- tures, the quenched lava contains a nearly constant 7 percent glass. At greater temperatures, the abundance of glass quenched from liquid increases smoothly to 1,070°C, the highest temperature core recovered from the lake. At 1,140°C, the projected maximum tempera- ture in the lake, the lava contained about 13 percent crystals, as shown by the crystal content of pumice air quenched at this temperature during the final phases of the eruption, August 22 and 23, 1963. Temperatures across the zone of partly molten crust and the underly- ing fluid lava down to a temperature of 1,100°C on December 30, 1963, are shown in figure 6, together with a curve showing the relative abundance of the melt fraction. This curve was derived by petrographic SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII 0 I I T I I THICKNESS OF UPPER CRUST, IN METERS 1 1 1 80 100 1 200 TIME, IN DAYS, AFTER AUGUST 22, 1963 Ficur® 5.-Thickness of the upper crust of Alae lava lake near the main drilling site (DH 1, 3, 4, 7, and 8) as a function of time in days after August 22, 1963 (dots and curve). The depths at which fluid lava was reached in drilling are shown by vertical and horizontal bars. study of core from several drill holes in Alae and in Makaopuhi lava lake (Wright and Peck, 1978; Wright and Okamura, 1977.) The temperature profile was measured in a ceramic tube emplaced in drill hole 4 through the crust and into underlying melt to a total depth of 5.5 m on December 2, 1963, 5 days after the hole was drilled through the crust. The profile shows the typical pronounced decrease in temperature gra- dient downward across the crust-fluid lava interface; this downward gradient was evident in each of the pro- files measured across the zone on 11 separate days. Computer analysis of the thermal data (p. 43; Peck and others, 1976, fig. 6) indicates that the decrease can be attributed to the release of latent heat of crystalliza- tion within the zone rather than by a pronounced downward increase in thermal conductivity. The de- parture of the temperature profile from a smooth curve at a depth of about 4.7 m takes place at the horizon where the base of the crust was penetrated by drilling more than 1 month earlier; very likely this departure is an artifact of chilling by coolant water used in the drilling, as discussed in the following section. The rate of crystallization with respect to change in temperature, measured by the decrease in the abun- dance of the melt fraction in samples quenched from different temperatures, was greater between 1,080°C and 1,050°C than at greater and lesser temperatures. The maximum crystallization rate was between 1,065°C and 1,060°C, where the abundance of melt de- creased nearly 8 percent. This relationship is shown by COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII +H TEMPERATURE, IN DEGREES CENTIGRADE 900 1000 1100 3 T T T T T T Crystallized basalt ® i u 41 Eu Partly molten crust bs < 3 - a. L3 u o \% |- Ya 5 \ Fluid lava \\ \ \ I 1 1 | t 0 20 40 60 80 WEIGHT PERCENT FiGur® 6.-Temperature profile across the zone of crystallization at the base of the crust on December 30, 1963. Closed circles indicate measured temperatures. The inflection of the profile at 4.7 m is an artifact of coolant water used in drilling, as discussed in the text. Curve showing the relative abundance of the melt fraction was derived from study of core from several drill holes in Alae and Makaopuhi lava lakes (Wright and Peck, 1978). study not only of drill core from Alae (Wright and Peck, 1978), but also of samples from a greater variety of temperatures from the 1965 Makaopuhi lava lake (Wright and others, 1968, fig. 12); the relationship dif- fers from preliminary estimates of Peck, Wright, and Moore (1966, p. 11) which suggested a nearly linear relationship between the abundance of melt and tem- perature. During further cooling, from 1,065°C to 980°C within the zone of partly molten crust, the melt fraction continued to decrease in abundance and its composition became more silicic, as shown by the re- fractive index of glass quenched from melt; the analyzed composition of quenched melt that oozed into drill holes; and the analyzed composition of residual glass separated from drill core from the crust (Peck and others, 1966, fig. 5 and table 1; Wright and Peck, 1978). The silica content increases from 50.3 percent at 1,065°C to 75.8 percent at 980°C. Because of the marked decrease in the proportion of melt upward across the zone of partly molten crust, and to a lesser extent because of the decrease in melt tem- perature and increase in silica content, the viscosity of the lava increases sharply upward across the zone. Judging from the analysis of Shaw (1969, p. 519-520 and fig. 4), which was based on field measurements of viscosity in Makaopuhi lava lake (Shaw and others, 1968) as well as on laboratory measurements of viscos- ity of basaltic melts and silicone fluid-glass sphere sus- pensions, the limiting viscosity in Alae lava lake in- creased upward across the zone of crystallization from about 10" at 1,065°C to 10'® at 980°C. EFFECT OF COOLANT WATER ON TEMPERATURES As described in the section on "Drilling," the pene- tration of fluid lava below the crust was followed, after removal of the drill stem, by a jet of superheated steam. After each steam blast, temperature profiles in the bot- tom of the hole showed a reversal; temperatures in- creased with depth to a maximum near the position of the 1,065° isotherm before drilling and then decreased with greater depth for %-1 m to a minimum near the depth of penetration into the melt during drilling. Shortly after the penetration and steam blast of November 27, 1963, for example, the temperature was | 975°C at a depth of 4.36 m but was only 700 °C at 4.66m. Temperatures in drill hole 3, after drilling was completed at 12120m on November 7, 1963, are shown in figure 7. The temperatures, which were measured at five different times during the 2-hour period of emplacement of a hollow probe into the melt, demon- strate the lower rate of thermal recovery with time near the base of the drill hole as compared with rates at more shallow depths. As shown in the figure, a minimum temperature was found at a depth of about 4.1 m, just below the depth of penetration into fluid lava. The steam blasts and temperature reversals took place only after drilling into fluid lava. Apparently, both were caused by vaporized coolant water. When drilling within the crust, the steam escapes up the drill hole or laterally through joint cracks in the crust. The fluid lava, in contrast, traps the steam against the drill stem until the latter is withdrawn. The temperature data from the lake discussed in the following sections were selected to avoid, as much as possible, the marked depression of isotherms caused by coolant water used in drilling. The effects of such water were marked after October 1, when we began pumping coolant water through the drill pipe with a portable engine during drilling instead of pouring it into the hole by hand. Usually about 100 liters of water were used for each linear meter of drilling, but at times as much as 600 liters/meter were used. After repeated drilling at the main site in the southeast part of the lake during October and November 1963, tempera- tures were probably decreased in a considerable vol- ume of crust beneath the site. After drilling, tempera- 12 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII 2.8 I T T 3.0 |- 3.2 |- 3 4}- 3.6 |- DEPTH, IN METERS 4.0 |- 4.2 - 1 | | 1 1 1 600 700 800 900 1000 _ 1100 TEMPERATURE, IN DEGREES CENTIGRADE 4.4 500 1200 FiGur® 7.-Temperatures near the bottom of drill hole 3 after drill- ing through the base of the crust at 12h20m November 7, 1963. Depth at which fluid lava was encountered during drilling is marked by hachured line at 4.068 m. tures in the lower part of the crust recovered to near predrilling levels within a few days, but those in the upper part required many weeks; temperatures at a depth of 30 em were still rising 32 days after the dril- ling on November 27. Temperatures at a depth of 0.915 m in drill hole 3 were monitored continuously after the hole was deepened October 24, 1963, from 2.38-3.48 m using about 150 liters of coolant water. The continuous monitoring was part of a cooperative program with a visiting team of scientists and technicians from Japan, described on page 15. The temperature at this depth, which was 570°C before drilling, climbed smoothly after drilling from 340°C at 23h October 24 to 517°C at 03h November 4. The recovery of temperatures at dif- ferent depths after drilling is shown in figure 8, on which are plotted the depths of isotherms in the crust after drilling on November 27, 1963. Because only about 200 liters of water were used in the drilling on this date, the isotherms were not as strongly depressed Drilling completed 14" Nov. 27 m| co SI&R 'n" ip £ 5 2) 3 § g s g ELE o O o o 0 I1 1 II 1 ; II [ ; | DEPTH, IN METERS 5 I 1 1 1 1 1 0 10 20 30 TIME, IN DAYS, AFTER 0" NOVEMBER 27, 1963 FicuUrE 8.-Depth of isotherms in drill hole 4 before and after drill- ing on November 27, 1963. Circles represent measured tempera- tures. Triangles represent extrapolated temperatures. as they were by other drilling operations. Increases in temperature with time were more closely monitored, however, because of repeated attempts to emplace a probe through the drill hole into molten lava and re- peated measurements of temperatures in the probe once it was successfully emplaced. The figure shows the different relative depression and recovery rates of isotherms at different depths in the crust. TEMPERATURES IN FLUID LAVA BELOW THE CRUST Temperature profiles extending to 1,100°C or more-that is, to depths well below the base of the crust-were measured on October 24, November 8, and December 2, 3, 5, 18, and 30, 1963 (table 3). The longest profile in fluid lava was measured in a hollow ceramic probe on November 8, 1963 (fig. 9), 1 day after drilling into fluid lava at a depth of 4.08 m. The maximum temperature measured was 1,135°C at a depth of 5.46 m, an estimated 2 m above the depth of the maximum temperature in the lake. The very low temperature COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 13 2.5 T T T T T Crystallized basalt 3.0 |- i Partly molten crust 3 5- = DEPTH, IN METERS p 0 T 1 Fluid lava 4.5 - = 5.0 |- L 5.5 1 1 1 1 1 900 1000 1100 TEMPERATURE, IN DEGREES CENTIGRADE Figure 9.-Temperature profile across partly molten crust and into underlying fluid lava on November 8, 1963. Open circles indicate measured temperatures. The zone of partly molten crust (980°C to 1,065°C) is shaded and the base of the crust (1,065°C) hachured. gradient in the melt shown by the profile, only 8.5°/m between 1,130° and 1,135°C, contrasts with the gra- dient of 240°/m measured in the crust between 710°C and 985°C before drilling on November 5. As in the profile measured December 30, 1963 (fig. 6), the November 8 profile shows some residual effect from coolant water used in drilling; the profile departs from a smooth curve at the approximate depth (4.1 m) where the drill broke into fluid lava on November 7. THERMAL HISTORY OF THE LAKE Temperatures in Alae lava lake during the 4 years when it solidified and cooled to less than 100°C are summarized in figure 10, in which the depths of isotherms in the lake, including the 1,065°C isotherm at the base of the crust, are plotted as a function of the square root of time after 06n August 22, 1963. The temperature profiles used in preparing the plot were selected to reduce to a minimum the effects of coolant water and are given and marked in table 3. The solid lines (and short-dashed extrapolations) represent the depths of isotherms near the center of the lake in drill holes 1, 3, 4, 7, 8, and 9. The long-dashed lines repre- sent the depths of isotherms as measured in drill hole 12, the only deep drill hole that remained open and accessible to measurements during the final stages of cooling of the solidified lake. Temperatures at shallow depths in drill hole 12 are not representative of the lake as a whole, apparently because of heating of the lava enclosing the drill hole by gases rising through a fracture intersected by the hole. Also shown in the figure are the depths of the base of the lake at the main drilling site and at drill holes 9 and 12, and the cumulative rainfall at the south rim of Alae crater. Contact temperatures at the lake surface were mea- sured by means of mercury thermometers near the col- lars of drill holes on many days when temperatures were measured in the drill holes. The measured day- time surface temperature during the first 2 years of cooling of the lake over the central part of the lake ranged from 80°C to 30°C and averaged about 70°C during the first 5 months and about 40°C thereafter. Surface contact and radiation temperatures were mea- sured systematically at 30-cm intervals along a 31-m base line near the center of the lake and at two "cold- base" stations at the thin margins of the lake in November and December 1964 when the maximum temperature in the lake was about 900°C. Contact temperatures were measured by means of ther- mocouple, and infrared radiation temperatures (>3 micrometers wavelength) by means of a radiometer that views a conic section of 28° and was held 30 cm above the ground. The results of the measurements and the calculated heat loss by radiation have been reported by Decker and Peck (1967). Radiation tem- peratures ranged from 15°C over the cooled margins of the lake at night to 85°C over fuming joint cracks near the center of the lake. The measurements show marked highs over fuming cracks and less marked highs over older cracks (Decker and Peck, 1967, fig. 3). During the day, radiation from most of the surface of the still hot center of the lava lake was not measurably different from the radiation from the sunlit basalt at the thin cooled margin of the lake. At night, however, radiation heat loss from internal sources over the central part of the lake was appreciable, constituting about 20 percent of the heat conducted to the surface. > The thickness of crust and the depth of isotherms in the crust and melt plot as nearly straight-line seg- 14 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII CcUumuLATIVE ments against the square root of time during the first 3 months. This relationship follows from heat- conduction theory for the conductive cooling of a semi- infinite slab (Carslaw and Jaeger, 1959, p. 285), as dis- cussed by Peck, Moore, and Kojima (1964, p. D4). The linear relationship is particularly well shown by tem- perature data collected during early and mid- September, 1963 when rainfall was relatively sparse and little coolant water was used in the drilling. The w G T 5 T ! NOV ! pec J TOF T mT ATMTJ T] TAT $TOINTD r Tore &© P ©1963" § 1964 1965 1966 1967 E I I I I I T I - int 2: I . "4 8°C is =d - 3 § at $ 2 L : a w ) | #4 :} a Nit avy. £ Mes only it; f. ) \ \ / \ \ \ C Na ! 7. A {1 3 / ) \ / f y XY 3 £ BASE OF LAKE (DH 12) __\ g g 14 |- APPROXIMATE BASE OF LAKE (DH 1, 3, 4, 7, and 8) 74 A Y BASE OF LAKE (DH 9) 4 77777777771777777171777 ( 16 1 | 1 1 L | | 0 5 10 15 20 25 30 35 40 SQUARE ROOT OF TIME, IN DAYS, AFTER AUGUST 22, 1963 FicurE 10.-Depth of isotherms (in degrees centigrade) in Alae lava lake from August 1963 to August 1967 as a function of the square root of time after 06h August 22, 1963, and recorded rainfall at the south rim of Alae crater. Tempera- ture measurements used in constructing the plot are marked in table 3. Rainfall is given in table 1, and the location of the drill holes plotted in figure 1. Temperature profiles (0-11) used in thermal modeling (see section on "Thermal Mod- eling") are marked above the zero depth line. depths of selected isotherms are shown in figure 11 for the first 25 days after the eruption as a function of the square root of time after 06h August 22, 1963. These depths lie on straight-line segments from September 3 to September 17. The excessive depths of the isotherms on August 30 were caused by coolant water used the previous day during the first drilling in the crust of the lake. Because of heavy rains on September 17, the depths of isotherms as measured 2 days later were de- COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 15 os" __ os" Aug. 22 Aug. 23 Aug.?0 Seet.3 Sept. 1|7'1963 | ol--o-ol-0-0o-_-_]Iy sss ccc --. 100° C a A x -= o b}\ s x: * Aja 300°C 0.5 |- =, w CC u wa w > z 1.0}- ~ T p a. u o TS5r- = 2.0 | 1 1 1 I 0 1 2 3 4 5 6 SQUARE ROOT OF TIME, IN DaYs, AFTER 06° august 22, 1963 FigurE 11.-Depth of isotherms (in degrees Centigrade) in Alae lava lake from August 22 to September 17, 1963, as a function of the square root of time in days after 06h August 22. Dashed lines are extrapolations of isotherms to zero depth on the basis of estimated thicknesses of crust during the August 1963 eruption on Alae cra- ter and measured surface temperatures during the March 1965 eruption in Makaopuhi crater. Short-dashed lines are linear ex- trapolations of straight-line segments defined by the depth of mea- sured temperatures from September 3 to 17. Vertical bars a, b, and c are thickness of crust estimated during the eruption, as discussed in the text. pressed as much as 0.15 m. During the following month and a half, the plotted depths continued to lie below the straight-line segments because of occasional heavy rains and the use of more abundant drilling water, but during periods of no rain or drilling, the depths tended to approach that of the straight-line segments. The choice of zero time for the plot of figure 11 is arbitrary, because the exact time of formation of the crust varied from place to place on the lake. Some crust near the vent was not formed until drain back at the end of the eruption at 08h10m August 23. Crust over linear ooze ups farther from the vent formed mostly during die-back of activity during the evening of Au- gust 22. Most of the crust over the central part of the lake probably formed by 061, August 22, when the rate of eruption dropped to nearly zero (Peck and Kinoshita, 1976). Estimates of the thickness of crust during the first several hours after its formation are also shown in figure 11. Thickness a of 8-10 cm is based on the aver- age thickness of pahoehoe slabs in the lava levee at the edge of the lake; the slabs are from crust that was rafted across the lake in about 1% hours during the night of August 21-22. Thickness b of 15-25 em was estimated during a trip to the edge of the lava lake at 18h30m on August 22; at that time the crust was barely thick enough to walk upon with caution. Thickness c of 25-30 cem is based on the thickness of slabby crust in tumuli on the lake that were formed during drain back of molten lava into the vent after 061, August 23. The depths of isotherms as measured in the drill holes are extrapolated backward in time as dashed lines to their inferred intersection with the zero depth axis on the basis of repeated optical pyrometer and thermometer measurements of surface temperatures during the March, 1965, eruption in Makaopuhi crater. Surface temperatures there dropped to 0.9 of the initial value after 10-20 seconds, to 400°C after 30 minutes, to 200°C after 4 hours, and to 100°C after 1 day. Linear projections backward in time of the segments established by the depths of isotherms in drill holes September 3-17 (dotted lines, fig. 11) do not intersect each other and the zero-depth axis unless an origin time of 121 August 24, well after the end of the erup- tion, is selected. The intersection in figure 9 is at V days equals 0.55, that is, 13115m, August 22. The exact time corresponding to the intersection has no real meaning; it reflects both the thermal constants and environment of the newly formed crust and also the somewhat arbi- trarily chosen origin time. Intersection to the right of the origin, however, is significant. It records slower cooling of the crust and upper part of the melt with respect to the square root during the first few days. For the most part this probably was the result of the higher surface temperatures, but it may also have reflected a lower thermal diffusivity of the thin, highly vesicular crust as well as heating of the crust by gases exsolved from the underlying molten lava. Temperatures near the base of the crust were moni- tored continuously at fixed positions in drill holes in the crust of the lake for several periods from mid- September to mid-December 1963. Temperatures were recorded as part of an investigation of possible varia- tions in temperature arising from convection of molten lava in the lake and were carried out as a joint effort with a visiting team of scientists and technicians from Japan headed by Prof. T. Minakami. Millivolt differ- ences were carried over wires to a millivolt chart re- corder at the southeast rim of the crater from ther- mocouples fixed near the bottom of drill holes and from a thermistor in the vacuum bottle cold junction. Re- 16 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII corded temperatures in the drill holes were checked periodically when temperature profiles were measured in the drill holes. From October 29 to November 5, 1963, for example, temperatures were recorded from a depth of 2.9 m in drill hole 3, which was deepened on October 24 from 2.38 m to 3.48 m, and the plotted temperatures were averaged over 1 hour periods. The temperature at this depth, which was initially within the zone of crystallization, fell smoothly during the monitoring period from 1,021°C to 958°C. No system- atic fluctuations were recorded that exceeded the preci- sion of the thermocouple. As shown in figure 10, the depths of isotherms deep in the crust and in the underlying fluid lava were nearly linear with respect to root time until early November 1963 and thereafter departed only slightly from linearity until the lake had solidified (cooled to less than 1,065°C) in early July 1964. Isotherms at shallow depths in the crust, however, departed mar- kedly from linearity starting in early January 1964 largely as the result of rainfall. During the first 5 months of 1964, 166 cm of rain were recorded at the south rim of the crater, in contrast to the 29 cm that fell during the preceding 4 months. Heating and va- porization of the rain water quenched a layer of crust immediately below the lake surface to steam tempera- ture (97°C at the altitude of the lake), depressing the 100°C isotherm from 0.09 m to a depth of 1.22 m over the 5 months of abundant rainfall. After the rains ceased, the 100°C isotherm slowly decreased in depth as heat flowed into the surface layer from the underly- ing crust, but it was depressed again by abundant rain- fall beginning in late February, 1965. Depression of the 200°C to 700°C isotherms by the early 1964 rains and their later partial recovery is also readily apparent in figure 10. As discussed in the section on "Thermal Modeling", measured temperatures can be closely matched by computed temperatures by withdrawing 620 calories (540 cal/g heat of vaporization and 80 cal/g for heating from about 20° to 100°C) for each centime- ter of measured rainfall. The effects of the heavy rains of January through May 1964 on the crust enclosing the upper part of drill hole 5 are shown in figure 12. The hole was first drilled to a depth of 8.9 m through the partly molten center of the lake near its margin on December 16, 1963, and the central 1.07 m redrilled on January 29, 1964, to clear it of ooze. Thereafter no drilling took place in or near the hole. Figure 12 shows the progressive formation of a chilled layer near the surface and the later temperature recovery. At depths in the lava lake greater than about 2.5 m the cooling effect of rain water is less obvious. Temperatures at a given depth varied significantly from one drill hole to another, even in the holes in the DEPTH, IN METERS \\| i TEMPERATURE, IN DEGREES CENTIGRADE 3 1 L 1 A_ 0 200 400 800 FicurE 12.-Temperature profiles in the upper part of drill hole 5 near the southeast margin of Alae lava lake, showing the effects of abundant rainfall in the first 5 months of 1964. central part of the lake. During the first month after the eruption, measured temperatures in the upper 0.6 m of crust in drill holes 1 and 2 differed by 10-30°C (data for Sept. 9 and 11, 1963, table 3). During the second month, measured temperatures in the upper 1.6 m in drill holes 1 and 3 differed by 10-25°C (Oct. 24 and Nov. 5, 1963, table 3). Measured differences in temper- ature in different drill holes were greater during the second year of cooling, as illustrated in figure 13, which shows temperature profiles in five drill holes on January 21, 1965, 4 months after complete solidifica- tion of the lake. Also shown is a computed temperature profile for the same day, based on a numerical model for a lake 49 feet (14.9 m) thick (fig. 39). The coldest hole, drill hole 5, was only about 30 m from the edge of the lake. Temperatures at a given depth within the upper 5 m of crust in three of the more centrally located holes (DH9, 10, and 11) varied within a range of 50°C- 100°C. Temperatures in the other drill hole (DH12) were consistently 25-100°C greater than those in any other drill hole at depths of 8 m and less. At greater depths in drill hole 12, temperatures were less than those in drill hole 9 because the lava lake was 1% m shallower beneath the drill site. The differences in temperature reflect lateral variation in porosity (and conductivity) at shallow depths in the lake and cooling COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII ¥7 Computed temperatures # "| % s |-II—J §/l w S Oy : 8 |- ’1 a y ® I rrziftrr & / & 10} 3 o 12- - 14 |- - 27777777 16 1 1 1 1 1 1 i 1 0 100 200 300 400 500 600 700 800 900 TEMPERATURE, IN DEGREES CENTIGRADE FiGur® 13.-Temperature profiles in five drill holes in Alae lava lake on January 21, 1965, and a computed profile for the same day based on a numerical model for a lake 49 feet (14.9 m) thick (fig. 39). Location of drill holes is shown in figure 1. Symbols indicate measured temperatures. Hachured lines beneath three of the pro- files indicate the approximate depth of the base of the lake at these sites. along joint cracks, although all holes were drilled as far as possible from known cracks. Lateral variations in vesicularity (porosity) in the lake were caused by differential entrapment of rising gases near the end of the eruption when the crust was 0.3-1.2 m thick (Peck and Minakami, 1968, p. 1154). Thermal conductivity decreases markedly with increased porosity as dis- cussed on p. 41. Near the end of the eruption, the crust of Alae lava lake was broken by a network of cracks outlining polygonal slabs of crust 3-4.5 m across (Peck and Kinoshita, 1976); some of these cracks were drawn open during drain back at the end of the eruption, forming gaping crevasses as much as 30 cm wide which diverged outward from the central part of the lake. Seven days after the eruption, the temperature at a depth of 1.5 m in an open crack 12 m southeast of drill hole 1 was 885°C. On this date, the base of the crust (at 1,065°C) below drill hole 1 was at a depth of only 1 m. Within a few weeks, the walls of these cracks cooled to less than red heat; by June 9, 1964, the temperature at a depth of 2 m in the crack southeast of drill hole 1 was only 35°C. Lava adjacent to the cracks was cooled by air circulating in the cracks and by rainwater flowing into them. As a result, isothermal surfaces in the crust were strongly depressed near the cracks. As the crust of the lake cooled, cracks split to become deeper, and new cracks opened. Some of these provided channel- ways for hot gases that streamed upward from the mol- ten and newly solidified basalt at depth in the lake and heated lava adjacent to the cracks. Drill hole 12 was split by such a crack; during drilling, water was turned to steam along the crack, making the drilling operation difficult. Rising gases along this crack caused tempera- tures at shallow depths in the drill hole to be greater than those in other holes. MAXIMUM TEMPERATURES AND TEMPERATURES AT THE BASAL CONTACT Maximum temperatures in the central part of Alae lava lake are estimated to have been 1,140°C on November 8, 1963, and 1,120°C on December 30, 1963, on the basis of 2 m extrapolations of temperature pro- files measured in the crust and underlying fluid lava on those dates (figs. 6, 9). These values compare with maximum extrusion temperatures which fell from an estimated 1,160°C at the beginning of the eruption on August 21, 1963, to 1,140°C near the end of the erup- tion during the evening of August 22. Maximum tem- peratures in the lake were first measured beneath its thin margin on December 18, 1963, when a value of 1,066°C was recorded at a depth of 5.5 m in drill hole 5 (table 4). Maximum temperatures beneath the central part were first measured on July 2, 1964, 10% months after the eruption, when a value of 1,050°C was re- corded at a depth of 8.2 m in drill hole 8. Beneath the main drilling site on the central part of the lake, maximum temperatures in the lake fell steadily during the following 3 years, decreasing to 672°C on August 19, 1965, to 476°C on May 27, 1966. On August 31, 1967, the maximum temperature measured was 90°C at the bottom of drill hole 12, 1.1 m greater than the assumed depth of the base of the lake at that site. The measured values are very similar to values computed using numerical models, as discussed on p. 47. Maximum temperatures in the lake beneath the main drilling site, (DH3, 4, 8, and 12), in the thin mar- gin (DHS5), and in the thickest part of the lake (DH9) are plotted in figure 14 as a function of the square root of time. The values beneath sites near the center of the lake plot as nearly linear curves at temperatures less than about 1,050°C, facilitating the estimates of maximum temperatures on intermediate dates. Also indicated in figure 14 are maximum temperatures computed for a lake 49 feet (14.9 m) thick using the numerical model discussed in the section on "Thermal Modeling" and illustrated in figure 39. The computed temperatures agree closely with the measured values 18 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII 4.-Depth and values of maximum temperatures and of basal contact temperatures in Alae lava lake Basal contact Depth of maximum Maximum Drill temperature in temperature temperature in °C' Date hole meters (most in °C (values in paren- values +0.15 m) theses are preferred) 1963 Nov. 8 3 estimated 7.5 - estimated 1,140 Dec. * 18; 5 5.5 1,066 30 3 estimated 7.5 - estimated 1,120 20-6 5.55 1,053 1964 Jan. ~290 : 6 5.6 960 Feb.: 17 _ 5 5.85 915 26 : -5 5.9 889 March 24 _ 5 6.0 812 April $ 6 TB 1,060 13 5 6.1 750 May 22 6 7.5 1,052 27 *:6 5.7 636 June 17 5 5.8 580 July 2..,.8 8.2 1,050 25 5.5 552 27.56 5.2 512 Aug. 14 5 4.7 494 Sept. 3 5 4.4 476 15. 9 9.3 1,005 21% -B 4.6 455 21'~ 9 9.1 995 Oct..: 16 5 4.0 440 16. 9 9.3 950 Dec. $ #5 4.0 892 8: - 9 9.3 902 725+50 (695) $ 12 8.5 872 690+30 (710) 1965 Jan:. 21 ~ 5 4.3 328 21% P 9.75 865 695+45 (670) 21 > 12 8.5 850 645+45 (675) March 2 . 5 4.3 285 2p 9.9. 833 685+50 (655) 2 8.7 817 630+40 (660) July 6" :s 4.9 95 6: : 9 11.0 725 640+30 (615) 6 12 9.175 720 585+30 (610) (19 5 4.9 95 19:9 11.6 688 620+30 (600) 19 12 9.75 672 570+80 (595) 1966 May. 27 12 11.9 476 460+10 (465) 1967 Aug. 31 12 13.1 88+22 88+2 (86) in DH9. The values for drill hole 5 plot more erratically than those for other drill holes, at least in part because of the effect of abundant rainfall on the relatively cool margin of the lake after the heavy rains in early 1964 and early 1965. The depth of the maximum temperature beneath the central part of the lake (table 4) fell systematically with time as the temperature decreased. Ranges of values for the proportionate depth (the ratio of depth of the maximum temperature to the thickness of the lake) in several drill holes are plotted as a function of the maximum temperature in figure 15. A large range of values for the proportionate depth for each hole is indi- cated in the figure because of uncertainties in the depth of the basal contact. This uncertainty for drill holes 9 and 12 can be greatly reduced if we can assume that the proportionate depth of the maximum tempera- ture was the same in each hole on any given day. On this basis, the lake thickness was 14.9+0.1 m at DH9 and 13.3+0.2 at DH12. The same values are obtained from similar assumptions as to the values of the con- tact temperatures. The smooth curve in figure 15 was constructed using these values for lake thickness and shows a decrease in the proportionate depth from about 0.6 when the maximum temperature was 1,050°% 1,000°C, to 0.75 at 680°, 0.9 at 475°, and 1.1 or more (below the lake) at 90°. Computed proportionate depths are also shown, based on the model illustrated in figure 39. Contact temperatures at the base of the central part of the lake were first measured on December 8, 1964 and are given in table 4. Large ranges of values are indicated because of uncertainties in the exact depth of the basal contact in each hole, but the ranges for drill holes 9 and 12 can be reduced by the assumptions dis- cussed above. On this basis, contact temperatures were as much as 700+10°C (60 percent of the beginning temperature of the lake) when first measured on De- cember 8, 1964, and decreased slowly to 600° 10°C on August 19, 1965. Computed contact temperatures at the base of a lake 49 feet (14.9 m) thick with a latent heat of 80 cal/g fell from 668°C to 606°C for the same period. From May 27, 1966, to August 31, 1967, the measured contact temperatures fell from 465°C to 86°C. Footnotes to table 4. Depth of basal contact estimated to be 9+1 m in DHS; 13+1 m in DH6, and 14+1 m in DHS, iased on vertical angle and level surveys after formation of the lake (table 6) and reeruption topography of Alae crater determined photogrammetrically (fig. 1). Depth of Basal contact in DH9 determined to be 14.5+0.8 m, on the basis of the recovery of drill core of 1840 lava between 14.6 and 15.2 m. Depth of basal contact in DH12 estimated to be 13.7+0.6 m on the basis of a cavernous layer penetrated in drilling between 13.1 and 14.3 m. If the assumption is valid that basal contact temperatures beneath drill holes 9 and 12 and the proportionate depths of the maximum temperature were the same on a given day, the depths of the basal contact are 14.9+0.1 m at DH9 and 13.3+0.2 m at DH12. Contact temperatures at these depths are shown in parentheses. *The maximum temperature measured was 90°C at the bottom of the drill hole 1.2 m below the assumed depth of the base of the lake. COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 19 1200 T T T T T T T T 1000 800 600 400 200 MAXIMUM TEMPERATURES, IN DEGREES CENTIGRADE T T I I I I I I | | 1 | I I I 1 | 1 | 1 1 1 1 1 1 1 0O 4 8 12 16 20 24 28 32 36 40 SQUARE ROOT OF TIME, IN DAYS, AFTER AUGUST 22, 1963 FiGurE 14.-Maximum temperatures at three sites in Alae lava lake as a function of the square root of time in days. Location of drill holes shown in figure 1. Crosses indicate computed temperatures derived from figure 39; black circles indicate measured temperatures in drill hole 9; open circles indicate measured temperatures in other drill holes. DENSITY OF ALAE LAVA AND VARIATION WITH TIME OF THE ALTITUDE OF THE LAKE SURFACE Molten basalt in Alae lava lake vesiculated because of the exsolution of gases as the lava cooled and solid- ified. The rate of vesiculation with time was recorded in the abundance of vesicles in solidified basalt newly added to the crust of the lake. The variation of vesicu- lation with temperature can be deduced from the size and shape of vesicles and the petrography of quenched partly molten basalt in drill core from near the base of the crust. Because of vesiculation, fluid lava in the lake expanded as it solidified, causing the surface of the lake to extend horizontally and to rise vertically. The solidified margin, in contrast, subsided because of thermal contraction. As a result, we were able to de- termine the outline of the lens of fluid and partly mol- ten lava and record its shrinkage with time by means of repeated surveys of the altitude of stations on the lake surface as well as the measurement of tempera- tures in drill holes. Data from the surveys and other measurements are used to calculate the density of fluid lava in the lake early in its cooling history. Repeated surveys of the stations after solidification of the lake reveal the pattern of continued cooling and provide data that are used to calculate the bulk coefficient of thermal expansion. This section presents data on the density and vesicularity of the lava and on the chang- ing altitude of the lake surface; the integration of this data and temperature data provides an analysis of the altitude changes. DENSITY AND VESICULARITY The measured density of individual pieces of drill core of solidified basalt from the lake ranges from 1.66 to 2.78 g/cc. In general density increased downward in the lake from an average value of 1.80 g/cc within 10 cm of the upper surface to a nearly constant 2.67 to 2.68 g/cc at depths below 3 m. Average values of den- sity at 0.3-m intervals from the surface to a depth of 3 m and at 1% m-intervals below that are given in table 5 and plotted in figure 16. Density was determined by planing flat the ends of each core 1 cm or more in length from the 12 drill holes and measuring its vol- ume and weight. A total length of 429 em of core was measured (almost all 1.43 cm diam), an average of 20 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII 0.5 % T T T T T PROPORTIONATE DEPTH 1.0 1 1 1 1 1200 1000 800 600 400 200 0 MAXIMUM TEMPERATURE, IN DEGREES CENTIGRADE FiGURE 15.-Proportionate depth of the maximum temperature in four drill holes in the central part of the lake as a function of the value of the maximum temperature. The smooth curve was con- structed on the assumption that the proportionate depths of the maximum temperatures in DH9 and 12 and values of the basal contact temperatures were the same in each drill hole. Crosses indicate computed depths and temperatures based on the thermal model illustrated in figure 39. about 30 percent for each interval. Because only a few drill holes penetrated the lower half of the lake, only a small amount of core was recovered below 6 m. Core recovery also was poor from the more vesicular zones in the lake, particularly from the highly vesicular and cavernous layer at a depth between 0.3 and 0.6 m. The density values listed in table 5 and plotted in figure 16 are probably biased towards greater density, particu- larly for depths less than 5 m. TABLE 5.-Bulk density and porosity of drill core [All values are for core collected from temperatures less than 1,000°C] Range in Depth Total length Average' measured Calculated interval (m) of core (cm) density density porosity (g/ce) (g/cc) (in percent) 0- 0.3 123 1.91 1.66-2.43 37 0.3- 0.6 7 2.12 1.86-2.43 30 0.6- 0.9 17 2.22 1.75-2.47 26 0.9- 1.2 17 2.36 1.91-2.64 22 1.2- 1.5 14 2.34 1.91-2.64 22 1.5- 1.8 17 2.47 2.13-2.56 18 1.8 2.1 25 2.50 2.19-2.60 17 2.1- 2.4 16 2.48 2.19-2.65 19 24- 2.7 30 2.59 2.87-2.171 14 2.7- 3.0 13 2.61 2.52-2.71 13 3.0- 4.5 51 2.68 2.54-2.78 11 4.5- 6.0 29 2.67 2.60-2.70 11 6.0- 7.5 20 2.65 2.61-2.65 12 7.5- 9.0 7 2.68 2.61-2.70 11 9.0-10.5 14 2.68 2.65-2.70 11 10.5-12.0 15 2.67 2.63-2.175 11 12.0-13.5 14 2.66 2.63-2.72 12 Weighted by length of each piece of core. *Porosity calculated using the average measured grain density of 3.01 g/cc. 1 Leg! hx Ty. lll 1 X 1 I " 3 | 40 30 fs a | : l 3 Porosity (~ vesicularity), in percent ; I H l l ‘ al i | DEPTH, IN METERS w T 1 1 11 73 12|- | a 13} 1 1 | 1 | 1 | I 1 1 1 1.8 2.0 22 2 4 2.6 2.8 CORE DENSITY, IN GRAMS PER CUBIC CENTIMETER 1 4 1.6 FIGURE 16. -Average core density and porosity (approximately equal to vesicularity) as a function of depth. Vertical bars indicate den- sity of individual pieces of core from the indicated interval. X's show average values for 0.3-m intervals from 0 to 3 m and 1% m intervals below. Density increased with depth in the lake because of a decrease in the abundance of vesicles. Because of the nearly constant mineralogic composition of the Alae basalt, the porosity, g, in percent, can be calculated from the bulk density, p,, and the average grain den- sity, p,, using the following equation (Daly and others, 1966, p. 21): ¢ = 100(1 42—5) The grain densities of eight samples of basalt from the lake range from 2.92 to 3.04 g/cc and average 3.01 g/ce. Porosity of the basalt is nearly equivalent to the ve- sicularity. A small proportion of the porosity of all the samples, however, probably is caused by microfrac- tures that formed as the result of thermal stresses in- duced by cooling. The microfracture porosity is COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 21 suggested by analysis of laboratory measurements of thermal conductivity of Alae and other Hawaiian basalt samples (Robertson and Peck, 1974) and by comparison of the coefficient of thermal expansion cal- culated from temperature and leveling data from Alae with values measured in the laboratory (Skinner, 1966; Richter and Simmons, 1974). As microfracture porosity probably is only about 0.3 percent or less of each sample (p. 33), it was disregarded in converting calculated porosity to vesicularity. Vesicularity so cal- culated agrees closely with modal vesicularity mea- sured in thin sections. In general, it decreased from 40 percent near the surface of the lake to a nearly con- stant 11 percent below 3 m. In detail, vesicularity did not decrease smoothly with depth. In at least some of the drill holes, it increased downward from the surface to maximum values in a highly cavernous layer at depths between 0.3 and 1.2 m and then decreased with greater depth. During the drilling of several holes at these depths, the drill fell abruptly as much as a tenth of a meter. Little core was recovered from this interval from any of the drill holes, particularly from a depth of 0.3-0.6 m. Caverns at shallow depth were detected by stamping on the sur- face during a detailed study of surface features and joint cracks (Peck and Minakami, 1968, fig. 3; Peck and Kinoshita, 1976, fig. 33). Fifteen cavities were detected in an area approximately 100 feet (30.5 m) square. The cavities were beneath the central parts of the broad hummocks that made up the surface of the lake and were formed near the end of the eruption by the en- trapment of rising gases below a thin surface crust. The appearance of similar features in older flows ex- posed on crater walls and roadcuts suggests that the cavities in Alae were lensoid openings a few tenths of a meter high and a meter or two in diameter. Below the cavernous layer to depths of about 5 m, the lava con- sisted of vague, alternating layers of more and less vesicular basalt a few centimeters or fractions of a cen- timeter in thickness. In general, the width of the ve- sicular layers decreased and the spacing between the layers increased with depth. They were sparse below 3 m and disappeared at a depth of about 5 m. The size of vesicles in the lava increased from a minimum at the surface to a maximum at depths of about 30 cm, and then decreased with greater depth. The average size of the 10 largest vesicles in each core interval increased linearly from 2 mm at the surface to 8 mm at a depth of 0.3 m, as shown in figure 17. Below 0.3 m, the average size of the largest vesicles de- creased, reaching a minimum of about 1 mm between 6 and 12 m and increasing slightly near the base of the lake. The vesicles were round to spheroidal in the upper meter of glass-rich basalt, but the more com- 0 T T 1.7 p. < "h“.‘h Zone of 1 F- I H ‘ large cavities a : 1h ~ 3 [- | l I 3 } H - p 1 5 F- I - DEPTH, IN METERS w T 1 $=" '; 4 a - | k - alel i | MAXIMUM VESICLE SIZE, IN MILLIMETERS, a i IN UPPER 30 CENTIMETERS 12 |- 2a .9 2 C 6 8. -t- 0 T T I'H|I T I& i | H $k. HP tn 13 - E o0|- "A1 } z 20 ‘IHI' | © 30 1 fa | 1 | | 1 1 1 1 0 2 4 6 8 MAXIMUM VESICLE SIZE, IN MILLIMETERS FigurE 17.-Maximum size of vesicles as a function of depth. pletely crystallized basalt at greater depths contained increasingly abundant minute angular vesicles, the voids characteristic of diktytaxitic texture. Below 6 m the vesicles were almost entirely of the minute angular type. Both types of vesicles are shown in the photomi- crograph of figure 18, which was taken of core collected November 17, 1964, from a depth of 1.5-2.1 m in drill hole 12. The large, round vesicles apparently formed at high temperatures in relatively fluid lava, most of them probably at temperatures of approximately 1,100°C or more, and their decrease in abundance with depth in the lake recorded the degassing and approach to equilibrium gas content of the Alae lava. The lava was supersaturated with gas on eruption because of the de- crease in confining pressure from the magma conduits to the surface. The relative rate of degassing with time can be estimated by comparing the change in vesicu- larity with depth to the change in depth with time of the 1,100°C isotherm. As shown in figure 19, this com- parison suggests that the rate of degassing decreased 22 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII B FigUrE 18.-Photomicrograph of vesicular basalt from a depth of 1.5-2.1 m in drill hole 12. A, plane-polarized light. B, Crossed nicols. The smooth round vesicle (V), a millime- ter in diameter, presumedly formed from gas exsolved from the supersaturated fluid lava at temperatures of 1,100°C or more; minute angular vesicles (A) formed at lower temperatures in the transient zone of crystallization from gas driven out of solution by crystallization. COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII B 0 t 0 1 VESICULARITY, IN PERCENT No =] T 1 (4 a | 1 (eZ | 1 1 40 8o 120 160 TIME, IN DAYS, AFTER AUGUST 22, 1963 as o 200 [=] FicurE 19.-Vesicularity of depth intervals in the upper part of the crust of Alae lava lake as a function of the time for each interval to cool to 1,100°C. Vertical bars indicate the vesicularity of depth intervals that are 0.3 m thick in the upper 4.5 m and 1.5 m thick below. The plot records the strong degassing of the Alae lava dur- ing the first 30 days after the eruption. drastically during the first 30 days, during which time the upper 2.5 meters of the lake cooled to less than 1,100°C; thereafter, gas from the slightly oversatu- rated fluid lava slowly exsolved until the lava reached an equilibrium gas content approximately 100 days after the eruption, when the upper 4.5 m had cooled to less than 1,100°C. Rapid degassing of the ponded lava in Makaopuhi crater during the first 4 months after the March 1965 eruption was demonstrated by Finlayson, Barnes, and Naughton (1968, fig. 5); they determined by chemical analysis the rapid depletion of carbon and sulfur in gas collected from drill holes in the crust of the lake. The size and shape of the minute angular pores and their rather constant abundance from top to bottom in the solidified lake suggest that they formed at lower temperatures than the spheroidal vesicles and near the crust-fluid lava interface, where growth of vesicles was constrained by a meshwork of crystals and higher vis- cosity melt. The vesicles apparently formed from gases exsolved from the melt because of the crystallization of anhydrous minerals-that is, "second boiling" of the melt. The rather constant 11 percent vesicularity at depths below 3 m would represent 0.0025 percent gas by weight, assuming the gas were water vapor. The relative abundance of the angular pores at different temperatures within the zone of crystallization could not be determined from the drill core because core from the zone also contains vesicles formed from vaporized coolant water. The natural abundance of the angular pores with temperature and the rate of vesiculation within the zone, however, can be estimated from the rate of crystallization with temperature. The estimated bo to 16 T [ TA T T T 1.6 f. (5 [2% < 14 |- s 41.4 &C 1s} : 7 O R | \\/Rate of vesiculation £ 2 32, | 112 z ® 2 .L | I 6 o T 2 I | c. w o t | \ C u 2 010 I \ 110 3 ¢ mew | \ 2 I w & - 8 \ 0.8 fl; 25 \ yz a s X 52 O 6F- 410.6 O S > R m E G z \ - a. 5 - y ~ I uw r 4 - /// \\ - 0.4 mg /// \\\\ 3 2 _// Cumulative abundance 1p. 2 8 / of vesicles z 0 1 1 | | | | L 0 980 1020 1060 1100 1140 TEMPERATURE, IN DEGREES CENTIGRADE FicurE 20.-Estimated rate of vesiculation (dashed curve) and cumulative abundance of minute angular vesicles (solid curve) as a function of temperature. rate of vesiculation and the cumulative abundance of vesicles (fig. 20) are based on petrographic observa- tions of crystal abundance and calculations from chem- ical analyses of samples from Alae and Makaopuhi lava lakes (Wright and Peck, 1978). The calcula- tions suggest that the maximum rate of vesiculation (1.6 percent vesicles formed for a 1°C drop in tempera- ture) took place between 1,060 and 1,065°C, and that two-thirds of the vesicles were formed between 1,100°C and 1,030°C. A strong peak of vesiculation at 1,060- 1,065°C is confirmed by calculations (p. 34) that show good agreement between the observed uplift during solidification at the main drilling site and the uplift calculated on the assumption that vesiculation took place at 1,065°C. ALTITUDE OF THE LAKE SURFACE Repeated surveys of the altitude of a grid of stations on the lake surface over the 4-year period of solidifica- tion and continued cooling reveal a complex, changing pattern of uplift and subsidence. During the year of solidification, the surface above the thick central part of the lake rose as much as 25 ecm and the margin subsided as much as 6 ecm. During the following 3 years of continued cooling, the entire lake surface subsided. Interpretation of the changes in altitude involves not only analysis of the leveling data but also integration of that data with temperature measurements, surveys of horizontal extension and contraction of the lake sur- face, and data on density and vesicularity of the crust of the lake. Accordingly, data on the altitude and hori- 24 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII A 20 -20 -100 -SO +20 August 30 to September 11, 1963 September 11 to September 24 to 26, 1963 100 METERS Values in cm/day x 100 > e\ o :s0 12 t - Ae t FicurE 21.-Rate of change in altitude of the surface of Alae lava lake during solidification of the lake from August 30 to September 24-26, 1963. Values are in cm/day x 102. A, August 30 to September 11, 1963; B, September 11 to September 24-26, 1963. zontal separation of stations are presented first, fol- lowed by interpretations of the recorded changes. CHANGES IN ALTITUDE DURING SOLIDIFICATION On August 30, 1963, 7 days after the eruption, the altitudes of nine stations on the lake were measured approximately. When they were remeasured 12 days later on September 11, seven of them had subsided, decreasing in altitude as much as 12.2 em, as shown in figure 21 A and listed in table 6. The full cross of 17 stations was leveled on September 11 and again Sep- tember 24-26. During the 2-week interval, most of the stations near the edge of the lake subsided, and most of those near the center were uplifted (fig. 21 B). The greatest uplift was 5.2 cm at station 5, near the center of the lake. The complete grid of stations was established and their relative altitudes measured September 24-26, and at monthly intervals for the following year while the lake continued to solidify. Changes in altitude are plotted and contoured as rates of change per day for six bimonthly periods in figure 22. The average rate of uplift of all stations was greatest during October and November 1963 (0.08 cm/day) and fell steadily thereaf- ter to 0.05 em/day for March and April 1964, 0.02 cm/day for June and July, and 0.003 em/day for August and September. The relative changes of altitude of the *As described in the section on methods of study, p. 7, the initial survey was conducted with a damaged transit, and the measured altitudes may be in error by as much as 1.5 cm. different stations on the lake showed a consistent pat- tern during each bimonthly period of the year. The margin of the lake subsided and the central part rose. The line between areas of uplift and subsidence mi- grated inward, very slowly during the first 6 months but more rapidly during the rest of the year. Uplift was not constant over the central part of the lake. During the first 8 months it was greatest in a belt part way between the center and the edge of the lake. During the last 4 months, the greatest uplift was near the center of the lake. Comparison of the patterns of uplift and sub- sidence with the contour map of the base of the lake (fig. 1) suggests a relationship between changes in al- titude and the depth of the lake. This relationship can be seen most clearly in the changes during April to June 1964 (fig. 22 D), when the ring of maximum uplift and the line between uplift and subsidence bent inward where the northern part of the lake thinned over a buried 1840 spatter ridge and where the southeast part thinned over a less pronounced ridge. Several stations were added during the year to de- termine whether differences in altitude changes be- tween stations of the grid took place smoothly or ab- ruptly with distance. Two stations (387 and 38) were marked in early December 1963 on the crests of small tumuli near station 4 at the center of the lake, and five stations (42 through 46) were added in early January 1964 on the north-central part of the lake between sta- tions 3 and 29. Repeated level measurements over the following 6 months showed no abrupt changes in al- titude between these closely spaced stations. COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 25 TaBue 6.-Altitude of stations on the lake surface [Altitudes given in feet above sea level (1 foot = 0.3048 m). Figures in parentheses show change in altitude between dates] Station Aug. 30, 1963 Sept. 11, 1963 Sept. 24-26, 1963 Pr e" lok, eee 2566.98 (+0.04) 2567.015 (+0.027) i e eae L cr 000 cakes 66.43 (+.02) 66.453 (+.117) Bbe er ae o Sco. yews 65.75 (+.04) 65.793 (+.041) A mail 2566.22 (+0.07) 66.29 (+.06) 66.346 (+.094) rear ie. 67.08 (-.06) 67.02 (+.17) 67.190 (+.153) b eel 66.85 (-.40) 66.45 (+.05) 66.498 (+.134) .. 66.93 (-.32) 66.61 (-.06) 66.545 (+.082) Bee l tio cf Ln Aes 67.93 (+.05) 67.981 (+.132) D rec nes, aon (ou. nake 66.70 (+.05) 66.751 (+.057) Tl 65.07 (+.03) 65.10 (-.07) 65.029 (-.062) 12 fel 2568.10 assumed 68.10 (-.11) 67.989 (-.009) To 67.05 (-.24) 66.81 (-.09) 66.721 (+.131) 14} X...... 67.14 (-.16) 66.98 (+.07) 67.050 (+.087) 104 66.96 (-.28) 66.68 (+.05) 66.726 (+.122) NOs 66.51 (-.06) 66.45 (+.05) 66.495 (+.097) Mi e evi Lo cala Pele 67.04 (-.01) 67.028 (+.030) _ Lac. " Puce? -s. 65.58 (+.02) 65.60 (+.032) 19 sec (e c (tC yey 2568.09 assumed hes: 2568.09 assumed evs BOP l n haes ec iim eate a -se 66.781 (+.100) eid ni ol oes nese r ag oe oe aus 66.649 (+.158) e merece n eed ne oop Fees: 66.513 (+.177) g e Ariat us Collet C eme 68.60 (0) 68.604 (-.052) Fp ol nn coe ae o h ae in oon wae swer 67.195 (+.116) o ese nene ean nolo a ao oal ea on aas 67.701 (+.069) RD seee o Alo tert, ege pei tte See. 67.086 (+.082) a ess t eves ie a ooo Seas 68.026 (+.113) BA sree i ike en e Lo LCC rossi es opal Sees 66.613 (+.016) e Eee -/ awe oon esu sie i eo eck 66.411 (+.227) B0 aries a l cl. o Peso eae A era aes 65.955 (+.073) or Uo ott ee ai arage uke 68.176 (-.088) anime "n __gc el a t ao ne 67.734 (-.052) BER c notte ent h an io anld se,. 66.868 (+.146) oder s sun) 1 it ec =c. 67.485 (-.113) LLL -" clt lth mais AQ CE sts 67.814 (-.005) Tae uas ee U2. sed Natu Advent sul mac aes nlf l mani amoeba eil oen rtg Ae eh mn vee MDa "" seven . tae fare net a noen? withe AT MALI L calico pis iol a fea aman ato oe sy Station Oct. 30, 1963 Dec. 4, 1963 Jan. 2, 1964 1: 2 cele... 2567.042 (+0.070) 2567.112 (+0.030) 2567.142 (+0.092) eines s. 66.570 (+.108) 66.678 (+.058) 65.731 (+.079) eA Lecco: 65.834 (+.078) 65.907 (+.025) 65.932 (+.088) A spar cl. 66.440 (+.129) 66.569 (+.019) 66.588 (+.083) Or 67.343 (+.062) 67.405 (+.002) 67.407 (+.053) Cia 66.632 (+.079) 66.711 (+.025) 66.736 (+.064) Tete ence.. 2. 66.627 (+.033) 66.660 (+.032) 66.692 (+.053) SHe esen. 68.113 (+.140) 68.253 (+.103) 68.356 (+.139) 9 cs 66.808 (+.092) 66.900 (+.076) 66.976 (+.104) 64.967 (+.001) 64.968 (-.011) 64.957 (+.010) 67.980 (+.023) 68.003 (-.010) 67.993 (+.006) 19 [fcc toll. c: 66.852 (+.132) 66.984 (+.082) 67.066 (+.127) 14 67.137 (+.049) 67.186 (+.037) 67.223 (+.056) 19 66.848 (+.056) 66.904 (+.018) 66.922 (+.057) 16: 66.592 (+.111) 66.703 (+.038) 66.741 (+.056) Tis ss sell... 67.058 (+.099) 67.157 (+.063) 67.220 (+.071) 18 "*-s:..l.3.l_ 65.632 (+.040) 65.672 (+.033) 65.705 (+.055) 19} 2568.09 assumed P oen 2568.09 assumed tues 2568.09 assumed nees 0 A AAIA... 66.881 (+.074) 66.955 (+.041) 66.996 (+.107) Vite no 66.802 (+.145) 66.947 (+.042) 66.989 (+.121) Pe 66.690 (+.137) 66.827 (-.008) 66.824 (+.095) 68.552 (-.032) 68.520 (-.022) 68.498 (-.016) Eee i 67.311 (+.115) 67.426 (+.040) 67.466 (+.078) o eet ls 67.770 (+.070) 67.840 (+.030) 67.870 (+.095) SO .s 67.168 (+.139) 67.307 (+.048) 67.350 (+.088) T 68.139 (+.106) 68.245 (+.074) 68.319 (+.117) re ete 66.629 (+.104) 66.733 (+.057) 66.790 (+.150) L 66.638 (+.119) 66.757 (+.012) 66.769 (+.111) G0 66.028 (+.107) 66.135 (+.051) 66.186 (+.102) ole: 66.088 (-.028) 68.060 (-.024) 68.036 (-.024) a Aae sk s 67.682 (-.008) 67.679 (+.002) 67.681 (+.027) Be es dien nol; 67.014 (+.101) 67.115 (+.038) 67.153 (+.101) SE CALL.: nl 67.372 (0) 67.372 (-.027) 67.345 (+.020) 9B. 67.809 (+.063) 67.872 (+.024) 67.896 (+.061) 26 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII TABLE 6.-Altitude of stations on the lake surface-Continued Dec. 4, 1963 Station 8T. NYs 68.667 (+.025) BB" seneca veel n ouly sss 68.335 (+.049) Af as c ORIC e ean ik r l ead cows AT aoc nre mses a L_ Ue et i y reeds yee? Station Feb. 7-10, 1964 Mar. 3, 1964 Terran 2567.234 (+0.036) 2567.270 (+0.048) 2M i 66.810 (+.032) 66.842 (+.063) § eee 66.020 (+.024) 66.044 (+.054) § 66.671 (+.023) 66.694 (+.053) BEER 67.460 (+.016) 67.476 (+.055) b 66.800 (+.036) 66.836 (+.069) J 66.745 (+.055) 66.800 (+.114) Droit na to 68.495 (+.074) 68.569 (+.153) ONA Ece 67.080 (+.078) 67.153 (+.093) 64.967 (+.004) 64.971 (-.032) 12 67.999 (+.013) 68.012 (-.019) 19 67.193 (+.106) 67.299 (+.092) T4 3}: 67.279 (+.053) 67.332 (+.043) 1h merce les 66.979 (+.032) 67.011 (+.049) 16 ~ 66.1797 (+.032) 66.829 (+.065) Tia AC... css 67.291 (+.042) 61.333 (+.082) 18 65.760 (+.042) 65.802 (+.046) T9 clt c 2568.09 assumed ews 2568.09 assumed sare POF raids 67.103 (4+:006) . .=! >= ore 67.110 (4.037): ._: " _:see. B2 NRO 66.919 (11052) = o SACs weak RJ CAL 68.482 (+.006) 68.488 (-.024) PA 67.544 (+.037) 67.581 (+.093) P5 e send 67.965 (+.052) 68.017 (+.095) - 67.438 (+.033) 67.471 (+.094) PT ll 68.436 (+.056) 68.492 (+.132) TB 66.940 (+.048) 66.988 (+.020) POSE Le 66.880 (+.027) 66.907 (+.060) BQ 66.288 (+.030) 66.318 (+.060) ST: 68.012 (+.015) 68.027 (-.010) Iles. 67.708 (-.012) 67.696 (+.007) BE esther ese 67.254 (+.047) 66.301 (+.147) 84 67.365 (-.015) 67.350 (-.014) BD 67.957 (-.003) 67.954 (-.026) 37 68.1769 (+.025) 68.794 (+.048) SB 68.480 (+.033) 68.513 (+.069) Melo 65.920 (+.010) 65.930 (-.035) A1 2 Acas. - oen - n cat Station May 12, 1964 June 9, 1964 I 2567.344 (+0.022) 2567.366 (+0.002) PCs ses 66.983 (+.082) 67.065 (+.035) B 66.146 (+.048) 66.194 (+.058) Cre R 66.804 (+.071) 66.875 (+.062) Bram 67.576 (+.073) 67.649 (+.055) C LAP 66.960 (+.133) 67.093 (+.032) T Vee 66.956 (+.017) 66.973 (+.007) Sil elsciclecc. 68.755 (+.031) 68.1786 (-.008) 67.219 (+.005) 67.224 (-.010) IY 64.936 (000) 64.936 (-.001) 12 RRI 67.992 (+.005) 67.997 (-.005) 19" 67.462 (+.007) 67.469 (-.003) 14 L 67.440 (+.097) 67.537 (+.109) 1B: 67.116 (+.103) 67.219 (+.104) 16: 66.964 (+.044) 67.008 (+.019) 17 sarc 67.481 (+.024) 67.505 (+.001) 18 H 65.851 (-.021) 65.830 (-.010) $8 ____________ 2568.09 assumed az 2568.09 assumed se oT f AE Xx cit a m ean Pais Te OB (~" poe mel o eevee EL no A 68.434 (+.002) 68.436 (-.005) Pd: 67.1732 (+.114) 67.846 (+.009) 2b Lolis 68.116 (+.002) 68.118 (-.021) 20 Hic 67.621 (+.124) 67.745 (+.035) PT 68.683 (+.007) 68.690 (+.006) P8: swell 67.022 (-.003) 67.019 (-.010) 2D 67.017 (+.085) 67.102 (+.071) Jan. 2, 1964 68.692 68.384 65.934 Apr. 13, 1964 2567.318 66.905 66.098 66.747 67.531 66.905 66.914 68.722 67.246 64.939 67.993 67.391 67.375 67.060 66.894 67.415 65.848 2568.09 assumed July 6, 1964 2567.368 67.100 66.252 66.937 67.704 67.125 66.966 68.778 67.214 64.935 67.992 67.466 67.646 67.323 67.027 67.506 65.820 2568.09 assumed (+.077) (+.096) (-.014) (+0.026) (+.078) (+.048) (+.057) (+.045) (+.055) (+.042) (+.033) (-.027) (-.003) (-.001) (+.071) (+.065) (+.056) (+.070) (+.066) (+.003) (-.030) (+.058) (+.004) (+.056) (+.059) (+.014) (+.150) (+.092) (-.017) (-.037) (+.040) (+.001) (-.001) (+.058) (+.059) (-.031) (-.012) (+0.009) (+.028) (+.070) (+.032) (+.052) (+.008) (+.003) (-.017) (-.010) (-.003) (-.006) (000) (+.053) (+.023) (+.008) (-.003) (-.017) (-.017) (-.010) (-.005) (-.015) (-.006) (+.003) (+.032) COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 27 TABLE 6.-Altitude of stations on the lake surface-Continued Station May 12, 1964 June 9, 1964 July 6, 1964 66.470 (+.049) 66.519 (+.029) 66.548 (-.006) 68.000 (-.014) 67.986 (-.027) 67.959 (-.004) 67.666 (-.008) 67.658 (+.004) 67.662 (-.001) 67.488 (+.003) 67.491 (-.003) 67.488 (-.004) LY. 67.337 (-.028) 67.309 (-.009) 67.300 (+.004) ote o. 67.919 (-.012) 67.907 (-.015) 67.892 (-.021) 68.900 (+.071) 68.971 (+.073) 69.044 (+.062) 68.641 (+.048) 68.689 (+.040) 68.729 (+.022) 65.864 (-.005) 65.859 (-.014) 65.845 (-.013) de- 67.435 (+.002) 67.437 (-.001) 67.436 (-.010) July 31, 1964 Aug. 27, 1964 Sept. 24, 1964 ____________ 2567.377 (-0.014) 2567.363 (-0.010) 2567.353 (-0.008) ____________ 67.128 (+.008) 67.136 (-.003) 67.133 (-.007) ____________ 66.322 (+.013) 66.335 (-.014) 66.349 (-.002) ____________ 66.969 (+.035) 67.004 (+.009) 67.013 (+.001) ____________ 67.1756 (+.015) 67.1771 (+.017) 67.788 (-.009) ____________ 67.133 (.000) 67.133 (-.005) 67.128 (-.021) ____________ 66.969 (-.012) 66.957 (+.004) 66.961 (-.017) ____________ 68.761 (+.001) 68.1792 (-.006) 68.756 (-.004) ____________ 67.204 (-.015) 67.189 (-.003) 67.186 (-.017) ____________ 64.932 (+.003) 64.935 (-.008) 64.927 (-.003) ____________ 67.986 (+.007) 67.993 (000) 67.993 (+.001) ____________ 67.466 (-.006) 67.460 (-.002) 67.458 (-.005) ____________ 67.699 (+.020) 67.719 (-.002) 67.717 (-.004) ____________ 67.346 (+.026) 67.372 (-.009) 67.363 (-.008) ____________ 67.035 (+.018) 67.053 (-.013) 67.040 (-.013) ____________ 67.503 (-.009) 67.494 (-.005) 67.489 (-.014) ____________ 65.803 (-.022) 65.781 (+.002) 65.783 (-.019) ____________ 2568.09 assumed vs 2568.09 assumed as 2568.09 assumed y ____________ 68.414 (-.004) 68.410 (-.005) 68.405 (-.008) ____________ 67.845 (-.004) 67.841 (-.007) 67.834 (-.012) ____________ 68.092 (-.025) 68.067 (+.006) 68.073 (-.029) ____________ 67.1765 (+.005) 67.770 (-.004) 67.766 (-.004) ____________ 68.690 (-.006) 68.684 (+.001) 68.685 (-.004) ____________ 67.012 (-.007) 67.005 (-.014) 66.991 (+.001) ____________ 67.205 (+.005) 67.210 (-.011) 67.199 (+.004) ____________ 66.542 (-.018) 66.524 (-.001) 66.523 (-.002) ____________ 67.955 (-.019) 67.936 (000) 67.936 (-.021) ____________ 67.661 (-.019) 67.642 (-.001) 67.641 (-.012) ____________ 67.484 (+.001) 67.485 (-.015) 67.470 (+.002) ____________ 67.304 (-.023) 67.281 (-.006) 67.275 (-.011) ____________ 67.871 (-.008) 67.863 (-.032) 67.831 (-.005) ____________ 69.106 (+.027) 69.133 (+.013) 69.146 (-.007) ____________ 68.751 (+.009) 68.760 (+.007) 68.767 (-.002) ____________ 65.832 (-.015) 65.817 (+.004) 65.821 (-.014) ________ 67.426 (-.010) 67.416 (+.004) 67.420 * Nov. 3, 1964 Jan. 21, 1965 May 11, 1965 frass... 2567.345 (-0.028) 2567.317 (-0.070) 2567.247 (-0.003) a eee dei ano 67.126 (-.014) 67.112 (-.038) 67.074 (+.004) cle 66.347 (-.017) 66.330 (-.036) 66.294 (-.002) d 67.014 (-.014) 67.000 (-.025) 66.975 (+.005) 67.779 (-.002) 67.777 (~-.023) 67.754 0 hor to lated 67.107 (-.001) 67.106 (-.023) 67.083 (-.005) eserves Lt 66.944 (-.005) 66.939 (-.016) 66.923 (+.002) Baer 68.752 (-.015) 68.737 (-.030) 68.707 0 o - 67.169 (-.035) 67.134 (-.027) 67.107 (+.001) icc. 64.924 (-.001) 64.923 (-.004) 64.919 2s T2 sel 67.994 (+.014) 68.008 (+.004) 68.012 t: 19 us 67.453 (-.015) 67.438 (-.034) 67.404 (+.015) 67.713 (-.008) 67.1705 (-.022) 67.683 (+.013) i (oen oats 67.355 (-.006) 67.349 (-.023) 67.326 (+.018) tO 67.027 (-.010) 67.017 (-.031) 66.986 0 Tics -l: 67.475 (-.019) 67.456 (-.031) 67.425 (-.003) 18 .o 65.764 (-.047) 65.717 (-.041) 65.676 (-.011) 10 s.. _X_. 2568.09 assumed 2g 2568.09 assumed a 2568.09 assumed 2s s . #s i ie a cece we "ike Arp to I ese an ion et s one meee arse ie a carne hen ae f l o a an tu A ap t a o tl tr te tti ote nota t cso i iy tap 68.397 (-.002) 68.395 (+013) 68.406 (+.002) TABLE 6.-Altitudes of stations on the lake surface-Continued SOLIDIFICATION OF ALAE LAVA LAKE, HAWAI May 11,1965 Station Nov. 3, 1964 Jan. 21, 1965 PA 67.822 (-.015) 67.807 (-.028) 67.779 (-.005) PB ) 68.044 (-.023) 68.021 (-.022) 67.999 (-.005) 20 HEC 67.762 (000) 67.1762 (-.014) 67.748 (+.006) 27. 68.681 (-.013) 68.668 (-.027) 68.641 (+.007) 2B 66.992 (-.030) 66.962 (-.059) 66.903 (-.001) 29: 67.203 (-.016) 67.187 (-.038) 67.149 (+.002) B0 2 sre 66.521 (-.022) 66.499 (-.028) 66.471 (+.002) GL s 67.915 (-.018) 67.897 0 67.897 (+.004) B2 s 67.629 (-.030) 67.599 (-.016) 67.583 (+.010) BQ ui 67.472 (-.028) 67.444 (-.047) 67.397 (+.005) 34 67.264 (-.031) 67.233 (-.081) 67.202 (+.002) BD 67.826 (-.030) 67.1796 (-.039) 67.1757 (+.022) 37 Asc s 69.139 maas seus alt agia L . x> 8B . luces 68.1765 sal a e seee aar e unis o whens $e 3,1] ____________ 65.807 (-.016) 65.791 (+.013) 65.804 tae Station Aug. 19, 1965 May 27, 1966 Aug. 31, 1967 2567.244 (+0.049) 2567.293 (-0.017) 2567.276 CA lea nile 67.078 (+.011) 67.089 (-.077) 67.002 8 i a 66.292 (+.007) 66.299 (-.108) 66.191 Ase e 66.980 e ai d . ae uue m 52 66.918 67.754 (+.032) 67.786 (-.082) 67.704 6 .l cess 67.078 (+.015) 67.093 (-.042) 67.051 Tre ceed 66.925 (+.019) 66.943 (-.037) 66.906 68.707 (-.038) 68.669 (-.048) 68.621 67.108 (-.041) 67.067 (-.029) 67.038 ______ y 65.059 (+.003) 65.062 ______ e 67.849 0 2567.849 assumed 67.419 (+.036) 67.455 (-.013) 67.442 67.696 (+.029) 67.1725 (-.078) 67.647 67.339 (+.031) 67.370 (-.068) 67.302 66.986 (+.020) 67.006 (-.067) 66.939 67.422 (+.003) 67.425 (-.043) 67.382 65.665 (+.021) 65.686 (-.037) 65.649 2568.09 assumed 5. 2568.09 assumed (-.043) 68.047 68.408 (+.001) 68.409 (-.014) 68.395 67.774 (-.024) 67.750 (-.043) 67.708 67.994 (-.009) 67.985 (-.033) 67.952 67.754 (+.022) 67.777 (-.044) 67.733 68.648 (+.002) 68.650 (-.032) 68.618 66.902 (+.029) 66.931 (-.034) 66.897 67.151 (+.022) 67.173 (-.094) 67.079 66.473 (+.031) 66.514 (-.065) 66.449 67.901 (+.019) 67.920 (-.027) 67.893 67.593 (+.060) 67.653 (-.009) 67.644 67.402 (+.017) 67.419 (-.106) 67.313 67.204 (+.077) 67.279 (-.033) 67.246 67.1779 i cases op a l ale Cumulative changes in altitude of five representat- ive stations on the lake are shown (fig. 23) as a function of time in days after the eruption. The stations include three (18, 23, and 35) within 30 m of the edge of the lake, one (29) about halfway between the center and edge, and one (15) near the center. Uplift of the station closest to the edge of the lake, 28, if it ever occurred, must have preceded the beginning of the survey. For the two other stations near the edge, subsidence fol- lowed uplift 200-300 days after the eruption. The two remaining stations show a pattern of changing altitude with time that was typical of almost all stations over the thick central part of the lake. During the second and third weeks after the eruption, station 15 subsided markedly, as did most of the few central stations sur- veyed. Thereafter, the stations were uplifted at a rate that decreased over the next several months to a nearly constant value. Shortly before the maximum uplift was reached, the rate of uplift increased to a high value for a month before quickly dropping to zero and to nega- tive values. The greatest cumulative uplift of the five stations was not at the stations near the center of the lake, but instead at station 29, part way between the center and COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII A B July 31 to September 24, 1964 e é? s 0 50 100 METERS beef cdc ed Contour interval 0.04 em/day Figure 22.-Rate of change in altitude of the surface of Alae lava lake during solidification of the lake from September 24-26, 1963, to September 24, 1964. Values shown in centimeters per day x 10°. Shaded areas were uplifted during the leveling interval. A, September 24 and 26 to December 4, 1963. B, December 4, 1963, to February 7 and 10, 1964. C, February 7 and 10 to April 13, 1964. D, April 13 to June 9, 1964. E, June 9 to July 31, 1964. F, July 31 to September 24, 1964. 29 to O o to o h o | 1 | 1 1 UPLIFT (+) AND SUBSIDENCE (-), IN METERS £ * 100 200 - 300 400 500 600 700 TIME, IN DAYS, AFTER AUGUST 22, 1963 800 FiGur® 23.-Cumulative changes in altitude of five stations on the lake after September 24-26, 1963, as a function of time in days after August 22, 1963. Location of stations plotted in figure 1. Stations 18, 23, and 35 are within 30 m of the edge of the lake; station 29 is about 75 m from the edge; and station 15 is near the center and deepest part of the lake. edge of the lake. A similar relationship for other sta- tions on the lake is shown in figure 24, on which is plotted the total cumulative uplift (later subsidence not subtracted) for the period September 24-26, 1963, to September 24, 1964, of all 29 main stations on the grid. The figure shows two areas of maximum uplift part way between the center and the eastern and west- ern margins of the lake. CHANGES IN HORIZONTAL DISTANCE DURING SOLIDIFICATION During the year that the lake solidified, distances SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII (ZN 20.4 - 20 7 * % 90 ,, 197 ~2 _ 194 b , 18.2 1320.4 C 4 4 17.0 100 METERS 0 50 Contour interval 5 centimeters FicurE 24.-Total uplift of the lake surface after September 24-25, 1963. between stations on the surface of the lake also changed in a consistent manner. The first measure- ments of the distances were on August 30, and Sep- tember 3 and 11, 1963. When the distances were re- measured on December 4 'and 17, 1963, the crust of the central part and north edge of the lake had expanded, and that of most of the lake margin had contracted, as shown in figure 25A. The greatest change was an ex- pansion of 32+1.8 ecm between stations 3 and 4. When the distances were remeasured on July 27, 1964, simi- lar but much smaller changes were found (fig. 25B). An A September to December 1963 B December 1963 to July 27, 1964 > #*x4 0 50 /7 100 METERS Lone Contour interval 10 centimeters 25.-Changes in horizontal distance between stations on the surface of Alae lava lake during solidification of the lake. A, Changes 1.5 cm and greater between September and December 1963. B, Changes greater than 0.6 cm between December 1963 and late July 1964. Areas that underwent expansion are more heavily shaded. COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 31 Contour interval 0.005 cm/day September 24, 1964 to January 21, 1965 Contour interval 0.005 cm/day January 21 to August 19, 1965 Contour interval 0.001 cm/day May 27, 1966 to August 31, 1967 Axes of trough of maximum subsidence <>\ 0 50 Exc nter) 100 METERS Ficur® 26.-Rate of subsidence of the lake surface after complete solidification in late September 1964. Values in centimeters per day x 10°. A, September 24, 1964 to January 21, 1965. B, January 21 to August 19, 1965. C, May 27, 1966 to August 31, 1967. D, Approximate positions of the axis of the belt of maximum subsidence between September 1964 and August 1965 and the point of maximum subsidence between May 1966 and August 1967. Areas that underwent subsidence of more than 0.005 cm/day are shaded. area of expansion in the central part of the lake was surrounded by a ring of contraction, which in turn was bordered on the north by a marginal area of expansion. The greatest increase was 2.7+0.6 cm between stations 3 and 4, and stations 7 and 26. The greatest decrease was 5.2+0.6 cm between stations 9 and 23. CHANGES IN ALTITUDE DURING FURTHER COOLING After the last interstitial melt in the lake solidified in late September 1964, the altitude of the surface con- tinued to change. The entire surface of the lake sub- sided, falling an average of 0.006 ecm/day between Sep- tember 1964 and January 1965, 0.004 ecm/day between January and August 1965, and 0.003 ecm/day between May 1966 and September 1967 (table 6).3 The rate of subsidence was not constant over the lake. From Sep- tember 1964 to January 1965 and from January to Au- gust 1965, subsidence was greatest in a belt part way between the center and the edge of the lake (fig. 264 and B). This belt migrated inward approximately 15 m between the successive surveys (fig. 26D). Presumedly, the survey in May 1966, if it had been successful, would have disclosed a similar ring of maximum sub- sidence closer to the center of the lake. During the final leveling period (fig. 26C), the contours of equal subsid- *The pattern and rate of subsidence between August 1965 and May 1966 is not known because the lake surface and base stations were shifted during an earthquake shortly before the survey of May 1966. 32 ence form a rude bullseye pattern having the maximum subsidence over the deepest part of the lake near station 3. INTERPRETATION OF CHANGES IN ALTITUDE SUBSIDENCE OF THE LAKE SURFACE AND CALCULATION OF THE COEFFICIENT OF THERMAL EXPANSION The processes causing subsidence of the surface after the lava lake had solidified will be discussed before taking up the more complex subject of altitude changes before complete solidification. Because fewer processes are involved-we can exclude lateral flowage of fluid lava and changes of bulk density upon crystallization and vesiculation-the interpretation is less ambiguous. Subsidence of the surface after solidification of the lake is attributed to thermal contraction of the cooling basalt. The thermal contraction also caused the open- ing of joint cracks (Peck and Minakami, 1968; Peck and Kinoshita, 1976). However, as prehistoric Kilauean flows and lava lakes exposed in section show joints that are mostly or entirely vertical or nearly vertical rather than horizontal, the altitude of the Alae lake surface probably was affected little, if any, by jointing. As will be shown later, microfracturing of the basalt very probably took place during cooling, resulting in a de- crease of both subsidence and thermal conductivity. The effective bulk coefficient of thermal expansion of the Alae basalt can be calculated by comparing the changes in average temperature of the lake (measured in drill holes penetrating the base) with the change in altitude of stations near the collars of the holes. The mean coefficient of linear thermal expansion (aL) is given by the following formula (Skinner, 1966, p. 94): - 4 (AL) "* ~ I. (AT) For the lava lake, L equals the thickness of the lake, AL equals the change in thickness (that is, the change in relative altitude of the surface) for a given interval of time, and Ar equals the change in average temperature during the same interval of time. Relevant data are available for two surveying inter- vals at the main drilling site (DH12) and for one inter- val at station 4, above the deepest part of the lake. Drill hole 12, which was equidistant from stations 6 and 26, pierced the base of the lake at a probable depth of 13.3+0.2 m (see table 4). Station 4 was 3.6 m east of drill hole 9, which penetrated 1,840 lava below the lake at a probable depth of 14.9+0.1 m. Changes in altitude and temperature at the two sites and the calculated linear coefficient of expansion are listed below. SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII Average change c @ Coefficient of Date, drill hole, and in temperature lhange in thermal expansion survey stations (°C) altitude (cm) (x108°C) January 21 to August 19, 1965 DH9 and Sta. 4 ______ -170 -0.61 2.4 DH12 and Sta. 6 and 26 ---__.___. -155 -.55 2.7 May 27, 1966 to August 31, 1967 DH12 and Sta. Gand 26 .:.... ... -295 -1.28 3.3 Taking into consideration the possible errors in de- termining changes in altitude, the effective bulk coeffi- cient of linear thermal expansion is thus 3+1 x 10 S/ °C.4 This value is approximately one-half the values reported from measurements on similar rocks in the laboratory. Skinner (1966, p. 94) gives an average value of 5.4 x 10~°/°C for 10 laboratory determinations of the coefficient of linear thermal expansion of basalt, diabase, and gabbro. Richter and Simmons (1974, table 2) give a measured value of 18.8 x 10 °/°C and a calcu- lated value of 19.0 x 10~°/°C for the coefficient of vol- ume thermal expansion of the chemically and mineralogically similar Frederick diabase; these val- ues are equivalent to a coefficient of linear thermal expansion of 6.3 x 10 °/°C. The difference between the value of the coefficient of thermal expansion determined in situ for Alae basalt and those determined in the laboratory on similar but coarser grained, more slowly cooled rocks can probably be attributed to the formation of microfractures across grains and along grain boundaries during cooling of the basalt as the result of stresses induced by differen- tial thermal contraction of the constituent mineral grains. This conclusion is supported by the experimen- tal work of Richter and Simmons (1974), which showed that high heating rates and heating to high tempera- tures (>350°C) induce microfracturing that reduces the observed thermal expansion and results in a per- manent lengthening of the samples. The two diabase samples studied by them lengthened 0.5 to 1.5 x 10 }, after repeated heating to temperatures of 550-660°C. Assuming that microfracturing reduced the effective bulk coefficient of linear thermal expansion in Alae lava lake by 50 percent, as indicated by comparison of the field and laboratory data, the expansion of the lake (actually, decreased subsidence) inferred to have taken place because of microfracturing during cooling from 980°C to 20°C was 3 x 10 °, the same order of mag- Thermal modeling of temperatures within and below the lake (p. 45-47) suggests that changes in altitude caused by cooling and thermal contraction of the lake from January to August 1965 may have been reduced by about 10 percent because of heating and thermal expansion of lava beneath the lake. In contrast, changes in altitude from May 1966 to August 1967 may have been increased by about 10 percent because of cooling and thermal contraction of the underlying lava. Taking this effect into consideration would bring the calculated coefficients for the two periods into better agreement with the average value of 3x 10 COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 33 nitude as the strains measured in the laboratory. The presence of relatively abundant microfractures in Alae basalt was inferred independently from the re- sults of thermal conductivity measurements on sam- ples from Alae lava lake and other Hawaiian lava flows. As reported by Robertson and Peck (1974) and discussed on p. 43 of this report, the effective conduc- tivity of the samples studied was found to be 40 percent less than the measured or calculated inherent conduc- tivity. On the basis of studies by Walsh and Decker (1966), the disparity was ascribed to microfractures, which were thought to form 0.5 percent or less of the solid portion of the samples. If the difference between the observed bulk coefficient of thermal expansion in Alae lava lake and the coefficient measured on similar rocks in the laboratory was caused entirely by micro- fracturing of Alae basalt, then microfracture porosity of the basalt was 0.3 percent. When the calculated value of 3 x 10 ~ °/°C is used for the coefficient of linear thermal expansion, the contour maps of rates of subsidence can be used to estimate average rates of change of temperature in the lake below the surveying stations. The plot of figure 27 was prepared in this manner from figure 26B, and shows the estimated daily average decrease in temperature from January to August 1965. Excluding the cold mar- gin of the lake, the calculated cooling rates during this period ranged from average values of about 0.7°C/day in the central part of the lake to a maximum of 2.7°C/ day below station 1, 30 m from the north margin. Cool- 0 50 100 METERS Contour interval 0.5 degrees centigrade FiGurE 27.-Estimated daily average decrease in temperature in degrees centigrade in the lake from January to August 1965, as determined from measured rates of subsidence dur- ing that period (fig. 26B) and the observed bulk coefficient of linear thermal expansion of 3 x 10 */°C. The sites of drill holes 9 and 12 are marked. ing rates were highest in a ring (or annular volume) underlying the belt of maximum subsidence shown in figure 26B. This ring migrated inward as the lake cooled from September 1964 to August 1965, as shown by the relative positions of the axes of maximum sub- sidence in figure 26D. Temperatures in the lake de- creased sharply across the ring. Midway in the January to August 1965 interval, for example, the maximum temperature in drill hole 12, 60 m inside the ring was 760°C, whereas the maximum in drill hole 5, 30 m outside the ring was only 95°C. The average cool- ing rate of 0.7°C/day for the central part of the lake inferred by the subsidence rates is approximately the same as the value for this period of 0.9°C/day computed from the one-dimensional thermal model illustrated in figure 39. Because inferred cooling rates greatly ex- ceeded this value in the ring of maximum subsidence and because the ring moved inward towards the lake in successive surveying periods, the ring is interpreted to be that part of the lake that was undergoing rapid cool- ing. Numerical analysis using two-dimensional models (p. 47) does not indicate the existence of such a zone of maximum cooling part way between the center and margin of the lake. Presumedly, the observed belt re- flects some cooling process that was not incorporated in the model-possibly convective heat transfer by cir- culating rainwater in the cooler margin of the lake. CHANGES IN ALTITUDE OF THE SURFACE DURING SOLIDIFICATION OF THE LAKE The most striking and consistent aspect of the con- tour maps showing rate of change of altitude during solidification of the lake (fig. 22) is the presence during each bimonthly period from September 1963 to Sep- tember 1964 of a central area that had undergone up- lift bordered by a marginal ring that had subsided. Subsidence of the entire lake surface after complete solidification suggests that before complete solidifica- tion the marginal ring of subsidence overlay the sol- idified outer part of the lake and that some process going on in the central lens of molten lava caused uplift of the overlying surface. This hypothesis can be tested by comparing temperatures in drill holes with rates of uplift and subsidence at the collars of the holes. Figure 28 shows rates of uplift and subsidence at stations on the lake compared with maximum temperatures in nearby drill holes. Each maximum temperature plot- ted is at the midpoint between the times of the corres- ponding level measurements. The temperatures were determined from plots of maximum temperature ver- sus time (fig. 14); the plots were based on measure- ments in drill holes at the main drilling site and also in drill holes 5, 6, and 9. The plot in figure 28 shows that uplift took place only over that part of the lake that to p T T T I I T I I 0 0 0 o & co T T | 1 PER DAY, o [=] B T 1 UPLIFT IN CENTIMETERS 0 ra hal fl & HEW 00.02— irt & 0.004 1 1 1 1 1 ; | | 1200 1000 800 600 400 MAXIMUM TEMPERATURE, IN DEGREES CENTIGRADE AT MIDPOINT IN LEVELING INTERVAL SsUBSIDENCE IN CENTIMETERS FicurE 28.-Rates of uplift and subsidence of stations on the lake compared with maximum temperatures in the lake in nearby drill holes. Maximum temperatures are at the mid- points between times of level measurements. Width and length of bars indicate the uncertainty of the data. had maximum temperatures of 1,000%+ 20°C or more, a value approximately the same as the solidus tempera- ture of the basalt (980°C) determined from petrog- raphic study of drill core from the lake. Uplift, there- fore, took place not only over the fluid core of the lake (T>1,065°C) but also over the relatively rigid, but partly molten part still undergoing crystallization (1,065°C>T>980°). Subsidence was limited to the com- pletely crystallized but still cooling margin. The line between uplift and subsidence corresponded to the outer margin of the 1,000° isothermal envelope, and the level measurements can be used to follow in detail the shrinkage of the lens of fluid and partly molten basalt as discussed on p. 39 and shown in figure 32. "Changes with time in the rate of uplift of stations above the thick central part of the lake during solidifi- cation (figs. 22 and 23) were analyzed by comparing the rate of change of uplift at the main drilling site with rates of uplift calculated on the assumption that uplift was caused by vesiculation of the lava during solidifi- cation. The calculations, which are discussed below, show that the observed uplift can largely but not en- tirely be duplicated on the basis of this assumption. The observed rates of uplift and subsidence at the main drilling site (DH1, 2, 3, 4, 7, 8, and 12) from August 30, 1963, to September 24, 1964, are shown as circles joined by solid line segments in figure 29. Rates are based on the average uplift and subsidence of sta- tions 6 and 26, which span the site, and were plotted at the midpoint of each surveying interval. Points joined by dashed line segments show calculated rates of uplift for the same time intervals based on measured tem- peratures in the drill holes and computed temperatures UPLIFT, IN MILLIMETERS PER DAY SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII -s I - 1 | =] I 1 1 1 100 200 300 400 DAYS AFTER AUGUST 22, 1963 o in sUBSIDENCE, IN MILLIMETERS PER DAY o FicurE 29.-Observed and calculated rates of uplift and sub- sidence at the main drilling site from August 30, 1963, to September 24, 1964. Circles joined by solid lines are ob- served rates, plotted at the midpoints of the surveying intervals. Dots joined by dashed lines are calculated rates, as described in the text, p. 34, 35. from a numerical model for a lake 47 feet (14.3 m) thick with a latent heat of 80 cal/g and a variable diffusivity, as discussed on p. 45. For each surveying interval, the uplift (AL) caused by formation of new crust was calcu- lated by subtracting the subsidence due to crystalliza- tion of the new crust from the uplift caused by simul- taneous vesiculation. In analogy with the relationship between linear and volume thermal expansion (Dane, 1941, p. 29) the net uplift was taken to be one-third the net expansion during solidification. In order to simplify the calculations, subsidence caused by thermal con- traction of cooling crust and fluid lava was neglected (the effect is approximately an order of magnitude less than the observed rates of uplift), and crystallization and vesiculation were assumed to take place at 1,065°C. The calculations are based on the following factors: the observed increase in thickness of the upper crust (AD,) during each leveling period (fig. 11); the computed increase in thickness of the lower crust (AD;,); the porosity (#, and &;) of solidified lava at those depths (table 5), as calculated from the measured bulk density of drill core; the density of 2.79 g/cc for fluid lava in the lake (p,,)), as calculated from leveling COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 35 and other data (p. 35-37); and the grain density (p,) of crystallized Alae basalt at 980°C, as calculated from the grain density at room temperature (3.01 g/cc) and the observed bulk coefficient of thermal expansion (p. 32). Expansion caused by vesiculation is propor- tional to the volume of vesicles and hence the porosity of newly formed crust, and equals $AD. Contraction caused by crystallization is proportional to the differ- ence between the density of fluid lava and the grain density of crystalline basalt, and equals pm—pg) (*) The formulation is as follows: AL = 4[AD,du+ADp$L + (AD, + AD, )C"-°*~ g Figure 29 shows a good correlation between observed and calculated rates from October 1963 (50 days) to July 1964 (300 days), indicating that uplift can indeed be attributed to vesiculation during solidification of new crust. The observed and calculated rates of uplift fell from an initial value of 1 mm/day at 50 days to a rather constant 0.5 mm/day after 100 days because of a progressive decrease in the vesicularity of new upper crust formed at greater depth and a decrease in the rate of solidification. The rates remained relatively constant until about 275 days when the margin of the lens of fluid lava passed beneath the site causing a short burst of rapid uplift. The rate thereafter rapidly decreased to zero as the entire lake beneath the site cooled to below the solidus. The divergence between calculated and observed rates of uplift at 275 days probably results from the simplifying assumption that all crystallization and vesiculation took place at 1,065°C. The divergence during the first month after the eruption shows that uplift from vesiculation at the site during this period was offset by another process- very probably by lateral flowage of lava. Between Au- gust 30 and September 11, subsidence was measured at all but two of the nine stations measured (fig. 214), and the pattern of uplift and subsidence suggests continued drainback of lava toward the vent. The changes mea- sured between September 11 and 26 (fig. 21B), form a more complex pattern and may record not only uplift due to solidification of vesicular crust and changes due to lateral flowage, but also structural warping caused by horizontal expansion of the crust. Because of the correlation in time between a late stage, high rate of uplift at a site and the passage of the outer margin of the lens of fluid lava beneath the site, the leveling data of the full grid of stations can be used to determine the approximate outline of the lens of fluid lava during each successive leveling period. This approach was used in constructing figure 31, which is presented later in this report. The results are not as accurate as the outline of the 1,000° isothermal envelope (fig. 32), be- cause the date of zero uplift at each station can be determined more accurately than the date of maximum uplift. If uplift at each station of the lake was caused solely by expansion of fluid lava upon solidification, as suggested for the main drilling site after late Sep- tember 1963 by the preceding analysis, then the ratio of lake thickness to total uplift at each station should be the same. Indeed, the total uplift at most stations after late September 1963 (fig. 24) was 1.4+0.2 percent of the lake thickness. Exceptions include the stations near the lake margin, all of which had smaller propor- tionate uplifts or even net subsidence. Some of these stations overlay parts of the lake that had largely or completely solidified by the date of the first complete survey. For other stations, uplift was diminished by seepage of the underlying lava toward the vent and central parts of the lake during the first several level- ing periods. Several stations nearer the center of the lake (8, 13, and 29) have proportionate uplifts substan- tially higher than average (1.7-1.8 percent), and sta- tion 7 had proportionately less (1.0 percent). Stations with greater than average uplift are in broad areas part way between the center of the lake and the east- ern and western margins. Inspection of plots of cumulative uplift for each station, such as that for sta- tion 29 in figure 23, reveals that the higher or lower than average uplift at these stations took place be- tween September 1963 and February 1964. The greater than average uplift may have been caused by the seep- age of lava from the lake margins into the fluid core beneath the stations. However, another possible ex- planation for the differences in total uplift is suggested by the measured lateral expansion of the crust over the central part of the lake between September 1963 and July 1964 (fig. 25). The outward thrust of the central crust may have caused flexure of the crust between the expanding center of the lake and the solidified margin, forming anticlinal ridges and troughs. After April 1964, the greater thickness of crust over the central part of the lake and the greater width of solidified mar- gin may have formed a container of solid basalt suffi- ciently rigid to prevent continued outward thrusting and the resulting warping. CALCULATION OF THE DENSITY OF FLUID LAVA IN THE LAKE Assuming that the average uplift of the surface above the lake after late September 1963 was caused solely by expansion of fluid lava during solidification, that is, that positive and negative variations in the uplift due to structural warping and lateral flow cancel 36 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII 0 T T (= i Crust 7963 2}- - A E- Yea \ I |- | un w 14 5 l m ui 5 "C- 14 lava 3 | z, F 17 < | T a Er 8p | a. u | o |- |{ | 10 |- I - | - F5 / & 12- - Crust & 1840 14 1 1 1 1 1 Lava [* 0 200 400 600 800 1000 1200 Sept. 25, 1963 TEMPERATURE, IN DEGREES CENTIGRADE FigurE 30.-Temperature profiles on September 25, 1963, and Au- gust 19, 1965, and a section through the lake on September 25, 1963. The 1963 profile was prepared from temperature measure- ments in drill hole 1 on September 24 and 26, 1963 (solid line), extrapolated to the base of the lake (dashed line) by computations based on thermal modeling. The 1965 profile was measured in drill hole 9. out, the density of fluid lava in the lake in late Sep- tember 1963 can be estimated from the measured changes in altitude of the surface during solidification. On September 25, 1963, the central part of the lake (using average values for all main stations on the grid except those very close to the margin-11, 23, 31, 32, and 34) consisted of 2.2 m of crust, 9.4 m of molten lava of average temperature, 1,130°C, and 1.8 m of crust at the bottom of the lake. These figures are derived from the contour map of figure 1 and temperature profiles in drill hole 1 on September 24 and 26, 1963 (table 3), extrapolated to the base of the lake by computations based on thermal modeling similar to that illustrated in figure 39. The measured and extrapolated tempera- ture profile is shown in figure 30, together with a pro- file in drill hole 9 on August 19, 1965, that was also used in the calculation. The calculations are based on the constancy of mass of the original lava and the derived solidified basalt. As mass equals density times volume, we can write the following equation: z psVs (1) 4 *~ vy. where p equals density, V equals volume, and sub- scripts f and s stand for fluid lava and solidified basalt. The density of the solidified basalt, p,, is known from measurements of drill core. The volume ratio, V,/V;, can be calculated from the measured uplift of the lake surface and the average depth of the lake as deter- mined by photogrammetry and vertical-angle surveys. In order to adjust for the effects of thermal contraction, temperature measurements in drill holes can be used to reduce the rock density measured at room tempera- ture to its value at the solidus (980°C) and to increase the observed uplift by the calculated thermal contrac- tion of the lake during the period involved. In the cal- culations, we envisage a vertical segment of unit width through the lake that is representative of the central part of the lake. The average density at 25°C of basalt solidified from the molten lava present on September 25, 1963-that is, basalt at depths in the lake between 2.2 m and 11.9 m-is 2.665 g/cc (table 5). The density at 980°C can be calculated using the coefficient of volume ther- mal expansion determined by the lake studies; ay equals 3a (Dane, 1941, p. 29), and thus equals 9 x 10 -~$/°C. The formulation is derived from the constancy of mass as follows: P, 25 V25 V. 980 Paso - o Pas Vas V25 (1 + avAT) l 1 Pes . (2) 1+eyAT where AT equals the change in temperature from 25° to 980°C. The ratio V,/V; (eq 1) of unit volumes of solidified basalt and fluid lava can be calculated from the change in thickness of the lake during solidification. In the lake, solidifying lava expanded in all three dimensions, as shown by changes in horizontal distance between stations on the lake surface (see section on "Changes in Horizontal Distance during Solidification"). Accord- ingly, the ratio of volumes, V,/V;, is equal to the cube of the ratio of thicknesses, L,/L;, of solidified and fluid lava. The formulation is as follows: vjV, = WIt;® 3 es Lf+ AL” -+ ALC Lf where L; equals 9.4 m, the average thickness of fluid COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII lava in the lake in late September, 1963; AL,, equals 0.1475, the average recorded uplift of 24 stations on the lake; and AL, equals 0.0233, the calculated decrease in altitude during this period caused by thermal contrac- tion of the unit segment of solidified basalt and fluid lava. This decrease was calculated using the coefficient of linear thermal expansion (3 x 10-8) and includes the following elements: 0.0018 m subsidence due to ther- mal contraction of 9.4 m of fluid lava as it cooled from 1,130°C to 1,065°C; 0.0161 m subsidence from thermal contraction as it continued to cool from 1,065°C to the average value of 495°C on August 19, 1965; 0.0039 m of subsidence due to thermal contraction of the original 2.2 m of upper crust as it cooled an average of 590°C during this period; and 0.0015 m of subsidence due to the inferred 280°C of cooling of the original 1.8 m of lower crust. When these values are used, the ratio V,/V; is 1.0555, and the calculated density of fluid lava is 2.79+0.03 g/cc. A fairly large uncertainty is indicated because of possible errors in several of the figures used in the calculation. The measured average density may be greater than the true density because of relatively poor core recovery from more vesicular layers in the crust. The measured average cumulative uplift may include an unknown contribution from structural warping or lateral flow. The value for the average depth of the lake may differ from the true value by as much as 0.3 m because of errors in the vertical angle measurements from the rim of the crater. The value of 2.79 g/cc reported here is 0.01 g/cc greater than the value reported earlier (Peck, 1969) from preliminary calculations. The calculated value for the density of fluid lava can be compared with values for the density of basaltic melt derived from the measured densities of basaltic glass or calculated from a chemical analysis of the glass. Fluid lava at 1,130°C in Alae lava lake contained an estimated 15 percent crystals, consisting of 11 per- cent olivine and pyroxene and 4 percent plagioclase (Wright and Peck, 1978). The content of gas bubbles is not known but is inferred to be % of 1 percent or less. Degassing of the initially oversaturated lava had largely finished by late September 1963 (p. 21-23 and figure 19). Vesiculation caused by crystallization over the temperature interval 1,140° to 1,130°C should have yielded no more than % of 1 percent gas bubbles, if vesiculation was proportional to crystallization, as il- lustrated in figure 20. As the mass of fluid lava (m;) equals the mass of the melt fraction (m,,) plus crystals (m,.) and gas bubbles (m,), the density of the melt frac- tion (p,,) can be calculated from the densities of fluid lava (9p), crystals (p,), and gas bubbles (Pg) from the following relationship: 37 mp = My +me +My va = Pav m *LeY¥ z +ngg Pm ~ Pfo—Pch _ngg Vm When one considers a unit volume of fluid lava, the velumes of melt, crystals, and gas bubbles (¥ ;,, V;. and V,) reduce to decimal proportions. The density of the exsolved gas is essentially zero, so that the term PV's can be dropped (the content of gas bubbles enters the calculations by reducing the decimal proportion of melt). The densities of the crystals at 1,130°C are 3.2 g/cc for olivine and pyroxene and 2.65 g/cc for plagio- clase, judging from their densities at room temperature as determined by mineral separations (Wright and Peck, 1978) and published coefficients of volumetric thermal expansion (Skinner, 1966, p. 87). When one assumes a zero content of gas bubbles, the calculated density of the melt fraction is 2.74 g/cc. With % of 1 percent gas bubbles the calculated density of melt is 2.76 g/cc. The measured density at room temperature of glass separated from two samples of pumice formed during the Alae eruption at 1,140°C and 1,160°C is 2.76 g/cc (Wright and Peck, 1978). Vesicle-and crystal-free glass from differentiated melt collected at a temper- ature of 1,100°C from drill hole 69-1 in Makaopuhi lava lake has a density at room temperature of 2.78 g/cc. (T. L. Wright and N. L. Hickling, oral commun., 1975). If these are reduced to the values at 1,130°C using the coefficient of volumetric thermal expansion of 9 x 10 ~S%°C, the densities of the melt fraction are 2.73 and 2.75 g/cc; these densities are in good agree- ment with the density of the melt fraction calculated on the assumption that the proportion of exsolved gas is negligible. The density of glass from Alae pumice can also be calculated from a chemical analysis of the pumice using the method of Bottinga and Weill (1970), in which density is calculated from the partial molar volumes of the oxide components. Molar volumes at 1,130°C were determined by linear extrapolations of the values listed for greater temperatures (1,250- 1,600°C, in Bottinga and Weill, 1970, table 5 and p. 175-176). The density for the melt at 1,130°C calcu- lated by this method is 2.704 g/cc, about 1 percent less than the values derived from the density measure- ments in the laboratory or calculated from the lava- lake data. THE PATTERN OF COOLING IN THREE DIMENSIONS Some aspects of the three-dimensional pattern of so- lidification and further cooling of Alae lava lake were revealed by repeated temperature measurements in drill holes in different parts of the lake and by repeated 38 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII 0 50 L 1 oad Contours in days after August 22, 1963 100 METERS FigcurE 31.-Approximate position of the outer margins of the 1,080°C isothermal surface in the lake as a function of time in days after August 22, 1963. The date when the margin passed beneath drill holes 3, 4, and 8 (marked by a dot in fig. 31) was calculated by extrapolation of a plot of maximum temperatures versus time (fig. 14), and the position of the margin was estimated elsewhere by determining the time interval during which each surveying station underwent a late-stage, high-rate of uplift (fig. 23 and table 6). surveys of the altitude of the grid of stations on the lake surface. In the following section, the position with time of the outer margin of the 1,080°C isothermal sur- face (approximately the fluid-crust interface) and the 1,000° isothermal surface (approximately the solidus) are shown in plan and those of the 1,000°C and 700°C surfaces are shown in cross section. The measured and inferred positions of the isothermal surfaces agree closely with the positions computed on the basis of two-dimensional numerical models, which are illus- trated in figure 40. The approximate position in plan of the outer margin of the 1,080°C isothermal surface is shown as a func- tion of days after August 22, 1963, in figure 31; the position is based on sparse temperature data from drill holes and on the surveying interval or intervals when each surveying station underwent a late stage, high rate of uplift. As discussed on p. 35, this phase of uplift at a station records the migration through the lake beneath the station of the outer margin of the lens of fluid lava and reflects the high rate of crystallization and vesiculation between 1,080°C and 1,050°C. Figure 31 clearly shows the effect on the rate of solidification of the thinning of the lake over preeruption ridges be- neath the north and southeast parts of the lake, as well as the rather rapid solidification of the lens of fluid lava COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 39 0 50 METERS h ncn l_ ] Values at drill sites (solid dots) and on contours are in days after August 22, 1963 32.-Outer margins of the 1,000°C isothermal surface (approximately the solidus) in Alae lava lake as a function of days after August 22, 1963. The maximum lateral extent of the surface on different dates is based on precise level measurements of stations on the surface of the lake (fig. 22 and table 6) and on temperature measurements in drill holes (solid dots) that penetrated to the depths of the maximum temper- ature at the time of complete solidification. after 250 days. The outer margin of the 1,080°C isothermal surface after 300 days outlines fairly closely that part of the lake that was 14 m or more thick. The position of the 1,000°C isothermal surface, ap- proximately the solidus, is shown in plan and longitu- dinal section in figures 32 and 33. The location of the surface was determined from interpolation and ex- trapolation of temperature data from drill holes 5, 6, 8, 9, and 12 and from the time at which each surveying station started to subside. The change from uplift to subsidence at a station reflected the completion of crys- tallization and vesiculation in the lake beneath that station (p. 33, 34 and fig. 28). Figure 32 shows, as does figure 31, the effect of prelake highs beneath the north and southeast parts of the lake and the rapid comple- tion of crystallization during the final 100 days. The position of the 1,000°C isothermal surface is known more precisely than the 1,080°C surface because more temperature data are available and because the time of the beginning of subsidence at each station can be ac- curately determined by plotting rates of change in al- titude versus time for each station and interpolating the intersection with the zero rate axis. Figure 33 shows the position of the isothermal surface in sec- tions, the lower with the vertical scale equal to the horizontal scale and the upper with a 10x exaggeration. In the lower part of the lake, the position 40 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII FEET| METERS 2580 - 2 |-785 - Z o z 2560- |- fi - 780 > O O < X w = L- X C “a; |- 775 % 8 2540- |- im - R < -770 - 2520- L - 800 f u 2600 E S 120 : x seo 240 <> > < 2500- a 750 s] o 50 100 METERS L 1 1 1 I Values in days after August 22, 1963 FiGur® 33.-Longitudinal cross section of Alae lava lake showing the positions of the 1,000°C isothermal surface as a func- tion of days after August 22, 1963. Isotherms dashed were extrapolated. Upper section, 10% vertical exaggeration; lower section, horizontal and vertical scales are equal. Numbered vertical lines are projected drill holes. of the isothermal surface (dashed) is largely inferred. Near the end of crystallization, the lens of partly mol- ten lava is at a depth 60 percent of the distance from the upper surface and within the thickest part of the lake. Figures 32 and 33 show that final crystallization took place more rapidly in the thinner southeastern part of the lake than in the thicker, spatter-covered northwestern part; the figures are in good agreement with the computed position of the 1,000°C isotherm in a generalized radial section of the lake after 300 days (fig. 40). The depth of the 700°C isothermal surface in both the upper and lower parts of the lake (fig. 34) is known from several drill holes in both the marginal (DH5 and 6) and central (DH1, 3, 4, 7, 8, 9, and 12) parts of the lake. The position through time of the outer margin, however, is rather poorly known at distances from drill holes, because the data from repeated surveys of the altitude of stations on the lake surface are of little help in determining its location. The only relevant data from the altitude surveys is the position of the axis of maximum subsidence after solidification of the lake; the position puts an outer bound on the margin of the surface during two surveying periods. As in figure 33, the section shows the effect of lake thickness on the position of the isotherms-the last part to cool to 700°C was in the thickest part of the lake and at a depth 75 percent of the lake thickness below the surface. Com- puted positions of the 700° isotherm at 300 and 600 days (fig. 40) are in close agreement. THERMAL PROPERTIES OF ALAE BASALT Several thermal properties of Alae basalt were mea- sured in the laboratory or were calculated from labora- tory data on the constituent minerals. COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 41 1 and 12 6 5 §and4 78nd? NORTHWEST FEET | METERS 2580 - |- 785 LJ _ o |- 780 pm i- i— - < 2550 - -775 |- 770 2520 - 700° C ISOTHERM 0 L 50 100 METERS 1 1 I Values in days after August 22, 1963 FiGurE 34.-Longitudinal cross section of Alae lava lake showing the position of the 700°C isothermal surface as a function of days after August 22, 1963 (10% vertical exaggeration). Isotherms dashed were extrapolated. The position was deter- mined from temperature profiles in drill holes, which are shown projected into the line of section as numbered vertical dashed lines. The thermal conductivity at ambient temperatures of Alae basalt can be determined from the graphs and formulation of Robertson and Peck (1974, figs. 7 and 8, eq 12), derived from detailed conductivity measure- ments at 35°C of vesicular basalt from Hawaii. The 61 samples studied included 4 from the crust of Alae lava lake (Robertson and Peck, 1974 or 1969, nos. 613, 618, 619, and 620). Conductivity was found to vary with porosity, olivine content, and the nature of the pore fluid (air or water). The conductivity of basalt samples which have the same pore fluid and similar olivine con- tents varies inversely with porosity, forming a linear plot against (1 - porosity)?. The conductivity (K) in air of Alae basalts can be calculated from the following equation (Robertson and Peck, 1974, eq 12, p. 4884): VK = a + b (WK, + qVK,) where a and b are experimentally determined con- stants, y = solidity (that is, 1 - porosity), & = porosity, K, = conductivity of fully solid basalt, and K,, = con- ductivity of air. When one uses the experimentally de- termined values of 6.1 x 10 ~* cal/em sec °C for K,, 0.063 x cal/em see °C for K,, 0.2 for a, and 0.7 for b, the equation reduces to: K = (1.929-1.554q¢). The conductivity of Alae and other Hawaiian basalts which have 0-5 percent olivine, as calculated from this equation, is shown as a function of porosity in figure 35. Alae basalt with 11 percent porosity, which forms the bulk of the lake, has a conductivity of 3.09 x 10~ cal/ecm sec °C, but the conductivity values for more ve- sicular basalt decrease to 1.71 x 10° cal/em sec °C for the basalt with 40 percent porosity that forms the thin surficial crust of the lake. An appreciable increase in conductivity with tem- perature for solidified, vesicular Alae basalt is shown by laboratory measurements reported in this section and also by the results of thermal modeling discussed later (p. 45-47). This variation differs from that of pre- viously published laboratory studies on the variation with temperature of the conductivity of basalt, appar- ently because of the effect of abundant vesicles. As summarized by Nafe and Drake (1968) and Robertson ip bo o SECOND DEGREE CELSIUS THERMAL CONDUCTIVITY, IN MILLICALORIES PER CENTIMETER o 0.6 POROSITY 0.4 0.2 0 aw 0 Ficur® 35.-Thermal conductivity at 35°C of air-saturated Alae basalt. Curve calculated using the formulation and experimentally determined constants of Robertson and Peck (1974, p. 4884), as discussed in the text. (unpub. data, 1976), most laboratory measurements of the variation with temperature of the conductivity of basalt indicate that conductivity values do not vary greatly with increasing temperature, although on the average they show a slight increase. Apparently the decrease in conductivity with increasing temperature of pyroxene and olivine is approximately counter- balanced by the increase in conductivity of plagioclase (Birch and Clark, 1940). Murase and McBirney (1973), however, found that the conductivities of Columbia River basalt and alkali olivine basalt from the Galapagos decrease about 50 percent from room tem- perature to 1,100°C. At greater temperatures conduc- tivities increased sharply because of increased photon conductivity (Murase and McBirney, 1973, p. 3584). The studies summarized above were made on sam- ples which have little or no porosity. In order to deter- mine the effects of abundant vesicles on the variation SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII of conductivity with temperature, Kaoru Kawada of the Earthquake Research Institute, Tokyo University, kindly measured the conductivity of four samples from Alae crater (written commun., 1966). Two of the sam- ples (A-5-1 and A-11-1) were coarsely vesicular drill core from near the surface of the 1963 lava lake, one sample (DPH-28) was a coarsely vesicular hand sam- ple from an open joint crack near the main drilling site, and the fourth sample (AM-51) was a finely vesicular hand sample from the petrographically similar prehis- toric Alae lava lake, collected on the cliff face of the mezzanine 9.8 m below the surface of the lake. Conduc- tivities were measured on cylindrical samples 2.5 cm in diameter and 5.0 cm in length in a furnace under controlled temperature. The temperature was in- creased in small increments by means of a heating filament inserted in a 0.25 cm diameter hole along the axis of each sample, and temperatures were recorded by thermocouples in two 0.18 em diameter holes drilled 0.3 and 1.2 cm from the axis, half the length of each sample parallel to the axis and in the same plane. The experimental apparatus and techniques are described by Kawada (1966). The results are listed in table 7 and are plotted in figure 36 as ratios of the measured conductivities at high temperatures to the conductivities at room tem- perature. For three of the samples, conductivities in- creased nearly linearly with increasing temperature to the highest temperatures at which conductivities were measured (673%950°C). The greatest rate of increase of conductivity with temperature, 0.09 percent/°C, was shown by the highly vesicular samples A-5-1 and A-11-1, which had bulk porosities of 28 and 25 percent respectively. A smaller increase, 0.03 percent/°C, was shown by the denser sample AM-51, which had a porosity of 7 percent. The fourth sample, DPH-28, which had a porosity of 24 percent, showed a nearly TaBLE 7.-Thermal conductivity of four samples of basalt from Alae crater as a function of temperature [Data from K. Kawada (written commun., 1966] AM-51° DPH-28¢ i 1 -A1-1* Temperature Cofidfgtlivit Temperature Cofiductivity Temperature Conductivit Temperature Conductivity (°C) (cal/em see °C) (°C) (cal/em see °C) (CC) (cal/em see °C) (°C) (cal/em see °C) 2B [elsa 1.19 x 107% 46 2.21 x 10" 32 2.07%: 107 40 2.86 x 10 T0 si 127 x10" ° 94 2.90 x 10°" 54 2.06 x 10% 74 2.17 18 LAr mn ias 1.23 x 10° 194 2.59% 10°" 62 2.05 x 10~* 149 2.170 x 10"~* 128 LSC 1.28'x 10% 430 2:01 x 10°" 123 1.99 x 10~ 185 2.36 x 10 % 199 A e- i 1.27 x10 ~* 551 B27 x 10" 274 2.17 x A0" 268 2.23 x 10" PIG rec lol e 1.52 x 10~" 699 3.64 x 10% 363 224 % 10 * 313 2.40 x 10 ~} 506: _-_» 1.66 x 10% 804 8.83 x 10 * 450 2895 x 10 ~* 370 2.21% 10% 489 _: SCEE elveul 1:70 x 10°% 946 4.05 x107* 451 2.25 x 10 ~* 495 2 11 x 10"% bar c: cgl 1.72 x 10% bes o Hea ms iu bane .C 528 227 x 10° s tan enc esas 65J Aen. 1.93 x 10% aa ain in os 673 2.49 x 10 ~ yeaa ao asan y TBA eved a eaid es 2.00 x 10~ eil pea aer ike o .o sad cosmic. 212 x 10~" t an n Sn rea aas fa t e eee er anes i ian ae aug s 950 ~ rescued 2.24.x 10°* n iol an «ye T a ates iri pani Drill core collected at a depth of 0-0.1 m in drill hole 5 in the 1963 Alae lava lake. Measured bulk density, 2.16 g/em®. Calculated porosity, 0.28. *Drill core collected at a depth of 0-0.15 m in drill hole 11 in the 1963 Alae lava lake. Measured bulk density, 2.25 g/em*. Calculated porosity, "Hand sample collected from the cliff face of the prehistoric lava lake 16 m below the mezzanine of Alae crater. Measured bulk density, 2.88] 0.25. g/em®. Calculated porosity, 0.07. "Hand sample collected from an open joint crack in the 1963 Alae lava lake near drill hole 1. Measured bulk density, 2.28 g/em*. Calculated porosity, 0.24. COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 2.0 T T T T T T T T T MLC 4 ~s in Tamu CcOoNDUCTIVITIES [=] RATIO OF THERMAL £. COE T Ct 0.5 | | 1 1 1 1 1 1 1 200 400 600 800 TEMPERATURE, IN DEGREES CENTIGRADE 1000 0 Ficure 36.-The ratio of conductivity at high temperature to con- ductivity at room temperature of four samples of basalt from Alae crater. Data from K. Kawada (written commun., 1966). Samples A-5-1 (circles), A-11-1 (dots), and DPH-28 (hexagons) are from the August 1963 lava lake; sample AM-51 (squares) was collected from a prehistoric lake exposed in the wall of the crater. The dashed line indicates ratios calculated on the basis of an increase in conductivity of 0.06 percent/°C, as deduced from thermal model- ing (p. 45). linear decrease in conductivity, 0.05 percent/°C, to 495°C, the highest temperature measured. Kawada suggested that the decrease in conductivity of this sample may record the temperature effect for dense, nonporous basalt. Although the sample was highly ve- sicular, the vesicles were large compared with the sample size, and none may have been situated along the axial plane containing the heating element and sensing thermocouples. Low porosity in the axial plane is suggested by the relatively high conductivity of 2.86 x 10~° cal/em see °C measured at room temperature. A disparity between bulk porosity of the samples and the porosity in the axial planes is also suggested for two other samples, A-5-1 and AM-51, by a comparison of the relatively low conductivities measured at room temperature in these experiments with the measure- ments of Robertson and Peck (1974) and the resulting formulation on page 41. The marked increase in con- ductivity with temperature of three of the samples was confirmed by H. R. Shaw (written commun., 1975) by laboratory measurements of the diffusivity of vesicular basalt from Makaopuhi laval lake as a function of temperature, but is in contrast with the small changes with temperature found by other workers on nonporous samples. The difference may be due to radiative heat transfer across the vesicles, convective heat transfer by gases in the vesicles, or progressive closing of micro- fractures as the result of thermal expansion of the samples during heating. In situ thermal conductivity in the Alae lava lake may have been greater at higher temperatures because of the effect of an additional process, a decrease in con- 43 ductivity with falling temperature in the cooling lava caused by successive formation of microfractures re- sulting from thermally induced stresses. The presence of relatively abundant (0.3 percent) microfractures in the crust of the lake is suggested by the disparity be- tween the measured bulk coefficient of thermal expan- sion in the lake and coefficients determined in the lab- oratory on basalt and diabase of similar composition (p. 32, 33). Microfractures across grains and along grain boundaries are also evident in thin sections under the microscope. The effect of microfractures on rock conductivity has been demonstrated in the labora- tory by measurements of conductivity with increasing pressure (Walsh and Decker, 1966) and was suggested as a possible factor in the conductivity of vesicular Hawaiian basalt by Naoyuki Fujii and George C. Ken- nedy (oral commun., 1974) after a review of the conduc- tivity data. Robertson and Peck (1974) found that the intrinsic conductivity of nonporous basalt, as measured on finely crushed samples or calculated from mineral conductivities and the mode, was 40 percent greater than the effective conductivity determined from labo- ratory measurements on discs. They suggested that the difference was due to microvesicles and microfractures formed during initial cooling of the parent lava. The thermal modeling of the temperature data in Alae lava lake described in the section "Thermal Mod- eling" also indicates that the conductivity of solidified basalt in the lake increased with temperature. The re- sults do not define the form of the relationship or the relative importance of radiative heat transfer, convec- tive heat transfer, or microfracturing. The results do indicate that conductivity (K) at a given temperature (T, in °C) can be related approximately to the conduc- tivity at room temperature (K,,) for temperatures less than 1,000°C by the linear relationship K x EKpj(1+0.000BT), where B equals 0.0006+0.0001. Val- ues calculated by this function are represented by the dashed line in figure 36. The conductivity of Alae basalt at temperatures of more than 1,000°C is not known. Thermal modeling using greatly different val- ues for the conductivity of solid and molten basalt re- sulted in very small differences in temperatures, less than the accuracy of the field measurements. Labora- tory studies by Murase and McBirney (1973) suggest that the conductivity of basalt at temperatures be- tween the solidus and the liquidus is less than the con- ductivity at lower temperatures. The heat capacity of Alae basalt at six temperatures from 300° to 1,400°K was calculated by R. A. Robie (written commun., 1964) using the modal mineral composition (p. 3) and experimentally determined values of the specific heats of the minerals later pub- lished by Robie and Waldbaum (1968). The results are 44 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII TABLE 8.-Specific heats of minerals and calculated heat capacity of Alae basalt [Data from R. A. Robie (written commun., 1964)] Heat Tempoeéature Specific heat cal/g* capacity Fayalite Fosterite Diopside Albite Anorthite Magnetite .iin;énite > gasralt 8 800 .s: il ece enan- aes 0.1564 0.2011 0.1742 0.1876 0.1789 0.1569 0.1572 0.1783 500 :::. .ccs 1890 2535 2331 2389 2319 1890 1854 2318 £00 .:: scc ev oda te rels 2108 2822 2617 .2688 2614 2610 2017 2608 1000 -. J re 2218 2949 2781 .2826 2747 2072 2092 2724 1200 eZ ALLE .- 2320 3060 2825 3029 2864 2072 2165 2849 __________________________ 2911 .3066 2974 2223 2418 .3165 tabulated in table 8. Values range from 0.1783 cal/g °C at 300°K to 0.2927 cal/g °C at 1,400°K. A smooth curve for heat capacity (c) as a function of temperature (fig. 37) was constructed on the basis of the calculated heat capacities and fitted to the following function of temp- erature (T in Kelvins) by R. A. Robie (written com- mun., 1975): c = 0.2356 + 4.3635 x 10 °T - 63443; 10° Thermal diffusivity in the cooling Alae lava lake un- doubtedly varied with time and depth. Diffusivity (D) is related to conductivity (K), density (p), and heat capacity (c) by the following relationship: D= Kipe Conductivity varies with vesicularity as well as tem- 0.30 T T T T T U 0.28 |- a & u Y 9.26 |- A ~ 2 o 3 - w 3 o T H - 0241 < -G > EA o C- s 0.22} o <4 C O 0 < 1 2. 0:20.(- - °C 0.18 - A 0.16 | | L I | 0 200 400 600 800 1000 - 1200 TEMPERATURE, IN DEGREES CENTIGRADE FicurE 37.-Calculated heat capacity of Alae basalt as a function of temperature. 2072 2927 perature, as discussed above, and hence with depth in the lake and probably with time at a given depth dur- ing solidification. The same is true of density. Both the heat capacity and conductivity vary with temperature, hence they vary also with time and depth. Increased conductivity with decreased porosity and increased temperature, however, is at least partly offset by the increased density of less porous basalt and the in- creased heat capacity at greater temperature. Changes in diffusivity of solidified basalt with time and depth in the lava lake are illustrated in figure 38; the figure shows computed diffusivities for solidified basalt on 2 days during solidification (Sept. 6 and Dec. 30, 1963) and 3 days thereafter. Diffusivity values were com- puted using a numerical formulation (p. 45-47) for a lake 49 feet (14.9 m) thick having a latent heat of 80 cal/g, a conductivity that increased 0.06 percent/°C and varied with porosity as described on page 41, a variable heat capacity described on pages 43, 44 and a density structure in the lake as tabulated in table 5. A low- conductivity vesicular cell at shallow depth was not included. Calculated diffusivities increased downward in the upper 3-5 m of the crust because of the increase in density and temperature with depth. On December 30, 1963, for example, the calculated diffusivity in- creased from 0.0049 cm*/sec at 0.15 m depth to 0.0064 cm/sec at a depth of 3.8 m. The average diffusivity of the upper crust (average of values at 0.3 m intervals) increased from 0.0055 cm*?/sec on September 6, 1963, to 0.0058 on September 3, 1964. The computed average diffusivity of the lake after complete crystallization in late September 1964 decreased from a maximum value of 0.0060 cm*2/sec on December 8, 1964, to 0.0058 em*/ sec in August 1965 and May 1966, but thereafter in- creased to 0.0060 on August 31, 1967, as temperatures continued to fall. The latent heat of Alae basalt can be deduced from numerical analysis of the thermal data (p. 45, Peck and others, 1977), and can also be calculated using esti- mated and experimentally determined heats of fusion of the component minerals. Unfortunately, the reliabil- ity of the calculated value is limited because of uncer- tainty in the values of heats of fusion of the major COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII DEPTH, IN METERS 10 |- 12 F- 1 1 4 5 6 2 DIFFUSIVITY, IN SQUARE CENTIMETERS PER SECOND x 108 Figure 38.-Calculated diffusivity of solidified Alae basalt (<1,000°C) as a function of depth and time. Curves are based on computed values at 0.3 m intervals, as described in the text. constituents augite and labradorite. Calculations were made using the modal data on page 3 and the norma- tive composition of the 6.75 percent residual silicic glass (Wright and Weiblan, 1968). Values for heats of fusion were taken from Robie and Waldbaum (1968) and Bradley (1962), and were supplemented by a value of 22.4 kceal/gfwt for augite Wo.,En., Fs., estimated by D. R. Wones (written commun., 1964, 1975) and two rather different estimates for anorthite-29.4 kceal/ gfwt (Wones) and 18.7 keal/gfwt (Yoder, 1975). When these values are used, the calculated latent heat of completely (100 percent) crystalline Alae basalt would be 94 cal/g or 85 cal/g, depending on the value used for anorthite. The calculated latent heat that would be re- leased during crystallization in the lake, however, is substantially less, 77 or 67 cal/g, because of the crystal- lization prior to the eruption of 13 percent olivine, 45 pyroxene, and plagioclase, and the retention of nearly 7 percent silicic glassy residuum at the end of crystalli- zation. The calculated values are in good agreement with the value of 80+10 cal/g deduced from numerical analysis of the temperature data (p. 45). THERMAL MODELING Temperature profiles in Alae lava lake were analyzed (Peck, Hamilton, and Shaw, 1977) using the one- and two-dimensional computer programs for cooling lava lakes devised by Shaw, Hamilton, and Peck (1977). The effort was an extension of a prelimi- nary analysis conducted with J. C. Jaeger (written commun., 1967) that was summarized by Peck (1974). Computed temperature profiles were compared with representative profiles measured in the central part of the lake on 12 separate days from September 6, 1963, to August 31, 1967. The effect of rainfall on tempera- tures in the lake was computed by withdrawing 620 calories from the 100°C isotherm for each centimeter of measured rainfall. The latent heat was distributed in three temperature intervals in accordance with modal data on crystal abundance within partly molten lava near the base of the crust-that is, 30 percent from 1,140° to 1,080°C, 41 percent from 1,080° to 1,050°C, and 29 percent from 1,050° to 980°C. The upper surface was held at 70°C for the first 5 months and at 40°C thereafter, and temperatures in the lake were reduced by conductive cooling to a surface at a temperature of 20°C at a depth 100 feet (30.5 m) below the bottom of the lake. A cell length of 1 foot (0.3 m) and a time interval of 3.237 x 104 seconds 0.385 day) were selected for the computations. During the initial runs using one-dimensional mod- els, thermal and physical properties of basalt in the lake were assumed to be constant, and a lake thickness of 48 feet (14.6 m) was chosen to simulate the deeper, central part of the lake. Ten measured profiles during the first 2 years of cooling were reproduced with aver- age deviations of 20°C using a constant diffusivity of 0.006 ecm*/sec and a latent heat of 90 cal/g. A better match, particularly of temperature profiles in the upper crust during initial cooling and at the base of the lake after complete solidification, was obtained using numerical models incorporating variable thermal and physical properties, as well as slightly different values for lake thickness and latent heat. The following mod- ifications were made to the initial model. 1. A variable diffusivity was included by assign- ing values for density and conductivity at room temperature to each 1-foot (0.3 m) cell in the lake; this density was based on the measured density (table 5) and the relation- ship between conductivity and porosity de- 46 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII 0 5 |- a g. 10 |- s 16 - Contact fl, A f } /f ff 20 |- - 0 I 360 f 6(|)0 l é 960 ' 1200 TEMPERATURE, IN DEGREES CENTIGRADE FicuUrE 39.-Computed (solid) and observed (symbols) temperature profiles in Alae lava lake. Temperature data are marked in table 3 and the dates indicated in figure 10. The dates for the profiles and the drill holes in which the temperatures were measured are as follows: 0, Sept. 6, 1963, DH 1; 1, Nov. 5 and 8, 1963, DH 3; 2, Dec. 30, 1963, DH 4; 3, Mar. 24, 1964, DH 4; 4, July 2, 1964, DH 8; 5, Sept. 3, 1964, DH 8; 6, Dec. 8, 1964, DH 9; 7, Jan. 21, 1965, DH 9; 8, Mar. 2, 1965, DH 9; 9, Aug. 19, 1965, DH 9; 10, May 27, 1966, DH 12; 11, Aug. 31, 1967, DH 12. COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 47 scribed on p. 41. Heat capacity, conductivity, and diffusivity were calculated for each cell by an iterative process, using the relation- ship between heat capacity and temperature given on page 43, and a linear increase of 0.06 percent/°C of conductivity with increas- ing temperature. 2. Heat loss by rainfall was computed using the measured density structure of the lake. 3. The lake thickness was increased to 49 feet (14.9 m), a better approximation of the deepest part of the lake near drill hole 9, and the latent heat for the basalt was decreased accordingly to 80 cal/g. 4. Low-conductivity and density values were as- signed to one of the cells to simulate the highly vesicular and cavernous zone found at shallow depth in the lake. The computed profiles for a lake 49 feet (14.9 m) thick and measured profiles in Alae lava lake are shown in figure 39. The measured profiles are based on tempera- tures measured at depth intervals of 0.3 m or less for the first four profiles and 0.9 m or less for the last eight. Temperatures read from smoothed plots of the profiles are shown by symbols at 0.3 intervals on the figure, and the dates of the profiles are indicated in figure 10. The last two profiles (10 and 11), which were measured in drill hole 12 on May 27, 1966, and August 31, 1967, are not strictly comparable with the computed profiles or with the other measured profiles. The lake was an estimated 1% m shallower at this site, and the upper part of the hole was heated by gases rising along a fracture intersected in the drilling. However, this was the only drill hole open through the lake during the last 2 years. Using the numerical model described above and illustrated in figure 39, average deviations (at 0.3 m intervals) between computed and measured temperatures for the 10 representative profiles during the first 2 years are less than 2°C. Measured tempera- tures for the first three profiles are matched with aver- age deviations of one-halfC. Measured and computed maximum temperatures after solidification (profiles 6 to 9) and basal contact temperatures for profiles 7 to 9 are matched within 2°C. The computed contact temp- erature for profile 6 is 27°C less than observed; this difference may reflect errors in the measurement of the values for depths or temperatures near the base of the lake for that profile. The poorest match of computed and measured temperatures is for profile 4, for which computed temperatures on the average are 55°C great- er than observed. The profile was measured in drill hole 8 only 2 weeks after the hole was drilled, and temperatures in the upper part of the hole were proba- bly still strongly depressed because of the coolant water used in the drilling. Computed and measured temperatures August 31, 1967 (profile 11), 4 years after formation of the lake, agree with an average de- viation of 0.5°C, and the computed and measured temperatures (at the base of the lake) differ by only 1°C. The value of the observed maximum of 86°C suggests that the lake cooled to less than 100°C several months before the date of the measurement. In the computer model the lake cooled to 100°C in mid-April 1967, 4% months before the final temperature mea- surements. Computed temperatures in the lake using a "no-rain" model are as much as 366°C greater than observed on August 31, 1967, and the lake takes a total of 19.5 years to cool to 100°C. Cooling at the margins of Alae lava lake was analyzed using a two-dimensional numerical model formulated by Shaw, Hamilton, and Peck (1977). The results are presented by Peck, Hamilton, and Shaw (1977) and summarized below. A generalized radial section of the southern quarter of the lake was divided into cells 15 m long and 1.5 m thick (Shaw and others, 1977, fig. 6), and the model cooled in 150 time steps of 8.318 x 105sec (9.63 days) duration. A constant value of diffusivity of 0.006 em*/sec and a latent heat of 80 cal/g were used in the calculations, and heat loss from rain was taken into account using a constant value for the average density of the lake. The results are illustrated in figure 9 of Peck, Hamilton, and Shaw (1977), which shows the positions of several isotherms (including the 1,080°, 1,000°, and 700°) after 300 days, and in figure 40, which shows the positions of the 1,000° isotherm after 120, 240, and 360 days. The com- puted temperature distribution in cross section, which is in close agreement with temperatures measured in drill holes and inferred from level surveys (figs 31-34), indicates that temperatures in the central part of the lake were not influenced by cooling at the margins. Throughout the first year, temperatures in the compu- ter model were decreased more than 1°C only in the outer 15 m of the lake, and in the fourth year only in the outer 45 m. The variation of 90 days in the dates of complete solidification in the central part of the lake (figs. 32 and 33) can be attributed to lateral variations of about 2 m in the depth of the lake. Heat transfer to the margins, however, may have been a more signifi- cant factor in the cooling of Alae lava lake than in the model because of factors that were not incorporated in the model. This conclusion is suggested by the exis- tence of a belt of maximum subsidence in the lake de- tected by repeated level surveys during the first year after solidification (fig. 26). The belt, which presumedly represents the part of the lake undergoing the most rapid cooling and thermal contraction, does not man- ifest itself in the two-dimension thermal model. Possi- 48 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII 3 METERS 0 30 METERS C_ Contours in days after August 22, 1963 FicurE 40.-Computed position of the 1,000°C isotherm in a generalized radial section of Alae lava lake after 120, 240, and 360 days. Computations based on a two-dimensional model having a latent heat of 80 cal/g and a constant dif- fusivity of 0.006 cm*/sec. Solid boundary shows the margins of the lake in a generalized section based on measured sur- face profiles and the topographic map of the crater (fig. 1). Dashed boundary shows the margin of cells in the simplified model used in the numerical calculations (Shaw and others, 1977, fig. 6B). bly, cooling in the lake itself was accelerated because of shallow circulation in the margin of the lake of water, less than 100°C, derived from rain that fell on the lake margin, bordering talus, and crater walls. CONCLUSIONS The 15-m-thick Alae lava lake solidified in 10% months and cooled to less than 100°C 3 years later. Study of the lake, combined with study of Makaopuhi lava lake (Wright and others, 1968; Wright and Okamura, 1977) and Kilauea Iki lava lake (Richter and Moore, 1966) provide a unique body of data on the cooling and crystallization of Kilauean tholeiite. Alae was the smallest of the three lakes, cooled in the least time, and showed the simplest pattern of solidification and cooling. Major conclusions of this study are as fol- lows: 1. Solidification of Alae lava lake took place by the slow increase in thickness of upper and lower crusts with time (fig. 10). Growth of the upper crust during the first 3 months was linear with respect to the square root of time; the thickness (in meters) was equal to 0.00132V'? - 0.18, where t equals time in seconds after the eruption. 2. Crystallization of tholeiitic basalt in the lake took place over a temperature interval of 160°C from the emplacement temperature of 1,140°C to the solidus temperature of 980°C. The interface be- tween fluid lava and solid crust was at 1,065°C. Measured temperatures at the base of the lake were as much as 700+10°C, 60 percent of the ini- tial temperature of the lake (table 4). 3. The lava vesiculated as it cooled and crystallized because of the exsolution of gases, raising the upper surface of the lake (fig. 22) and forming porous basalt that decreased in porosity (and in- creased in density) downward from the surface (fig. 16, table 5). Initially, gas was exsolved be- cause of supersaturation of the newly erupted lava; later vesiculation was caused by crystalli- zation and resulted in minute, angular pores (fig. 18). Repeated leveling surveys combined with temperature measurements in the drill holes permit delineation of the approximate outline in plan of the lens of fluid lava (fig. 31) and the outline in plan and section of the fluid lava and partly molten crust (figs. 32, 33). Comparison of these figures with computer runs based on two- dimensional thermal modeling (fig. 40) indicates that solidification rates in the central part of the lake were not affected by cooling at the lake margin. 4. The lake continued to cool for 3 years after solidifi- cation. Thermal contraction of the cooling basalt caused subsidence of the lake surface (fig. 26). The bulk linear coefficient of thermal expansion (3+ 1x10-8/°C) determined by the lake's studies is markedly less than values measured in the laboratory on similar rocks, presumedly because of microfracturing caused by differential thermal contraction. 5. Measured temperatures in the cooling lake can be matched within the limits of error of the mea- surements and of the lateral variability of tem- perature in the lake by computations based on numerical models (fig. 39), indicating that the lake cooled largely by conductive heat transfer. Vaporization of the abundant rainfall, however, strongly affected temperatures in the upper crust during solidification and throughout the lake thereafter. Near joint cracks, temperatures were affected by downward circulation of cool air from the upper surface and by upward flow of hot gases from the interior of the lake. 6. Thermal properties of the Alae basalt measured in the laboratory or calculated from laboratory data are in good agreement with values deduced from the thermal modeling. Values for the properties are as follows: COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII (a) Conductivity (K, in cal/em see °C) increased with decreasing porosity (¢) and with increasing temperature (T in °C), and can be represented by the following expression; K = (1.929-1.554¢)%(1+0.00067) (b) Heat capacity (c) can be represented by the fol- lowing function of temperature (T, in Kelvins): e = 0.28356 +4.3635 x 1057 - 6.3440 x 10°/T* (c) Thermal diffusivity varied with time and depth in the lake from about 0.005 to 0.006 em*/sec with changing conductivity, density, and heat capacity. (d) The latent heat for 80 percent crystallization of the basalt was 80+10 cal/g. The results of this investigation are pertinent to in- terpretations of the cooling histories of other basaltic lakes and flows, but they should be applied with cau- tion. Such factors as mineralogical composition and vesicularity strongly affect thermal properties. Differ- ent modes of formation of the lake or flow may result in different initial temperature distributions. Tempera- tures in the upper part of Makaopuhi lava lake, for example, were decreased markedly by repeated crustal foundering during drain back of lava into the vent at the end of the eruption. Separate lakes and flows are subjected to different amounts of rainfall. Further- more, the cooling effect of rain may be markedly less in thicker lakes, where a significant fraction of the va- porized rainwater condenses at shallow levels in the crust rather than escaping to the atmosphere. REFERENCES CITED Ault, W. U., Eaton, J. P., and Richter, D. H., 1961, Lava tempera- tures in the 1959 Kilauea eruption and cooling lake: Geol. Soc. America Bull., v. 72, no. 5, p. 791-794. Ault, W. U., Richter, D. H., and Stewart, D. B., 1962, A temperature measurement probe into the melt of the Kilauea Iki lava lake in Hawaii: Jour. Geophys. Research, v. 67, no. 7, p. 2809-2812. Birch, Francis, and Clark, Harry, 1940, The thermal conductivity of rocks and its dependence upon temperature and composition: Am. Jour. Sci., v. 238, no. 8, p. 529-558, no. 9, p. 613-635. Bottinga, Y., and Weill, D. F., 1970, Densities of liquid silicate sys- tems calculated from partial molar volumes of oxide compo- nents: Am. Jour. Sci., v. 269, no. 2, p. 169-182. Bradley, R. S., 1962, Thermodynamic calculations on phase equilib- ria involving fused salts, Part 2, Solid solutions and applications to the olivines: Am. Jour. Sci., v. 260, no. 7, p. 550-554. Carslaw, H. S., and Jaeger, 1959, Conduction of Heat in Solids: Lon- don, Oxford at the Clarendon Press, 510 p. Daly, R. A., Manger, G. E., and Clark, S. P., Jr., 1966, Density of rocks, in Clark, S P., Jr., ed., Handbook of Physical Constants (rev. ed.): Geol. Soc. America Memoir 97, p. 19-26. Dane, E. B., Jr., 1941, Density at high pressure; thermal expansion, in Birch, Francis, Schairer, J. R., and Spicer, H. C., Handbook of Physical Constants: Geol. Soc. America Spec. Paper 36, p. 27-37. Decker, RW., and Peck, D. L., 1967, Infrared radiation from Alae lava lake, Hawaii, in Geological Survey research 1967: U. S. Geol. Survey Prof. Paper 575-D, p. D169-D175. 49 Finlayson, J. B., Barnes, I. L., and Naughton, J. J., 1968, Develop- ments in volcanic gas research in Hawaii, in The crust and upper mantle of the Pacific area: Am. Geophys. Union Geophys. Mon. 12, p. 428-438. Kawada, Kaoru, 1966, Studies of the thermal state of the earth. The 17th Paper: Variation of thermal conductivity of rocks, Part 2: Bull. Earthquake Research Inst., Tokyo Univ., v. 44, p. 1-21. Murase, Tsutomu, and McBirney, A. R., 1973, Properties of some common igneous rocks and their melts at high temperatures: Geol. Soc. America Bull., v. 84, p. 3563-3592. Nafe, J. E., and Drake, C. L., 1968, Physical properties of rocks of basaltic composition, in Hess, H. H., and Poldervaart, Arie, eds.; Basalts, v. 2: New York and London, Interscience Pub., p. 483- 502. Peck, D. L., 1969, Density of molten lava in Alae lava lake, Hawaii [abs.]: EOS, v. 50, no. 4, p. 339. 1974, Thermal properties of basaltic magma; results and prac- tical experience from study of Hawaiian lava lakes-summary, in Colp, J. L., and Furumoto, A. S., eds., The Utilization of Vol- cano Energy: Sandia Laboratories, Albuquerque, p. 287-295. Peck, D. L., Hamilton, M. S., and Shaw, H. R., 1977, Numerical analysis of lava lake cooling models; Part II, Application to Alae lava lake, Hawaii: Am. Jour. Sci., v. 277, p. 415-437. Peck, D. L., and Kinoshita, W. T., 1976, The eruption of August 1963 and the formation of Alae lava lake, Hawaii: U. S. Geol. Survey Prof. Paper 935-A, p. A1-A33. Peck, D. L., and Minakami, Takeshi, 1968, The formation of colum- nar joints in the upper part of Kilauean lava lakes, Hawaii: Geol. Soc. America Bull., v. 79, no. 9, p. 1151-1165. Peck, D. L., Moore, J. G., and Kojima, George, 1964, Temperatures in the crust and melt of Alae lava lake, Hawaii, after the August 1963 eruption of Kilauea volcano-a preliminary report, in Geological Survey research 1964: U. S. Geol. Survey Prof. Paper 501-D, p. D1-D7. Peck, D. L., Wright, T. L., and Moore, J. G., 1966, Crystallization of tholeiitic basalt in Alae lava lake, Hawaii: Bull. Volcanol., v. 29, p. 629-655. Rawson, D. E., 1960, Drilling into molten lava in the Kilauea Iki volcanic crater, Hawaii: Nature, v. 188, p. 930-931. Richter, D. H., and Moore, J. G., 1966, Petrology of the Kilauea Iki lava lake, Hawaii: U. S. Geol. Survey Prof. Paper 537-B, 26 p. Richter, Dorothy, and Simmons, Gene, 1974, Thermal expansion be- havior of igneous rocks: Int. Jour. Rock Mechanics and Mining Science, v. 11, p. 403-411. Robertson, E. C., and Peck, D. L., 1974, Thermal conductivity of vesicular basalt from Hawaii: Jour. Geophys. Research, v. 79, no. 32, p. 4875-4888. Robie, R. A., and Waldbaum, D. R., 1968, Thermodynamic properties of minerals and related substances at 298.15°K (25.0°C) and one atmosphere (1.013 bars) pressure and at higher temperatures: U. S. Geol. Survey Bull. 1259, p. 1-256. Shaw, H. R., 1969, Rheology of basalt in the melting range: Jour. Petrology, v. 10, no. 3, p. 511-535. Shaw, H. R., Hamilton, M. S., and Peck, D. L., 1977, Numerical analysis of lava lake cooling models; Part I, Description of the model: Am. Jour. Sci., v. 277, p. 384-414. Shaw, H. R., Wright, T. L., Peck, D. L., and Okamura, R., 1968, The viscosity of basaltic magma: An analysis of field measurements in Makaopuhi lava lake, Hawaii: Am. Jour. Sci., v. 266, no. 4, p. 225-264. Skinner, B. J., 1966, Thermal expansion, in Clark, S.P., Jr., ed., Handbook of Physical Constants (rev. ed.); Geol. Soc. America Mem. 97, p. 75-96. Stearns, H. T., and Macdonald, G. A., 1946, Geology and ground- water resources of the Island of Hawaii: Hawaii Div. Hydrog- 50 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII raphy, Bull. 9, 363 p. Swanson, D. A., Duffield, W. A., Jackson, D. B., and Peterson, D. W., 1973, The complex filling of Alae crater, Kilauea volcano, Hawaii: Bull. Volcanol., v. 36 (1972) no. 1, p. 105-126. Swanson, D. A., Jackson, D. B., Duffield, W.A., and Peterson, D. W., 1971, Mauna Ulu eruption, Kilauea Volcano: Geotimes, v. 16, no. 5, p. 12-16. Swanson, D.A., and Peterson, D. W., 1972, Partial draining and crus- tal subsidence of Alae lava lake, Kilauea Volcano, Hawaii, in Geological Survey research 1972: U. S. Geol. Survey Prof. Paper 800-C, p. C1-C14. Walsh, J. B., and Decker, E. R., 1966, Effect of pressure and saturat- ing fluid on the thermal conductivity of compact rock: Jour. Geophys. Research, v. 71, no. 12, p. 3053-3061. Wright, T. L., Kinoshita, W. T., and Peck, D. L., 1968, March 1965 eruption of Kilauea volcano and the formation of Makaopuhi lava lake: Jour. Geophys. Research, v. 73, no. 10, p. 3181-3205. Wright, T. L., and Okamura, R. T., 1977, Cooling and crystallization of tholeiitic basalt, 1965, Makaopuhi lava lake, Hawaii: U.S. Geol. Survey Prof. Paper 1004, 78 p. Wright, T. L., and Peck, D. L., 1978, Crystallization and differentia- tion of the Alae magma: U.S. Geol. Survey Prof. Paper 935-C (in press). Wright, T. L., Peck, D. L., and Shaw, H. R., 1976, Kilauean lava lakes: natural laboratories for study of cooling, crystallization, and differentiation of basaltic magma: Am. Geophys. Union Geophys. Mon. 19, p. 375-390. Wright, T. L., and Weiblen, P. W., 1968, Mineral composition and paragenesis in tholeiitic basalt from Makaopuhi lava lake, Hawaii [abs.]: Geol. Soc. America, Spec. Paper 115, p. 242. Yoder, H. S., Jr., 1975, Heat of melting of simple systems related to basalts and eclogites: Carnegie Inst. Wash. Year Book 74, p. 515-519. TABLE 3 52 TABLE 3.-Temperatures in Alae lava lake [Temperatures measured with thermocouples of chromel-alumel unless noted otherwise. Profiles used in constructing fig. 10 are marked by an asterisk (*). Square root of time in days after 06h Aug. 22, 1963, in parentheses] Drill hole ; Depth? Temperature No. Time' (g) ds August 30, 1963 (2.87) DH1__ - 10h58m-11h46m -__- 0. 351.60 272 496 668 817 925 954 ppprer. ~Naonouo September 3, 1963 (3.49) 489 623 748 846 868 DH1__ _ 10h31m-10h58m -__- ~ o ao 1: 1: 2. 2. 2. September 4, 1963 (3.63) 264 439 578 709 822 914 993 5 1,022 - 9h50m-11h20m _--- go £o 5D bo po i- I-! & i September 6, 1963 (3.90) 274 440 569 691 794 882 951 1960 #071 *DHL__ 10n29m-llhl1m _--- co go 60 £0 bo DD i- 1 O i oor September 9, 1963 (4.27) 246 397 528 643 742 826 898 967 991 1097 1263 4425 4564 1656 4681 DH1__ - 11n25m-13h38m -__- or ppprh. miki ffr fo mo mo h b o t o o o t See footnotes at end of table. SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII TABLE 3.-Temperatures in Alae lava lake-Continued DrllNloljlde Time! Dig)“ Temxégcrinure September 11, 1963 (4.50) *DH1__ 10n30m-12h27m -__- 0.0 67 .5 235 1.0 392 1.5 517 2.0 626 2.5 717 3.0 799 3.5 877 4.0 942 4.25 965 4.25 1968 DH2:.: .5 1266 1.0 4414 1.5 4537 2.0 1637 2.15 1662 September 17, 1963 (5.12) *DH1__ 9n39m-10h47m __-- 1.0 1339 2.0 4581 3.0 "736 3.5 4806 4.0 4870 4.2 4899 September 24, 1963 (5.76) *DH1__ 10h05m-12h06m -__- 0.5 4200 1.0 4312 1.5 417 2.0 4511 2.5 4594 3.0 1668 3.5 "734 4.0 4795 4.5 4854 5.0 1907 5.4 1939 September 26, 1963 (5.93) DH1__ 10h05m-11h49m -__- .5 1167 1.0 1295 1.5 403 2.0 503 2.5 4584 3.0 1657 3.5 4724 4.0 4784 4.5 4839 5.0 4891 5.4 4923 October 1, 1963 (6.30) *DH1L__. On35m-1O0h4§m ___. 0.9 1262 1.4 4357 1.9 4456 2.4 4541 2.9 1614 3.4 1680 3.9 4740 4.4 1798 4.9 1848 5.4 4889 COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 58 3.-Temperatures in Alae lava lake-Continued TaBLs® 3.-Temperatures in Alae lava lake-Continued Drilllloljmle Time! Dig}? Temgoeé‘?ture Drirlllofmle Time! flat)“ Temlégérure October 2, 1963 (6.42) October 24, 1963 (7.95) DH3__ 11h31m-138hl14m ___ - 7.5 906 DH3._ O9n55m-10h22m ___ _- 0.8 255 8.0 985 1.3 336 8.5 1,026 1.8 409 9.0 1,056 2.3 481 9.5 1,050 3.3 602 1{hSBm-....i...... 7.5 926 3.8 655 8.0 1,000 4.3 7OL 8.5 1,037 4.8 745 9.0 1,063 5.8 824 9.5 1,053 6.3 865 7.5 955 6.8 903 8.0 1,018 7.8 939 8.5 1,047 DH1I-. 10h40m-..__....... 3.4 589 9.0 1,069 3.9 639 9.5 1,058 4.4 691 7.55 983 4.9 740 8.05 1,031 DH3__ - 15h00m-16h04m _____ 10.0 41,043 8.55 1,056 11.0 11,077 9.05 1,074 11.5 141,079 9.55 1,062 12.5 41,108 12H22m_............ 7.6 1,002 13.0 41,113 8.1 1,039 13.5 41,117 8.6 1,063 14.0 41,121 9.1 1,078 ione o nna, sm. i...... oe for November 5, 1963 (8.68) 8.1 1,044 +> =- 8.6 1,067 *DH1_.' 15h00m-._._...__._ 1.4 291 g}; 1,079 2.4 427 P 1,066 y 534 7.6 1,016 ii 637 8.1 1,046 5.4 726 3113 {833 *DH3__ 13h46m-l4hl16m ___ - 1.0 185 k F 2.0 365 9.6 1,065 3.0 509 +t 1,022 4.0 616 8.2 1,051 5.0 710 8.7 1,070 5.8 778 9.2 1,078 6.8 855 9.7 1,064 7.8 925 7.9 1,033 8.8 987 8.4 1,058 9.8 1,035 8.9 1,075 | 3 : 9.4 1,076 9.9 1,066 November 7, 1963 (8.79) Tokl2m.c:ll......s 8.1 1,043 > s ies 8.6 1,065 ..."... 9.2 596 9.1 1,075 10 2 688 9.6 1,071 2 823 10.1 1,075 (22 192 19HI14m............ Z BA 1,045 129 1.050 8.6 1,066 | : w 1'079 9.2 634 36 1070 10.2 718 10.1 1.084 11.2 882 f > 12.2 1,011 13.2 1,053 9.4 680 October 17, 1963 (7.50) 10.4 747 11.4 874 *But.. _... os 141 (2s Pose 1.0 239 ¢ : 1.3 2 9.5 708 20 Te 10.5 776 35 192 11.5 893 $3 r 12.5 1,044 > bor 13.5 1,045 4 aos 12hBBbm-........... 9.5 724 13 Mos 10.5 796 52 Fos 11.5 908 + 12.5 1,049 18.5 1,047 See footnotes at end of table. 54 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII TABLE 3.-Temperatures in Alae lava lake-Continued TABLE 3.-Temperatures in Alae lava lake-Continued Drilllolfiole Time! Dfigfihz Temfieéfiture Drilllohde Time! Dist)“ TemEleETture DH3-___ 183n00m.-_______---- 13:2 g‘l‘g November 8, 1963 (8.85) 11.5 924 = 12.5 1,053 DHS.. : 1.0 94 13.5 1,048 2.0 97 13hO5m :._.:_;-.. = 9.5 754 3.0 94 10.5 832 4.0 269 11.5 934 5.0 436 12.5 1,057 6.0 593 13.5 1,051 7.0 741 1Sh10m:.::,;.: ...s. 9.5 765 8.0 858 10.5 845 8.9 944 11.5 948 9.0 944 12.5 1,059 9.4 979 13.5 1,052 9.9 1,008 J8kibm..._ .. 9.5 776 10.0 1,005 10.5 858 10.4 1,034 11.5 958 10.8 1,053 12.5 1,061 10.9 1,055 13.5 1,054 11.4 1,071 9.5 784 11.8 1,085 10.5 866 11.9 1,086 11.5 964 12.4 1,095 12.5 1,063 12.8 1,1t2? 13.5 1,056 12.9 1,099, 1,100 A3h26m....:: :\ 9.5 796 13.4 1,104 10.5 879 13.8 1,110 11.5 974 13.9 1,114; 1,115; 1,115 12.5 1,065 14.8 1,123 13.5 1,057 14.9 1,124; 1,126; 1,125 9.5 807 15.9 1,129; 1,130; 1,131 10.5 887 16.9 1,134; 1,132; 1,132 11.5 981 17.9 1,136; 1,185 12.5 1,069 13.5 1,062 ¥ a * males 18h81m.-_________---- g? 11823 November 20, 1963 (9.50) IShBBm :: 12.5 1,070 f oy s 13.5 1,062 *DH3.__ - 10nl6m-10h25m ___ - 1.1 171 A3hB4m:.....}.l.00. 9.5 811 2.6 408 10.5 894 4.1 550 11.5 986 5.6 694 12.5 1,071 7.1 803 13.5 1,063 8.6 908 12.5 1,071 10.1 983 13.5 1,063 9.5 817 f £ f b> Ts 332 November 27, 1963 (9.86) 12.5 1,072 has § tes aa > 13.5 1,064 DH4.;. 11.3 810 ..it" 12.5 1,073 123 897 13.5 1,064 13.3 948 ..." 2. 9.5 824 14.3 974 10.5 907 15.3 702 11.5 999 T4h4fim............ 11.8 827 12.5 1,074 12.3 918 13.5 1,065 13.3 969 ..is}: 12.5 1,074 14.3 997 13.5 1,065 15.3 760 11.55 1,001 12.55 1,075 £ 'a 18.55 1,065 November 28, 1963 (9.91) 13rd im :.:}: olo 12.55 1,075 3 wean 2 3 i325 izggg DH4__ 10h42m-10h57m -__- 6.0 481 13.6 1,067 7.0 586 Tih49m...!...:.} 12.6 1,078 8.0 693 13.6 1,067 9.0 797 12.7 1,080 10.0 886 13.7 1,067 11.2 981 1ji8im 13.0 1,081 12.2 1,035 14.0 1,064 13.2 1,065 14.2 1,068 15.2 1,058 See footnotes at end of table. COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 55 TaBLs 3.-Temperatures in Alae lava lake-Continued TABLE 3.-Temperatures in Alae lava lake-Continued Drfilfole Time! Difighz Temffcr‘Tture DrlNllofiole Time! Digghz Temgfé‘zaxture December 2, 1963 (10.11) she o 120 197% 15.3 1,079 DH4__ - 12n25m-13h30m -__- 3.0 328 15.8 1,087 3.5 377 16.3 1,089 4.0 425 16.8 1,094 4.5 470 17.3 1,095 5.0 514 17.8 1,098 5.5 559 17.0 11,096; 1,086 6.0 612, 598 o 2 December 18, 1963 (10.79) 7.5 740 ar Puri r r coc." - homo. espr 8.0 781 *DH4-_ lin2in-l2h21m .... 0.0 67 8.5 820 1.0 172 9.0 849, 853 2.0 289 9.5 884 3.0 394 10.0 917 4.0 484 10.5 946 5.0 567 11.0 974 6.0 643 11.5 1,001 7.0 713 12.0 1,013; 1,022 8.0 779 12.5 1,033 9.0 837 13.0 1,048 10.0 904 13.5 1,059 11.0 950 14.0 1,066 12.0 995 14.5 1,071 13.0 1,031 15.0 1,062; 1,075 18.5 1,049 15.5 1,071 13.85 1,062 16.0 1,079 14.0 1,058 16.5 1,088 14.5 1,069 17.0 1,094 14.85 1,078 17.5 1,099 15.5 1,083 18.0 1,100 15.85 1,091 16.5 1,084 16.85 1,101 December 5, 1963 (10.26) 175 1.100 # 17.85 1,106 DH4__ _ 10h12m-13h22m ____ 1.0 127 nema n nels pne e Horna cea oen eae £8 jig? December 30, 1963 (11.41) 4.0 150 ..: is . Meer noe cee tries. 5.0 1546 *DH4._ 10nl6m-13h25m ___ _- 0.0 - 258 6.0 +635 10 : 196 7.0 1715 2:0 ~ 302 8.0 +793 8.0 _ 895 9.0 +864 40 ' 177 10.0 1926 5.0 - 547 11.0 4981 6.0 639 12.0 11,010 7.0 700 13.0 41,048 8.0 758 14.0 11,069 9.0 812 15.0 11,086 10.0 861 16.0 41,102 11.0 _ 926, 934 - ~ 11.75 968 12.0 970, 980, 975 December 16, 1963 (10.78) 12.25 990 nya" 12.5 - 999 DH4__ _ 10n45m-15h28m ____ 1.0 158 12.75 1,005 2.0 284 13.0 1,005; 1,018; 1,015; 1,009 3.0 399 13.25 1,024; 1,025 4.0 493 13.5 1,031; 1,032; 1,036 5.0 573 13.75 1,036; 1,043 6.0 653 13.85 1,042 7.0 724 14.0 1,042; 1,048; 1,044; 1,043 8.0 794 14.25 1,053; 1,050 9.0 857 14.5 1,057; 1,056; 1,060 10.0 911 14.75 1,062; 1,064 11.0 953 14.85 1,064 12.0 995 15.0 1,063; 1,063; 1,064; 1,063 13.0 1,034 15.25 1,065; 1,068 13.3 1,045 15.5 1,069; 1,074; 1,073 13.8 1,060 15.175 - 1,079; 1.077 14.0 1,061 15.85 1,076 14.3 1,064 16.0 _ 1,083; 1,084 See footnotes at end of table. 56 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII TABLE 3.-Temperatures in Alae lava lake-Continued Drill hole No. ; Depth*® Temperature Ts. to (°C) December 30, 1963 (11.41)-Continued *DH4__ _ 10n16m-13h25m-__- 1,085; 1,086 1,088; 1,089; 1,089 1,091 1,092 1,097 17.25 1,098 17.5 1,100; 1,100 17.175 1,101 17.85 1,103 45 128 217 292 365 434 526 594 655 713 766 842 887 928 963 993 1,021 1,032 1,038 1,045 1,049 1,052 1,050 1,050 1,044 1,041 January 29, 1964 (12.70) 16.25 16.5 16.175 16.85 17.0 DH5__ 14h09m-15hl14m -__- 0 [= t go go 1 ~1 g» @» on Ou a Co ho i © tD 0D ~1 @) C0 bo i- C INQ bk ot heut heut Jok foot fut Jout foud foot fest fut jet fd 0 w w o co 1 1 on on on *DH4__ _ 12h11m-14h48m -__- 95 161 272 374, 362 453, 436 533, 528 602, 596 683, 670, 660 743, 728 765 796, 781 818 846 867 891 912 953 b tor Ap: tort woowocowetweccec0c0c0c0c0 | nt bo got t dt bet bet to bo bo - - & o w w go ~1 February 17, 1964 (13.43) 100, 97 101, 97 101, 97 441 563, 555 641, 633 711, 704 767, 760 817, 809 853, 844 879, 877 DH5__ _ 10n35m-13h25m -__- bored ket bt bet on or to bo © (D 1 n & G0 o uo uou ou oo | | I | | | | I | | | | | See footnotes at end of; table. Drill hole No. DH5__ _ 10h35m-13h25m-__~ *DH4__ _ 14h05m-15h28m -__- March 24, 1964 (14.71) DH5__ _ 10n51m-12h07m -__- *DH4__ - 12n35m-13h1l5m -__- to bo 1 @ @ c D go ~1 &) Ot i t ho i- bibbbobo0c0000000 bok bok jet kes pa D RQ bo RD RY RD -t - ket fut ked ked pa pea ro - go 1 Sue No ~ to go on on on to ho 0 tD 1 @ in to i- in 0 in o in o in o in 0 0 0 & & & ip in ip in ip in ip an io n 0 in o in © tt 0 &n O n 0 o w w 0 to 1 1 ® & ouk co ies jes *DH4__ - 12h42m-13h23m -__- mmpktobr s s0000000 TABLE 3.-Temperatures in Alae lava lake-Continued April 13, 1964 (15.37) Temperature °C) COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII TaBLs 3.-Temperatures in Alae lava lake-Continued 3.-Temperatures in Alae lava lake-Continued ' § 2 Dri 4 2 r DrllNlo]?0]e Time! Digh Temg’eé‘jnure rgllobole Time! Dig?! Temyfcinure *DH4_- 12h42m-13h23m.__- 8.0 451 9.0 517 June 17, 1964 (17.36) 10.0 578 oue. $ G 199 DHS.! fenibm tintem __. _ 00 252 . tl, 6.0 331 3.0 236 9.0 492 6.0 356 12.0 608 9.0 456 15.0 669 12.0 513 16.3 725 16.4 578 22.3 729 22.4 557 28. 598 a ra y fg prt ananas aln ae et rain oo July 2, 1964 (17.79) April 16, 1964 (15.47) -- oes. omage -- | *DHS-. - 11h00m-12n56m .-.. 1.5 96 DH6__ _ 12n28m-14h17m ____ 1.5 111 3.0 99 3.0 153 4.5 110 4.5 111 6.0 231 6.0 177 7.5 322 7.5 307 9.0 407 9.0 428 10.5 483 10.5 530 12.0 556 12.0 638 13.5 621 18.5 703 15.0 690, 706 15.0 808 16.3 765 16.5 897 18.0 843 18.0 959 19.3 894 19.5 988; 1,004 21.0 956 20.5 1,023 22.3 996 21.0 1,015 24.0 1,032 22.5 1,023; 1,032 25.3 1,047 283.5 1,039 27.0 1,049 24.0 1,028 28.3 1,044 25.5 1,019; 1,027 3 26.5 1,023 27.0 1.007 July 27, 1964 (18.48) 28.5 982, 989 29.5 978 *DHS-_ - ___ 0.0 43 30.0 955 3.0 102 31.5 926 6.0 256 32.5 905 9.0 431, 436 e r casas ar 12.0 573, 580 May 22, 1964 (16.59) 15.0 698, 710 ermine alek ihe d. o r inn Ln 2. Asa: 18.0 827 bte... 0.0 139 : sl Tsp 1.45 96 4.45 97 September 3, 1964 (19.49) 3.45 97 10.45 554 % hor as DHS.. . 0.0 53 17.25 908 3.0 112 20.25 996 15 174 23.25 1,046 6.0 279 26.25 1,044 715 377 29.25 1,005 f eres. a 10.5 513 13.5 632 June 9, 1964 (17.13) 16.5 742 f 19.5 834 *DH7__ 13hl15m-15h10m -__- 0.0 "37 .8 78 3.3 161 y December 8, 1964 (21.81) 6.3 274 7.8 367 DH12 10h05m-10h57m ___ 0.0 236 9.3 423 3.0 278 9.35 449 6.0 414 10.8 523 9.0 544 12.35 609 12.0 639 13.8 686 15.0 716 15.35 761 18.0 T74 16.8 828 21.0 828 18.35 889 24.0 863 21.35 981 27.0 871 See footnotes at end of table. 58 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII TABLE 3.-Temperatures in Alae lava lake-Continued TABLE 3.-Temperatures in Alae lava lake-Continued DriNlloyole Time! Dig‘ghg Temileéiature Drillllobole Time! Dfigfihz Tem;(>°ecrz)ature December 8, 1964 (21.81)-Continued DH11 3.8 242 is . 145 DH12._ 10h05m-10n57m-__- _ 830.0 867 6.0 250 33.0 861 9.0 355 36.0 822 12.0 473 39.0 719 15.0 579 42.0 730 DHS5__ - 15n15m-15h40m -__- .0 $30 45.0 690 3.0 159 *DH9__ 12h07m-13h12m ____ .0 830 6.0 231 3.0 152 9.0 289 6.0 291 12.0 322 9.0 420 15.0 326 12.0 537 16.3 321 15.0 644 19.3 297 21.0 800 22.3 264 24.0 852 25.3 229 27.0 888 28.3 192 30.0 902 33.0 897 > P me paar a e ere oni 32-2 222 March 2, 1965 (23.65) 41.8 814 44.8 776 *DH9__ 10n20m-11h50m ---- 0.0 "31 47.8 718 3.0 154 DHSG._ 15hl2m-15h87m _-_ - 8.0 221 6.0 231 6.0 318 9.0 316 9.0 369 12.0 393 12.0 391 15.0 525 15.0 384 21.0 701 16.3 377 24.0 761 19.3 347 25.5 788 22.3 310 27.0 805 25.3 270 28.5 818 28.3 229 30.0 828 81.5 832 33.0 833 January 21, 1965 (22.79) 34.5 831 bait -_... . 00 250 179 $2 3.0 221 39'0 806 6.0 351 f 9.0 479 42.0 773 12'0 594 45.0 733 15'0 682 48.0 676 y DH12 12h45m-14h15m -__- .0 334 18.0 749 3.0 237 21.0 801 6.0 367 24.0 833 9'0 457 27.0 848 12'0 557 30.0 845 . 34.7 820 15.0 640 37'7 783 18.0 708 40'7 735 21.0 764 43'7 673 22.5 782 46-7 610 24.0 798 / Meppm anion. .l _ 0 "40 are 3.0 131 F 6.0 214 28.5 817 9'0 283 30.0 815 12'0 441 31.5 811 15'0 562 34.5 792 18.0 670 34.7 791 21'0 751 37.7 756 24'0 804 40.7 709 27'0 843 43.7 655 30'0 860 46.7 598 3 DHS5__ - 15n30m-15h55m -__- .0 $22 35.7 855 3.0 136 38.7 829 6-0 209 41.7 791 9'0 256 44.7 745 12'0 281 47.7 692 15'0 280 DH10 .0 340 15.8 274 3.0 96 a 6.0 295 18.8 250 9'0 403 21.8 220 12'0 552 24.8 189 15'0 660 27.8 187 See footnotes at end of table. COOLING AND VESICULATION OF ALAE LAVA LAKE, HAWAII 59 TABLE 3.-Temperatures in Alae lava lake-Continued TaBus 3.-Temperatures in Alae lava lake-Continued Drgllcfxole Time! Dips)“ TemyoeéTture Drlllllowle Time! Dffigh’ Temffé?ture *DH12.. 'm-14h05m.. _. 14. 1 August 19, 1965 (27.01) $409 17g 3358) - -- ei 20.8 355 * 23.8 377 DH9. _ 10n05m-12h20m ...- _ 6.0 125 aas 9; 12.0 249 pyar rk 15.0 339 as os 18.0 406 a s n 21.0 474 rs bof 24.0 538 ano 171 27.0 596 rs % 30.0 637 dua ? $1.5 658 22 12. 33.0 664 we. 7, 34.5 679 cs A 35.9 684 ers 122 37.5 687 as (y 38.9 686 pi 404 40.5 682 as Nap 41.9 675 129 451 43.5 666 46.8 448 44.9 653 *DH9__ - 14n20m-15h30m ___. .6 62 47.9 616 2.6 20 pus. 30 51 6.6 40 6.0 56 3 u s 7a 11] 68 12.0 88 12.6 110 15.0 94 15.6 121 15.8 94 18.6 116 18% as 21.6 164 21.8 85 24.6 . 165 24.8 75 (Blocked by sublimates at 24.6 ft) 27.8 66 s. g *DH12 - 14h40m-16h12m -__- 2.8 $23 August 31, 1967 (38.35) 9.0 319 & 12.0 390 DHQ.-- ~. s.: 9:00 cs lt.}. 14.6 74 15.0 456 17.6 74 20.0 559 20.6 74 23.0 606 23.6 72 24.5 629 26.6 70 26.0 642 *DHI12 . .ll 5.0 63 27.5 656 8.0 63 29.0 663 11.0 65 30.5 670 14.0 65 32.0 669 17.0 68 33.5 671 20.0 68 35.0 669 23.0 68 36.5 658 26.0 68 38.0 654 29.0 68 41.0 626 32.0 76 44.0 585 35.1 76 47.0 543 38.1 80 namie atts tan cf o ol 41.1 84 May 27, 1966 (31.78) 171 # a x. DH12 11h00m-14h05m -__- 22 $23 'All times are given in hours and minutes, Hawaii Standard Time. 8.8 273 21 foot equals 0.3048 m. 1 1' 8 295 Measured with a mercury thermometer. C % f Measured with a thermocouple of platinum-platinum plus 10 percent rhodium. x U.S. GOVERNMENT PRINTING OFFICE: 1978 - 789-107/65 9-II § ain tR" [View * afm sup r die s P A io. uP A s Crystallization and Differentiation of the Alae Magma, Alae Lava Lake, Hawai By THOMAS L. WRIGHT and DALLAS L. PECK $0 LLPDIE ELC A L1O0ONX OF AL AE LAVA CAKE, HAW AJ I GEO L Q G 1ICAL_S URV E YP ROFESSIONAL PAPER 95 5-C An account of the crystallization history and chemical differentiation of tholeiitic basalt from one of Kilauea's lava lakes UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1978 UNITED STATES DEPARTMENT OF THE INTERIOR CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY H. William Menard, Director Library of Congress Cataloging in Publication Data Wright, Thomas Llewellyn. f Crystallization and differentiation of the Alae magma, Alae lava lake, Hawaii. (Solidification of Alae lava lake, Hawaii) (Geological Survey professional paper; 935-C) Bibliography: p. 1. Lava-Hawaii-Alae lava lake. 2. Phase rule and equilibrium. I. Peck, Dallas Lynn, 1929- joint author. II. Title. III. Series. IV. Series: United States. Geological Survey. Professional paper; 985-C. QE461.W84 552'.2 T7-17801 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 Stock Number CONTENTS Page Abstract e erie me ae wore me or mn an e tel le pe mi se m Wee e on mal ae ae wee C1 2 2... . 21. belo 2 nel wn ann onlien niin s ull ak - a ola n alee ae an oo eminem mene bo a a alee o. 1 Acknowledements 2 3422.00 > o as n ae e ee a cs o o aire a a a a aoe fee ail ie n a a a mnm hee onc tie a acm iew s 2 Methods of study oe oe een e are gi a ane dow ie maas be a io 2 Drilling f a ma fei i at ar mei rl anld is n mie al in o - afe d Temperature. measurements "22 . A2 gucci toon col neenee ede ean ain n a ena ane 4 Laboratory. SHUUIGS .. >. 2%. .. o. c aL cn o, ule ». - -- io - a on a at s o he is io in ie fae i Hai Pace Rl ie mae oo 4 Petrology of 'Alac )_._~ .. .. LEC UXL ue an ona nn do cwakne ans ren beatae. s 7 Bulk chemical Composition: -2L.._LL___: - G2 _ 4 Chemical MOGES _... .. . c cus c uunns ds o on oe in andin mee = 11 - . el acad enses 13 Disc USSIOM S 22. 2... ce cn ane l Hie ue e cls a hice d ame i's aie ie ma ad mig ie a fie in it a it iene i o mn ar we al ae aan e alee ie aoe a ea ad 16 Mineral PHTASENCSIS -.. .. 1. . 2... - m 2a ue oa le ule o ld s id he cae r an toe anais ie ae a in an aa aad i a hae a se 16 Rates of crystallization of minerals -_- 17 Variation of liquid composition during crystallization ._______________________ 18 Natural differentiates of Kilauean tholelite 18 SUMMARY 7% 322. .. SLL. , 2 u whe i a ar +. - mile ninae i ie ale mld io a a hie + oo in ine apa aree a Rie cane aii a n meni igen 19 References: Cited" - ust .. ch aus one b en ea rane ane. Hire o acme s roel a an nari a n thoi ack s 19 ILLUSTRATIONS Page FIGURE 1. Index map showing topographic contours and location of drill holes __ C3 2-4. Graphs showing: 2. Modal data for Alae lava 14 3. Weight percent glass plotted against temperature _________ 15 4. Liquid line of descent 17 TABLES Page TABLE 1. ' Record of drill holes in- Alae lava lake C2 2+. Drill cores from Alag IAVA ARE 2. .-. ne none neon nein nan nanan mines ak nee s 5 3. Chemical analyses of undifferentiated and differentiated samples, Alae lava lake L. 2. . os .co rrr ll PLT LCE Sil n Laue cen 8 4. Chemical analyses of mineral and glass separates, Alae lava lake .. 10 b.> Composition of residual glass, Alae lava 10 6. - Computer-calculated chemical modes, Alae lava lake __________________ 11 7. Mineral analyses used in mixing calculations ._______________________ 12 8. Comparison of observed and calculated values of relatively incom- patible trace elements in Alae lava lake ________________________ 13 9. Petrographic modes for selected oozes from Alae lava lake .___________ 16 III IV CONTENTS METRIC-ENGLISH EQUIVALENTS Metric unit English equivalent Metric unit English equivalent Length Specific combinations-Continued = - 0.03937 inch (in cubic meter per second figtléznitgf) ton) = " s88 feet é“) 1't(m3/s) a it's) = 35,33~3 culfiic {geett per secondd c had 2 i mi iter per secon s e 0353 cubic foot per secon kilometer (km) % C cubic meter per second i per square kilometer Area [ (m/s) /km*] = ~D1L.47 cubic feet per second per square mile [ (ft?/s)/mi?] square meter (m*) z 10.16 square feet (ft") meter per day (m/d) - -' 8.98 feet per day (hydraulic square kilometer (km?) = .386 - square mile (mi?) megar/Eg) kilometer i fest?!)ductiixi'ltsgt/(fggd) = y acres = C er mile mi hectare (ha) ed 3 kilometer per hour C (km/h) = 9113 foot per second (ft/s) Volume meter per second (m/s) = 3.28 feet per second meter squared per day cubic centimeter (cm) = - 0.061 - cubic inch (in?) (m*/d) = 10.764 _ feet squared per day (ft"/d) liter (L) = 61.08 cubic inches . (transmissivity) cubic meter (m*) = 85.81 cubic feet (ft?) - cubic meter per second Jis cubic meter == .00081 acre-foot (acre-ft) (m/s) = 22.826 - million gallons per day cubic hectometer (hm?) _ = 810.7 acre-feet 4 (Mgal/d) liter = ~ £118 pints (pt) cubic meter per minute liter = 1.06 quarts (qt) & (m?/min) = 264.2 gallons per minute (gal/min) liter = .26 gallon (gal) liter per second (L/s) ="45.85 gallons per minute cubic meter = 00026 million gallons (Mgal or liter per second per 10° gal) _ meter [(L/s)/m] = - 4.88 gallons per minute per foot cubic meter = - 6.290 - barrels (bbl) (1 bbl=42 gal) kilometer per hour [ (gal/min) (tkm/h) a's = 912527 mge per hhour (mi/h) P meter per secon m/s & A miles per hour Weight gram per cubic C centimeter (g/cm) = 62.48 pounds per cubic foot (Ib/ft?) gram (g) = 0.035 ounce, avoirdupois (0% avdp) gram per square gram x .0022 pound, avoirdupois (Ib avdp) centimeter (g/cm) = 2.048 _ pounds per square foot (Ib/ft?) tonne (t) =. 11 tons, short (2,000 Ib) gram per square {tonne s '98 ton, long (2,240 1b) centimeter 0142 pound per square inch (Ib/in?) Specific combinations Temperature i er square o kllggggfieger (fig/cm?) =' 0.986 atmosphere (atm) degree Celsius (°C) =< A.Ss degrees Fahrenheit (°F) kilogram per square degrees Celsius % centimeter = .98 bar (0.9869 atm) (temperature) = [(1.8%°C) +32] degrees Fahrenheit SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII CRYSTALLIZATION AND DIFFERENTIATION OF THE ALAE MAGMA, ALAE LAVA LAKE, HAWAII By Tomas L. Wricut and L. PrEcK ABSTRACT The 15-m-thick lake of lava, which erupted into Alae pit crater in August 1963, has been sampled by core drilling through the crust and by emplacing ceramic probes in the melt. Temperature profiles were obtained by inserting thermocouples in open drill holes. Petrographic studies of pumice and drill cores, combined with chemical analyses of crust, melt, filter-pressed liquids (oozes), and mineral and glass separates, show the sequence of crystallization in a tholeiitic basalt. Olivine (Fow) is the first mineral to crystallize at about 1,190°C, followed closely by augite(WouEnis.sFsu,) at 1,180°- 1,185°C and plagioclase (An») at 1,165°-1,170°C. Imenite (IIm.Hem..) first appears at 1,070°C, followed by pigeonite (composition unknown) at about 1,050°C, magnetite at about 1,030°C, and apatite at 1,010°C. The solidus is 980°C. Plagioclase changes composition during crystallization from An» in the pumice to an average composition of Ans in the crystallized basalt. Traces of sodic plagioclase and alkali feldspar were detected during the separation of the residual glass for analysis. Bulk clinopyroxene in the crystal- lized basalt is calculated to have an "FeOQ'"/MgO ratio of 1.3. Olivine was zoned to at least Fo before it ceased crystallizing at about 1,100°C. The Alae basalt is a tholeiite that has a very uniform bulk composition as follows: Si0O.:=50.47; Al.0;=18.67; Fe:O: =1.30; FeQ=9.80; MgO=7.55; CaO=11.11; Na.0=2.38; K.0 =0.54; TiO.=-2.74; P:0,=0.27; MnO=0.17. During crystal- lization of silicates, but before the appearance of Fe-Ti oxides, the liquid changes composition with, overall, a slight decrease in SiO» and decreases in Al:0O;, MgO, and CaO, and increases in "FeO," Na:0, K.:0, P.0O;, and TiO. Following the appearance of opaque phases, Al:0; shows little change, MgO and CaO continue to decrease, SiO», Na:0O, K0, and P:.0; increase; TiO; and "FeO" reach a maximum, then de- crease. P.0;, which is progressively enriched in the liquid throughout most of the crystallization, eventually goes into apatite. Among minor elements, Cr and Ni are depleted and Ba, Cu, Li, Nb, Rb, Y, Yb, and Zr are enriched in successively lower temperature liquids. Sr remains nearly constant, Ga shows a slight increase in late liquids, and V reaches a maxi- mum and then is depleted in lower temperature liquid. The residual liquid, represented by a colorless glass making up about 6 percent by weight of the crystallized basalt, con- tains more than 75 percent SiO» and can be calculated to consist of more than 95 percent quartz plus feldspar. Its composition corresponds to a calc-alkaline rhyolite. Liquids (oozes) that flow into open drill holes were found to have brought in less than 10 percent crystals from the enclosing partly crystalline crust. Within this crystal frac- tion the plagioclase/pyroxene ratio tends to be higher than in the host rock, suggesting that plagioclase and pyroxene were separated during flow. Extensive use has been made of "chemical modes," the computation of mineral and liquid proportions from knowl- edge of the chemical composition of a rock and its constituent phases. It is shown that optical modal data for Alae basalt are grossly in error; pyroxene is overestimated by 20 per- cent and opaque minerals by a factor of two, and plagioclase is correspondingly underestimated. Using the chemical modal data, the crystallization curve is shown to be S-shaped and to have a horizontal discontinuity at the temperature of the crust-melt interface. INTRODUCTION The eruption of Kilauea Volcano in August 1963 ponded a basaltic flow in the bottom of Alae Crater, forming a stagnant lava lake. Six days after the eruption we drilled a hole 0.9 m into the crust, an estimated 8 ecm less than the thickness of the crust. The following day we measured a temperature pro- file in that drill hole. During 1963 and 1964 we drilled 11 additional holes from which 17 kg of core were recovered. Additional samples of molten basalt were collected in ceramic combustion tubes and on thermocouples. Repeated temperature measurements were made in the open drill holes during 1963 and 1964 while the lake solidified and for another three years while the lake continued to cool. The center of the lake reached ambient temperature by August 1967. Petrographic, mineralogical, and chemical studies of samples of melt, drill core, and quenched samples of pumice provide data that contribute to our knowledge of the crystallization and, particu- larly, differentiation of tholeiitic basalt. C1 C2 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII The following preliminary reports describe aspects of the cooling history. Peck and others (1964) re- port temperature measurements in the lake through February 1964; Peck and others (1966) describe the crystallization of basalt in the lake ; and Decker and Peck (1967) give an account of infrared tem- perature measurements in the lake. Skinner and Peck (1969) report on an immiscible sulfide liquid found in one sample of melt. Recently Shaw and others (1977) and Peck and others (1977) report results of finite-element modeling of the cooling history. In October 1968 the floor of Alae Crater was over- run by new lava. Subsequently, beginning in May 1969, the crater was filled by lava from the long- lived Mauna Ulu eruption (Swanson and others, 1971). As of this writing, the spot once marked by Alae Crater now stands high as part of the Mauna Ulu shield volcano. This report is the third of a series of chapters on the solidification of Alae lava lake. Chapter A (Peck and Kinoshita, 1976) describes the eruption and the formation of surface features of the lake. Chapter B (Peck, 1978) presents and interprets temperature measurements made in the drill holes, the density and vesicularity of recovered core, and also the results of repeated level surveys of the lava surface, all of which provide information on the pat- tern of solidification of the lake. The present report extends the field studies of crystallizing basalt in Kilauea lava lakes and benefits significantly from earlier study of the two lava lakes, one prehistoric and one historic, in Makaopuhi pit crater (Wright and Okamura, 1977). A general sum- mary of studies of Kilauea lava lakes, including an annotated bibliography, is given by Wright and others (1976). ACKNOWLEDGMENTS This study of Alae lava lake has been a team effort by the staff of the U.S. Geological Survey's Hawaiian Volcano Observatory. For the gathering of temperature data we are particularly indebted to the following staff members: George Kojima, who constructed the thermocouples; William Francis, John Forbes, Reginald Okamura, Elliot Endo, and Willie Kinoshita, who helped in the drilling and in the temperature measurements ; and Burton Loucks, who made much of the special equipment used in the study. The continuous recording of temperature in drill holes in the lake was a joint effort with a visiting team of scientists and technicians from Japan, which included Profs. T. Minakami, D. Shimozuru, S. Aramaki, and K. Kamo, and Messrs. T. Miyasaki and S. Hiraga. Dr. Aramaki also helped in the drilling. During this study, we have received much helpful advice from Joel Swartz, Herbert Shaw, and Arthur Lachenbruch, all of the Geological Sur- vey. Rosalind T. Helz and Bjorn Mysen reviewed the manuscript. METHODS OF STUDY DRILLING Holes were bored in the solidified crust of the lava by means of a portable core drill with tungsten TABLE 1.-Record of drill holes in Alae lava lake Drill hole Date No. Depth of interval drilled (ft) 8/20/63 "4 0 - 2.83 9j3/é3 FH 1 2.67- 3.15 9/6/63 __ 1 3.62- 4.6 (to base of crust) 22 0 _- 2g 9/17/08 1 4.20- 5. 16/1463 and 10/2/63 _- 3 0 - 9.33 (to base of crust) ti 10/24/68 gens 3 7.80-11.40 (to base of crust) MITOS ««- os 3 9.9 -13.40 (to base of crust) 11/20/68) . .u _ 44 0 _ -15.3 (to base of crust) ___ _u. 4 11.3 fig; Sh bustleI 02f gustamelt 12 55 0 . ~20. roug 3 lower crust and base of lake at approxi- mate]?y 28. 5) ft) 1/20/04 :.. el 5 17.5 -21.0 (plug of ooze f Lw! 8 0 -32.25 (through 6 ft of melt 4/1/64 and 4/2/64 . 6 htc lgwexl' crust?) } 17.15-42.8 through plug of ooze SMA sen g (and lower crust to near base of lake) C. n one eg 6 19.4 -35 (through plug of ooze) 4122704 . onc cee TT 0 - 0.83 (mast hole) 4/20/04: _ .- JA LOI. T 0 _ -20 6/4/64 ... i T 19.7 -21.8 6/10/6422... 20 .. ucual 7 21.35-23.4 6/AT/OL) cucu ce lued aas. 88 0 _ -20.15 6/18/04) __ 8 20.75-28.65 (through 1 ft of melt into lower crust) _________ 9 0 _ -11.7 f 9 11.7 -12.7 - (EX hole to 24.5 ft) 9 12.1 -22.7 (SP hole 24.5-34.5 ft) 9 22.1 -84.5 (through partly molten center of lake into lower crust) 10/20/04 _u 9 34.5 -52.9 (through base of lake at approximately 48.0 ft and into 1840 lava (cored from 48-50 ft)) 10/22/64 .._... 10 0 _ -16 411 0 - 9 11/17/64 11 9 -16 1212 0 _ -15.7 11/19/64: Lous 12 15.7 -47 (through base of lake at approximately 44.5 ft) ' Near the major axis of the lava lake on the opposite side from the vents, 165 ft S. 44 E. of the center of the lake (fig. 1) ; altitude of collar 2,567.2 ft on Feb. 10, 1964. 216 ft N. 75 1. of drill hole 1; altitude of collar 2,567.3 ft on Feb. 10; “156 it S. 15 E. of drill hole 1; altitude of collar 2,567.3 ft on Feb. 104 ll'Jt"(l;’,4N. 75 E. of drill hole 3; altitude of collar 2,567.3 ft on Feb. "2309 It &. 49 E. of drill hole 1; altitude of collar 2,565.8 ft on Feb, 10, 1964. # 240 fto8. 17 E. Sept. 24, 1964. 136 ft 8. 85 W. May 12, 1964. 81 ft SW. of drill hole 7; altitude of collar 2,566.9 ft on May 12, 1964. ©2060 ft N. 42 W. of drill hole 1; altitude of collar 2,567.5 ft on July 31, 1964. 310 ft N. 16 E. Nov. 3. 1964. "330 ft S. 87 W. of drill hole 1; altitude of collar 2,566.9 ft on Nov. 3, 1964. 2 ft due east of drill hole 2; altitude of collar 2,567.5 ft on Jan. 21, 1965. of drill hole 1; altitude of collar 2,567.4 ft on of drill hole 1; altitude of collar 2,566.9 ft on of drill hole 1; altitude of collar 2,568.0 ft on CRYSTALLIZATION AND DIFFERENTIATION OF THE ALAE MAGMA EXPLANATION ©10 Drill hole C3 0 100 200 FEET 0 50 METERS CONTOURS IN METERS ABOVE SEA LEVEL FIGURE 1.-Index map of Alae lava lake, showing topographic contours at the base and margin of the lake and location of the numbered drill holes. carbide bits. The drill was powered by a 9-horse- power gasoline engine. Most of the holes were 2.86 em in diameter, but the upper part of one hole (drill hole 9) was cut to 3.8-em diameter by means of EX bits. During the first 15 months after the eruption 12 holes were drilled. Dates and depths of drilling are listed in table 1, and the locations of the num- bered drill holes are plotted on figure 1. All except the first two holes were made with a drill mounted on a portable mast anchored in the crust. The bits were cooled with water pumped through the drill pipe by a portable 31,-horsepower gasoline engine. The water was fed through hoses from the rim of the crater and stored in a 50-gallon (190-L) steel drum near the drilling site. About 10 gallons of water were used for each linear foot of drilling. The first two holes were drilled by holding the drill manually and pouring water into the hole by hand. A core barrel was used during the drilling, except C4 near the base of the crust, where the barrel impeded the flow of the cooling water and had to be removed. For most of the drilling, the drill stem was pulled at 1- or 2-foot (30- or 60-em) intervals and the core was collected. Core samples were numbered consec- utively in each hole; thus sample A-9-5 is the fifth core sample from drill hole 9 in the Alae flow. Core recovery averaged between 25 and 50 percent for all holes. Typically, recovery was good from that part of the crust that was at temperatures of 700°C or less before drilling. In several holes, however, re- covery was poor from the cavernous and highly vesicular zone between depths of 0.3-1.2 m. Core recovery was also poor from the interval of crust that was at temperatures between 700°C and 950°C before drilling, apparently because some cores shattered when quenched by the drilling water. For- tunately, abundant core was recovered from the in- terval of crust that had been at temperatures of 950°C and 1,070°C, the interval spanning the zone of crystallization. No core was recovered from lava that was at temperatures of more than 1,070°C, before drilling. TEMPERATURE MEASUREMENTS Temperatures in drill holes were measured on 47 different days during the four-year study of the cooling lake. Most measurements were made with thermocouples of chromel-alumel, and, a few with thermocouples of platinum-platinum plus 10 percent rhodium, using a portable millivolt potentiometer and a 0°C reference junction in an ice-filled vacuum bottle. Most of the temperature measurements have an accuracy of + half a percent at 500°C and +1 percent at 1,000°C. Details of thermocouple con- struction and calibration are given in chapter B. In several drill holes, temperatures were measured both in the crust and in the underlying melt. This was done by pushing a ceramic probe into the melt through the bottom of a drill hole after boring through the base of the crust. Temperatures were measured with a thermocouple inserted in the probe. Probes as long as 3 m were constructed from 91.5- cm lengths of 2.54-em OD mullite tubes (the lower tube closed at one end), joined with 10.16-em sleeves cemented with silica cement. The probe was attached to the drill steel with a machined fitting. The whole probe unit was placed in the completed drill hole and the temperature was then measured at the bot- tom of the hole. After the temperature had reached 1,065°C to 1,070°C,' the unit could be pushed slowly on into the melt. Sometimes, but not always, it could also be pulled out. Excessive pressure caused SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII the ceramic probes to crack, allowing melt to enter. An unbroken probe was not emplaced in the melt until December 2, 1963. Six attempts were unsuccess- ful during the preceding three months. Later work in the 1965 lava lake in Makaopuhi Crater has shown that stainless steel probes can be used successfully for long periods in the melt, and can be emplaced with less hazard of breakage. Several different methods were used to estimate the temperature of core samples immediately before drilling. Sometimes the temperatures of cores re- covered from drill holes adjacent to existing holes were determined by measuring the gradient in the existing holes immediately before drilling. Other times the holes were deepened only a few inches at a time, particularly in the zones of crystallization ; and the core temperatures were estimated by extrapolat- ing the gradient measured in these holes before renewed drilling. In other holes, the temperatures were estimated from a plot of the depth of isotherms versus the square root of time (see fig. 10, chapter B), extrapolating the depths of isotherms back to the time of drilling from a plot of temperatures measured in each hole over a period of time after drilling. Temperature intervals assigned to cores from the upper crust of the lake in the range 500°C to 1,070°C are probably accurate to +2 percent, estimated from the uncertainty in reconstructing isotherms for a single drill hole (see fig. 10, chapter B). Temperatures assigned to cores in the range below 500°C are less accurate, perhaps +10 percent, because of greater differences between holes, marked effect of rainfall, and slower thermal recovery after drilling. Temperatures assigned to cores from the lower margins of the lake are highly interpretive and may be in error by as much as +20 percent. Table 2 is a complete list of all samples collected and an estimate of their temperatures at collection. LABORATORY STUDIES All drill cores were inspected under a binocular microscope, and a thin section from each core inter- val was examined under the petrographic micro- scope. Modal counts were made with a point counter. Selected pieces of core were analyzed chemically by the rock analysis laboratory of the Geological Survey in Denver, Colo. A total of 22 standard and 48 partial 'This temperature marks the interface between crust and melt in all of the lava lakes. The transition from crust to melt takes place when the lava is approximately 50 percent crystallized, and is marked by an abrupt change in physical properties-particularly shear strength-of the basalt. Below the transition temperature, the basaltic crust behaves as a rigid solid. Above the transition temperature, the melt, either in a liquid or supercooled form, gives little resistance to either a drill or a blunt-ended probe pushed by hand. CRYSTALLIZATION AND DIFFERENTIATION OF THE ALAE MAGMA C5 TABLE 2.-Drill cores from Alae lava lake TABLE 2.-Drill cores from Alae lava lake-Continued Drill hole No. 1 Drill hole No. 6 Estimated Estimated Sample temperature Weight of Sample temperature Weight of - Com- No. Depth (ft) Date - of collection core (g) Comments No. Depth (ft) Date of ggggction core (g) ments (°C) A- 141." ._=>0.8 = 12 8/29/63 520- 650 11 A-.634. .. 0 ~:04 4/ 1/64 _ 40- 50 127 Boos sA 9/ 3/63 8s80- 950 4 _- May include P4, ' A - 5B do 50 72 ooze from D ree .D -if do 50- 60 29 2.7-2.8 tf €. 2% .T s 9 do 60 89 (tempera- §... 0. - 1.0 do 40- 60 89 ure= b:. 1.0 - 2.0 do 60- 90 12 1,040°- T .- 2.0 - 80 do 90- 140 34 1.060°C). 8. Z.. 8.0 - 5.0 do 140- 250 13 8 .~ §1 - Af5 do 950—1940 2T 9 -_ 5.0 - 6.0 do 250- 320 66 4 Sl 3.75 040 5 10. __ 6.0 - 1.0 do 320- 410 65 ses $4 ay" . n/, 6/63 ©099-1,015 T 11 __ 8.0 -10.0 do 480- 610 144 6- -_ '4.25- 4.6 1,040-1,075 T 12 __ 10.0 11.9 do 610- 670 62 T Jus' 4:8 -i5.0 9/17/63 960- '980 T 18 ._11.0 -12.0 do 670- 730 100 8 "_s: 5.0. -'5.2 do 980-1,000 18 14 __ 13.0 -15.0 do 790— 890 112 15 __ 15.0 -16.0 do 890- 930 42 fiat =s" ts if Ts § 17* _ 19.0 -20. 1,0 0 Drill hole No. 2 18 __ 22.0 -23.0 do - 1,060-1,070 17 foci tat ~ Ak it 20 ._ 24.0 -25 0 - "1,07 Are arte "ho- meant . "Roo 13 21+ _ 25.0 -26.0 do - 1,070? 27 3 2 "10 _ 1.8 10 190 15 g2 ._ 27.0 -28.0 do _ 1,070? <1 1 "¢ _- ts Rel 580 23 23 __ 28.0 -29.0 do _ 1,060-1,070 26 Sy 20 -_ S15 do 720 3 24* _ 31.0 -32.0 do __ 1,030-1,010 30 Ke .. =f f 25*® _ 19.0 -19.7 4/15/64 1,010-1,030 23 Ooze. 26 __ 20.0 -27.5 do _ 1,030-1,060 52 . 27 ._ 35.0 -40.0 do 970- 900 61 Drill hole No. 3 28 __ 19.4 -21.0 4/22/64 1,000-1,020 ? 17 i o 0 os.. "1% oze. Std ~- (Titik "as"" Roo ts 31 ._ 28 -887 a/ 3761 1,045-1,060 ? 19 Ooze on s 2 do _ 620- 740 25 s 4 __ 4.8 -'5.8 do 850- 940 12 I 5 I. 6.8 - T8 do 1,020-1,075 16 6 IL T5 - 9.75 10/24/63 '950-1,080 <1 Drill hole No. 7 ooze 9.9 11/ 6/63 1,040 <1 Contaminat- ed from old _ A- 7-1 __ 0 - 0.4 4/22/64 _ 40- 60 159 drill pipe. 2 _.c;.4 - .0 do 60- 70 121 I -.- _. 06'=".8 7 80 85 g ay 07 = 2.3 4/29/64 _ 50- 75 g; - ott ~ 120 Drill hole No. 4 6 .. 2:0~ 5.0 do 120- 150 89 cf ot #81 ~f a- 4-11... 0 ~- b.25° 11/20/65 - 50-70 19 -- 4.0 - 8. 0 190- 2 $# cL lesa ls ldo/ 70- 130 56 9 .._. 5.0 --6.0 do 230- 2900 88 5 a. $ <21 do 130- 290 5 10 __ 6.0 - 7.0 do 290- 360 95 4 . 21 - 805 do 290- 410 20 11 __ 7.0 - 8.0 do 360- 430 140 5* __ 3.05- 4.45 do 410- 580 38 12... 8.0 - 9.0 do 430- 500 172 6‘ LD 4.45- 5.95 do 580- 715 57 13 __ 9.0 -10.0 do 500- 560 47 LL ©5.95- 6.95 do 15- 790 T 14 __ 10.0 -10.6 do 560- 600 50 8. 9.9 -10.9 do 970-1,010 33 15 __ 12.0 ~13.0 do 640- 700 64 10* _ 10.9 -12.0 do 1,010-1,040 28 16 __ 13.0 -14.0 do 700- 770 46 11 __ 12.0 -12.85 do 1,040-1,060 3 § i; E4 fig fig g go $838— 358 33 *~ 14. 19 oze on -- 14.0 - 0 - f s tho 1%/81/98, 1,001 therhio- 19 __ 18.0 -19.0 do 950- 990 140 al 20.0 - o/ bee 9i5- 955 1 18* _ 14.2 -20.5 - 12/ 2/63 -1,125 500 _ Filled -_- 19.7 - 945- < / 2/63. 1,065-1,125 C 22 °- 20.0 -21.0 0 955-990 127 combus- 28 2. B1.0 ~21L.8 do 990-1,010 53 tion tube: 24 _. 21.8 -22.1 6/10/64 - 995-1,015 3 25* _ 22.7 -28.4 do 1,015-1,030 47 Drill hole No. 5 Drill hole No. 8 & $41 <.: 0 6/18/64 _ 40- 100 56 A- 5-1 __ 0 _- 0.35 12/16/63 40- 70 116 2 - 40 do 100- 170 211 2... .00- 05 do 10- . 90 73 3 __ 4.0 - 6.0 do 170- 260 122 lr G-: 17 do 90- 100 61 4 __ 6.0 - 7.45 do 260- 340 59 4 .1.0 -. do 40- ~55 17 5 __ 8.0 -10.0 do 370- 470 T2 9 =a, 2 - 1.0 do 55- 130 23 6 _©10.0 -12.0 do 470- 570 166 6 __ 1.0 - 2.0 do 130- 220 27 7 2218.9 ~14'0 do 670- 720 26 7. .. 2.0 > 8.0 do 220- 320 34 8 L. 14:0 -15.9 do T20- 770 12 8 __ 8.0 - 5.1 do 320- 510 177 9 __20.75-21.85 - 6/19/64 960- 990 20 9 61 = TA do 590- 660 30 10* _ 22.85-23.35 do 1,020-1,030 20 10°. 10.2 -11.2 do 840- 890 26 11 __ 23.85-24.85 do 1,040-1,050 25 11}. 11.2 -12.2 do 890- 940 45 12 __ 24.85-25.85 do 1,050-1,060 8 12 ;. 18.2 -14.2 do 980-1,010 30 13 __ 25.95-26.85 do 1,060-1,070 9 18 [_ 14.2 16.2 do 1, 010 1,060 89 14 __ 26.85-27.85 do 1,060-1,050 24 14 _: 18.23 -20.2 do © 1.0 10 ( i . 2 - o _ 1,040-1, i % 1g ie Egg ggg do $a S78 i? eaton Drill hole No. 9 18 __ 28.2 -29. 870-84 thermo- A- 9-1¢ __ 0 - 0.3 7/28/64 _ 40- - 60 137 19 __ 17.8 -18.3 12/30/63 1,050+2 23 couple. 19 .s. .8 -! (B8 ldo/ 60- 70 102 1b < ..55-. .85 do T7O- 90 141 20* _ 19.0 -20.0 do _ 1,045+5 55 - Do. a 2 ao Jo - "ufos 1897 400° gg 21 .2 ca. 18.5 do 1,050+ <1 Do. 4 ____ 2.3 as 3.55 do 130- 170 zs 16 5 __ 4.55- 5.55 do 200- 240 172 See footnote at end of table. 6 __ 5.55- 6.7 do 240- 290 217 C6 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII TABLE 2.-Drill cores from Alae lava lakee-Continued Drill hole No. 9-Continued Estimated Sample temperature Weight of No. Depth (ft) Date of (cggection core (g) Comments A- 9-7 c_ 6.17 _ 1.1 do 200- 360 134 8 L- T.T - 8.6 do 360- 410 88 9 __ 8.6 - 9.7 do 410- 470 146 10 __. 9.7 -10.7 do 470- 520 249 11 __ 10.7 -11.7 do 520- 570 150 12 ._ 117 -12.1 7/29/64 570- 610 278 18 ._ 12.7 -13.7 9/11/64 560- 600 155 14 __ 18.7 -14. do 600- 620 37 15 __ 14.2 -15.7 do 620- 670 427 16 __ 15.7 -16.7 do 670- 710 135 17* _ 16.17 -17.7 do 750 161 18 __ 17.7 -18.1 do T50- 780 130 19 __ 18.7 -19.1 do T80- 800 32 20 __ 19.1 -20.1 do 800- 830 44. 21 __ 20.7 -21. do 850- 870 35 22 __ 24.5 -26.0 9/14/64 940- 970 114 23* _ 26.0 -27.0 do 970- 990 47 24 __ 28.8 -30.0 do 1,000-1,010 13 25* _ 30.0 -31.0 do 1,010+ 24 26 __ 32.0 -33. do 1,010- 990 115 27 __ 42.0 -45.0 10/20/64 900- 860+ 144 28 __ 48.0 -50. do 810- 770+ 8 1840 lava. Drill hole No. 10 A-10-1 __ 0 - 0.10 10/27/64 29 2 z= al - do 30- 80 216 8 -_ ~ ;b5- .10 do do T8 4. .._ f + :.80 do do 41 5.» B+ .0D do do 73 6° .:.. 95- L10 do do 45 8 __ 0 - 2.0 do 30- 125 125 9 __ 2.0 - 8.0 do 125- 180 92 10 __ 8.0 - 4.0 do 180- 230 88 11 __ 4.0 - 5.0 do 230- 285 112 12 __ 5.0 - 6.0 do 285- 340 159 13 __ 6.0 - 7.0 do 340- 395 159 14 __ 7.0 - 8.0 do 395- 445 118 15 __ 8.0 -10.0 do 445- 540 228 16 __ 10.0 -12.0 do 540- 620 218 17 __ 12.0 -14.0 do 620- 685 155 18 __ 14.0 -16.0 do 685- 745 69 Drill hole No. 11 A-11-1 __ 0 _- 0.5 10/27/64 30- 75 222 2 t 47 do do 119 O ~eum id '- s B do do 27 d .. "0 .% 20 do 30- 150 110 5 __ 2.0 - 4.0 do 150- 255 215 6 __ 4.0 - 54 do 255- 320 195 1 -- 54 - T.0 do 320- 390 244 §. sk AAO. s 9.0 do 390- 460 226 9 __ 9.0 -10.0 11/17/64 455- 485 108 10 __ 10.0 -12.0 do 485- 550 228 11 __ 14.0 -16.0 do 610- 675 89 Drill hole No. 12 A-12-la __ 0 - 0.1 11/17/64 40- 150 18 Ib ../ A1 - 0.8 do do 69 16... « 8 + 0.7 do do 186 14... .7 - 0.9 do do 110 22-00 *~. 1.0 do 40- 150 54 8 __ 1.0 - 3.0 do 150- 280 73 4 __ 8.0 - 5.0 do 280- 375 51 5 __ 5.0 - 1.0 do 315- 465 182 6° __ 7.0 - 9.0 do 465- 545 322 T __ 9.0 -11.0 do 545- 610 171 8 __ 11.0 -13.0 do 610- 675 272 9 __ 13.0 -15.0 do 615- 725 250 10 __ 15.0 -19.0 11/19/64 - 725- 815 380 11 __ 19.0 -23.0 do 815- 880 425 12 __ 28.0 -27.0 do 880- 905 38 13 .-27.0 _31.0 do 905- 895 212 14 __ 31.0 -85.0 do 895- 860 263 15 __ 35.0 -89.0 do 860- 810 222 16 __ 39.0 -43.0 do 810- 750 398 * Samples have complete chemical analyses (table 3). analyses were made of core; and four additional complete analyses were made of samples of pumice, spatter, and crust. The constituent minerals and glass in two samples of pumice and one core sample were separated after grinding to -400 mesh size, using heavy liquids, an ultrasonic vibrator, a magnetic separator, and an elutriator. The separates were studied under the microscope in oil immersion, by X-ray diffraction, and by chemical analysis. Glass from the pumice samples was separated by repeated centrifuging in BR-DMF (Bromoform-Dimethylforamide) mixtures of specific gravity 2.74+ and 2.78. The specific grayv- ity of the glass was determined as 2.761+0.005. The purity of the glass was first tested by studying the separates on an X-ray diffractometer and com- paring the resulting pattern with previously cali- brated mixtures of glass and known amounts of plagioclase, pryroxene, and olivine. The calibration showed that is little as 2 percent of plagioclase or olivine and as little as 3 percent of pyroxene could be detected as crystalline contamination of a glass. The final separates showed no X-ray diffraction peaks for the three silicate phases. Next, a standard mixture of 1-percent crystals in crystal-free glass was prepared as an optical standard. The crystal content of the separated glasses was estimated to be less than that of the standard and hence the glasses submitted for chemical analysis were con- sidered to be 99+ percent pure. Plagioclase from the pumice was concentrated (but could not be purified) by flotation in a BR-DMF mixture of specific gravity 2.71. Separation of olivine and clinopyroxene from glass in samples of pumice was achieved by first collecting the -270 to +400 mesh fraction which sunk in methylene iodide (MI) diluted with DMF to a den- sity of 3.20. This sample was then resieved and put on a Frantz magnetic separator. When a side slope of 15° and a forward slope of 20° were used, the olivine was found to be magnetic at 0.65 amperes, and the pyroxene to be nonmagnetic at 0.67 amperes. Final purification of the pyroxene was achieved by crushing the nonmagnetic fraction to -400 to +500 mesh and centrifuging in MI-DMF mixtures, the specific gravity of which was increased from 3.20 in steps of 0.01 until the sink fraction consisted of essentially glass-free clinopyroxene as determined by inspecting it in oils with a petrographic micro- scope. Pyroxene separated in this way was submitted for chemical analysis. The composition of olivine (not more than 90 percent pure) was determined using the X-ray diffractometer and the calibration of Murata and others (1965). CRYSTALLIZATION AND DIFFERENTIATION OF THE ALAE MAGMA CT One sample of core collected at subsolidus tem- peratures was crushed and sieved to -400 to +500 mesh for separation of residual glass and plagioclase. Pyroxene could not be separated effectively because of its included opaque phases. The plagioclase was separated by repeated centrifuging in BR-DMF liquids of specific gravity 2.725 and 2.671. Because of the fine grain size, further magnetic separation was not attempted. Plagioclase submitted for chemi- cal analysis was estimated from optical examination to be essentially free of glass and pyroxene but to contain a small but unknown percentage of opaque mineral as inclusions. The residual glass was separated after elutriation of the -400 to +500 mesh sample that formed the float fraction in the liquid of specific gravity 2.671. The elutriation produced size fractions estimated to be -44 to +25 microns (x) and -254 to +104. These were then centrifuged in BR-DMF mixtures of specific gravity 2.515, 2.400, 2.345, and 2.285. The specific gravity of the glass was found to be be- tween 2.285 and 2.345. The following fractions were collected for chemical analysis: 25-441 ___ 2400 42.06 15.46 13.05 17.60 eal 1.72 1.00 * Le 129 17. 42 17.66 4.53 14.19 10.26 .08 .08 NaO ® _s "ases .57 Set- 3.87 5.67 Prec 810 : | ol ses 09 sens "21 .21 awaz wine TOA * nung cc 02 1.66 .50 .52 Tess eus 48.14 22.52 PMQs *t. R tike .06 eod uas e= ~ - ae MnQ_ 27 17 .50 51 uk. A4 .37 Nature of samples : D. Pigeonite from Makaopuhi lava lake. A. Stoichiometric Faso with 1.5 percent Cr-spinel included. 0.3 per- E. Stoichiometric NaAlSisOs and CaAlsSiO« with 0.5 percent KO cent CaO included as B. Analyzed augite (see table 4). Fe:Os and FeO adjusted (see text). C. Most iron-rich augite from Makaopuhi lava lake. solidus temperatures. The calculation is very sensi- tive to Fe,0;,/FeQ ratios used in the pyroxene, glass, and bulk rock composition. The Fe,0;,/ (Fe.0;,+FeQ) contents in the pyroxenes and the differentiates were adjusted to fit the following two assumptions: 1. The parent was taken as the average composi- tion showing the least oxidation in table 3 (see average (s), table 3A ; Fe,0,=1.20, FeO=9.84) . This represents the composition of the lava before sub- solidus oxidation took place (see Sato and Wright, 1966, and Wright and Okamura, 1977, for further discussion of oxidation). 2. Magnetite was assumed to appear at a lower temperature than ilmenite and the magnetite/ ilmenite ratio was assumed to increase slightly dur- ing crystallization. Two consequences of these assumptions are, first, that the Fe,0;,/(Fe.0,+FeQO) ratio in the analyzed DPH-77 augite is too high to fit the calculations for the least oxidized bulk composition. We need values of -Fe;0,=0.88 and FeQ=7.55 compared with analyzed values of Fe,0,=1.78, FeQ=6.51. The calculated mineral formula (table 4) for DPH-T77 shows a cation deficiency relative to six oxygens and this imbalance could be improved by reducing the amount of Fe.0O;. Thus we feel this analysis is probably in error; an adjusted analysis that fits the rock composition is given in table 7. The second consequence of the assumptions is that A-6-25 is the only ooze that has an analyzed Fe,0;,/(Fe,0;,+ FeQ) ratio consistent with its degree of crystalliza- tion if one assumes no subsolidus oxidation. The calculated and observed Fe,.0,/(Fe,0,+FeOQ) ratios are shown at the bottom of table 6. They show that samples A-4-13, A-4-12, and A-5-20 are reduced included as KAISisOs. F. Ilmenite and magnetite from Makaopuhi lava lake. relative to what is predicted from the chemical mode calculation. This is consistent, as noted above, with the environment of collection where the molten ooze was collected in contact with drill steel or thermocouple sheathing. A-6-30 and A-6-29 are slightly oxidized relative to their predicted value, which is also consistent with their appearance in thin section. These samples and A-6-25 were drilled out when solid. The slight oxidation may have taken place because of the presence of air-contaminated drilling water when the ooze flowed into the drill hole. A-9-23, a natural segregation, is essentially unoxidized. The mineral percentages calculated in table 6 are generally consistent with the inferred paragenesis, except for the inclusion of pigeonite and ilmenite in samples A-4-13 and A-4-12. TiO, is lower in these samples than would be predicted on the basis of crystallization of silicates only. If ilmenite is re- moved from the calculation, a TiO, residual of 0.15 results, far too high to ascribe to analytical error in the rock, ooze, or pyroxene. Pigeonite is probably an artifact of ilmenite being included in the calcu- lation and is taken in order to balance the SiO,. Calculations made without pigeonite have a SiO, residual of about 0.2 if Na.0 is balanced to give the same plagioclase composition. We do not know the reason for the anomalous compositions. Two possi- bilities are (1) reaction with steel resulting in loss of TiO, (and FeO?) or (2) admixture of lower temperature ooze during collection. The latter seems more likely as samples in Makaopuhi lava lake that were collected in stainless steel picked up small amounts of iron and showed no TiO, gain or loss. Mineral compositions change in a regular way as CRYSTALLIZATION AND DIFFERENTIATION OF THE ALAE MAGMA crystallization increases. (Note: These represent the average composition of each phase erystallized before collection of the ooze or glass sample. As the minerals are normally zoned, the calculated com- positions of feldspar and mafic silicates are more calcic and magnesian, respectively, than the actual composition of the mineral in contact with the liquid represented by the glass or ooze. Electron microprobe study would be necessary to specify the exact compositions of coexisting mineral-liquid pairs.) Olivine and pigeonite are taken as fixed compositions so that the changing composition of the mafic assemblage is represented by the MgO/ (MgO+FeQ) ratio calculated for augite and given in parentheses in table 6. These range from 0.66 in the pumice (H-186) to 0.45 in the subsolidus as- semblage (A-12-10, -11). Plagioclase changes com- position from An,, in the pumice to An;, in the sub- solidus assemblage. In summary, the data of table 6 represent a nearly quantitative representation of changing mineral proportions and compositions during crystallization of Alae lava lake. These data are plotted in figures 2 and 3. Trace-element content also varies along the liquid line of descent. Unfortunately, the collection of melt samples and ooze in proximity to drilling steel con- taminated some samples, especially with regard to Cr and Ni. The overall variation in trace-element content is less regular than for major elements (see table 3), precluding an accurate assessment of distribution coefficients between silicates and liquid along the liquid line of descent. Table 8 summarizes data on relatively incompatible elements. The values for an undifferentiated bulk composition are taken from average (3), table 3A. Maximum enrichment factors, assuming that each element is totally parti- tioned to the liquid, are calculated from the results of mixing calculations given in table 6. For example, sample A-6-29 represents 57.7 percent crystalliza- tion and a consequent maximum enrichment factor C13 of 11 The mixing calculations of table 6 assume that P.;0;, was the most incom- patible minor element. The data of table 8 indicate the following: 1. Nb is the most incompatible of the trace ele- ments during silicate crystallization; in fact, it appears to be anomalously enriched in oozes A-6-25 and A-6-30 beyond what is possible by complete partitioning to the liquid. One possible explanation is that P.,0, is partitioned in part to phases other than augite or liquid (for example, plagioclase). A very small amount of P.0; in plagioclase would in- crease the enrichment factors to a point where the Nb data would fit. Nb is well determined (D. Gott- fried, written commun., 1966), yet it still varies over a 10 percent range in the undifferentiated samples and even in the pumice. Also, Nb may possibly be redistributed by a process as yet unrecognized for any other element in the lava lakes. At lower temper- atures Nb is concentrated by ilmenite and magnetite , and the rest of the samples show depletion of Nb relative to the maximum calculated enrichment. 2. Li and K,0 show parallel variation indicative of slight partitioning to plagioclase at high tem- peratures, increasing toward lower temperatures. 3. Zr is not well determined but shows significant partitioning to crystallizing phases (probably pyrox- ene) in the two lowest temperature samples. Other trace elements show significant partitioning to crystallizing phases at high temperatures and either remain constant or are depleted in lower temperature samples. PETROGRAPHY The modal composition of Alae lava lake has been summarized in an earlier paper (Peck and others, 1966, fig. 4, p. 635-641). For the present study, new thin sections were cut for several samples of partly molten core. Counts of 1,500 points were made on these and the older sections, using a 0.3 X 0.3-mm grid. All modal data, obtained in volume percent, TABLE 8.-Comparison of observed and calculated values of relatively incompatible trace elements in Alae lava lake [Calculated results are obtained by multiplying the trace-element values for average (3) in table 3A by the maximum enrichment factor given in table 6] Sample K.O (weight percent) Li (ppm) Nb (ppm) Zn (ppm) No. Obs. Cale. Obs. Cale. Obs. Cale. Obs. Cale. AA 1B nemen un om 0.80 0.90 9 6.6 30.1 26.6-31.7 240 247 AIM Ieee .88 .95 6 6.9 29.9 27.8-88.2 210 258 6=20 ___ "Ci l te .96 1.03 T 7.5 38.7 30.3-36.1 280 280 6 -BU te- a ao- ae dbl aire mn s 1.07 11% 7. 8.5 43.8 84.5-41.1 280 320 6B ALL Ee cain nec 1.16 1.30 8 9.5 44.3 38.3-45.6 320 355 _ ecu tl cic 1.57 1.86 11 13.5 47.4 59.7-65.2 340 507 * Range calculated from the maximum variation in Nb in undifferentiated core (table 3A). C14 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII 100 y 1190 \ A & 's % 90 |- L \\+ +\ 11150 so |- 11100 70 |- a S a ¥ + : a 60 -i * * w E w F * E ® \. + ‘l’ \_ . § [-] 5 "| \ \ \ \ z a | X N \ ud z H . . s meLt & us 40 -+ + 1 1065 _ & 4 w & cRUst & Fol C. 30 + % & M + 3050 5 1 1000 10 h —_}F f}; qt- 0 * . O - 980 o 1 1 1 1 1 T 1 T i 1 i 1 0 - 5% -o 10 20 30 40 50 0 10 20 30 40 0 10 20 OLIVINE, CLINOPYROXENE, PLAGIOCLASE, Fe-Ti OXIDE, IN WEIGHT IN WEIGHT PERCENT IN WEIGHT PERCENT IN WEIGHT PERCENT PERCENT FiGurE 2.-Graph showing modal data for Alae lava lake. Crosses show modal data obtained in transmitted light con- verted to weight percent (see text). The length of the bars represents the estimated counting error of the mode according to the method of Van Der Plas and Tobi (1965). The dots (*) indicate weight percent modes calculated from analyzed glasses and oozes (data of table 6) and represent the modal abundances consistent with the bulk chemistry. Note that pyroxene and opaque minerals are overcounted and plagioclase undercounted in the optical modes. were converted to weight percent by assuming the following specific gravity for minerals: olivine and pyroxene (sp gr=3.3) ; opaque phases (sp gr=4.5). The specific gravity of glass used in the conversion from volume to weight percent was estimated from the refractive index of the glass, assuming linear variations between the refractive index (R.I.) and specific gravity (sp gr) for the separated glasses: (sp gr=2.3, R.I.--1.488; sp gr=2.16, R.1.=1.0602); The new modal data are summarized in figure 2 along with the curves defined by the chemical modes. An important finding is that there is significant bias in the optical modes relative to the chemical modes, marked by high values for pyroxene and opaque phases and by low values for plagioclase. This is evidently the result of the fine grain size of the basalt, in particular the very thin (less than 0.01 mm) laths of plagioclase. From figure 2, we can calculate that the pyroxene is consistently over- estimated by approximately 24 percent and the opaque minerals by 50 percent. If these percentages are subtracted from the optical estimates of pyrox- ene and opaque mineral, respectively, and are added to the plagioclase percentage estimated optically, a close approximation to the chemical mode is ob- tained. The glass content is estimated satisfactorily by optical methods. Figure 3 shows the modal percent- age of glass plotted against temperature, using the chemical modes and all of the optical modes of partly molten samples for which accurate tempera- ture data were available. A single curve can be drawn through most of these data, indicating that the optical estimate of glass content is satisfactory. Serious discrepancies exist for two samples of ooze for which chemical modes were calculated CRYSTALLIZATION AND DIFFERENTIATION OF THE ALAE MAGMA 1200 LIQuIipus 7 /. H-186 GL " 1150 |- a DPH-77 GL t F f @ 1100 |- i = u I a ed MELT 3 I ane ASL CRUST ud * A_ -4-1 a 1050 ig c" «A- 4-12 w A-6-301 & 3 I , 3 & | A-6-25 g 1000 E4 fe sOoLIDUs A-12-1011 RESIDUAL GLAss 950 |- - | 1 | | I | | | | 0 10 20 30 40 50 60 70 80 90 100 GLASS, IN WEIGHT PERCENT FIGURE 3.-Graph showing weight percent glass plotted against temperature. The S-shaped line is a best-fit by eye, drawn with the modal data for undifferentiated samples in figure 2. Data for glass content of differentiated samples (for example, A-6-25) are taken from table 6 and plotted against temperature of collection. Discrepancies for samples A-6-25 and A-6-30 are discussed in the text. (A-6-25 and A-6-30). Both samples suggest a temperature of formation higher than the tempera- ture at the time of collection. Drill hole 6 was the first hole to penetrate the lower crust of the lake (tables 1, 2). The samples A-6-25 through A-6-30 were samples of core that filled the open drill hole adjacent to the partly molten crust between the two 1,000°C isothermal surfaces. The easiest way to rationalize the temperature discrepancy shown on figure 3 is to infer that all the ooze came in from near the hottest zone (1,070°C, see table 2) and moved up or down the drill hole as it flowed in. Thus A-6-29 was collected at a depth corresponding to the maximum temperature in the hole and close to where it entered the drill hole; and A-6-25 and A-6-30 were collected at depths respectively above and below where the ooze entered the drill hole. Hence the estimated temperatures at the depth of collection are too low. The temperatures of the pumices (H-186 and DPH-77) were estimated from temperature meas- urements made during the eruption. The liquidus temperature of the whole rock (100 percent liquid) is estimated from the Mg/Fe ratio using the method of Tilley and others (1964). The shape of the curve that relates glass content to temperature is very similar to that obtained for Makaopuhi lava lake, for which a broader range of temperature data is known (Wright and Okamura, 1977; Wright and others, 1976, fig. 12). The temperatures obtained using figure 3 are also shown on figure 2. Figures 2 and 3 together give a quantitative description of crystallization of the liquid line of descent (fig. 4) for Alae lava lake. Petrographic modes of selected samples of ooze are given in table 9. Some of these (for example, A-5-20 and A-4-12) contain less than 10 percent crystals, commonly clotted together, placing an up- per limit on the amount of previously crystallized material brought in with the liquid. Other oozes, notably those from drill hole 6 (A-6-25, A-6-29, and A-6-30), are highly crystalline because they C16 SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII TABLE 9.-Petrographic modes for selected oozes, in volume percent, from Alae lava lake [A, crystals brought in with liquid; B, crystals growing in the liquid] Mineral A-4-12 A-6-25 A-6-30 A-6-29 A-5-20 % 2 1 2 1 2 1 2 Olivine: ______________________ mmcim - ee 0.1 awl 0.1 0.0 0.0 0.0 0.0 s 3 s... -s Po 0.0 § ~ 0.0 oA 2G 0 . 0 2.0 Pyroxene: _____________________ 3.0 4.8 3.6 6.4 4.4 3.0 2.1 2.7 27 ______________________ tels 26.9 28.7 26.1 29.0 18.9 12.9 A .8 Totals 3.0 31.7 32.8 32.5 38.4 21.9 15.6 3.1 3.0 Plagioclase: ) . L en ss 4.4 3.1 3.0 5.8 4.0 4.0 2.7 1.5 1.2 Boo cece s da rennes aah myy 11.2 12.8 14.8 12.9 8.1 5.9 2.0 1.1 _< cae 44 14.3 15.8 20.1 16.9 12.1 8.6 3.5 2.8 Ilmenite: ______________________ rasa $7. ««2% .8 n a=~- .8 saas -T. 1.0 _____________________ 1.3 13.6 15.8 13.4 14.1 11.4 6.9 A 1.0 'TObAL, 1.8 14.3 15.8 14.2 14.1 12.2 6.9 1.1 2.0 Magnetite: pA c _ OAs y {ug 9 C27 A Als. oe Tee. Total 1 2 .9 2.7 7A 1.0 1.8 A .0 Glagg { 91.2 39.6 34.7 30.3 28.1 52.8 67.1 92.4 92.7 Total; fa oal I[ l————o-o".‘—._. .——"“~\\\ j] Ti0Og Feds: PC x I E ye C . _ \ 4 0 @- -- --- \ - g- --* _ < 0 - € [ at K ] - -=- =G-O X - - of F* S 2s 410 / : \\ : P i s- ost SF y'. 'I. reo s Ys. a f #" A-1,190°C). -Where olivine is the only mineral crystallizing, successive liquid compositions will lie along olivine control lines. The best description of an eruption in which the changing composition of lavas was controlled by olivine is that of 1959 (Murata and Richter, 1966; Wright, 1973). Removal of olivine results in depletion of the liquid in mag- nesia, Cr, and Ni, and to a small extent in iron ; but it also results in enrichment of the liquid in every other chemical constituent. 2. Crystallization of pyroxene and plagioclase in addition to olivine (1,180°C>T>1,070°C).-Once pyroxene and plagioclase begin to crystallize, the liquid is depleted in CaO and SiO, in addition to MgO and is further enriched in TiO, and alkalis relative to the olivine control lines of stage 1.* Total iron begins to increase markedly. The variation of Al,0; in the liquid depends on the ratio of plagio- clase to pyroxene. Generally the Al,O, decreases once plagioclase begins to crystallize in significant amount. Among the trace elements, Ba, Sr, and Ga are partitioned to plagioclase, and Cu, Y, and Yb are partitioned to pyroxene. All of these elements are eventually depleted in liquids collected at tempera- tures below 1,070°C. The bulk composition of Alae lies near the low- magnesia terminus of olivine control. As shown in figures 2 and 4, all three silicate phases appear very close on the liquidus. Nonetheless, the differences between the H-186 and DPH-77 glasses and the bulk composition show the effects of olivine control, particularly in Al,O; and SiO,, as well as the lesser effect of removal of pyroxene and plagioclase (main- ly lower CaO). 3. Crystallization of oxides (1,070°C>T>980°C). -Once ilmenite begins to crystallize, the rate of increase of total iron and TiO, in the liquid dimin- *The analyzed plagioclase and pyroxene from Kilauean basalts, including those from Alae lava lake, are always more siliceous than the bulk composition of the magma. SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII ishes. As crystallization continues, maximum values for TiO, and "FeQ'" are reached, after which both oxides decrease markedly in the liquid throughout the rest of the crystallization history; CaO and MgO continue to decrease. SiO, and alkalis, particu- larly K,0, continue to increase. P,0; is enriched in the liquid to slightly beyond the point where apatite begins to crystallize. V and Nb, and perhaps Zr, are concentrated in the Fe-Ti oxides but are enriched in the lowest temperature liquid (A-5-20) over their initial value. The residual liquid, formed as the end product of tholeiitic differentiation, is of rhyo- litic composition; 95 percent of the chemical com- ponents are those of feldspar and quartz. The- analyses of the residual liquid from A-12-10, -11, recalculated in table 5, can be calculated to the following mineral assemblage: Percent Fe-rich pyroxene 1.2 MaAPNCLILE! LCC Pal ecru cect ca- aka stank se snes 1.3 IImenite ~:~. 25.24... ln rack bear 0.7 Feldspar (OrwAbmARNs) 63.3 QJUATEL . 2222 Ee aas h ann s a mme a i he - male a tie p ce tt oe akai on on te at ue 33.5 This assemblage is like that found in a hedenber- gite granophyre or in an anhydrous rhyolitic pitch- stone. NATURAL DIFFERENTIATES OF KILAUEAN THOLEITE The known natural differentiates of Kilauean tholeiite all correspond in their chemistry to those low-temperature filter-pressed differentiates that are similar to the Alae oozes (Wright and Fiske, 1971, appendix 3). Commonly these appear as fracture fil- lings (segregation veins). The natural differentiates generally correspond to liquids produced near the boundary separating stages 2 and 3 of the liquid line of descent outlined above. Differentiation temperatures range from about 1,040° to 1,065°C. Two factors are believed to operate in restricting the temperature range of formation. First, below 1,065°C the crystal mush provides a stable framework through which the liquid can move. Secondly, open fractures are re- stricted to the brittle crust and cannot penetrate deeper than the crust-melt interface. Hence, the pressure differential necessary to trigger filter pres- sing of liquid is lacking at temperatures higher than that of the crust-melt interface. Although the filter-press mechanism is well estab- lished, the differentiates produced by this mecha- nism will vary in composition according to the effi- ciency of the filter press in separating liquid from crystals. Many natural differentiates show the effects CRYSTALLIZATION AND DIFFERENTIATION OF THE ALAE MAGMA of incomplete separation of liquid and crystals and separation of minerals in a ratio differing from their abundance in the host rock. One example is the natural segregation vein from drill hole 9 (A-9-23, table 6). The chemical mode for A-9-23 may be compared with the ooze A-6-29 in which the calculated amount of liquid is ap- proximately the same (41.4 percent, 42.3 percent). The calculated An content of plagioclase and the Mg/(Mg+Fe) content of augite are greater in A-9-23 than in A-6-29. This is interpreted as the result of small amounts of sodic plagioclase and Fe- rich pyroxene being brought in with the liquid. The data of table 9 show that the higher tempera- ture Alae oozes brought in crystal aggregates which have a higher ratio of plagioclase to pyroxene than was in the liquid before filter pressing. This "plagio- clase" effect, and any other differences in the ratios of minerals brought in with the liquid to those avail- able in the liquid, produce a differentiate which does not, strictly speaking, lie on a "liquid line of descent." This must be taken into account in postulating a parent magma for any basaltic differentiate. At the lowest temperature (1,045°C) and highest crystal- linity (7O percent) of any Alae ooze (A-5-20), the plagioclase effect apparently did not operate. Per- haps in this case, the crystal mush was sufficiently tight that fragments of the crystal framework were brought in directly with no opportunity for differ- ential movement of plagioclase and pyroxene. sUMMARY Study of Alae lava lake provides information on changing amounts and composition of liquid and crystals during crystallization. Modes computed from chemical data are more accurate than modes obtained optically in transmitted light. Mineral pro- portions and compositions vary in a regular and predictable manner. Likewise, the composition of liquid in terms of both major and trace elements also varies smoothly. Among the trace elements, Nb, Li, and Zr are partitioned to the liquid at high temperatures and are increasingly partitioned to crystals at lower temperatures. Cr and Ni are de- pleted at high temperatures by concentration in olivine and spinel. The rest of the trace elements (Ba, Sr, V, Y, Yb) are partitioned between crystal- lizing phases and liquid at all temperatures but all tend to become slightly enriched in the low-tempera- ture liquids relative to their value in the bulk com- position. Differentiation takes place by filtration of liquid C19 through a erystal mush with variable but small retention of the crystal framework in the separated differentiated liquid. REFERENCES CITED Anderson, A. T., and Greenland, L. P., 1969, Phosphorous fractionation diagram as a quantitative indicator of crystallization differentiation of basaltic liquids: Geo- chim. et Cosmochim. Acta, v. 33, no. 4, p. 493-505. Decker, R. W., and Peck, D. L., 1967, Infrared radia- tion from Alae lava lake, Hawaii: U.S. Geol. Survey Prof. Paper 575-D, p. D169-D175. Evans, B. W., and Moore, J. G., 1968, Mineralogy as a function of depth in the prehistoric Makaopuhi tholeiitic lava lake, Hawaii: Contr. Mineralogy and Petrology, v. 17, no. 2, p. 85-115. Lindsley, D. H., 1962, Investigations in the system FeQ- Fe:O;-TiO:»: Carnegie Inst. Washington Yearbook 61, 1961-1962, p. 100-106. Moore, J. G., and Evans, B. W., 1967, The role of olivine in the crystallization of the prehistoric Makaopuhi tholeiitic lava lake, Hawaii: Contr. Mineralogy and Petrology, v. 15, no. 3, p. 202-223. Murata, K. J., Bastron, Harry, and Brannock, W. W., 1965, X-ray determination curve for Hawaiian olivines of com- position Fom-«s: U.S. Geol. Survey Prof. Paper 525-C, p. C35-C37. Murata, K. J., and Richter, D. H., 1966, Chemistry of the lavas of the 1959-60 eruption of Kilauea Volcano, Hawaii: U.S. Geol. Survey Prof. Paper 537-A, p. Al- A26. Peck, D. L., 1978, Cooling and vesiculation of Alae lava lake, Hawaii: U.S. Geol. Survey Prof. Paper 985-B, 59 p. (In press). Peck, D. L., Hamilton, M. S., and Shaw, H. R., 1977, Numerical analysis of lava lake cooling models: Part II. Application to Alae lava lake, Hawaii: Am. Jour. Science, v. 277, p. 415-437. Peck, D. L., and Kinoshita, W. T., 1976, The eruption of August 1963 and the formation of Alae lava lake, Hawaii: U.S. Geol. Survey Prof. Paper 935-A, 38 p. Peck, D. L., Moore, J. G., and Kojima, George, 1964, Tem- peratures in the crust and melt of Alae lava lake, Hawaii, after the August 1963 eruption of Kilauea Volcano-A preliminary report: U.S. Geol. Survey Prof. Paper 501-D, p. D1i-D7. Peck, D. L., Wright, T. L., and Moore, J. G., 1966, Crys- tallization of tholeiitic basalt in Alae lava lake, Hawaii: Bull. Volcanologique, v. 29, p. 629-656. Peck, L. C., 1964, Systematic analysis of silicates: U.S. Geol. Survey Bull. 1170, 89 p. Powers, H. A., 1955, Composition and origin of basaltic magma of the Hawaiian Islands: Geochim. et Cosmochim. Acta, v. 7, nos. 1-2, p. 77-107. Sato, M., and Wright, T. L., 1966, Oxygen fugacities directly measured in magmatic gases: Science, v. 153, p. 1103- 1105. Shaw, H. R., Hamilton, M. S., and Peck, D. L., 1977, Numerical analysis of lava lake cooling models: Part I. Description of the method: Am. Jour. Science, v. 277, p. 384-414. C20 Skinner, B. J., and Peck, D. L., 1969, An immiscible sulfide melt from Hawaii: Econ. Geology Mon. 4, p. 310-322. Swanson, D. A., Jackson, D. B., Duffield, W. A., and Peter- son, D. W., 1971, Mauna Ulu eruption, Kilauea Volcano: Geotimes, v. 16, no. 5, p. 12-16. Tilley, C. E., Yoder, H. S., Jr., and Schairer, J. F., 1964, new relations on melting of basalts, Carnegie Inst. Washington Yearbook 63, 1963-1964, p. 92-97. 1967, Melting relations of volcanic rock series: Carnegie Inst. Washington Yearbook 65, 1965-1966, p. 260-269. Van Der Plas, L., and Tobi, A. C., 1965, A chart for judging the reliability of point counting results: Am. Jour. Sci., v. 263, no. 1, p. 87-90. Wright, T. L., 1971, Chemistry of Kilauea and Mauna Loa lava in space and time: U.S. Geol. Survey Prof. Paper 735, 40 p. Wright, T. L., 1973, Magma mixing as illustrated by the 1959 eruption, Kilauea Volcano, Hawaii: Geol. Soc. America Bull., v. 84, no. 3, p. 849-858. SOLIDIFICATION OF ALAE LAVA LAKE, HAWAII Wright, T. L., and Doherty, P. C., 1970, A linear program- ming and least-squares computer method for solving petrologic mixing problems: Geol. Soc. America Bull., v. 81, p. 1995-2008. Wright, T. L., and Fiske, R. S., 1971, Origin of the differen- tiated and hybrid lavas of Kilauea Volcano, Hawaii: Jour. Petrology, v. 12, no. 1, p. 1-65. Wright, T. L., and Okamura, R. T., 1977, Cooling and crystallization of tholeiitic basalt, 1965, Makaopuhi lava lake, Hawaii:; U.S. Geol. Survey Prof. Paper 1004, p. 1-78. Wright, T. L., Peck, D. L., and Shaw, H. R., 1976, Kilauea lava lakes: Natural laboratories for study of cooling, crystallization, and differentiation of basaltic magma: Am. Geophys. Union, Geophys. Mon. no. 19, p. 375-390. Wright, T. L., and Weiblen, P. W., 1967, Mineral composition and paragenesis in tholeiitic basalt from Makaopuhi lava lake, Hawaii (abs.): Geol. Soc. America ann. mtg. New Orleans.