(College 
 
 This book voas presented 
 bq 
 
J.. .^J'-IJ/) 
 
 5 
 
 This BOOK may be kept out TWO WEEKS 
 ONLY, and is subject to a fine of FIVE 
 CENTS a day thereafter. It is due on the 
 day indicated below: 
 
 i^l^fl^^^ cjg. 
 
THE EUSPORANGIATAE 
 
 THE COMPARATIVE MORPHOLOGY 
 
 OF THE OPHIOGLOSSACEAE 
 
 AND MARATTIACEAE 
 
 BY 
 DOUGLAS HOUGHTON CAMPBELL 
 
 PROFESSOR OF BOTANY, LELAND STANFORD JUNIOR UNIVERSITY 
 
 WASHINGTON, D. C. 
 
 Published by the Carnei^ie Institution of Wasiiinjj;ton 
 
 1911 
 
CARNEGIE INSTITUTION OF WASHINGTON 
 Publication No. 140 
 
 Copies of this Book 
 wertt first issued 
 
 Allb2919n 
 
 PRESS OF ISAAC H. BLANCHARD COMPANY 
 NEW YORK 
 
PREFACE. 
 
 The great importance of the eusporangiate ferns as the nearest existing rela- 
 tives of the Paleozoic ancestors of the higher types of Hovvcring plants invests 
 them with an especial interest tor the student of plant evolution; and no apology 
 is necessary for presenting at length a summary of our present knowledge of the 
 structure and development of these important plants. 
 
 For more than twenty years the writer has been much interested in the study 
 of the Eusporangiatre and during this time has published a number of works deal- 
 ing with them. He has had rather unusual opportunities for collecting these forms, 
 and an extensive and representative collection of materials comprising a good many 
 species has been accumulated, so that the time seemed ripe for a comparative study 
 of the group, for the purpose of determining, as far as might be, the relationships 
 existing between the different genera, as well as for throwing some light upon the 
 question of their position in the great series of ferns. 
 
 The present memoir is an attempt to present the results of these studies, based 
 mainly upon the writer's own materials, but supplemented by a careful study of 
 the work of other investigators who have described the structure and development 
 of the EusporangiatsE. 
 
 The writer's grateful acknowledgments arc due to a number of colleagues who 
 have rendered assistance in various ways. Especially is he indebted to Prof. E. C. 
 Jeffrey, through whose kindness a large number of admirably preserved prothallia 
 and young plants of Botrychnim virginimmm were sent the writer, and in addition 
 a number of valuable slides of the same. Without this material the work on 5o- 
 trychium would have been very incomplete. 
 
 To my colleague in Stanford University, Prof. L. L. Burlingame, thanks are 
 due for valuable assistance in the preparation of the photographic plates, as well as 
 for the use of a number of important sWAts oi Ophioglossum and Helminthostachys. 
 
 To Prof. J. C. Willis, of the Botanic Gardens at Peradeniya, the writer would 
 express his appreciation of many kindnesses and assistance in collecting during his 
 stay in Ceylon. 
 
 It was the good fortune of the writer to enjoy the unequaled facilities for 
 collecting material offered by the great gardens at Buitenzorg and Tjibodas in Java, 
 where, through the interest and courtesy of the distinguished director. Professor 
 M. Treub, whose recent death was such an irreparable loss to science, means were 
 afforded for securing the most valuable materials used in the preparation of the 
 present work. 
 
 Douglas Houghton Campbell. 
 
 Stanford University, April, 1910. 
 
 50568 
 
TABLE OF CON'lENTS. 
 
 Pkefack iii 
 
 Introuuction 3-4 
 
 PaKT I. Till-: Ol'HIOGLOSSALES 
 
 I. Thh (Jamhtophv,. b-ii 
 
 Germination in Ophioglossum 7 
 
 The Adult Cianietophyte of Ophiogloisum 10 
 
 The Histology of the Camctophyte 15 
 
 The Gamctophyte of fici^rvi/""'" 16 
 
 The Histology of the Gametophyte of Z?o?,v. /"",/, 18 
 
 The Gametophyte of f/f/m;n?/(oi/ii( 7/ vx 19 
 
 The Endophyte 21 
 
 The Sexual Organs 22 
 
 The Antheridium 22 
 
 The Antheridium of Of>liioglnssunt 22 
 
 The Antheridium of i?o// J. /hh/h 24 
 
 The Antheridium of Hrlnunthostiuhys 25 
 
 Spermatogenesis 26 
 
 The Archegonium 28 
 
 The Archegonium of O/j/dVvf/ojjHm 28 
 
 The Archegonium of 5ofr)i(7(/«m 30 
 
 Fertilization ^I 
 
 Fertilization in 5o/ry(7(n/m r;;f;H/V;HH?H 31 
 
 Significance of the Endophyte 32 
 
 II. The Embryo 34-54 
 
 The Embryo of Ophioglossum 34 
 
 The Development of the Primary Bud in 0/)/»oi;/oj-X!i/« mo/wi I (»H("; 40 
 
 The Embryo o( Ophioglossum vul^atum 4^ 
 
 The Anatomy of the Young Sporophyte of 0/)/;/Q^/ojJum 44 
 
 The Embryo of 5o^rvi7i/um 46 
 
 The PZmbryo of Hehmnthostachys 54 
 
 in. The Young Sporophyte 55-84 
 
 The Young Sporophyte of Ophioglossum 55 
 
 The Young Sporophyte of Botryrhium 59 
 
 The Young Sporophyte of //<7m(H(/if)jM<7n'.f 67 
 
 Comparison of the Young Sporophytes of the Ophioglossaceae 82 
 
 IV. The Adult Sporophyte 85-116 
 
 The Sporophyte of Ophioglossum 86 
 
 The Anatomy of £Mo/)/i/o^/oj.fwm 89 
 
 The Root in Euophioglossum 93 
 
 Anatomy of OphtoJerma , 94 
 
 Anatomy of Cheiroglossn 99 
 
 The Sporophyte of 5o(ry<:/"«m 99 
 
 The Sporophyte of Hclminthost/uhys 104 
 
 The Sporangiophore of the Ophioglossales 108 
 
 The Development of the Sporangiophore . . 1 10 
 
 The Development of the Sporangium 112 
 
VI CONTENTS 
 
 Part II. The Marattiales. 
 
 I. The Gametophyte 119-134 
 
 The ProthaWium of KduljnssKi 122 
 
 The Prothallium of Z)rtn<f(i .... 124 
 
 The Endopliyte of the Marattiaccs 127 
 
 The Sexual Organs 128 
 
 The Antheridium 128 
 
 Spermatogenesis 129 
 
 The Archegonium 130 
 
 Fertihzation 134 
 
 II. The Embryo i35-'59 
 
 The Embryo oi Marattta 136 
 
 The Embryo of Angiopterts ijg 
 
 The Y.mh\yo oi Kauljussia 141 
 
 The Embryo of Z)a««a 142 
 
 The Anatomy and Histology of the Young Sporophyte 145 
 
 The Cotyledon .... 146 
 
 The Stem of the Young Sporophyte 153 
 
 The Root 155 
 
 The Second Leaf 156 
 
 III. The Older Sporophyte 160-208 
 
 The Development of the Vascular System in Z),;nrSrt 160 
 
 The Adult Sporophyte of Dtiiura 175 
 
 The Anatomy of the Leaf 177 
 
 The Apical Growth of the Roots 178 
 
 The Sporophyte of Kauljussia i yg 
 
 The Sporophyte of Marattta 1 88 
 
 The Adult Sporophyte of A/rtra«(a Ig^ 
 
 The Sporophyte of Angiopterts igj 
 
 The Adult Sporophyte of Angiopterts 200 
 
 Archangioptcris 203 
 
 Macroglossum 204 
 
 Tissues of the Marattiaceas 204 
 
 The Sporophyll of the Marattiaceae 204 
 
 The Sporangium of the Marattiaceae 205 
 
 Part III. 
 
 The Origin and Relationships of the Eusporangiatae 209 
 
 Conclusion ... 217 
 
 Bibliography 219 
 
 List of Plates 223 
 
 Inde.x 225 
 
THE EUSPORANGIATAE 
 
 THE COMPARATIVE MORPHOLOGY 
 
 OF 
 
 THE OPHIOGLOSSACEAE AND MARATTIACEAE 
 
INTRODUCTION. 
 
 It is pretty well ao;rcf(l ammio; botanists that most of the seed plants, perhaps 
 all of thcni, are dcscemlants of some fern-like Paleozoic ancestors. The geological 
 record is remarkably perfect in many respects, and our knowledge ofniany of these 
 Paleozoic fossils is extraordinarily complete. Among the most important con- 
 tributions made of late years to our knowledge of these Paleozoic fossils is the fact 
 that many of these Paleozoic "ferns" were really seed-bearing plants intermediate 
 in character between the true ferns and the more highly develo|)ed Howering plants. 
 
 At best the fossil record is very incomplete in regard to many extremely impor- 
 tant structural details, and therefore it is especially necessary that these points 
 should be thoroughly studied in such of the existing ferns as, for any reason, seem 
 to be at all closely related to the ancient Paleozoic types. 
 
 The name luisporangiane was proposed by Goebcl to include the two very 
 peculiar families of fern-like plants, the Marattiaceae and Ophioglossacc.x, which 
 differ in several important respects from the much more numerous and specialized 
 Leptosporangiatac, the predominant ferns of the present day. The Eusporangiatas 
 comprise lOO or more species of widely distributed ferns, of which the Marattiaceae 
 are mainly tropical in their distribution, while the Ophioglossaceae include a good 
 many species of temperate regions as well. The Kusporangiatae are distinguished 
 primarily by the character of the sporangium, which is always much more massive 
 than in the typical ferns, the Leptosporangiatae. In the latter, the sporangium can 
 almost always be traced back to a single mother cell which usually arises fnim the 
 surface of the leaf, while in the Eusporangiatse the sporangium is already multi- 
 cellular when it is first recognizable. 
 
 The Marattiaceae are in general appearance much like the typical ferns, and 
 there is no question of their relationship to the Leptosporangiatae. The resem- 
 blances are less obvious in the case of the Ophioglossacea?, and some students of 
 the ferns have expressed the opinion that the Ophioglossacea? should be separated 
 entirely from the ferns and placed in a distinct class. (See Bower 9,) A careful 
 comparative study, however, of the two families included in the Eusporangiata' shows 
 so many close correspondences in structure, both of the sporophyte and gameto- 
 phyte, that their association together is amply justified. 
 
 The MarattiaceiP are known to be very ancient forms, unmistakable members 
 of this family occurring abundantly in Palet)zoic formations. I he Ophioglossacea", 
 on the other hand, are very unsatisfactorily represented in a fossil condition, and 
 for this reason doubt has been thrown on their antiquity, although their structures 
 show strong evidences of an extremely primitive character. Certain of the oldest- 
 known fossil ferns, the Botryopteridea-, may possibly prove to be relateil to the 
 Ophioglossacex, but the evidence at present is not entirely conclusive. 
 
 In spite of the unsatisfactory nature of the geological evidence, I am neverthe- 
 less strongly inclined to believe that the Ophioglossaceae, on the whole, represent 
 the most ancient type among the living ferns, this conclusion being based upon a 
 very careful study of the structure of both gametophyte and sporophvte. 
 
 There are many practical difhculties in the way of studying these interesting 
 ferns, and these difficulties undoubtedly account for the comparatively small number 
 of researches that have been made upon their development. A few species, like 
 Ophioglossiiiii vtilgattnii and several species of Butrwh/tnii, arc widespread in their 
 
 N. C. State CoUego 
 
^ INTRODUCTION 
 
 range and occur in the temperate regions of Europe and America; but very much 
 the larger number of these, including all of the Marattiace.ne, are tropical or sub- 
 tropical plants and are very seldom seen in cultivation, hence material for their 
 study must be collected in their natural habitat. 
 
 The gametophytes, or sexual plants, are seldom encountered unless very 
 special search is made for them, and they have been but rarely collected; and even 
 where the spores can be made to germinate, the long time necessary for the growth 
 of the gametophyte under artificial conditions, and the difficulties of rearing them 
 to maturity, readily explain the small number of successful attempts to study these 
 plants under artificial conditions. 
 
 As the result of several visits to the tropics of both the Old and New Worlds, 
 the writer has collected material of all of the eusporangiate genera, except the 
 recently discovered Archangiopteris from southwestern China and Macroglossiim 
 from Borneo. It has seemed desirable, with this material as a basis, to make a some- 
 what comprehensive comparative study of the whole group of the Eusporangiatae. 
 In this study, especial attention will be devoted to the structure and development of 
 the gametophyte and embryo and to the development of the young vegetative organs 
 of the sporophyte. The general anatomy and histology of the mature sporophyte 
 have already been pretty satisfactorily studied and described for most of the genera 
 and a particularly satisfactory account of the development of the sporangium has 
 been given us in the comprehensive memoirs of Professor Bower (Bower 5, 6), and a 
 detailed investigation of these points has hardly seemed necessary, although, so far 
 as possible, the results of the work of previous investigators have been verified. 
 
 The writer has already published various papers upon the development of 
 both the Marattiaceae and Ophioglossaceae, and these have been freely drawn upon 
 in the preparation of the present monograph; but these earlier studies have been 
 materially extended by further work upon the species previously studied and in- 
 vestigations have been made upon a number of species which hitherto have not 
 been examined at all, or have been studied only very incompletely. This is partic- 
 ularly true of the genus Daneea, which has received comparatively little attention 
 hitherto. An especially fine lot of material of several species of Daneea was col- 
 lected in Jamaica in 1908 and an extended study of the gametophyte and young 
 sporophyte has been made, the complete results of which are now published for 
 the first time. Attention has also been given to the development of the young 
 sporophyte in the other eusporangiates, particularly in the peculiar genus Kaul- 
 fussia, which, like Daneea, has received less attention from the students of these 
 plants than have the genera Marattia and Angiopteris. 
 
 Of late there has been a tendency, especially among students of the fossil ferns, 
 to lay great stress upon the importance of the structure of the vascular skeleton of 
 the ferns in the study of their phylogeny. It is therefore highly important that a 
 careful examination should be made of the origin and development of the vascular 
 system in those forms, i. e., the Eusporangiatae, which, there is good reason to sup- 
 pose, are the nearest relations among living plants of the ancient Paleozoic ferns. 
 
 The results of the writer's studies on the development of the fibro-vascular 
 bundles of the sporophyte in all of the eusporangiate ferns have led to some rather 
 unexpected conclusions as to the real nature of the vascular system in these ferns, 
 conclusions decidedly at variance with the views generally accepted at present. 
 These results have seemed sufficiently important to warrant a more extended treat- 
 ment of this subject than was at first contemplated; and this will explain what 
 might otherwise seem to be a disproportionate amount of space devoted to the origin 
 and development of the fibro-vascular system in the young sporophyte. 
 
PART 1. THi-: oi^hio(;lossales. 
 
 IIk' Ophioglossales, the "adder-tongue" ferns, include three genera (of which 
 two arc practically cosmopolitan) and embrace numerous species, being sometimes 
 further divided into several subgenera; the third genus is a monotypic one, confined 
 to the tropics of the Old World. These three genera are evidently (juite closely 
 related among themselves and may without hesitation be referred to a single family, 
 the C)phioglossace;e. The genus Ophioglossum includes several species of the 
 temperate regions of the whole world, and some of the species, like O. vulgatum, 
 are very widespread; the tropical species are much more numerous, but the number 
 of these is very uncertain and much confusion exists as tt) the limits of certain species. 
 The second genus, Botrycliiinu, belongs mainly to the temperate regions, only a 
 small number of species occurring in the tropics, and these are principally confined 
 to the cooler mountain regions. 1 he third genus, Helmmthostnchys, with a single 
 described species, H. ze\liiiii\a, is a not uncommon fern of the lowland forests of 
 the Indo-Malayan region. 
 
 rhe Ophioglossacea; are for the most part quite glabrt)us plants, of small or 
 moderate size. The smallest species, like Botrychium simplex and some of the 
 smaller species of Ophioglossum (e. g., 0. californicum, O. bergtannm, and some of 
 the smaller forms of O. nioliicc/iniim) may be only 5 or 6 centimeters in height. The 
 largest species of Botrychium and Hclminthostachys sometimes attain a height of 
 50 centimeters or more, with ample, much-dissected leaves, while the long, ribbon- 
 like, pendent leaves of the epiphytic Ophioglossum pendulum, the giant of the family, 
 may reach a length of 1.5 meters. 
 
 Except for two species of Ophioglossum belonging to the sections O phioderma 
 and Cheiroglossa, the Ophioglossaceae are terrestrial plants, usually growing in soil 
 abounding in humus, and the gametophyte in all cases is a subterranean structure 
 quite destitute of chlorophyll. 
 
 The stem is a rhizome, which is short and upright in the terrestrial species t)l 
 Ophioglossum and in Botrychium, but is dorsiventral in Helminthostachys and the 
 epiphytic species Ophioglossum pendulum. In most of the species of the temperate 
 regions only a single leaf is developed each year, but there are some exceptions to 
 this rule in Ophioglossum, especially in the tropical species, where there is no 
 interruption of the growth. 
 
 The leaves are usually ample, and may be (juite undivided, as in most species 
 of Ophioglossum, or they may be dichotomously divided in O. pnlmnium and some 
 forms of O. pendulum; or they may be much dissected, usually in a ternate fashion, 
 in Botrychium and Helminthostachys. The sporangia are sometimes very large, 
 and are borne upon characteristic spikes, or "sporangiophores," whose morpho- 
 logical nature is a matter of some controvers\'. The venation of the leaves is 
 reticulate in all species o( Ophioglossum, but in the other genera ir resembles that 
 of the typical ferns. 
 
 With the exception of Ophioglossum (Cheiroglossa) palmatuni, the sporangio- 
 phore is normallv attached to the adaxial side of the leaf, usually near the junction 
 of the laminii and prtiole; but sometimes it is inserted much lower down, or it 
 
THE OPHIOGLOSSALES 
 
 may arise close to the base of the petiole and appear to be quite independent of 
 the sterile leaf segment. The sporangia range in number from about half a dozen, 
 in some of the smaller forms of Botrychtum simplex and Ophioglossum, to many 
 hundred in such large species as Botrychium vtrgtnianiim and Helminthostachys. 
 There is a most interesting series of forms connecting the smaller and simpler types 
 with the large and complicated ones. With the increasing complexity of the sporan- 
 giophores, there is usually a reduction in the size of the sporangia, which, however, 
 become better differentiated than in the simpler types. 
 
 There is no question about the close relationships existing between the dif- 
 ferent genera of the Ophioglossaceae, but there is some difference of opinion as to 
 their connection with other Pteridophytes. While they are probably sufficiently 
 distinct to be relegated to a special order, Ophioglossales, their peculiarities hardly 
 warrant separating them entirely from the ferns. As will be seen later, they offer 
 many points of resemblance to the Marattiales, both in regard to the structure of 
 the adult sporophyte and that of the gametophyte or sexual plant, and the early 
 phases of embryonic development. These resemblances are too numerous and too 
 exact to make it at all likely that the two orders are unrelated. 
 
 I. THE GAMETOPHYTE. 
 
 The first discovery of the gametophyte of the Ophioglossaceae was made by 
 Hofmeister (Hofmeister 1) who, in 1854, found the gametophyte of Botrychium 
 liiuaria; two years later Mettenius (Mettenius 1) described much more fully the 
 prothallium of Ophioglossum peduncuhsum, which was cultivated in the botanical 
 garden at Leipzig. No further additions were made to the subject until the writer 
 (Campbell 4) succeeded in obtaining the first germinating stages of O. pendulum, 
 collected in the Hawaiian Islands, and those of Botrychium virginiatium; and the 
 older gametophyte of the latter species was also described. In the year 1898 Jeffrey 
 (Jeffrey 1) published a complete account of the gametophyte and embryo of Bo- 
 trychium virgmianum. Later contributions to the subject are those of Lang 
 (Lang I) on the gametophyte of Ophioglossum pendulum and o{ Helminthostachys; 
 and the papers of Bruchmann (Bruchmann 1, 2) on the prothallium of Botrychium 
 lunaria and Ophioglossum vulgatum. In 1905, Lyon (Lyon 1) published a brief 
 account of the embryo of Botrychium ohliquum. 
 
 In 1906 I collected in Ceylon a few prothallia of a species of Ophioglossum 
 which may have been O. reticulatum, and shortly afterward, during a stay in Java, 
 a number of prothalliaof O. moluccanmn were found, as well as a few belonging to 
 an undetermined species associated with the latter. During my stay in Java I was 
 fortunate enough to obtain also a large number of prothallia of O. pendulum. An 
 account of these has already been published (Campbell 8). 
 
 All species of the Ophioglossaceae that have yet been examined agree in the 
 underground habit and saprophytic nature of the gametophyte, which always has 
 associated with it a peculiar endophytic fungus, or "mycorrhiza," which is un- 
 doubtedly connected with the assimilation of organic food. Mettenius states that 
 chlorophyll may be developed if the prothallium appears above ground, and Bruch- 
 mann found that this was also true in O. vulgatum; but as a rule the prothallium 
 remains subterranean and quite destitute of chlorophyll. 
 
 The spores of the Ophioglossaceae are always of the tetrahedral type, and when 
 ripe possess a moderately thick sculptured outer membrane, which is usually colorless 
 or pale yellow, so that the masses of spores are either white or a pale sulphur-yellow 
 tint. The ripe spore as a rule is packed with granular matter, which make.s the 
 
CAMI-lOl'in IF 
 
 contents appf;ii ()|i;i(|iii- iiiul olisi'iiits mi>ic or kss ihv i(nii;ill\ |)l;R-ed nucleus. 
 A marked e.\ct|ni(>n to the oidinar) tv|>i- ot'spdii- \\;is ih;ii Iduiul in .m luuietermined 
 species o( Opliioglossuin collected at Hiiitenzorg in Java. 1 his was supposed to be 
 O. moliiccfimini, with which it was {^rowinj^;, hut a comparison with typical specimens 
 of the latter species showed marked tlitterences, the most stiiking being the spores, 
 which were larger than in the type, had much less dense contents, and were espe- 
 cially notable in that they had regulaily two nuclei, a condition unique, so far as I 
 know, among the ferns. (See Campbell 8, fig. 157.) The granular contents of the 
 spores include numerous albuminous granules, together with more or less starch 
 and oil. 
 
 (;krmination in oimiioglossum. 
 
 1 he first successful attempts to germinate the spores of Ophioglossiini were 
 made by me in 1892, when ripe spores of the epiphytic (). pendulum were collected 
 in Hawaii and brought to California. 1 his species was found in Hawaii, growing 
 usually upon the trunks of tree ferns, and the spores were sown upon bits of the 
 bark-like masses of roots, which in the commoner tree ferns of Hawaii (species of 
 Cihotiuni) cover the trunk with a thick, felted mass. These masses of roots were 
 kept in jars, and the spores were sown upon them, a good many of them germinating 
 in course of time. Fhe germination was very slow, the spores often remaining 
 unchanged for months, and none of these young prothallia developed beyond the 
 stage with three cells. This failure to develop further was undoubtedly due to the 
 fact that they did not become infected with the mycorrhiza, which is essential to the 
 full development of the prothallium. 
 
 In 1906 ripe spores of the same species were collected in Ceylon and Java. 
 In Ceylon spores were secured at the botanical garden in I'eradeniya and in the 
 Barrawa Reserve Forest near Hanwella, where Lang obtained his material. In Java 
 the spores were mostly collected at Tjibodas. In all of these later experiments the 
 spores were sown in humus collected from about the base of the spore-bearing 
 plants. These masses of wet humus were kept in stoppered bottles. 
 
 As in the earlier experiments, 
 the germination was slow, the first 
 germinating stages being found only 
 after a month or more, and in some 
 cases the spores remained un- 
 changed for a very long period. The 
 bottles containing -the spores sown 
 upon the humus were brought back 
 to California, and a recent exami- 
 nation (September 1909) showed a 
 considerable number of apparently 
 normal spores, as well as living 
 three-celled prothallia, presumably 
 the result of comparatively recent 
 germination. 
 
 Spores sown in Tjibodas on 
 April 18, 1906, were first found 
 germinating at Buitenzorg on May 
 24, germination at this time being pretty well advanced. On June T, a number of 
 these had three cells. The germination (fig. i ) in all cases coi responded closely with 
 the writer's former observations and closely resembles that which we shall see pres- 
 ently in OpIiKiolfissuru uKilucniiiuiii. In no case could any chloroplull be detected. 
 
 (icrmin.iting spore of Ophioglouum pendulum. X360. 
 Optical section of A. 
 
 Three views of a very young gametophyte of 0. pendulum, showing 
 infection by mycorrhizal fungus. X360. 
 
8 THE OPHIOGLOSSALES 
 
 and apparently the prothallium of 0. pendulum is strictly saprophytic throughout 
 its existence. 
 
 The extremely favorable conditions for plant growth at Buitenzorg made this 
 an unusually promising place for studying germination, and very soon after my 
 arrival there a quantity of plants of Ophioglossum moluccaiiutn, or what was sup- 
 posed to be this species, was secured, and the spores were sown. Subsequent study 
 showed that at least three species grew together at Buitenzorg, so that it is not at 
 all certain that all of the germinating spores and the prothallia which were col- 
 lected later really belonged to O. mohiccanum. 
 
 Since in all previous experiments with the Ophioglossaceae germination had 
 been very slow, it was with much surprise that the first lot of spores that were sown, 
 when examined a week later, were found to be germinating very freely and had 
 evidently been growing for some days. New sowings were made, and in some cases 
 the first germination stages were evident within three days from the time the spores 
 were sown. Inasmuch as the spores contain no chlorophyll, this rapid germina- 
 tion is really remarkable. In the most favorable instances the greater part of the 
 spores germinated and many thousand germinating spores were studied. 
 
 The first sowings were made upon earth taken from where the plants were 
 growing; the earth was placed in small glass dishes and flooded with water; the 
 spores were then scattered over the surface of the water, some sinking, but the 
 greater part floating on the surface. In later experiments cavities were hollowed 
 out in the earth and these were filled with water, while the rest of the earth was 
 left wet, but not flooded. It was found that the spores germinated more promptly 
 in the water than on the wet earth, and this suggested that probably under natural 
 conditions germination occurs where the spores fall in slight depressions which are 
 filled with water for a time after heavy rains. The prothallia oi Hehninthosiachyi, 
 to judge from the locality where they were found in the Barrawa Forest, occurred 
 only where the forest was subject to inundation, and it may be that immersion in 
 water is necessary for the germination oi Ophioglossum moluccanum, or at any rate 
 facilitates germination. The older prothallia of this species which were found in 
 Buitenzorg were growing in low ground between the projecting roots of trees, where 
 water might very well stand for some days in wet weather. 
 
 The first sign of germination consists in the enlargement of the spore contents, 
 which soon burst the rigid outer membrane along the three lines upon its ventral 
 face, and through the cleft thus formed the spore contents, surrounded by the col- 
 orless inner membrane ("endospore" or "intine") protrude as a blunt papilla 
 (plate I, fig. 2). The first division wall is transverse and is soon followed by a 
 second wall in the upper cell (i. e., the one turned away from the opening). The 
 second wall is at right angles to the first, and the young prothallium now consists of 
 three cells, the basal one, which is partly exposed through the cleft at the upper side 
 of the spore, and the two upper cells, which have more dense contents than the 
 basal one. The basal cell in position corresponds to the first rhizoid in the germi- 
 nating spore of the typical ferns, but it was never found to become extended into a 
 true rhizoid, and no rhizoids were seen in any of the very young prothallia. 
 
 While Ophioglossum moluccanum and O. pendulum agree very closely in the 
 early stages of germination, in the latter species no trace of chlorophyll can be 
 detected at any time, but in O. moluccanum it is not uncommon to find from one to 
 three chloroplasts present. These chloroplasts are very pale in color, but it is 
 certain that a small amount of chlorophyll is present in some cases. 
 
 While most of the young prothallia oi Ophioglossum moluccanum do not advance 
 beyond a three-celled stage, a few were found in which there were four cells, but all 
 
THr-: c;ami-toi>hvti-; y 
 
 attempts to cany tluiii 1)c\()ik1 this stage failed and after exhaiistinir the spore con- 
 tents they soon died. Ihe small amount of chlorophyll (jccasionall)- found in these 
 young prothallia is evidently insufficient for their independent growth, and after a 
 few weeks the granular contents all disappear and the cells soon collapse, showing 
 that the young gametophyte has died from starvation. 
 
 A number of specimens of the rare Ofiliioglossiun intcniircltiini were collected 
 near IJuitenzorg and a few ripe spores were also secured. I'hese spores contain 
 somewhat less dense contents than those of the other species described and have a 
 more delicate epispore. Spores were sown on March 30 in Buitenzorg, and when 
 examined about two weeks later no germinations were found, nor did a second 
 examination about the end of April show any further results. On May 21, however, 
 two three-celled prothallia were found, and subsequently a small number of others, 
 but no later stages were discovered. The young prothallia appeared in every 
 respect like the similar stages in the other species (plate i, fig. 6). 
 
 In Ophioglossum periclulum a number of young prothallia were found which 
 had increased very much in size and undergone further division. 'Ihe first of 
 these were observed on April 3, and had developed from spores sown in Peradeniya 
 on February 9. These young prothallia (fig. i, ,) had from four to six cells. It 
 was found that in each case the mycorrhiza had connected itself with the young 
 prothallium, and evidently had caused a stimulus in its growth. In every case 
 where the young prothallium had more than three cells there was found associated 
 with it the mycorrhiza, which could be easily seen to penetrate into the basal cell. 
 The infection was in all cases due to fragments of mycelium, and in no case to any- 
 thing which could be interpreted as spores. The fungus was apparently growing 
 free in the humus where the spores were sown. This soil, as we shall see, was taken 
 from about the roots of the sporophytes which furnished the spores. 
 
 The free surface of the basal cell has its wall decidedly thickened, and it was 
 here that the infection took place in all the specimens seen. The branching mycelium 
 of the mycorrhiza was closely applied to the surface of the cell and a haustorium was 
 sent down through the cell wall into the basal cell (fig. i, m). This haustorium is 
 pointed at first, but after it penetrates into the cell it enlarges and assumes the form 
 of a somewhat thickened worm-shaped body, much thicker than the free mycelium 
 outside. In the cell infected with the fungus, the contents show the peculiar aggre- 
 gated appearance characteristic of the infected cells of the older prothallia. On 
 the 6th of April a specimen with seven cells was found. About a month later, a 
 number of other specimens were observed also, some of these having as many as 
 thirteen cells (plate i, fig. 9). This was the largest number found in any of the 
 young prothallia. 
 
 The divisions of these young prothallia are mainly in the lower cells, so that 
 the apex, as in the prothallium of the true ferns, develops mainly from the lower of 
 the two original prothallial cells. The basal cell, however, al.so undergoes divisions, 
 and there is no very marked difference between the lower and upper ends of the 
 prothallium. At this stage there is a marked resemblance, except for the absence 
 of chlorophyll, to the early germination stages o( Lvropocliiini ceriniinu (Treub I). 
 The mycorrhiza penetrates the cells adjacent to the one first infected, but leaves the 
 apical region free, and this region probably remains permanently free from the 
 endophyte, as it does in the adult prothallium. 
 
 'ihe number of young prothallia found was too small to make it possible to 
 determine exactly what may be considered to be the normal succession of cell divi- 
 sion, and whether at this early stage there is a definite apical cell could not In- decicUd. 
 As will be seen from the figures, there is evidently a good deal of variation in the 
 
10 THE OPHIOGLOSSALES 
 
 early divisions. In these larger prothallia there is already the beginning of an axial 
 tissue. Whether the cell v (plate i, fig q) is to be regarded as an apical cell, it would 
 be hard to say. 
 
 Owing to my departure from |ava about three months after the first observa- 
 tions were made, it was impossible to trace the development of the prothallia further, 
 but this much is certain — without the infection of the fungus, growth will not pro- 
 ceed beyond the three-celled stage, and apparently in 0. pendulum no chlorophyll 
 will develop under any conditions, and the prothallium from its earliest stages must 
 be considered saprophytic in its nutrition. Whether the oval body described as the 
 product of germination is to be considered as a sort of tubercle, such as is found in 
 Lycopodium cernuum, must be decided by further investigations. Lang's descrip- 
 tions and figures of the smallest specimens which he discovered would indicate 
 that this is not the case in 0. pendulum; but the tuberous body usually found at the 
 base of the older prothallia in 0. moluccauuni (and this is true also in O. vulgatum) 
 would indicate that in these species it is not impossible that a primary tubercle 
 is first formed and subsequently the fertile branch. 
 
 THE ADULT GAMETOPHYTE OF OPHIOGLOSSUM. 
 
 In 1856 Mettenius (Mettenius 1) found the gametophyte of O. pedunculosum 
 Desv., a tropical species, growing spontaneously in the pots where the plants had 
 been cultivated in the botanical garden at Leipzig. He did not succeed, however, 
 in making the spores germinate. These prothallia were slender, subterranean 
 bodies, sometimes branched, sometimes without branches. They ranged in length 
 from 1.5 lines to 2 inches (fig. 3, A, B). There was usually present a basal enlarge- 
 ment or tuber, from which the fertile part of the prothallium extended. The older 
 portions were brownish in color; the growing tips of the branch white. From the 
 surface there grew numerous short brown rhizoids. Archegonia and antheridia 
 grew more or less intermingled and were formed in large numbers. Except for the 
 greater size, these prothallia closely resemble those of O. moluccanum collected by 
 me in Buitenzorg; and as O. pedunculosum Desv. has been held to be a synonym 
 of O. moluccaniim-^chXtcht, it is possible that Mettenius's plants were the same as 
 those found by me growing in Buitenzorg in Java.* 
 
 The next account of the prothallium oi Opliioglossum is that of Lang (Lang 1). 
 He collected in Ceylon specimens of the prothallia of Opliioglossum pendulum. 
 These were found in the Barrawa Reserve Forest, not far from Colombo, and were 
 buried in the humus accumulated between the leaf bases oi Poly podium quercifolium, 
 an epiphytic fern to which O. pendulum is often attached. I visited this same locality 
 in February, 1906, but was unsuccessful in collecting the prothallia, although I 
 obtained numbers of the sporophytes. Some time after, however, when in Java, I 
 found a very large number of prothallia which were growing in much the same way 
 as those collected by Lang, except that in this case the fern to which the Ophio- 
 glossum was attached was the widespread bird's-nest fern, Asplenium nidus. 
 
 Bruchmann has given a detailed account of the prothallium of the widespread 
 O. vulgatum, which agrees closely in its essential details with 0. pedunculosum and 
 O. moluccanum. Bruchmann's specimens were collected in the Thuringian Forest, 
 in a depression that w*as subjected at times to overflow, a condition paralleled by the 
 locations where the prothallia of Helminthostachys were collected by Lang and 
 
 * Cristensen in his recent Index Filicum (1906) regards 0. moluccanum as a synonym of 0. pedunculosum. Through the kind- 
 ness of Professor K. Goebel, I recently had an opportunity of examining some specimens of O. pedunculosum growing in the 
 botanical garden in Munich. These plants were the descendants of the specimens in Leipzig described by Mettenius and certainly 
 very closely resembled the typical O. moluccanum from Buitenzorg. 
 
THE GAMF.TOPHYTK 11 
 
 mysfir, ;in(l ;il.s(i liki- the l()calit\' wlun- inosr of my s|Hi.iiiun.s of (J. tnoliimniiirn 
 were iouiul in |ava. 1 liis, in (.onnection uitli my expLiimciits in gciininating the 
 spores of O. nwliKcamitti, makes it not unlikely that actual submersion is a necessary 
 condition for the germination in the terrestrial species ot Ophio^losstitn. 
 
 In April 1906, about fifty prothallia of O. nioluccanuni were collected by me 
 at Buiten/org, where, as we have seen, this species is extremely abundant. Only a 
 small number of these were young enough to show the young reproductive organs, 
 and most of them had already developed the young sporophyte. The greater part 
 of these prothallia were found growing together in a slight depression between the 
 projecting roots of a tree. 
 
 These prothallia were slender bodies, from 5 to 10 millimeters in length, and none 
 of them branched. They showed a more or less conspicuous basal tuber, like that 
 described by Mettenius for O. peJunculosum, and indeed they very closely resembled 
 his figures of the simpler forms of that species, but are very much smaller. The 
 youngest specimen found (plate i, fig. 10) consisted of a small, irregular tuberous 
 body of a brownish color, from which grew a short appendage or branch, the tip of 
 which was white. The older ones also showed this basal tuber, but the cylindrical 
 branch was much longer. Owing to their slender form, the prothallia are not always 
 easily distinguishable from roots, and in some cases a microscopic examination is 
 necessary before one can be sure of their real nature. Growing from the surface are 
 scattered short brown hairs like those described by Mettenius for 0. peJunculosum. 
 Archegonia and antheridia are formed at an early period, and can be traced to the 
 base of the fertile branch, or in some cases may be found even upon the tuber. In 
 these specimens the reproductive organs seemed to be formed in rather smaller 
 numbers than is the case either in 0. pednnciilosum or O. vulgatum. Among the 
 prothallia was one very much larger than the others which had very large numbers 
 of old archegonia. It is highly probable that this represents a second species, but 
 unfortunately there was no means of determining to which of the two or three 
 forms associated under the name 0. moluccmnnn it belonged. 
 
 The cells of the tuber and the lower part of the branch contain the character- 
 istic endophytic fungus, but the greater part of the fertile branch is quite free from 
 this, and the cells appear almost transparent, but they contain numerous small starch 
 granules. The endophyte is much more abundant in the cells of the tuber. The 
 hyphre, which stain readily with gentian violet, are irregular in outline and branch 
 freely. Very often branches can be seen piercing the walls of adjacent cells. 
 
 In the living condition the pointed apex of the prothallium is pure white, and 
 even with a hand lens the projection of the antheridia is clearly evident. Mettenius 
 noticed this "varicose"' appearance t)f the smaller prothallia in 0. pedunculosuvi. 
 A median section of the branch shows that the end is decidedly pointed and has a 
 clearly defined apical cell. Owing to the very small amount of material available, 
 no satisfactory transverse section of the apex could be made, and it can not be told 
 whether the apical cell in transverse section is three-sided, as described by Hruch- 
 mann for O. vulgatum, or is four-sided, as it is in O. pendulum. In the smaller speci- 
 mens antheridia were more numerous than archegonia, although several of the latter 
 were present. The antheridia arise, in general, in acropetal succession, but it is 
 not unlikely that secondary ones may be formed also. The archegonia are scattered 
 among the antheridia apparently without anv definite order. In some specimens, 
 especially in the larger ones, the old archegonia were found in great numbers, many 
 more than the antheridia. In other specimens a considerable part of the protliallium 
 was (juiie destitute of either archegonia 01 antlu 1 iilia, and in this respect the pro- 
 fhalluini <il O. moliiccauum differs fioni that ot (). f^eduih uh^sum or O. vulgatum. 
 
12 
 
 THE OPHIOGLOSSALES 
 
 The gametophyte of (). moluccaiiiun shows a greatei or less miinber of rhizoids 
 both at the base and along the fertile branch (plate i, fig. lo). In the more slender 
 forms, however, these rhizoids are few. They are in some cases two-celled, but more 
 commonly consist of a single elongated cell. It is not unusual to find within this the 
 penetrating filament of the mycorrhiza, as has been described for other species 
 of Ophioglossiim. The rhizoids are much longer relatively than in O. pciididum. 
 According to Bruchmann, the rhizoids are quite absent from the prothallia of O. 
 vulgatum. 
 
 Mettenius states that in O. pcduncidosum the prothallia often appear above the 
 surface of the earth, and they then become somewhat flattened and sometimes 
 divided into several small lobes, and in such cases chlorophyll is developed. Met- 
 tenius, however, does not note any further development of these green lobes. 
 Bruchmann found that chlorophyll might also develop in 0. vulgatum, when the 
 prothallia were exposed to the light, but he did not find any flattening of the apex. 
 Owing to the very small number of growing prothallia found by me, I could not 
 test the power of developing chlorophyll in 0. mohiccanum, but the occurrence of 
 chromatophores in the germinating spores makes it highly probable that chlorophyll 
 may be developed in the older prothallia under the stimulus of light. 
 
 Gametophytes, /ir, and young sporophytes of Ophioglossiim mo/i/rranum and allied species, slightly enlarged. C and F from 
 Hakgala, Ceylon; the others from Buitenzorg, Java; ky bud on primary root; /, primary leaf; /", secondary leaf; /, tuberous 
 enlargement at base of gametophyte. 
 
 The small size of the prothallium in O. r?ioluccanum and the cessation of growth 
 after the sporophyte is formed indicate that the gametophyte lives only for one 
 season, and this is probably the case also in Helminthostachys. In this respect 
 O. moluccanum and its allies differ markedly from O. pendulum and 0. vulgatum, 
 where the gametophyte lives for many years. 
 
 At Hakgala, in Ceylon, an undetermined species of Ophioglossum of the type 
 of O. reticulatum is common. After careful search, a few prothallia which closely 
 resembled those of O. mohiccanum were collected (fig. i, C, F). The material was 
 too scanty to make a detailed study possible, but from the external appearance it is 
 likely that the structural details would closely resemble those of 0. mohiccanum. 
 
 The prothallium of Ophioglossum vulgatum, according to Bruchmann (Bruch- 
 mann I), closely resembles that of O. mohiccanum, but is very much larger, some- 
 times reaching a length of 6 centimeters, and it is not infrequently branched 
 
13 
 
 Fig. 3. 
 A, B. Gametophytcs of Ophioglossum peduncutosum (after Mettcnius). 
 C, D. Gametophytcs of 0. vulgatum (after Bruchmann). 
 E-G. Gametophytcs of 0. pendulum, sp, young sporophytc; r, secondary root. 
 
 (Hg -5. <".'). Iluic is .1 mou- or less ci)iis|Mcii(iiis l);is;il nilKi, like tli;it Inuiul in 
 O.nioluranuiin, ami t'lom iliis extends a braiuli iKaiint; tlit- iciMocliictivc 
 but, unlike the piothal- 
 liuni of 0. riiohiccaniini, 
 there seems to be a 
 complete absence of 
 ihizoids. In general, 
 the form is mt)re irreg- 
 ular than that of O. 
 itioluicaiium and more- 
 over it is very long- 
 lived. The branches 
 may become detached 
 and thus form new in- 
 dividuals. From the 
 rate of growth in speci- 
 mens kept under ob- 
 servation for several 
 months, Bruchmann 
 concluded that they 
 might live for twenty years or more. Where the apical growth of a branch is 
 interfered with, there may be a formation of adventitious buds, a phenomenon 
 which is also common in 0. pendulum. 
 
 Ophioglossum {Ophwderma) pendulum, a remarkable epiphytic species wide- 
 spread through the tropics of the Old World and reaching to Hawaii, is the giant 
 of the order, the pendent ribbon-shaped leaves sometimes attaining a length of 
 1.5 meters. The sporophyte grows rooted in masses of humus, either upon the 
 rough trunk of a tree fern or palm, or hanging from the mass of humus accumu- 
 lated about the base of certain epiphytic ferns. The hivd's-nest fern. As plenium 
 nidus, furnishes the favorite substratum for this species in the forest of Tjibodas in 
 Java, where my material was collected. 
 
 The young prothallium in 0. pendulum is an ovoid bod\', the somewhat smaller 
 forward portion corresponding to the fertile branch in O. moluccanum (see Lang I, 
 page 25). The older ones become very much larger than in any other species that 
 has been described, and they branch freely in all directions, except as their growth 
 is controlled by their position. They are always found buried in a mass of humus 
 between the imbricated leaf bases of the fern, and are often much flattened by the 
 pressure of the leaves between which they are confined. Branches extend in all 
 directions, but their growth is to some extent controlled by the inclosing leaf bases 
 and also by the tangled mass of roots ot the fern, which grow in all directions through 
 the mass of humus and among which the branches of the prothallium of the Opliio- 
 glossum extend. In one instance several hundred were collected from one large 
 plant of Jsplennirn nidus. The prothallia closely resemble Lang's figures and 
 descriptions, but in many cases are very much larger and more extensively branched 
 tlian :in\ of the specimens collected by him in Ceylon (plate i, figs. 11-14). The 
 largi r prothallia may be stellate in form, but they are usually very irregular. The 
 branches penetrate in all directions between the dense tangle of roots which the 
 Asplenium sends into the humus between its persistent leaf bases, and on pulling 
 these back a mass of fine hunuis is found, held together by the mat of roots so that 
 it can be removed intact. Tlu- protliallia aie excessively brittle, and it is practically 
 impossible to remove tiie larger ones without a loss of some of the numerous branches. 
 
14 
 
 THE OPHIOGLOSSALES 
 
 These break oft at the least touch, ami no tloubt serve to propagate the ganietophyte, 
 which is apparently capable of unlimited growth in this way. It is often impossible 
 to say whether the smaller gametophytes that are found loose in the humus are 
 anything more than branches which have become spontaneously separated from 
 the larger prothallia. 
 
 The older parts of the gametophyte are dark brown in color, but the tips of the 
 branches are white, as in the other species. The branching, as we have seen, may 
 be very irregular, and old fragments kept moist often send out great numbers of 
 adventitious buds (plate i, fig. ii) which apparently in time develop into normal 
 prothallia. A number of the commoner forms are shown in the figures. Plate i, 
 fig. 14, represents the largest one met with. This is by no means complete, as a 
 
 A. Longitudinal section of apes of gametophyte 
 
 ftrij two-celled embryo. 
 
 B. Apex more highly magnified, showing apical cell, x. 
 
 C. Transverse section of a branch of the gametophyte of Ophi 
 
 D. Apical cell of same. 
 
 E. F. short hairs from gametophyte of 0. pendulum; in E m; 
 
 iglossum pendulum. 
 
 y be seen the endophytic fungus fila 
 
 number of branches were unavoidably broken off^ in removing it from among the 
 tangle of roots in which it was embedded. This specimen measured about 15 
 millimeters in diameter — more than twice the size of the largest specimen secured by 
 Lang. The surface of the older parts of the prothallium shows a slightly roughened 
 appearance due to the numerous very short papillate hairs which occur upon 
 it. These are never of the slender form found in 0. moliiccanum and perhaps 
 are not properly to be considered as rhizoids. Dotted over the surface are pale 
 brown spots, easily seen with the naked eye, and these on examination are found to 
 be the large, empty antheridia. The branching is sometimes dichotomous, but 
 lateral branches may arise at almost any point, and old fragments of the prothallia, 
 as already indicated, often develop many adventitious buds, a condition of things 
 which apparently obtains also in the long-lived prothallia of 0. vidgatum. The 
 rate of growth of prothallia kept by the writer for more than a year, as well as their 
 
THK GAMKTOPHYTI' 
 
 15 
 
 position in rlu' lunnns alioiit the pl.mt, wluu iluy .iii' louml in tin- nldtr parts 
 between ie.ivcs which nnist ii.nc hccn tliaii lor man)' ytais, imlii.ite that they are 
 veiy long-livcti anti, as we iiave seen, hy the leady sejjaiation of tlie blanches they 
 aie easily piopagated. 
 
 THK IIISIOI.OGY Ol- THI-; OAMICI'OI'HYTE. 
 
 The younger portions of the growing branches aie composed of thin-walled, 
 colorless parenchyma, whose cells have a conspicuous nucleus, and usually numerous 
 small starch granules. In the older portions ot the gametophyte the tissue becomes 
 infected with the often-described endophytic fungus. Practically all of the cells of 
 the basal tuber arc thus infested, but in the fertile branches the cential tissue is 
 usually free fiom the fungus, and this mcdulla-likc central tissue is surrounded by a 
 more or less definite mantle of cells, in which the endophyte is especially luxuri- 
 ant. The outermost cells ate piactically free fiom fungus, although new infections 
 probably may take place there through these outer cells. In O. moluccaniim the 
 endophyte is much less developed than in O. pendulum, this no doubt being corre- 
 lated with the much briefer duration of the gametophyte in the former species. 
 According to Bruchmann, the endophyte is strongly developed also in O. vulgatum. 
 Ihe endophyte, as in the other forms that have been studied, is quite absent from the 
 apical region of the branches. The limits of the infested zone are not very clearly 
 marked and any cell of the older tissue of the gametophyte may harbor the fungus. 
 
 Fig. 5. 
 
 A. Longitudinal section of gametophyte apei of phiofloaur 
 
 B. Transverse section of gametophyte apex, a, the apical co 
 
 C. Archegonium. X225. 
 
 0. Two free spcrmato/oids. X550. (All figures after Bruchii 
 
 The apical cell in O. pendulutu is usually a four-sided pyramid, and not tetra- 
 hedral, as it is in 0. vulgatum. In longitudinal section (fig. 4, D), the apical cell 
 appears triangular with a fairly regular segmentation, but there is also active division 
 in the adjacent tissue and apparently the segmentation of the apical cell is not very 
 rapid. Cross-sections show the apical cell to be approximately four-sided (fig. 16, //), 
 but the sides are not always of equal length and .sometimes it is almost triangular in 
 outline; possibly it may be that in some cases, as in 0. vulgatum, it is tetrahedral. 
 There seems to be no absolute rule, however, as to the succession of divisions in the 
 young segments. A inore or less definite superficial layer arises from the first 
 periclinal divisions, but anticlinals follow rapidly. 
 
16 
 
 THE OPHlOGLOSSALli.S 
 
 II IK GAMETOPHYTE OF BOTIO'CIIIUM. 
 
 Hofmeister (Hofmeister I), in his studies of the Archcgoniates, has included 
 Botrychiiim huiaria, whose prothaHium he discovered in 1854. No further contri- 
 butions to our knowledge of the subject are recorded until 1893, when the writer dis- 
 covered at Grosse He, Michigan, a number of old prothallia of 5. virgin! an tim, with 
 the young sporophytes attached. The earliest germination stages of this species 
 were also secured. In 1895, Prof. E. C. Jeffrey collected at several points in Canada 
 a large number of prothallia of this species, and his account (Jeffrey I) is much the 
 most satisfactory one which had appeared up to that time. In the spring of 1903 
 Lyon (Lyon 1) discovered prothallia of 5. ohliqiiitm in Minnesota, and the following 
 year secured a few specimens oi B. simplex and B. matricaricEfolium, but he has only 
 published a brief note in regard to these species. He states that the reproductive 
 organs in B. obliquuni "differ essentially from those oi B. virgin i an inn," but he does 
 not explain in what these differences consist. Bruchmann, v^'ho has added so much 
 to our knowledge of the gametophyte of the European Lycopodiaceae and Ophio- 
 glossaceae, has recently given a very satisfactory account of the prothallium and 
 embryo of B. lun aria, which he(found in various parts of Germany and Switzerland. 
 He corrected certain errors made by Hofmeister in his account of the same species. 
 (Bruchmann 2.) 
 
 i^^M' 
 
 A, B. Germinating spore (B, optical section). 
 
 C. Three gametophytes, X3; em, embryo. 
 
 D. Section of gametophyte, X12; the shaded region is that occupied by the endophyte. 6 antheridia. 
 
 E. Apical region of gametophyte, X150. 
 
 F. short multicellular hair or paraphysis. 
 
 Owing to the. kindness of Professor Jeffrey, a large number of prothallia and 
 young sporophytes of B. virginianum, together with several slides showing the se.xual 
 organs and embryos, were put at my disposal. This has made it possible for me to 
 make a very satisfactory study of the reproductive organs and embryo in this species. 
 
 The spores of Botrychiiim, like those of Ophtoglossum, are of the tetrahedral 
 type and are quite colorless, their contents showing the usual granular appearance, 
 but without any trace of chlorophyll. The early stages of germination are exactly 
 like those of Opiiioglossiim, the first division wall being transverse, this being then 
 followed by a second wall in the inner cell at right angles to the primary wall (fig. 6, 
 A, B). A few chloroplasts were seen in some of the cells, but this does not seem to 
 be a constant character and perhaps was an abnormality due to the spores having 
 
AMI 
 
 17 
 
 fji-miin.itnl in tin li!:lit. UiiKlim.inn i Im lulimiinn 2. p;i<;c 207) also found the 
 earliest stages in 11.' Iiiii.nia, ami these oHus|).m(led rxaetly to those in B. vir- 
 giniamim. 
 
 The gamet(ii)h\te of" /^ I'lri^inininnn is a large, tiiher-like ho(l\-, which grows 
 at a depth of ahoiit 10 centimeters below the surface of the earth. The smallest 
 of the specimens found by Jeffrey were 2 millimeters in length by 1.5 millimeters 
 in breadth, exceeding in size the fully developed prothallia of li. liinaria. The 
 older ones may reach a length of 20 millimeters. The young prothallia are 
 quite smoothly oval in outline (Hg. 6, C), while the older ones are more or less 
 
 A, YounR sporophytc of Botrychium vir 
 
 mum attached to gametophytc, pr, 
 
 B. A pamctophytc with small sporophyte. 
 
 A. Two young gamctophytcs. 
 
 B. An older one, with sporophyte, sp, attached 
 
 X16. 
 
 C. Surface view of two antheridia, showing opcrcu' 
 
 lar cells (shaded). X lOO. 
 
 D. A spcrmatozoid. X500. 
 
 iire^iihn, ami oceasiomillv the anterior end is divided into two equa' 
 |iinl)al)l\- as a lesuli of dichotomy. .'\s in Ophioglossitm, the older parr.' 
 gametophyte are brownish in color, while the young apex appears white, 
 earlier stages they are covered with numerous rhizoids, which disappear 
 less completely as the prothallium becomes older. The forward end is 
 pointed and there is a definite growing point. The antheridia make their 
 ance first, and together with the archegonia are formed only upon the upper 
 of the dorsiventral thallus, which is noticeably different in form from the cyl 
 radially constructed prothallium of Ophioglossiim. 
 2 
 
 1 lobes, 
 i of the 
 In their 
 more or 
 slightly 
 a p pear- 
 surface 
 
18 THE OPHIOGLOSSALES 
 
 The antheridia form a more or less evident row, even in the very young 
 prothallium, and later this median row of antheridia is raised upon an elevated 
 ridge, upon whose sides the archegonia later make their appearance (fig. 6, D). 
 
 The gametophyte of B. lunaria closely resemhles the early stages of B. virgin- 
 ianum. It is very much smaller and rarely exceeds a length of 2 millimeters, and, 
 like the young stages in B. virginimium, it is covered with many long rhizoids 
 (fig. 8, A, B.) The smaller prothallia are globular or oval in shape; the larger ones 
 somewhat heart-shaped. Hofmeister states that the antheridia occur upon the upper 
 surface and the archegonia below; but Bruchmann found that both archegonia and 
 antheridia were confined to the upper surface, as in B. virginianum. As in the 
 latter, also, the dorsal ridge is present, bearing the antheridia upon its crest and the 
 archegonia upon its flanks. 
 
 The large prothallia of 5. virginianum live for many years, and Jeffrey even 
 found a plant bearing spores, which was still connected with a prothallium. It is 
 probable that the very much smaller prothallia of 5. lunaria have a shorter duration, 
 but in this species also the prothallium persists for some time after the young 
 sporophyte is established. 
 
 THE HISTOLOGY OF THE GAMETOPHYTE OF BOTRYCHIUM. 
 
 In B. virginianum the forward part of the prothallium is made up of colorless 
 tissue, which extends for some distance beyond the youngest antheridia. The apical 
 meristem in this species (fig. 6, E) lies on the upper side of the prothallium; in a 
 vertical longitudinal section it shows a group of columnar cells, one of which is 
 
 ion of a gametophyte of Ophioglo. 
 endopfiytic fungus. X20. The shaded region 
 Endophyte from prothaUium of Bolrychium virgin 
 Sporangium-like enlargements of endophyte. 
 
 m fcnihil 
 
 probably the apical cell; but this is not absolutely certain. In B. lunaria (see 
 Bruchmann 2, page 209) the apical meristem occupies the middle region ot the 
 anterior surface of the prothallium, and is of very limited extent. As in B. vir- 
 ginianum, its cells appear columnar in form, but neither in longitudinal section nor 
 in a surface view could a single initial cell be recognized. 
 
 The rhizoids in B. virginianum are from i to 4 millimeters in length and may 
 be multicellular, especially those which arise from the crest and flanks of the pro- 
 thallium, while those which originate from the base are unicellular and longer than 
 these multicellular ones. It is not unlikely that these multicellular structures are 
 rather of the nature of paraphyses than true rhizoids. In B. lunaria the rhizoids 
 seem to be, usually at least, unicellular. In both species the rhizoids soon turn 
 brown and their walls become strongly cutinized. Both Jeffrey and Bruchmann 
 
THE GAMETOI'HYTK 19 
 
 observed the penetration of fungus hyph:E into the rhi/oids, and they believe that 
 the fungus whieh occurs within the thallus is, mainly at least, due to this method of 
 infection. ^>om a comparison with the early infection of the young prothailium 
 in Ophioglossinn, it seems to me more probable that the young prothailium in 
 Botrychiitm also is infected at a very early period and that the endophyte, once 
 established within its tissues, grows with the development of the gametophyte, the 
 secondary infection through the rhizoids of the older gametophyte being of minor 
 importance. 
 
 In both species of thjirycluiim the infected region comprises the greater part 
 of the central tissue, leaving only a comparatively narrow peripheral region free 
 from the endophyte. This uninfected area is thicker upon the upper surface and 
 comprises the whole of the meristematic region, together with the developing sexual 
 organs. As in Ophioglossum, this uninfected tissue contains small starch granules 
 in considerable numbers, but not much else in the way of granular contents. The 
 invasion of the fungus results in the destruction of the starch and the accumulation 
 of large amounts of oil. This oil, according to Jeffrey, is not readily soluble in 
 alcohol, and the cells containing it, both in fresh and stained sections, appear dark- 
 colored. 
 
 THE GAMETOPHYTE OF HELMINTHOSTACHYS. 
 
 The monotypic Heltnitithostachys zeylanica is not uncommon throughout the 
 lowlands of the Indo-A4alayan region and often occurs in large numbers. The only 
 account yet published of the gametophyte is that of Lang (Lang 1). The prothallia 
 which he describes were collected in part by himself in the Barrawa Reserve Forest 
 in Ceylon in March 1901. Other material studied by Lang was collected at the same 
 place by Mr. Coomara Swamy. I made a visit to the same locality in February 
 1906, and also found a considerable number of prothallia, but all of these had been 
 fertilized and had attached to them the young sporophyte, so that no young repro- 
 ductive organs were found. The forest where they were collected is subject to 
 inundation from a river which runs through it, and it was in the parts that had been 
 overflowed that the young plants were discovered. This makes it not unlikely that, 
 as in the case of Ophioglossum mohucanuin, germination is favored by having the 
 spores immersed in water, and this may be a necessary condition for the first stages 
 of germination. 
 
 A quantity of ripe spores were collected and various attempts were made to 
 germinate these, but without success, and, as none of the specimens obtained by 
 Lang were very y^ung, the early history of the prothailium still remains to he 
 investigated. 
 
 As in the other Ophioglossace;e, the prothallia are subterranean, occurring at a 
 depth of from 5 to 6 centimeters in the earth. In form (figs. 10, 1 1) they are some- 
 what intermediate between Botrychium and Ophioglossum, but are on the whole 
 more like the latter. They are somewhat irregular in outline, with a broad base, 
 recalling the basal tuber of O. moluccanum, but this tuber is relatively larger and 
 more lobcd. From this basal tuberous portion a short upright branch extends, 
 much as in Ophioglossum moUnconum, but it is relatively thicker and shorter. The 
 whole gametophyte is radial in structure, as in Ophioglossum, and thus differs 
 strikingly from the dorsiventral gametophyte of Botrychium. The basal enlarged 
 portion is brown in color and covered with rhizoids, which are mostly absent from 
 the upper, more slender part upon which the reproductive organs are borne. 
 
 Lang found that there is a tendency to dicecism in the prothallia, some pro- 
 ducing only antheridia (figure 11, A), while in others archegonia predominate, 
 
20 THE OPHIOGLOSSALES 
 
 although in the hitter a few anthcridia are almost always developed before the arche 
 thallium the basa 
 
 gonia appear. 
 In the mal 
 
 1 "vegetative" portion is relatively small 
 and often is irregularly lobed, these lobes 
 being not of the nature of true branches, 
 but merely the result of unequal growth. 
 I'he female gametophyte (fig. lo) has the 
 basal region relatively larger and is even 
 more irregular in form than the male, 
 while the fertile region is shorter and 
 wider. Some of the more elongated types 
 of the male gametophyte are quite similar 
 to the simple prothallium of Opiuoglossiini 
 moluccainim. As in the other genera, the 
 endophytic fungus is always present and 
 occupies much the same position as in 
 Botrychium. 
 
 Lang found that the growth is due 
 to the activity of a single apical cell having 
 the form of a four-sided pyramid, much 
 like that found in Ophioglossiim pendulum. 
 1 he lateral segments divide by periclinal 
 walls, and from the superficial cells thus 
 formed originates the layer of tissue from 
 which the sexual organs arise. The inner 
 cells contribute to the axial tissue of the 
 gametophyte. The antheridia in the male 
 plants are evenly distributed about the 
 periphery, so that in cross-section they form 
 a nearly uniform circle. The axial tissue, 
 ^ especially in the male prothallium, has the 
 
 cells much elongated, and Lang thinks 
 nduction of plastic material from the basal region of the 
 prothallium to the growing point. 
 
 The outer part of the basal region consists of two or three layers of somewhat 
 flattened cells, which, like the corresponding tissue in 5o/rv( /;;»»/, are free from the 
 endophyte. From some of these superficial cells there are developed unicellular, 
 elongated rhizoids, with markedly cutinized walls. This cutinization extends to 
 the outer walls of the superficial cells, while the inner walls of these cells, as well as 
 those of the inner tissue, show the cellulose reaction. The formation of the lobes 
 of the basal region are due, not to definite apical growth, but to irregular cell divi- 
 sions in the outer layers, which remain free from the fungus which occupies the 
 greater part of the central region of the lobes, as well as the axis ot the main shoot 
 of the prothallium. 
 
 The central tissue is made up of about equal parts of infected and uninfected 
 cells. The latter, as in the other cases investigated, contain starch granules which 
 are absent from the cells harboring the fungus. Fungus hyphae were seen in many 
 cases to penetrate the rhizoids, but it is highly probable that in Helmmthostachys 
 also there is a primary infection at an early stage in the germination of the spores. 
 
 Lang notes that in the older male prothallia the fungus is dead and the further 
 growth of the prothallium is dependent upon the amount of reserve food (mainly 
 
 achys zeylanica, 
 riniary root. X 
 I'veloped second 
 
 they are useful in the 
 
1HI-: c;,\Mi 
 
 )1'1IVI 
 
 21 
 
 starch) \vhi(.h is lift in tlu- iills. It is not clear wliar mic the enddpliyte pla\s in 
 the inaniifactiiie ot" the reserve food materials upon which the further growth of the 
 prothalliiim depends. It is highly prohaiile, however, that some of the necessary 
 organic elements are derived from the destruction of the fungus tissues which serve 
 as food foi rhi' tinther development o{ the prothallium. While no data are at hand 
 to jiroxe this, tlu- lack of permanent growing tissue in the prothallium and the com- 
 plete destruction of the fungus make it highly probable that the life of the gameto- 
 phyte in H rIiuiiitli<K<tiii hxs is restricted to a single season, as it is in Ophioglossum 
 moluccnnutn. 
 
 of Helmhlho^tmh- 
 
 :. X8. C Surface view of ripe antlieriJium, showing opercular cells 
 U. An old archegoniiuii. X about 200. 
 (Figs. A-C. after Lanj;.) 
 
 rilK F.MXJl'inTK. 
 
 Ihe endophytic fungus which inhabits the gametoph\re in all of the Ophio- 
 glossaceae is very much alike in all of the species, differing only slightly in size in 
 different forms (fig. 9). A special study was made of this endophyte in Ophio- 
 glossum pendulum, where it is especially well developed. As we have already seen, 
 the endophyte is absent from the younger parts of the prothallium, but in the older 
 parts it is exceedingly conspicuous. Sometimes fiagments ot a fungus are found 
 outside the prothallium, growing in the humus, antl these are evidently the same 
 forms that infect the very young prothallium when it arises from the germinating 
 spores. These external hyphre in some cases have an occasional septum, and this 
 is also the case in those forms which infect the young prothallium. In all of the 
 hyphae seen within the prothallium, however, these septa seem quite absent. The 
 infection of the prothallium through the rhizoids has been noted in all the species, 
 but, as we have already stated, it is probable that the endophyte is derived mainly 
 from the first infection of the very young gametophyte. An examination of the young 
 cells before the entrance of the fungus shows a conspicuous nucleus and numerous 
 starch grains, which stain ver\- strongly with gentian violet. The invading mycelium, 
 whither from the outsiiK' of the prothallium or tVom the adjacent cells, peiieriates 
 the cell wall and laiiiihes within the cell, the growth being entirel\- intracilhilar. 
 The h\ph;e are noticeabU' thicker than those of the external mycorrhi/a. 
 
 The endophyte is perfectb' fixed with I per cent chromic acid antl it stains 
 well with the double stain of gentian violet and safranine. the walls assuming a 
 violet color and the niinuTous niulei staining deep red. In the \dungei h\ ph;e, 
 which are of \ar\ing sizes, tin prot()|ilasni is lieiisely granular, but in the older ones 
 
22 THE OPHIOGLOSSALES 
 
 the granular appearance disappears to a considerable extent, though the nuclei 
 continue to stain strongly. The fungus is quite variable in form. Sometimes the 
 filaments are nearly straight, running from cell to cell and branching sparingly. 
 Sometimes a cell is completely filled with a dense tangle of hyphae, while in other 
 cases there are sack-like vesicles of very irregular form. Not infrequently, quite 
 regular, nearly globular bodies are seen, recalling the oogonia of the Peronosporeae. 
 These at first contain comparatively few nuclei, scattered through the granular 
 cytoplasm, but in the older ones the nuclei are very numerous and decidedly larger. 
 It looks sometimes as if this were a preparation for the formation of spores, but no 
 certain evidence of the formation of such spores could be seen, although in several 
 instances there was an appearance which might point to this. Structures resembling 
 the "conidia," described by Jeffrey for Botrychium (Jeffrey I, page 12), were seen 
 and are probably the same thing. 
 
 As the fungus invades the young cells, irregular, strongly-staining clumps are 
 formed by the aggregation and breaking down of the starch grains. The nucleus 
 of the host cell appears to be but slightly affected by the growth of the fungus, and 
 can usually be found quite unchanged, even in those cells which are almost com- 
 pletely filled by the endophyte. Finally, the thin-walled vesicular growths of the 
 fungus are quite broken down and probably serve to nourish the cells of the gameto- 
 phyte, which is thus parasitic upon the cells of the fungus. The systematic position 
 of the fungus is doubtful. The varicose, swollen hyphae found at certain stages 
 closely resemble a parasitic fungus, Completoria complens, which is sometimes 
 extremely destructive to green fern prothallia. The obgonium-like organs, often 
 present, suggest Pythium, a parasitic fungus belonging to the Peronosporeae. 
 Jeffrey (Jeffrey 1, page 13) thinks the endophyte may fairly be regarded as an 
 intermediate form between the two genera, Completoria and Pythium, and says that 
 in this case Completoria should be placed in the Peronosporeae instead of in the 
 Entomophthoreae, where it has been placed by some students of the Fungi. 
 
 THE SEXUAL ORGANS. 
 
 THE ANTHERIDIUM 
 
 The antheridium in all of the Ophioglossaceae is of much the same type and 
 closely resembles that of the Marattiaceae and also Equisetum and Lycopodium. 
 The mother cell does not project at all above the level of the adjacent tissue, and in 
 all cases the first division separates a superficial or cover cell from an inner one, the 
 latter by further divisions giving rise to a mass of sperm cells which may be very 
 numerous. The cover cell also divides, sometimes only in one plan, sometimes 
 (e. g., in Botrychium) having also periclinal divisions. At maturity the antheridium 
 may project more or less strongly, but this is not always the case. 
 
 THE .ANTHERIDIUM OF OPHIOGLOSSUM 
 
 The antheridium was first described by Mettenius in Ophioglossum pedunculo- 
 sum, and later Lang described it for O. pendulum and Bruchniann for O. vulgatum. 
 The present account is based mainly upon my own studies of 0. moluccanum and 
 0. pendulum. 
 
 The mother cell of the antheridium, which may arise very close to the growing 
 point of the prothallium, lies flush with the neighboring cells, but later it may 
 become more or less elevated above the surface, forming a prominence which, in 
 slender prothallia, like those of Ophioglossum moluccanum, may give an irregular 
 undulate outline to the branch which bears them (fig. 4, A). In (). pendulum 
 
THE GAMETOPHYTE 
 
 23 
 
 this projection is less marked, although this may be somewhat decided in the 
 ripe antheridium. The first division in the inner cell is usually vertical, but may be 
 transverse, at least in O. pendulum (fig. 13, //). The second walls intersect the 
 first at right angles, and there are always four nearly equal inner cells resulting 
 (fig. 13, D). The third set of walls is vertical, and the next, in some cases at least, 
 is in the same direction. This is not true, however, of the deeper and narrower type 
 of antheridium. Further divisions occur until the number of spermatocytes is very 
 large. The fully developed mass of spermatocytes is plainly visible to the naked 
 eye, and in O. pendulum may reach a diameter of more than 325 /x. 250 cells and 
 upward may be counted in a single section of a large antheridium, and this would 
 mean that there are several thousand in the whole antheridium — perhaps more than 
 in any other Pteridophyte. In O. moluccanum (fig. 12), the number of spermatocytes 
 is much smaller, while in this regard the antheridium of O. vulgtitum appears to be 
 intermediate between that of 0. moluccanum and O. pendulum. 
 
 I'lG. ll.— Oph, 
 
 A. Ripe antheridium. Xi8o. 
 
 B, C. Young antheridia. X320. 
 
 D. Surface view of antheridium, showing opercular cell. 
 
 The nuclei of the young cells show a conspicuous nucleolus, which becomes less 
 marked in the later stages of the antheridium. According to Mettenius, the outer 
 wall of the antheridium in 0. pedunculosiim is composed of two layers of cells; but 
 both Lang and Bruchmann found that the central part of this outer wall of the 
 
 I in Ophioglossum pcnduiun 
 
 A. Median section of gametophyte, apei, showing apical cell, x, and two youn 
 
 B. Transverse section of apex; x, apical cell. 
 
 C. D. Median sections of young antheridia. 
 
 E. Median section of a nearly ripe antheridium. 
 
 F. Surface view of ripe antheridium; o, opercular cell. 
 
 antheridia, 3. X175. 
 
24 
 
 THE OPHIOGLOSSALES 
 
 antheridium is but one cell in thickness, and I have verified this for both of the 
 species under consideration. 
 
 The first wall to be formed in the cover cell is a nearly median one and vertical 
 (fig. 12, B), and this is followed by a second wall which intersects it, as well as one 
 of the lateral walls of the primary cover cell, so as to include a nearly triangular cell. 
 In this triangular cell there are later formed, as both Bruchmann and Lang showed, 
 a varying number of segments arranged spirally in the fashion of the segments of a 
 three-sided apical cell (fig. 13, F). The same thing occurs in Lycopodium (Treub 1). 
 I have also found a similar condition in the antheridium both of the Marnttiaceae 
 and q{ Equisetuw. The last-formed triangular cell is the opercular cell (fig. 13,^,0). 
 From the prothallial tissue which adjoins the sperm cells are cut off flattened cells 
 which surround the sperm cells with a more or less definite layer of "mantle" cells. 
 The limits of the original cover cell are usualh' plainK' visible in both longitudinal 
 and surface sections. 
 
 THK ANTHERUJU'M OF BOTRVCHIUM. 
 
 The antheridia in Botrychiuw occur only upon the dorsal surface of the gam- 
 etophyte, which is always monoecious. The first antheridia in B. virginianiim (see 
 Jeffrey 1, page 8), form a small cluster which is not noticeably raised above the 
 general level of the prothallium, and from this primary cluster extends a single 
 median line of antheridia toward the apex of the gametophyte. Later this median 
 region becomes raised and forms a conspicuous ridge along whose crest the anthe- 
 lidia are borne. In B. Innarin, according to Bruchmann, the arrangement is 
 much the same. 
 
 A-D. Longitudinal sections of young antheridia. 
 
 E-H. Transverse sections. F, G show only the 
 
 cell, 0. I, section of two ripe antheridi; 
 
 H, surface view showing opercul.i 
 
 The development of the antheridium is very much like that of Ophioglossum. 
 The first division, as in the latter, is a periclinal one, separating the primary cover 
 cell from the mother cell of the spermatocytes. The divisions in the cover cell, 
 however, (lifter somewhat from those in OphiooJossinn, in that periclinal walls 
 
 Library 
 N. C, State College 
 
THE GAMI-TOPHV 
 
 25 
 
 are formed in some of" tluse, so thar tlic eover Ikcoiiks to some extent double. 
 Jeffrey states that the cover is two layers throui^liout, hut Hruchmann found that 
 some of the cells — two and sometimes three — remain undivided and that one of 
 these undivided cells functions as the opercular cell. I have examined this point 
 carefully in Botrycliinm virgiiiianum and find that there is an opercular cell in this 
 species also. The division walls in the primary cover cell usually intersect at right 
 angles, so that the opercular cell is four-sided, instead of triangular, as in Op/iio- 
 glossum (fig. 14, H); but exceptionally it may have the triangular form. In the 
 dehiscence of the antheridium only a single opercular cell is destroyed, not two 
 superimposed cells, as Jeffrey supposed to be the case. 
 
 The divisions in the central cell are very much the same as in Opiiioglossiim. 
 Whether the first division is regularly transverse or vertical could not be determined, 
 as no examples of the first division were found. The earliest stages seen after the 
 first separation of the cover cell had four inner cells arranged quadrant-wise, but 
 it was not clear which was the primary wall. The quadrant division is, usuall)' at 
 least, followed by an octant division, and these earh' divisions may sometimes be 
 recognized for a long while, as there is not nuich displacement of the cells due to the 
 subsequent divisions. In the later stages the primary divisions can no longer be 
 made out clearly. The number of sperm cells finally formed may be very large — 
 more than a thousand in the largest antheridia thus nearh' or quite equaling the 
 number found in Opliioglossiun pcndiiliiin, although the speini cells are very much 
 less in size than in the latter and the antheridium correspondingly smaller. Some- 
 times the antheridium is divided into irregular chambers, apparently due to the 
 persistence of some of the early division walls (fig. 15, B). 
 
 Fig. 15. — Bolrychium virgwu 
 
 A. Longituilin.-il section of a nearly ripe, but rather small, antheridium. X about 200. 
 
 B. Section of a ripe antheridium which has discharged the spermatozoids; some cell-walls have remained, 
 
 forming irregular chambers, and retaining some of the spermatozoids. 
 
 C. A surface view of a ripe antheridium, showing square opercular cell. 
 
 The antheridium of B. lunaria, according to Bruchmann, is smaller than that 
 of 5. virginianiini, but the size of the spermatocytes and spermatozoids seems to be 
 about the same (fig. 8, D). 
 
 THE ANTHERU)UM Ol' HLI.MIXTHOSTACHVS. 
 
 The antheridium of H clmitilhostdrhys has been studied by Lang, and it is 
 evident from his account that it resembles that of BcArychimn much more than it 
 docs that of Opiuoglossnni. As in the former, the primary cover cell undergoes 
 periclinal divisions as well as anticlinal oiu s, bur there are from two to four cells 
 which do not divide by periclinal walls, and oiu of these cells becomes the opercular 
 cell (fig. I I, (J). Thise four umlividt-d cells stain more strongh' than the other cells 
 of the cover, bur onU oiu' of rluni is broken down at ihr time of deliisceiuT. These 
 
26 THE OPHIOGLOSSALES 
 
 cells, as in the case of Botryihiiini Itiiuuia, are nearly square in outline, most of the 
 walls in the cover cells being at right angles to each other, as in BrArychium. 
 
 The divisions in the central cell are very similar to those in Ophioglossum and 
 Botrychium, and the original quadrant divisions are perceptible for some time. 
 The number of spermatocytes, as in Ophioglossum and Botrychium, may be very large. 
 The spermatozoids were not seen by Lang, but from a comparison of the nuclei of 
 the spermatocytes they are apparently of about the same size as those of Botrychium. 
 
 SPERMATOGENESIS. 
 
 Ophioglossum is especially suited to a study of spermatogenesis, owing to the 
 very large size of the spermatozoids. Those of 0. pendulum are probably the 
 largest known among the Pteridophytes. In material fixed with i per cent chromic 
 acid, or weak Flemming's solution, and stained with gentian violet and aniline 
 safranine, the coloring is beautifully clear and the blepharoplast stains with extra- 
 ordinary sharpness. The spermatozoids of O. pendulum are larger than those of 
 O. moluccanum or O. vulgatum, but the development is very much the same. 
 
 If the sperm cells are examined previous to the final division to form the 
 spermatocytes (plate i, fig. i6), the nucleus will be seen to have a small but distinct 
 nucleolus and a dense reticulum. The whole nucleus stains strongly with safranine. 
 The cytoplasm is fairly dense, with granules of various kinds in it. In material 
 fixed with Flemming's solution there are often small black specks, probably fatty 
 bodies, which sometimes interfere somewhat with the study of the cytoplasm. In 
 well-stained sections the blepharoplasts may be seen as tw^o small, rounded bodies 
 of a violet color (plate i, fig. i6, hi) lying near the nucleus. Several cases of the final 
 nuclear divisions were met with, but all of these were in material fixed with chromic 
 acid, and the blepharoplasts were not very well difl^erentiated (plate 2, fig. 22). The 
 nuclear spindle is very distinct, and the very numerous chromosomes are so crowded 
 that it was quite impossible to determine their number exactly, but it is very large. 
 In the young spermatocytes (plate 2, fig. 26) the nucleus shows a more or less con- 
 spicuous reticulum, but the nucleolus has disappeared, as it usually does at this 
 stage in all the forms that have been studied. The further development of the 
 spermatozoid coTresponds very closely with that of Equisetum (see Campbell 4 and 
 BelajefF 1, 2). One of the blepharoplasts in the primary spermatocyte goes with 
 each daughter cell (the definitive spermatocyte) and can be seen as a distinct rounded 
 body lying near the nucleus (plate 2, fig. 24). In some cases what appeared to be 
 the blepharoplast was lying in a depression at the periphery of the nucleus, and 
 looked very much like a nucleolus. 
 
 Before the nucleus undergoes any marked change, the blepharoplast begins 
 to elongate (plate 2, figs. 26, 27) and assumes the form of a pointed, slender, strongly 
 staining body lying near the nucleus. This body is really somewhat ribbon-shaped, 
 and more pointed at one end than the other, and it is also somewhat curved, even 
 in its earlier stages. A transverse section appears crescent-shaped. The nucleus 
 now elongates slightly and the reticulate structure becomes very conspicuous (plate 
 2, fig. 25). There are large, strongly stained granules, which are probably composed 
 of several more or less coherent chromosomes, as the number of these granules is 
 very much less than the number of chromosomes in the nuclear plate of the dividing 
 nucleus. The blepharoplast continues to elongate, and in favorable cases the young 
 cilia can be seen growing from it, but none of my preparations showed the cilia 
 nearly so plainly as Belajeff's figures show them in Equisetum and Gymnogramme. 
 There is no doubt, however, that the cilia arise in much the same way as BelajefF 
 describes. The nucleus now becomes indented on one side and assumes a crescent 
 
THE GAMETOPHYTE 27 
 
 shape, elongating and also becoming nioif or kss flattencti. One end becomes 
 narrower and sharply pointed, the other remaining thicker and rounded. The 
 reticulum at this time stains with great intensity and shows a tendency to coalesce, 
 which in the final stage results in an almost homogeneous, deeply-staining mass, 
 composed apparently of the completely fused chromosomes. In successful prepara- 
 tions the nucleus at this stage stains a clear carmine red, in strong contrast to the 
 bright violet of the blepharoplast. With the complete coalescence of the chromo- 
 somes the volume of the nucleus is noticeably decreased (plate i, fig. 19). The 
 blepharoplast forms a spirally coiled narrow band, from which the cilia can be seen 
 to grow, following its curve, but the blepharoplast is not in close contact with the 
 nucleus. 
 
 i'he spermatocytes and the nuclei are rather smaller in (Jphioglossiirn moluc- 
 camtm than in O. pendulum. In the older stages of the spermatozoid, the nucleus 
 in the former species is decidedly more elongated and more sharply pointed at both 
 ends (plate i, fig. 18). In this respect it more nearly resembles the spermatozoids 
 of Botrychiurn and those of the true ferns; while in the larger, comparatively short 
 nucleus, as well as in some other respects, the spermatozoid of O. pendulum is strik- 
 ingly like that of Equisetum. Ihe number of cilia is large, but the exact number 
 could not be determined. 
 
 Surrounding the spermatozoid and included in its coils is a considerable amount 
 of cytoplasm. I was not successful in obtaining any living spermatozoids, although 
 many attempts were made to do this, but in a number of sections of the opened 
 antheridium some were found in which the spermatozoids had been retained. While 
 these were usually more or less distorted, some were very well fixed and gave a good 
 idea of the structure of the free spermatozoid (plate 2, fig. 28). The cilia were 
 very much distorted in some cases, but in some they were clearly seen. Ihere is 
 one thick posterior coil mainly composed of the nucleus, which is very much larger 
 than that of the sperm cell before it is discharged from the antheridium. The 
 nucleus has the form of a slightly coiled thick band, tapering somewhat at both ends, 
 but more markedly so in front. Beyond this extends a second coil, composed, 
 apparently, mainly of cytoplasm. 1 his second coil extends into a third much smaller 
 one, which seems to be a flattened band along whose upper edge the blepharoplast is 
 closely applied (plate 2, fig. 28). The resemblance of the spermatozoid to that of 
 Equisetum is very strong, but the nucleus is even more shortened than in the latter. 
 
 The spermatozoid ofOphioglossum pendulum is larger than that of any Pterido- 
 phyte that has yet been described. The cytoplasmic envelope and vesicle are not 
 very clearly separated and probably are similar to those of the free spermatozoids 
 of O. vulgatum or o{ Equisetum. Sometimes this protoplasmic envelope completely 
 surrounds the lower part of the spermatozoid and reminds one somewhat of the 
 peculiar spermatozoids of the Cycads. Owing to their large size, the spermatozoids 
 were often sectioned, and in some of the sections the blepharoplast was free from 
 the body of the spermatozoid and the attachment of the cilia was very easily seen 
 (plate 2, fig. 29, a). 
 
 Mettenius figures the free spermatozoids of O. pedunculosum, but his figures 
 are certainly not accurate. Bruchmann, who has figured those of (). vulgatum 
 (fig. 5, D), shows that they closely resemble the spermatozoids of the true ferns, but 
 are more massive, and the vesicle which envelops the posterior coils adheres more 
 closely to the spermatozoid than is usual in the terns. In these respects it resembles 
 the spermatozoid of (). pendulum. 
 
 Jeffrey examined the development of the spermatozoids in Botrychium vn- 
 ginidnum, which closely resemble those of Ophioglossum. Owing to the method of 
 
28 THE OPHIOGLOSSALES 
 
 Staining which he employed, the blepharoplast was not clearly evident. I have 
 examined the developing speimato/oids in this species, using the material which 
 Professor Jeffrey was so kind as to send me, but employing the double stain of 
 safranine and gentian violet, which was used in the study of Ophioglossum. While 
 the differentiation was inferior to that of Ophioglossum, where the material was 
 fixed with Flemming's solution, nevertheless the presence of the blepharoplast was 
 unmistakable. 
 
 The spermatocytes are only about half as large as those of Ophioglossum 
 pendulum, and the spermatozoids correspondingly smaller; hence less favorable 
 for the study of details. It is evident, however, that the changes in the nucleus and 
 blepharoplast are essentially the same as in Ophioglossum (plate 2, figs. 34 to 37). 
 The nucleus, however, becomes much more elongated and the spermatozoids are 
 more slender, in both of which respects it shows a resemblance to the spermatozoids 
 of the typical ferns. 
 
 The study of the free spermatozoids was confined to those which happened to 
 have become fixed within the open antheridium and within the venter of recently 
 opened archegonia. These were in some cases very satisfactorily fixed and stained, 
 and showed very well indeed the general form of the spermatozoid. The cilia, 
 however, were not as a rule very well fixed, and while they are evidently numerous 
 their number could not be made out. 
 
 Bruchmann's figures (fig. 8, D) of the free spermatozoid o'i Bo1r\ihiinu lunaria 
 agree closely in form and size with those of B. virgiuianum. In both species the 
 vesicle attached to the spermatozoid is more distinct than in Ophioglossum, and 
 in this respect also the spermatozoid shows a likeness to the true ferns. 
 
 TH1-; ARCHEGONIl'M. 
 
 The archegonium in the Ophioglossaceae is very much like that of the typical 
 ferns in its general development. The mother cell of the archegonium, like that of 
 the antheridium, is first divided by a periclinal wall into an outer cover cell, which 
 later gives rise to the four rows of neck cells, and an inner cell which, as in the typical 
 ferns, divides again by a transverse wall into a lower or basal cell and a central cell, 
 the latter subsequently giving rise to the egg cell and the canal cells. A marked 
 feature m the archegonium of the Ophioglossaceae is the inconspicuous character 
 of the ventral canal cell, which very often it is impossible to show, and it is possible 
 that in some cases the ventral canal cell may not be formed at all, although I think 
 this is doubtful. In general the archegonium is much alike in the three genera, but 
 the neck is much shorter irr Ophioglossum than in Botr\chium or Hrlmiiitiiostaihys. 
 
 THE ARCHEGONUM OF OPHIOGLOSSUM. 
 
 My own studies of the development of the archegonium of Ophioglossum were 
 based mainly upon a study of Ophioglossum pendulum. Only a very small number of 
 the young archegonia was secured in O. moluccanum, and so it was impossible to 
 make a complete study in this species. Lang accurately figures several stages of 
 the archegonium in Ophioglossum pendulum, and Bruchmann has described quite 
 completely its development in O. vulgatum. Bruchmann failed to see the two 
 nuclei of the neck canal cell which Lang correctly figures for 0. pendulum. Fhese 
 two nuclei are invariably present in both 0. pendulum and 0. moluccanum, and it 
 is to be expected that they will also be found in O. vulgatum, as they are constantly 
 present also in Botrychium virgiuianum and in all of the ferns that have been 
 accurately examined. Neither Lang nor Bruchmann saw the ventral canal cell, 
 which is exceedingly difficult to demonstrate. 
 
TH1-: GAMKTOI'IIV 
 
 In ()i^lii<,olnssinu tin- \cum<; ;iirlHi;nni;i m;i\ lie tniiiul lu.ii thr ,i|nx ot tin- 
 pi<.tlialli;il hi;nuh, Inir tluy m;iv als.i ."iiisc- ;it .1 coiisidtiiilik- ilistaiur liack nf" it. 
 In mniial, liki- the antheiidia, tlu\ ailsc in atrc)|>ital succession. 
 
 Ilu- niiitlifi cill, like that of tin- antlu mliuni, is sometimes lirnail, sonutinus 
 nairow and deep, and the cover cell is conespondini^ly shallow or dee|i. 1 In- hist 
 division of the inner cell takes place shortly after the cover cell is cut off. The next 
 division is in the cover cell, which Hrst divides hy a vertical wall, a second \ertical 
 wall being formed almost immediately, intersecting the first and dividing the cover 
 cell into four practicalh e(pial cells, arranged quadrant-wise (fig. 16, A). The 
 middle cell next divides by a transverse wall into the primary neck canal cell and 
 the central cell (fig. 16, /•"). The canal cell pushes up between the four primary 
 neck cells, which presently divide by nearly horizontal walls, so that there are two 
 tins of neck cells. One 01 both of these di\ ide again later, so that each row of neck 
 cells consists of thiee or four. Rareh' the re ma\- be Hve cells in one 01 more of the 
 rows. 
 
 — Development of the archegonium in Ophiogloisun, 
 
 \. Transvcriic section of gametophyte apex, showing two young archegonia,? , and apical cell, x. 
 
 B-G. Successive stages in the development of the archegonium; seen in longitudinal section; «, neck canal cell; b, b.isal cell. 
 
 H. Recently fertilized arcliegonium; 5/>, a spcrmatozoid within the egg nucleus. 
 
 The neck canal cell is very conspicuous, its base being broad, and the upper 
 part narrower and extending to the uppermost neck cells. The large and conspicu- 
 ous nucleus soon divides into two, but as a rule there is no division wall, although 
 occasionally two distinct neck canal cells may be present. Sometimes both of the 
 nuclei remain in the broad basal part of the cell, and sometimes one is at the base 
 and one nearer the apex. This arrangement seems to depend upon the direction 
 in which the nuclear division takes place. 
 
 1 he basal cell divides by a vertical wall at about the same time that the primary 
 canal cell is cut off from the central cell. Ihe basal cell subsequentU undergoes 
 further divisions, but its limits are readily distinguishable up to the time that the 
 archegonium is mature (fig. 16, G). 
 
 In its earlier stages, the archegonium of Ophioglossitrn pendiduni bears a 
 striking resemblance to that of the Marattiacea?, which the mature archegonium 
 more nearly resembles than it does that oi Botr\chiiim. According to Bruchmann's 
 account, 0. vulgatiini has a neck somewhat longer, and this is also true of O. mohn- 
 ciiniini. Even when mature, the neck projects but little above the surface of the 
 prothallium. although there is some elongation of the cells at the time of dehiscence 
 (fig. 1^'. (!)■ 
 
 The venrial canal cell is ver\- iliffcult to demonstrate, and one is sometimes 
 inclined to ilouht whether it is formed at all in some cases; neither Lanij nor Bruch- 
 
30 
 
 THE OPHIOGLOSSALES 
 
 nianii was able to detect it in (). pcndiihini and O. vulgatiiiii. It is probable that 
 its apparent absence in most ot the archegonia is due to the fact that it is formed very 
 late and is extremely inconspicuous. The same apparent absence of a ventral canal 
 cell in the Cycads and some Conifers has been shown, on more critical study of the 
 material, to be due to the small size of the ventral nucleus and to its very evanescent 
 character. In nearly all of the archegonia examined, just before they opened there 
 was present a vesicular body above the egg, which was probably the ventral canal 
 cell much distended with fluid preparatory to the opening of the archegonium. A 
 small nucleus, or what looked like one, could sometimes be seen, but it must be said 
 that its nuclear nature was not above suspicion (plate 2, fig. 33). 
 
 Just before the archegonium is ready to open, the egg cell, which up to this 
 time is compressed above by the basal wall of the neck canal cell, becomes distended 
 and pushes up the base of the canal cell, which thus becomes concave below. It is 
 probably about this time that the ventral canal cell is cut off. Unfortunately, no 
 cases were found showing mitosis in the central cell, but there seems no good reason 
 to doubt that a ventral canal cell is, usually at least, cut off. 
 
 Fig. 17. — Development of tlie archegonium in j 
 
 A-D. Longitudinal sections. X320. 
 
 K. Ripe archegonium, showing ventral canal cell i'. X 
 
 F. Recently fertilized archegonium, showing spcrmatt 
 
 65. (After Jeffrey.) 
 ;oids within venter. X32 
 
 The neck canal cell does not show the complete disorganization which is com- 
 mon, but retains its form up to the time that the archegonium opens. With the 
 opening of the neck there is some elongation of the outer neck cells, but there is 
 decidedly less projection above the surface of the prothallium in O. prmhilutn than 
 is the case in O. viilgatnm. The nucleus of the egg cell is large, but it does not always 
 stain readily, except the nucleoli, and it may be that the same resistance to stains 
 is the reason why its sister nucleus in the ventral canal cell is so difficult to sec. 
 
 THF, ARCHEGONIUM OF BOTRYCHUmf. 
 
 The archegonia in Botrychltim are found usually upon the flanks of the median 
 ridge, upon whose summit are borne the antheridia. In their early stages they 
 closely resemble those of Ophioglossum (fig. 17). As in the latter, there is some 
 variation in the form of the mother cell, which may be quite deep and narrow, or 
 relatively broad and shallow. The subsequent divisions in the central cell and in 
 
THK GAMKTOPHVTI-; 31 
 
 the covL-i cell agree very closely with those in Ophioglossinn; hut the neck cells arc 
 more numerous and at maturity the neck projects much more strongly than is the 
 case in Ophioglossum. There may be as many as seven or eight cells in each row 
 of the neck, which, except that it is quite straight, resembles that of the typical 
 ferns. It is especially like that of OsmunJd, in which the neck is also straight. The 
 neck canal cell (fig. 17, D) becomes much elongated, and the nucleus divides, as in 
 Ophioglossum, but in most cases at least there is no division of the canal cell itself. 
 As in Ophioglossum, the ventral canal cell is very inconspicuous, and often 
 impossible to detect. Jeffrey (fig. 17, E) figures a very evident ventral canal cell, 
 but the nucleus is much smaller than that of the egg or neck canal cell. I have also 
 found what seemed to be a ventral canal cell in the archegonium shortly before it 
 opened, but, as in the case of Ophioglossum, this was not absolutely certain. As 
 in the corresponding stage in Ophioglossum, above the egg cell, with its large and 
 conspicuous nucleus, there is a clear space containing a small round body, which 
 showed no evident nucleolus, but otherwise stained very much like the nucleus of 
 the egg, and was probably the nucleus of the ventral canal cell (fig. ij,D,v.) 
 While no certain cases of mitosis of the egg nucleus for the cutting off of the ventral 
 canal cell were encountered, in one case the egg nucleus looked as if it were in the 
 early prophase of division. Bruchmann's figures of the archegonium of Botrychium 
 lunaria closely resemble that of B.virginianum, but he was unable to find a ventral 
 canal cell, nor did he apparently see the division of the nucleus of the neck canal cell. 
 The archegonium ofHchnint}iostachys{fig. 11, D) closely resembles that of Botrychiutn. 
 
 FERTILIZATION. 
 In only a few cases were spermatozoids seen within the neck and venter of the 
 open archegonium in Ophioglossum, but twice a spermatozoid was seen within the 
 nucleus of the egg; as there were no other stages obtained, however, the details of 
 nuclear fusion could not be followed (fig. 16, H). The spermatozoid penetrates 
 the nucleus of the egg, where for a time it can be seen distinctly. It is probable 
 that its fusion with the egg nucleus is much like that described by Shaw for Ouoclca 
 (Shaw I, 2). In the mature egg cell the nuclear reticulum is often decidedly con- 
 tracted, but whether this is normal or the result of reagents can not be said. 
 
 FICKTILIZATtON IN BOTRYCHIUM VIKCINIANl'M. 
 
 Jeffrey observed a single spermatozoid within the venter of the archegonium 
 and noted that at the time of fertilization the egg developed what he called a 
 " receptive prominence." I have observed the same phenomenon in several cases 
 (plate 2, fig. 46), and in a good many instances have also found one or more sper- 
 matozoids within the venter. In one case, what looked very much like a sperma- 
 tozoid was seen within the egg nucleus itself, and the process of fusion is probably 
 very much the same as that so fully described by Shaw for Ouoclca. In the specimen 
 shown in fig. 17, /•, oneof the spermatozoids looks as if it has partially penetrated 
 the egg, but it is mipossible to be certain that it was not simpl\' lying against it. In 
 this case, just above the nulceus of the egg there was a slight bicak in the gianular 
 cytoplasm that looked as if it might be a receptive spot, bur the speimatozoitl was 
 not entering the egg at this point. 
 
 In another instance a curious thing was noted, but whether it was normal 
 or not it is impossible to sav (plate 2, fig. 46). A spermatozoid had just entered 
 the neck of the open archegonium and the egg had developed a ver\- conspicuous 
 receptive prominence. The nuclear contents were very strongly Contracted and 
 deeply stained and a portion apparently projected beyond the nuclear membrane 
 
32 THD OPHIOGLOSSALES 
 
 into the receptive prominence. A similar synapsis was noted in several cases. In 
 most instances where free speimatozoids were present in the venter of the arche- 
 gonium the egg nucleus presented a curious appearance (plate 2, figs. 47, 48). A 
 single large nucleolus was present, but scattered through the nucleus there were 
 sometimes a dozen or more of intensely staining round bodies, which at first sight 
 looked like nucleoli, but on more careful examination were seen to differ from 
 the large nucleolus in that they appeared less homogeneous and generally showed a 
 central vacuole-like structure. These granules stain very much like the body of 
 the spermatozoid, and it was thought that possibly they might be derived from a 
 fragmentation of the body of the spermatozoid that had entered the nucleus, but 
 this could not be satisfactorily demonstrated and the nature of this phenomenon 
 must remain for the present uncertain. 
 
 Fertilization in HchnintliostaL-hys has not been obseived. 
 
 SK;N1F1CANCE of the KNUOI'HYTE. 
 
 That the presence of the endophyte is essential to the existence of the sapro- 
 phytic gametophyte of the Ophioglossaceze is indicated by the failure of the germi- 
 nating spores to develop unless they become associated with the fungus. Moreover, 
 the universal occurrence of a similar endophyte in all humus saprophytes among 
 the seed plants indicates that in all of these chlorophyll-less plants the presence of the 
 fungus is necessary for the existence of the host. Although it has not been directly 
 proved, it is generally assumed that one role of the endophyte is the elaboration of 
 some of the carbonaceous constituents of the humus. The infrequent communica- 
 tion between the external hyphae and the internal mycelium makes it unlikely that 
 the nutritive products are directly absorbed by the fungus, and it seems much more 
 probable that the rhizoids of the gametophyte are the direct agents of absorption. 
 How the humus constituents are changed by the action of the fungus so that they 
 are available for the cells of the host is not clear, and it is by no means impossible 
 that some at least of the necessary carbon may be derived from the fungus itself, 
 in the digestive process to which it is subjected in the cells of the host. This seems 
 plausible from the fact that in the green prothallia of certain ferns, where presumably 
 the gametophyte is entirely able to supply its own carbon compounds through 
 photosynthesis, these digestive cells appear to be wanting; at any rate they were not 
 observed in a number of forms that I have studied. 
 
 The experiments of Ternetz (Charlotte Ternetz I), show that certain fungi, 
 including endophytic mycorrhizae, are able to assimilate free nitrogen and confirm 
 the assumption of^ earlier observers that the fungus is useful to the host in supplying 
 it with nitrogen compounds; but, while this is probably a very important part of 
 its functions, it seems to me that it is not perhaps the only one, and that the carbon 
 also is supplied, directly or indirectly, through the agency of the fungus. 
 
 In an extended study of the endophytic mycorrhiza of the saprophytic orchid, 
 Neoitin, W. Magnus (W. Magnus 1) has shown that two types of mycelium ex- 
 hibited by the endophyte are of very different nature. The slender, cylindrical 
 hyphae constitute the active portion of the fungus, which behaves like a parasite 
 toward the cells which it invades, destroying the starch and probably other constitu- 
 ents of the cells, but not attacking the nucleus. The swollen vesicular mycelium, 
 however, is a degenerating structure and is itself destroyed by the cells of the host, 
 which actually digest these fungus mycelia in much the same way that the cells of 
 the leaf of Dnsera digest their prey. Magnus has very graphically shown that the 
 relation of the two symbionts is mutually antagonistic, each one acting as a parasite 
 on the other; nevertheless the presence of the fungus is essential to the higher 
 
THE GAMETOPHYTE 33 
 
 ort^aiiisni, so long as tin- latter is destitute of chlorophyll. Tht- explanation of the 
 widespread saprophytism exhibited by so many of the higher plants may be sought 
 in this attempt to defend themselves against what was probably at first a strictly 
 parasitic organism. Having acquired the power to attack and to feed upon the 
 parasite, the photosynthetic functions were more and more subordinated, until a 
 state of true parasitism (or saprophytism) resulted. The numerous semi-saproph}tes 
 like most of the green Ericales and many of the green ()rchidace;e, are good examples 
 of transition stages, while the characteristic leafless humus saproph\tes, such as 
 iVIonotropa and Corallorhlza, represent the fully developed phase of this peculiar 
 form of symbiosis. We might say that such green prothallia as those of the Marat- 
 tiaceiE and Gleichenia, which contain an endophytic fungus, bear somewhat the 
 same relation to the subterranean prothallia of the Ophioglossaceic that the green 
 Ericales do to Monotropa. 
 
 The occurrence of a similar endophyte has also been noted in a number of 
 liverworts. Cavers (Cavers I) has studied this association with some care in the 
 common liverwort, Fegatella, as well as in some other Hepatic;E. He found in 
 Fegatella that the endophyte is beneficial to the growth of the host, which was more 
 vigorous when the fungus was present. He assumed that this was due to the assist- 
 ance given by the fungus in the assimilation of organic matter from humus or other 
 organic substrata. (See also Humphrey 1). 
 
 This occurrence of an endophyte in the Hepatica- makes its occurrence in the 
 green prothallia of ferns readily comprehensible. Whether in the latter it is an 
 advantage to the host to have the endophyte present remains to be seen, but it 
 is highly probable that such is the case. Once having acquired the habit of 
 associating itself with the fungus, the gradual evolution of the purely saprophytic 
 subterranean gametophyte of the Ophioglossaceae from green forms similar to those 
 of the Marattiace;e is readily comprehensible. 
 
34 
 
 THE OPHIOGLOSSALES 
 
 II. THE EMBRYO. 
 
 The development of the embryo in the Ophioglossaceae has been more or less 
 completely studied in Ophtoglossum pednnculosum, O. vulgatmu, O. moluccainim, 
 O. pendulum, Botrychium virgintaniim, B. liinarta, and B. oblifjinini (Mettenius I, 
 Bruchmann 1 and 2, Lang I, Campbell 8, Jeffrey 1, Lyon I). 
 
 The first division in the young embryo in all of these is usually approximately 
 transverse, although there may be a good deal of variation in this respect. It is 
 probable that in all cases the primary root, the stem ape.\, and the foot all arise from 
 the epibasal region. The embryo reaches a very large size before the root emerges 
 from the overlying prothallial tissue and all of the organs of the young sporophyte 
 are very late in developing, so that it is not easy to trace their origin back to the 
 early cell divisions in the young embryo. Much the most conspicuous organ of the 
 young sporophyte is the root, which may reach a very large size and an advanced 
 stage of development before any evidence of the other organs is apparent. Indeed, 
 several roots may be developed before the shoot is established. In Ophtoglossum 
 vulgatum and Botrychium lunarta, according to Bruchmann's account, the young 
 
 Fig. \%.—0[,hioglossum ,r 
 
 An old archegonium. Xl8o. 
 Two-celled embryo within the archegonii 
 Two sections of an older embryo. Xl8< 
 
 sporophyte remains several years under ground before the first green leaf appears 
 above the earth, and it is probable that in Ophtoglossum pendulum there is also a 
 long period of underground existence before the first green leaf is developed. In 
 Ophtoglossum moluccanum, however, and in Botrychtum virgintanum, the first 
 leaf developed is a green foliage leaf, which grows rapidly and soon appears above 
 the surface of the ground. 
 
 THE EMBRYO OF OPHIOGLOSSUM. 
 
 The first figures that we have of the embryo of Ophtoglossum are those of 
 Mettenius, but his figures of the embryo of O. pedunculosum are not at all satis- 
 factory, although he shows correctly sections of the older sporophyte. Lang figures 
 only one embryo, a somewhat advanced one of 0. pendulum. Bruchmann figures a 
 two-celled stage and a single more advanced embryo of O. vulgatum, but he describes 
 and figures several stages of the young sporophyte. My own study of Ophtoglossum 
 was based mainly upon the development of the embryo in O. pendulum, where a 
 
Tin-: iMiuoo 35 
 
 fairly comjilcte series of embryos was olnaiiud and rhr (k\(lo|imtiir lollowitl quite 
 satisfactorily. 
 
 In (). rnohh, iiiiiiin 1 foiiiul two <|iiirc \oiin<i cmhrNos, rhr \-oungest consisting 
 of two cells (tig. iS, /?), the other iiuich more advanced, with the cotyledon already 
 differentiated and the beginning of the primary root recognizable. In the first 
 embryo the basal wall was nearly transverse to the axis of the archegonium and, to 
 judge from the arrangement of the organs of the older embryo, it looks as if the whole 
 of the hypobasal region formed the foot, the epibasal half giving rise to the cotyledon 
 and root, ihe older embryo (fig. i8, 6') was cut longitudinally in the plane of the 
 cotyledon, which at this time comprises pretty much the whole of the upper part of 
 the embryo, the hypobasal region being occupied mainly by large cells which con- 
 stitute the foot. Unfortunately, in this series the section containing the ape.x of the 
 leaf was missing, and so it is impossible to say whether at this stage a definite apical 
 cell is present in the cotyledon; but, as in somewhat older stages such a cell seemed 
 to be always present, it is probable that there was an apical cell in the cotyledon of 
 the embryo in question. At the base of the leaf, and almost in the center of the 
 embryo, there was a group of actively growing cells, evidently marking the position 
 of the root apex, which arises deep in the tissue of the embryo, very much as we 
 shall see to be the case in the Marattiaceae; and it seems probable that in O. tnoliic- 
 canum, as in Dancea, the root grows downward through the foot and in a direction 
 coincident with the axis of the young cotyledon. 
 
 19. — phioglossum vulgatum (after Bruchmai 
 
 A. Median section of a young embryo. Xi6o. B. An older embryo. 
 C, D. Older embryos, showing beginning of apical bud, i>; r, root. X 
 
 The cotyledon grows upward and soon ruptures the overlying prothallial tissue, 
 while the root grows down in the opposite direction and pierces the prothallium at a 
 point some distance below where the leaf emerges. As the root grows downward 
 through the foot the latter becomes unrecognizable, its outer cells remaining, 
 however, as a zone of large cells encircling the equatorial region of the very much 
 elongated bipolar embryo. 
 
 The young root very early develops a tetrahedral apical cell, like that of the 
 later roots. This cell is probably cut out from the central tissue of the embryo, close 
 to the base of the leaf. The cell r in fig. 18, C, to judge from its position and the 
 arrangement of the cells around it, is probably the initial cell of the primary root. 
 
 The young cotyledon grows rapidly and has first a conical form (fig. 23, B), 
 terminating in a definite apical cell. As it grows it develops a small oval lamina 
 and a slender petiole, and presently the little green leaf appears above the surface 
 of the earth. 
 
 1 he embr\o of (). vulgatum differs remarkably from that of 0. mohiccanum in 
 the late development of the leaf. Bruchmann was unable to obtain the younger 
 stages of the embryo, so that the origin of the different members is still somewhat 
 obscure. In this species it is the root which develops first, and it soon becomes 
 
36 
 
 THK OPHIOGLOSSALES 
 
 exceedingly conspicuous. Bruchmann considers that the root and toot are both of 
 hypobasal origin, but he bases this on a comparison with the true ferns rather than 
 upon actual study of embryos, as he was unable to obtain embryos sufficiently young 
 to demonstrate this, and all trace of the original divisions disappears before any 
 sign of the stem and leaf is evident. It may be well questioned whether, as in 
 Botrychium and 0. pendulum, the foot does not take up the whole hypobasal region. 
 It is not impossible that the position of the basal wall may also vary in O. vulgatum. 
 From a comparison with the embryo of 0. pendulum, I am inclined to assign more 
 of the embryo of the former to the foot region than is done by Bruchmann. 
 
 In O. vulgatum there is, finally, a differentiation of the stem apex from the 
 epibasal region, as in Botrychium and the true ferns, while in O. moluecanum there 
 is no trace of a stem apex in the very young sporophyte, this developing later as a 
 bud upon the first root. 
 
 From a study of O phioglossum moluccanum and also ot 0. pendulum it is 
 evident that the history of the young sporophyte in these species differs strikingly 
 from that of the other Pteridophytes. In both of these species the definitive sporo- 
 phyte always arises secondarily, as a bud upon the root, in the same way that 
 adventitious buds are commonly formed upon the roots of the adult sporophyte. 
 Bruchmann notes in 0. vulgatum the very precocious development of the primary 
 
 A. Transverse section of very young embryo. 
 
 B. Longitudinal section of young embryo. Xi8o. 
 
 C. Three transverse sections of an older embryo; l-i, basal wall; r, primary root. XlSo. 
 n. An older embryo; r, primary root. 
 
 root and the late appearance of the stem apex and first leaf; but in this species the 
 shoot apex, according to his statement, is derived directly from the epibasal half of 
 the embryo, as it is in most Pteridophytes. 
 
 In Ophioglossum pendulum, where the development of the embryo seems to 
 offer no check to the further growth of the prothallium, the position of the arche- 
 gonium varies a good deal and it is impossible to tell from a section just what the 
 position was in the living state, as the branches of the gametophyte extend in all 
 directions and archegonia may be formed at any point upon their surface. To 
 judge from the youngest stages of the embryo that were met with (fig. 20), the basal 
 wall in this species is not necessarily transverse. In both of the cases figured it 
 was oblique, and more nearly longitudinal than transverse. It is likely, however, 
 that it is horizontal, or approximately so. In the four-celled embryo shown in 
 fig. 20, J, the quadrant walls were at right angles to each other, and this was also 
 the case in the five-celled embryo shown in fig. 20, B. Somewhat older embryos (C) 
 show that there is a pretty regular octant formation, and Bruchmann states that 
 this is also the case in O. vulgatum. 
 
 While in the typical ferns, and in Botrychium, all of the organs of the young 
 sporophyte can be traced to certain regions of the young embryo, in Ophioglossum 
 
THK KMHRVO 37 
 
 pendulum only one of the definitive organs, the root, arises in this way, and this 
 becomes differentiated at a very early period. One of the octants next to the arche- 
 gonium at once becomes the apical cell of the young root. This cell is very soon 
 recognizable by its size and shape, and quickly begins its regular segmentation. 
 The primary cap-cell is soon cut off (fig. 20, D), and from now on the young root 
 is very conspicuous. The octant divisions are vc-ry clearly marked in this case, 
 and in section 3, which is the uppermost of the three, the large triangular apical 
 cell of the root is very evident. 
 
 Two types of the embryo were seen in this species. One ot these (fig. 21, A) 
 is nearly globular in form; the other (fig. 21, /i, 6") is elongated. The former looks 
 as if it originated from an embryo in which the basal wall was transverse to the a.\is 
 of the archegonium; in the other it was probably more or less vertical. It is probable 
 that in the former instance the root initial is one of the epibasal octants, while the 
 whole of the hypobasal portion gives rise to the large foot. In the second type it is 
 difficult to say which half should be considered epibasal and which h\p()basal, hut, 
 as in the other case, one half ma\' be considered to be root, thc' other halt tout, the 
 
 C. Older embryo, with single root. 
 
 D. Embryo with two roots; /)r, gametopliyte. Shaded region occupied bv the mycorrhir.a. 
 
 growth of both being nearly in a plane at right angles to that of the archegonium 
 axis, andl suggesting the relative positions of cotyledon and foot in the embr}o of 
 Ophioglossum moluccanum. 
 
 In form the first type somewhat recalls the embryo of Boirychium virginianum. 
 The whole lower portion, or hypobasal half, forms the very conspicuous foot, while 
 from the epibasal region the primary root is developing, and already the rudiment 
 of the second root is visible. Whether the latter arises directly from the primary 
 root, or whether it arises independently from the second epibasal quadrant, is not 
 quite certain. The cells of the epibasal region are evidently actively growing, having 
 abundant protoplasm and conspicuous nuclei. The cells of the foot are larger and 
 much more transparent. 
 
 The second type of embryo (fig. 21, B, C) resembles very closely the second 
 state of the embrxo of Opinocrlossinn viilguturii, where, as is said b\' Hruchinaiin, 
 rlu- embr\() is "all root." 
 
38 
 
 THE OPHIOGLOSSALES 
 
 In O. molticcanuni the prothallium probably lives for a single season only, and 
 the formation of the sporophyte stops its further development; but in 0. pendulum, 
 where embryos are much less frequently found, the large gametophyte continues 
 its growth apparently unchecked by the development of the attached sporophyte, 
 which retains its connection with the gametophyte for a very long time, as in 0. 
 vulgatum and Botrychium virgmianum. 
 
 The embryo reaches a very large size before it breaks through the prothallium. 
 The primary root then emerges as a conical point (fig. 3, F, sp). The second root 
 remains short for a time. There seems to be a good deal of difference as to the time 
 of the appearance of the latter. In the globular type of embryo the second root 
 appears very early, and it looks as if it might have been formed quite independently 
 of the primary root. Sometimes, however, the primary root may attain a length of 
 several centimeters, and may even begin to form rootlets before the second root 
 emerges; while in other cases the two roots grow in opposite directions and seem to 
 be of about equal size (fig. 3, E). 
 
 The leafy shoot in Ophioglossum pendulum 
 does not appear until the root system is well 
 advanced. The primary root, although attaining 
 a length of 10 centimeters or more, in no cases 
 showed any signs of leaf-bearing buds in the 
 specimens that I collected. Rootlets were devel- 
 oped in some cases, and it is not impossible that 
 there may be an extensive development of the 
 root system before the first leafy bud is formed. 
 This is quite in harmony with the large develop- 
 ment of the roots in O. vtilgatum, where Bruch- 
 mann believes that it may be eight or ten years 
 before the first foliage leaf appears above the 
 ground. In O. vulgatum, however, this leaf arises 
 from the original stem apex derived directly from 
 the embryo, and not from an adventitious bud. 
 Ophioglossum moluccanum and the similar 
 species that grow with it differ very much from 
 Ophioglossum pendulum in the character of the 
 young sporophyte, which, as we have seen, at once 
 develops a green foliage leaf or cotyledon. The 
 sporophyte very closely resembles that of O. /jf </;//;- 
 culosum described by Mettenius; in fact, the 
 resemblance is so close that it would seem to 
 confirm the close relationship of this species and 
 ""*■ possibly its identity with some of the forms asso- 
 
 ciated by Raciborski under the name 0. moluccanum (Raciborski I). 
 
 As Mettenius correctly showed in O. pedunculosum, the first organ to be 
 developed is the cotyledon, which soon pierces the earth and appears as a green 
 foliage leaf This primary leaf is continued directly into the primary root, but no 
 stem apex is developed, nor is any sheath formed about the leaf base in the young 
 sporophyte, which consists practically of leaf and root alone. The latter often 
 penetrates for some distance into the prothallial tissue before it emerges, so that the 
 central portion of the young sporophyte is surrounded by a sheath formed by the 
 prothallial tissue. A longitudinal section of the sporophyte (fig. 22, A) shows that 
 the tissues of the leaf are continued directly into those of the primary root. A 
 
 Fig. 11.— Ophioglossum moluccanum. 
 
 A. Median section of young sporophyte before 
 
 formation of bud; /, cotyledon; r, root, 
 pr, the gametophyte. X15. 
 
 B. Central region of same, more enlarged; /r, 
 
 first tracheids. 
 
 C. An older sporophyte, showing bud, fc, de- 
 
 veloping from primary root. X15. 
 
 D. E. Lamina of cotyledon, showing venation. 
 
THE EMBRYO 
 
 39 
 
 single axial vascular bundle traverses the whole of the young sporophyte without 
 interruption, and there is no recognizable boundary between the tissues of the leaf 
 base and those of the root. The central region of the embryo is somewhat thickei 
 and its outer cells are enlarged, these outer cells probably belonging to the foot 
 through which the root has penetrated and which at this period can not be clearly 
 recognized. The strictly bipolar character of the young sporophyte and the way 
 in which it perforates the gametophyte resemble most nearly the corresponding 
 stages of Equisetum and the Marattiace;e, and it may be said that Lyon's recent 
 studies on Botrychium ohli(juum indicate that in this species also there is a similar 
 bipolar arrangement of leaf and root. 
 
 Mettenius pointed out in O. pedunculosum that the definitive sporophyte arises 
 as an adventitious bud upon the primary root, either close to the leaf base or, more 
 commonly, at some distance from it. This is also the case in O. nioluccanum and 
 the other allied species. In one case observed by me two independent sporophytes 
 were found growing from the same prothallium, but this is unusual, and in most of 
 the cases where two sporophytes seem to be present, one of these is realh' the sec- 
 ondary sporophyte growing from the primary root. 
 
 Fig. 23. — Ophioglosium moluccanum. 
 
 A. Young cotyledon, longitudinal section; /)r, gametophyte. x8o. 
 
 B. Apex of cotyledon more highly magnified; x, apical cell. 
 
 C. Young sporophyte in which a rudimentary second leaf (/") is present. XSo. 
 
 D. Cross-section of apex of primary root. X200. 
 
 While in O. vidgatum the sporophyte stays under ground for several years, 
 in O. moluccanum there is every reason to believe that there is only a brief interval 
 between the first formation of the leaf and its appearance above ground. The small 
 size and the character of the gametophyte, as well as the quick germination of the 
 spores and the rapid growth natural to a tropical climate, indicate that the gameto- 
 phyte is an annual and that it dies as soon as the young sporophyte is established. 
 
 In 0. pendulum the primary root as it breaks through the prothallium elongates 
 rapidly, but just how far it grows before the bud is formed upon it could not be 
 determined. These roots are very brittle and easily broken off, and in no cases were 
 young buds found upon roots which were still connected with the prothallium. The 
 first root sometimes reaches a length of 3 or 4 centimeters before the second root can 
 be seen at all. The growth of the roots in the young sporophyte is in all respects 
 like that occurring in the older ones. There is a large tctrahcdral apical cell, the 
 divisions ot which are quite regular, and the conspicuous axial vascular cyliiuK 1 is 
 
40 THE OPHIOGLOSSALES 
 
 developed at an early period. The stele of the second root joins that of the first 
 where the latter joins the foot (fig. 21, D). The primary root in O. pendulum is, 
 usually at least, diarch. 
 
 DEVELOPMENT OF THE PRIMARY BUD IN OPHIOGLOSSUM MOLUCCANUM. 
 
 The several terrestrial species of Ophioglossum growing at Buitenzorg and 
 associated under the name O. moluccanum differ strikingly from O. vulgatum in the 
 further history of the young sporophyte as well as in the early development of the 
 first functional leaf. This difference is also found in the young sporophyte of O. 
 reticulatum ( t) collected at Hakgala in Ceylon. All of these tropical species agree 
 with O. pedunculosum, which was correctly described by Mettenius (Mettenius 1), 
 in the origin of the definitive axis which arises as an adventitious bud from the 
 primary root of the young sporophyte. A similar condition of things probably is 
 true also in O. pendulum, although in the latter no leaf is at first developed from the 
 young sporophyte, which is composed exclusively of one or two roots in addition to 
 the foot, so that the first leaf in O. pendulum is also an adventitious structure. 
 Whether or not the primary leaf in 0. pendulum, and probably at the same time the 
 stem apex, arises as a bud from the primary root, or whether (which seems more 
 likely) it is developed from a secondary root, can not be stated, as no young leafy 
 buds were found in connection with the gametophyte. 
 
 Fig. 24. — Ophioglossum moluccanum, 
 
 A. Tangential section of primary root, showing young, endogenous bud, b. X about 50. 
 
 B. C. Two sections of tfie same bud. B, tfie first leaf; C, stem apex. X150. 
 
 D. Two transverse sections of a young bud; l passes through stem apex st; 2, through leaf. 
 
 Mettenius gives no detailed study of the development of the bud upon the root, 
 nor does he state whether he recognized its endogenous origin. The bud arises 
 much in the same way that a secondary root does, and is not visible upon the out- 
 side of the root until it is far advanced in its development, the young leaf breaking 
 through the overlying tissues in much the same way that a young root emerges. In 
 two instances the rudiment of the young bud could be seen close to the apical 
 meristem of the root, in a position which is exactly the same as that described for the 
 leafy buds formed at the apex of the root in O. vulgatum. More commonly, the 
 bud originated nearer the base of the root, but it may arise at any point between 
 the base and the apex. Figure 24 shows sections of the youngest buds that were 
 found. // is a tangential section of the primary root, in which the young bud has 
 formed at a point not very far back from the root apex, r. The first leaf of the young 
 
THE FMHRVO 
 
 41 
 
 hml aiul tlu- young stem :ipt\ can alix'acl\- Ik- iccogni/A-il, and tlit-se lia\c appaic-ntly 
 developed quite independently of each other. Fhe leaf rudiment, B, lies nearer the 
 bundle of the main root and may perhaps have taken its origin from a single endo- 
 dermal cell, but this could not be certainly determined. The young leaf soon forms 
 a slightly projecting conical body, composed of a few cells with conspicuous nuclei, 
 separated from the adjacent root tissue by a small;|space. At this stage it is not quite 
 clear whether the apical cell has been developed, but later stages show the presence 
 of an unmistakable apical cell of approximately tetrahedral form. Lying nearer 
 the periphery of the root is the young stem ape.x, composed of a small group of 
 meristematic cells, one of which is unmistakably the initial cell for the stem. This 
 apical cell is truncate below and the base is broader than the outer free face. Above 
 the stem apex can be seen the cavity which separates the outer tissues of the young 
 bud from the cortical tissue of the root. 
 
 Fig. 24, D, shows two transverse sections of a but! of about the same age as the 
 one just described. No. i passes through the stem apex and shows the conspicuous 
 triangular apical cell; No. 2 shows a section of the leaf above the level of the stem 
 apex with the basal tissue of the leaf extending around the stem region. This is the 
 beginning of the hollow stipular sheath which incloses the stem apex of the bud and 
 
 A. Bud about 
 
 B. Apex of le; 
 
 ) break through primary root; 
 more highly magnified. 
 Transverse section of a young bud passing 
 
 D. Section of same passing through stem apex, .v. 
 
 E. Transverse section of stem apex from an older sporophyt 
 
 stem apex. X about 80. 
 gh base of leaf, /. X 1 50. 
 
 The apical cell is four-side 
 
 which is so conspicuous a feature in the stem apex of the older sporophyte. Probably 
 the large central cell in the section of the leaf shown in No. 2 is the apical cell, but 
 this is not quite certain. 
 
 A transverse section of a somewhat older bud (fig. 25, C) shows a nariow cleft 
 at the base of the leaf in front, opening into a cavity within which the stem apex lies. 
 It is impossible to say exactly how much of this stipular sheath really belongs to the 
 leaf base and how much is derived directly from the adjacent cortical tissue of the 
 root, since the tissues of the leaf base merge insensibly into the latter. From this 
 time on, each new leaf develops this conical stipular sheath, which incloses the stem 
 apex and the next youngest leaf, and which has so ofttn Ikiii discrihcd in the oUKr 
 sporophyte. 
 
 If a median section of the young bud is examined just before the bud breaks 
 through the overlying tissues of the root (fig. 25, A), it can be clearly seen that the 
 leaf, which now has the form of an elongated cone, lies above the level of the young 
 stem apex, and by this time the differentiation of the vascular bundle of fhe leaf is 
 well advanced. This bimdle connects with the vascular bimdle of the primary root 
 
42 THE OPHIOGLOSSALES 
 
 and there is no connection between it and the tissues of the stem apex of the bud. 
 The latter consists of a shallow mass of tissue with the conspicuous apical cell in 
 the center, but below it there is no sign of the development of any procambium. 
 The stem apex lies in a depression formed by a shallow ridge, which encircles it 
 and forms the beginning of the stipular sheath belonging to the base of the leaf. 
 The exact limits of the basal tissue of the leaf, as we have already stated, can not 
 be clearly defined. The cavity above the stem apex is still very evident, but how 
 much of the tissue lying above the cavity belongs to the leaf base, and how much to 
 the cortical tissue, is not clear. The section through the apex of the leaf (fig. 25, B) 
 shows a single large terminal cell which, without doubt, is the apical cell. 
 
 The young bud is now ready to emerge and, very soon after, the rapidly elongat- 
 ing apex of the young leaf pushes through the outer root tissue and emerges upon 
 the outside. The stem apex remains buried within the root tissues and the sheathing 
 base of the leaf. The leaf base is surrounded by a ragged sheath, formed by the 
 ruptured outer tissues of the root. 
 
 The development of the bud upon the root in O. pendulum unfortunately could 
 not be followed. The smallest leaves found were 10 centimeters or more in length, 
 and, although these were probably the primary leaves, they were not recognized as 
 such at the time the plants were collected, and so there was no opportunity of tracing 
 their connection with the original root. 
 
 The fully developed cotyledon in what may be considered the typical form of 
 O. moluccaniim is more or less lanceolate in outline (fig. 22, D). There is a central 
 vein from which branch secondary veins on either side, connecting with the central 
 vein by anastomosing branches which inclose elongated meshes. In the form with 
 broader leaves, probably another species, the mid-vein is more obscure and the 
 meshes are broader and more numerous. 
 
 Three types of the embryo may be recognized in Ophioglossum, represented 
 respectively by O. mohiccaiium, O. vulgatum, and O. pendulum. If, as seems not 
 unlikely, O. moluccanum is the most primitive of the three, some interesting points 
 arise as to the significance of the peculiarities exhibited by the embryo, which shows 
 only two organs aside from the foot, viz, the cotyledon and primary root, these 
 growing in an almost exactly opposite direction, without any clear line of demarca- 
 tion between them. I have ventured to draw a comparison between the sporophyte 
 of Ophioglossum and that of Anthoceros, assuming that the former has arisen from 
 some bryophytic type not unlike Anthoceros, by the development of a root from the 
 base of the sporogonium and of a special foliar organ from the basal meristem of 
 such a sporogonium. The embryo of 0. moluccanum approaches this hypothetical 
 form, as it consists only of leaf and root, and no stem apex is developed from it, 
 its growth being of limited duration. In this case the definitive sporophyte is a 
 secondary structure developed as a bud upon the primary root. In O. vulgatum, 
 however, the definitive stem apex, although of very late origin, is apparently a 
 product of the original embryonic tissue, but the first foliage leaf is of much later 
 origin. In 0. pendulum the formation of the leafy sporophyte is also secondary, 
 but neither stem apex nor leaf is developed from the embryo itself. 
 
 If, as the writer believes, Ophioglossum represents the most primitive type of 
 the fern series, it is quite conceivable that in O. moluccanum and its allies the embryo 
 represents the condition existing in the ancestral type from which these have sprung. 
 On the supposition that the leafy sporophyte is derived from a large bryophytic 
 sporogonium resembling that oi Anthoceros, there must have been a stage when the 
 sporophyte consisted of two parts only, the upper sporogenous portion, which later 
 developed into a sporophyll, as represented in Ophioglossum, and the root. Of 
 
THE EMBRYO 43 
 
 course, it is quite possible that the pecuhar origin of the detinitive sporophyte in 
 O. moluccanum and O. pedunculosiim is secondary, but this is by no means neces- 
 sarily the case. However, it seems highly probable that the extraordinary develop- 
 ment of the roots in O. vulgatuni and O. prmiuliim and the protracted subterranean 
 life in these species constitute a secondary phenomenon associated with the pro- 
 nounced saprophytic life of the gametophyte. 
 
 The apex of the shoot lies in a narrow depression between the base of the leaf 
 and a narrow ridge which extends around it on the side opposite the'leaf base. This 
 ridge is the beginning of the conical sheath characteristic of the shoot apex of the 
 older spf)rophyte. A transverse section of the young apex at about this time shows 
 that the apical cell is triangular in outline. 
 
 The first root of the young bud does not emerge until the leaf is nearly complete. 
 From this time on the further growth is due to the activity of the stem apex, from 
 which new leaves and rt)ots presumably are developed in the same way as in other 
 species that have been studied. How long it is before fertile leaves are formed was 
 not ascertained, but in the rapidly growing species of a tropical climate it is likely 
 that this takes place before long. The occurrence of very small fertile individuals 
 (fig. 55, B) points to this. 
 
 THK KMBRYO OF OPHIOGLOSSUM VULGA'lUM. 
 
 The embryo of O. vulgatuni (Bruchmann 1) differs from that of both O. 
 pendulum and O. moluccanum in its development, but in the earlier stages it closely 
 resembles the former species. As in O. pendulum, the root is the first organ to be 
 developed and it reaches a large size before there is any indication of the formation 
 of a stem apex or cotyledon. Bruchmann states that the root apex arises from the 
 hypobasal half of the embryo, but his figure (fig. 19, A) of the young embryo, where 
 the root apex is just visible, closely resembles a corresponding stage in O. pendulum, 
 where the root is certainly of epibasal origin; and, to judge from the elongated form 
 of the two-celled embryo which Bruchmann figures, one is inclined to believe that 
 the basal wall is really at right angles to the long axis of tlie embryo, and not parallel 
 with it, as l>e figures, so that the whole of the hypobasal part might be interpreted 
 as the foot, while the root would be of epibasal origin, as it is in O. pendulum. 
 
 The next stage figured by Bruchmann (fig. 19, B) is very much like the type of 
 0. pendulum, where only one root was developed at first, and in this case also we are 
 inclined to consider the whole of the enlarged base ot the embryo as the foot. 
 
 Before the cotyledon and stem apex can be recognized the first tracheary tissue 
 of the root has already been developed, and about the same time the apex of the 
 second root can be seen arising near the base of the primary root (fig. 19, C). 
 
 The cotleydon arises near the base of the root, upon the upper side of the 
 embryo, and forms a small conical protuberance, which grows from a definite apical 
 cell. At the base of the cotyledon, on the side turned away from the root apex, a 
 shallow depression is formed, and in this there may be seen a single large superficial 
 cell, which is apparently the apical cell of the very limited stem apex. Inclosing the 
 cotyledon and the stem apex is an elevated ridge which grows up about them and 
 finally forms a sheath, which rapidly grows so as to include the bud within a cavity 
 completely closed except for a narrow canal which opens outward. 
 
 While the cotyledon and stem apex are thus formed from superficial tissue in 
 (). vulgatuni, their early inclosure in the sheath suggests that we have to do with a 
 condition intermediate between the completely endogenous shoot apex in O. moluc- 
 canum and the entirely exogenous shoot in Butrychtum. 
 
44 
 
 THE OPHIOGLOSSALES 
 
 The second root arises quite independently of the stem apex, being developed 
 below the insertion of the primary root, and its vascular bundle joins that of the 
 first root near its base. 
 
 A rudimentary vascular bundle is developed within the cotyledon, and also 
 joins the bundle of the main root near its base, but no bundle at all is developed 
 within the stem. The cotyledon reaches only a very small size, but the second leaf 
 soon appears nearly opposite the cotyledon. 
 
 The second leaf also develops at an early period a vascular bundle, which 
 connects with the vascular bundle of the primary root near its point of junction with 
 the second root, and this exactly resembles the arrangement of the bundles in the 
 bud formed upon the root in O. moluccanum. It is evident that in the young sporo- 
 phyte of O. viilgatuni, as in O. moluccanum, the vascular system is made up entirely 
 of the bundles of the roots and leaves, the stem itself having no proper stele. 
 
 The second leaf finally emerges and appears above ground as the first sterile 
 green leaf of the young sporophyte. Bruchmann states that the second leaf does not 
 appear above ground until five years after it is first formed, and he believes that the 
 sporophyte, at the time its first leaf appears above ground, is nine to ten years of age. 
 
 The third leal may be fertile, but this is not always the case. 
 
 THE ANATOMY OF THE YOUNG SPOROPHYTE OF OPHIOGLOSSUM. 
 
 In Ophtoglossum pendulum the young primary root soon breaks through the 
 prothallium and elongates rapidly, but owing to its brittleness it is easily broken 
 off, and it is impossible to state here just how far it develops before the bud is 
 formed upon it. The development of the second root varies a good deal. The first 
 root may reach a length of 3 or 4 centimeters before the second root can be seen 
 at all. The growth of these earlier roots is in all respects similar to that of the later 
 ones. There is a large tetrahedral apical cell whose divisions are quite regular, 
 
 and there is soon visible the axial vascular 
 bundle which extends for some distance 
 into the foot, where it ends blindly. The 
 vascular bundle of the second root joins 
 the first at the junction of the latter with 
 the foot (fig. 21, D). 
 
 The first tracheary tissue appears at 
 the point of junction, and is made up of 
 short, somewhat irregular, pointed tra- 
 cheids with reticulate thickenings. From 
 this point the development of the tracheary 
 tissue proceeds toward the apex of the roots. 
 The bundle is diarch, as is plainly seen in 
 cross-sections (fig. 27, E'). The endodermis 
 is very clearly defined and the characteristic 
 radial markings are extraordinarily clear, 
 especially in sections treated with a double 
 stain of safranine and gentian violet. The 
 tracheary tissue is also beautifully differen- 
 tiated by this stain. The bundle is slightly 
 elliptical in form and the protoxylem ele- 
 ments appear at the foci of the elliptical 
 section. 1 he first appearance of the tracheary tissue is some distance back of the 
 apex and the development proceeds rather slowly. In the oldest part of the roots 
 
 older bud of Ophiogh. 
 of the bud. 
 
THK I;MI!KV() 
 
 45 
 
 exaniiiud rlic tun wlcm masses wcic iiiU(|Lial in si/.r, tin- lai<ii-i showing ahmit 
 half a dozen tiacheids in cross-section, the smallei two or three. Whether the two 
 protoxylems are uhimately joined hy intermediate tracheal tissue, so as to form a 
 continuous plate, as in the older roots of the adult plant, can not now be stated, hut 
 in no cases examined was this true, and it is not unlikely that in the primary root 
 the two xylem masses are permanently separated. The cells of the foot, as usual, 
 are more or less papillate where they are in contact with the tissue of the gameto- 
 phyte. They early become infected with the endophyte, which probably makes its 
 entrance from the prothallial tissue, and not from the outside. i"his point, however, 
 is not perfectly clear. The infected area follows the giowtli of the young root, but 
 leaves the apical tissues tree. 
 
 in (). moluccanum the leaf is the Hrst part of the young sporophyte to develop. 
 In the larger embryos the leaf forms a conical body, merging into a nearly globular 
 basal portion, which is the foot, and within this,' probably near the junction of the 
 epibasal and hypobasal halves of the embryo, the apical cell of the root is developed. 
 The leaf now shows a definite apical cell, triangular in .section and exhibiting a 
 regular segmentation. The inner cells of the segments form the axial strand of 
 tissue, which is continued through the embryo into the root. 1 he limits of the two 
 
 Section of petiole of tcttyledon o 
 Section of young root. 
 Vascular bundle of median regie 
 Vascular bundle of primary root 
 
 of O. pcn.lulu 
 
 primary organs, the leaf and the root, remain quite indistinguishable. The central 
 region, which remains surrounded by the prothallial tissue, is somewhat larger in 
 diatneter and the whole of this functions as a foot, although it is composed in part 
 of tissue belonging to the root and the leaf. The young leaf elongates rapidly after 
 it has ruptured the calyptra, and its apex begins to widen out, but still shows a 
 single apical cell. As the upper part widens out there is a division of the original 
 vascular cylindci, and there is developed within the leaf a reticulate s\stcm of 
 vascular bundles or veins (fig. 22. D, E). 
 
 As already indicated, the vascular stranil of the \(iung spoioplnte is continuous 
 through the cotyledon and root, and sections at difttrent points show essentially the 
 same structure. The petiole of the cotyledon (fig. 27, A), which is traversed by two 
 large lacuna, shows that the axial bundle is decidedly collateral in structure. The 
 xylem consists of a group of about half a dozen tiacheids at the inner limit of the 
 bundle, and no endodermis can be recognized. As a section of the bundle is made 
 
46 THK OPHIOGLOSSALES 
 
 in the mid-region of the sporoph}te (fig. 27, C), the only ditterence noted is a slightly 
 greater development of the xylem. The section of the root (fig. 27, D) presents 
 almost exactly the same appearance as that of the leaf. Whether we should call the 
 root bundle "monarch" or "collateral" is merely a question of terms. In the mid- 
 region the endodermis can be clearly seen, and it is then evident that the xylem is 
 separated from it by a single layer of pericycle cells. 
 
 I of Botrychium virginianumj containing a two-celled embryo 
 B. Four longitudinal sections of an embryo with 7 cells. X275. 
 
 THE EMBRYO OF BOTRYCHIUM. 
 
 The following account of the development of the embryo of Botrychium is 
 based mainly upon material of B. virginianiim furnished me through the kindness 
 of Professor Jeffrey. 
 
 In this species, as in the other Eusporangiatae, there is a marked increase in the 
 size of the fertilized ovum before the first division takes place. At the time of the 
 
 nbryos of Botrychi 
 
 first division the ovum is generally more or less elongated, but this is not always the 
 case. This elongation is less marked in B. hinaria (fig. 36), while in B. ohltquum 
 the ovum becomes much elongated before the first division occurs (see Bower 9, 
 fig. 266). 
 
THE EMBRYO 
 
 47 
 
 Till' first division wall, tlu' basal wall, as in tin- otlur J.iispoiangiatjc, is, usually 
 at least, transverse to the axis of the archcgoniuni (Hg. 28, A), ilach of" the cells thus 
 formed is next divided by a vertical wall, so that the embryo is dividc<l into approxi- 
 mately equal quadrants. Jeffrey states that the quadrants are next divided by a 
 third (transverse) wall, so that there is a regular octant formation (Jeffrey I, page 
 16), but adds that this segmentation presents various irregularities and this is 
 confirmed by my own studies of the young em- 
 bryo. In B. lunaria, according to Bruchmann, 
 the octant divisions are very regular. 
 
 In B. vtrglniiitiiim the young embryo is 
 often somewhat pointed above, and transverse 
 divisions of such embryos may show that the 
 octant formation is suppressed in the epibasal 
 region, and this tapering upper part of the young 
 embryo suggests a suspensor similar to that de- 
 veloped' tin Daima, but less clearly defined, and 
 probably not determined by the first division of 
 the embryo. Whether the suspensor found by 
 Lyon in B. obliquinn (Lyon 1) originates in a 
 similar fashion remains to be seen. This elonga- 
 tion of the upper part of the embryo, whether it 
 is considered as a suspensor or not, is doubtless, as Jeffrey has suggested, useful 
 in pushing the developing embryo deeper down into the prothallial tissue. 
 
 For some time, only anticlinal walls are formed, this showing especially plainly 
 in transverse section (fig. 31). Sooner or later periclinal walls are also formed, and 
 it is possible that in some cases the first periclinal wall in one of the large epibasal 
 cells may establish the initial cell of the root, but this is hard to decide. The early 
 stages of the embryo are extremely difficult to embed without shrinkage, and this, 
 together with the difficulty of securing a sufficient number of the right stages and 
 the impossibility of regulating the direction in which they are cut, makes the deter- 
 mination of the origin of the primary organs of the young embryo a peculiarly 
 difficult problem. 
 
 , 30. — Three longitudi 
 embryo of Bolrychi 
 root; /, the foot. ) 
 
 Neither |effrey nor Bruchmann was able to recogni/e a root initial until the 
 embryo had attained a large size, and its origin could not be determined. In the 
 embryo shown in fig. :50, I am inclined to believe that the cell r is really the root 
 initial. Unfortunately, stages between this and the one shown in fig. 32, where the 
 root initial is unmistakable, were wanting, but the position and state of develop- 
 ment of the root in the latter harmonize with this supposition. My fig. 32 is evi- 
 dently of about the same size as Jefl^rev's fig. 46, and probably an examination of 
 other sections of the series from which Jeffrey's figure was drawn would show the 
 root initial, as well as traces of the stem apex and cotyledon. 
 
48 
 
 THE OPHIOGLOSSALES 
 
 All of the organs of the young sporophyte arise, as Jeffrey showed, from the 
 epibasal region, and in this respect Botrychium virgtnianum agrees with the Marat- 
 tiaceae and with Ophioglossum. According to Bruchmann, the embryo in B. lunaria 
 remains quite undifferentiated up to the time it breaks through the calyptra, even 
 the root apex being unrecognizable at this time. It is therefore impossible to say 
 what relation the organs of the young sporophyte bear to the primary divisions of 
 the embryo. 
 
 b h, basal wall; /, foot; 
 
 In B. virginianum traces of the quadrant formation are still evident at a com- 
 paratively late stage, and there seems no reason to doubt the correctness of Jeffrey's 
 conclusion as to the epibasal origin of both the root and the cotyledon. As in 
 Ophioglossum pendulum and O. vulgatum, the root is especially conspicuous and 
 reaches a large size, while the stem apex and cotyledon are still inconspicuous. 
 Indeed, in B. lunaria, according to Bruchmann, no trace of either stem apex or 
 
 cotyledon can be made out 
 until the root has broken 
 through the calyptra. At 
 this stage (fig. 36, C, D) 
 the embryo of B. lunaria 
 bears a striking resem- 
 blance to that of Ophio- 
 glossum vulgatum. 
 
 It was found by Jef- 
 frey in B. virginianum 
 that the stem initial was 
 developed before the co- 
 tyledon could be seen, but 
 I have failed to verify this 
 in the specimens I have 
 examined. Fig. 32 shows 
 three sections of a series 
 taken from an embryo of 
 about the age of Jeffrey's fig. 46. I'his embryo was cut transversely to its long axis, 
 and the root apex is thus seen in cross-section. 1 he root here probably comprises 
 the whole of one of the epibasal quadrants, from the other of which, or from part of 
 it, perhaps a single octant, the stem apex and cotyledon arise in close proximity. It 
 is probable that part of this quadrant goes to form the " suspensor," or that epibasal 
 tissue which is not concerned in the formation of the young organs of the embryo. 
 Fig. 32, B shows a section passing through the young cotyledon, whose single apical 
 cell is already differentiated. This in section is triangular, and the cell is probably 
 of tetrahedral shape. Fig. 32, C, shows the section passing through the stem apex. 
 
49 
 
 Piobahlv the large cell {st) is the apical cell, hut this is not ahsolutely certain. In 
 foini and position this cell resembles the apical cell figured by JcttVey (Jeffrey I, 
 fig. 48) in an older embryo. It may be said that in Jeffrey's figure the beginning of 
 the cotyledon, situated between the stem apex and the root, although not lettered, is 
 evidently present. 
 
 As 'the embryo grows the loot rapidl) increases in length, and together with the 
 foot comprises the greater part of the embryo. The foot is very large, usually 
 nearly hemispherical in foim, but not infrequently a good deal elongated (fig. 35), 
 and penetrates deep into the prothallial tissue. 
 
 The apical cell of the root is very conspicuous and can be made out without 
 difficulty as a large cell, triangular in form, whether it is viewed in longitudinal 01 
 transverse section. Segments are cut off in regular sequence from all of the four 
 faces, and as these are relatively large and contain but little granular contents the 
 apical meristem sho\\s\ei\ cltarly against the smaller celled and more deeply stained 
 
 A-C. Three horizontal 
 D. The stem region 
 
 adjacent tissue. In the earlier stages the divisions in the segments cut oH fmni the 
 apical cell do not show absolute uniformity. The first wall in the young segment is 
 probabU- anticlinal, and seems to be followed by a periclinal wall or walls, cutting 
 off inner cells which contribute to the very large plerome cylinder ot the young root. 
 The root cap, which is very massive, is derived in part from segments cut oft directl\ 
 fiom the apical cell, and in part from cells separated by periclinal walls from the 
 outer part of the lateral segments of the apical cell. 
 
 The development of the plerome begins very early, and it soon forms a con- 
 spicuous massive strand of procambium cells extending from tlu' root apex to the 
 junction of the root antl foot, where it ends abruptly on the lower side, but bends 
 upward on the upper side, and is extended as a simjile axial strand of tissue into the 
 cotyledon. 
 
 The cotyledon, which is first recognizable at an early stage in the development 
 of the embryo, at the stage in question (fig. 35, J) projects slightly as a broad, flat- 
 tened cone, strongly bent away from the root. Fig. 34. //, shows the cotyledon cut 
 
50 
 
 THE OPHIOGLOSSALES 
 
 through horizontally. At the slightly projecting apex there is a single apical cell, 
 although it is not always quite certain which is the apical cell and which is its 
 youngest segment. The broad base of the cotyledon is extended laterally, like the 
 stipules of the older leaves, and there is thus inclosed a slightly depressed area, very 
 much as in the embryo of Ophtoglossum vidgatum, and within this is situated the 
 stem apex, close to the base of the cotyledon, which bends over it so as to leave 
 only a narrow space above the stem apex. A section across the base of the cotyledon 
 shows a group of small cells indicating the section of the young vascular bundle 
 which lower down (fig. 34, C) joins the young bundle of the primary root. 
 
 The relative position of the young organs just before the root ruptures the 
 calyptra is best seen in a longitudinal section. Such a section (fig. 35, A) shows the 
 very large foot, occupying approximately half of the embryo, and deeply embedded 
 in the prothallium. Above this is the root, occupying the major part of the epibasal 
 region of the embryo and varying somewhat in position, perhaps due to the early 
 divisions in the young embryo. It may lie in a plane almost coincident with the 
 cotyledon, or it may make a marked angle with the latter. In the former case the 
 
 Fig. 35. 
 
 Median section of a young sporophyte of Botr 
 An older stage; /', second leaf; r^, second i 
 
 young vascular cylinder is continued almost straight into the cotyledon; in the 
 latter it bends sharply upward to meet it. 
 
 The stem apex, which is usually more or less oblique, is small and incon- 
 spicuous, but usually a single initial cell may be made out without difficulty, although 
 sometimes it is not easy to distinguish this from the younger segments which have 
 been cut off from it, as the apical meristem is very shallow, and in longitudinal 
 section has the appearance of a columnar epithelium. There seems to be, at least 
 in the earlier stages of development, some variation in the form of the initial cell, 
 which may have a pointed base, or may be truncate below, and while later it 
 usually shows the form of a three-sided pyramid, in these earlier stages it often 
 approximates the truncate form found in Ophioglossum moluccanum. 
 
 The apical meristem of the stem is very limited in extent and, so far as could 
 be made out, the segmentation at first does not follow any regular scheme. While 
 in both root and cotyledon a central strand of procambium is developed at a very 
 early period, no trace of anything which can he interpreted as a "stele" is formed in 
 the young stem. The tissues derived from the further segmentation of the apical 
 meristem remain undifferentiated parenchyma and contribute only to the central 
 pith of the hollow woody cylinder which later traverses the axis of the young 
 
THI-: EMBRYO 51 
 
 sporophyte. In short, the strand of vascular tissue in the young sporophyte at the 
 time the first root emerges is composed solely of the cf)alescent strands of the root 
 and leaf exactly as is the case in the young sporophyte of Ophioglossinn nioluicinnitti, 
 and there is absolutely no trace of a cauline stele. 
 
 The foot is composed of large parenchyma cells, which become smaller 
 toward the periphery, where they usualh' sliowmoreor less dense contents and some- 
 times form a quite definite laycrof epithelium-like cells, which stain much more deeply 
 than the inner tissue of the foot. This outer layer of cells is presumably active in 
 the absorption of nutriment from the adjacent prothallial tissue. Bruchmann states 
 that previous to the emergence of the root of the young sporophyte of B. liuuiria 
 the cells of the embryo are densely packed with granular matter of apparently 
 albuminous nature; but after the emergence of the root starch is developed abun- 
 dantly in the cells of the foot. This appears to be true also of i^. virgtmanuni. 
 
 Compared with B. virgininniim, the embryo of 5. lunaria (fig. 36) is character- 
 ized by the lesser development of the foot and the later appearance and rudimentary 
 character of the stem apex and the cotyledon. 
 
 The stem apex in B. liuiana (Bruchmann 2) is first evident as a slight sujierfi- 
 cial depression near the base of the massive root. A single superficial cell becomes 
 differentiated as the apical cell of the shoot, and a small group of cells is formed 
 before the rudimentary cotyledon develops. Both stem apex and cotyledon remain 
 very inconspicuous. The cotyledon is developed as a scale-like rudiment, which 
 never develops into a foliage leaf. Bruchmann states (Bruchmann 2, page 223) 
 that from seven to nine of these rudimentary leaves are developed before the first 
 green leaf appears above the earth. Hofmeister (Hofmeister 1) believed that three 
 of these rudimentary leaves were developed during the first year, and that the 
 second year the first spore-bearing leaf appeared above ground. Bruchmann thinks 
 that this is not the case and that the first spore-bearing leaf requires five years for its 
 development, as it does in the older sporophytcs, and that only o\w leaf is formed 
 each year, as in the older plant. 
 
 The first organ to penetrate the calyptra is, as we have seen, the root, which in 
 B. virgin I aiitit?! is very large and quite overshadows the relatively inconsjiicuous 
 bud at its base (fig. 7, B). At the time the root first emerges the vascular bundles 
 still have the form of procambium, and it is not until the young root has a length of 
 several millimeters (5 to 20 millimeters, according to Jeffrey) that the first tracheary 
 tissue is developed. The first tracheary tissue arises at the base of the root and the 
 development, as usual, proceeds toward the apex. The primary root in Botrvclinnn 
 virginiauiim is usually diarch, but triarch roots are sometimes found, and leffrey 
 says that the size of the root has no relation to the number of proto.xylems formed. 
 The first tracheids are short and reticulately marked, as they are in Ophioglossum. 
 The surface of the root is quite destitute of root hairs, as is the case in the adult 
 sporophyte. 
 
 In B. limaria the puclominance of the root over the shoot is even more marked 
 than in B. vngiiuaiunu, for tin re are se\eral roots, sometimes four to fi\e, ilevelopcd 
 before the first foliage leaf is formed, and the bud remains extremely small and 
 inconspicuous. 
 
 A longitudinal section of the cot\ledon of i?. virginiaiiiitru just before it breaks 
 through the overlying cahptra, shows it to be strongly curved awa\- from the root 
 and overhanging the cleft within which lies the stem apex. If the section is exactly 
 a median one, there may be seen near the tip, but lying somewhat toward the lower 
 side, a marginal cell of triangular outline, which, from the arrangement of the cells 
 about it, is evidentU' the apical cell. The cotyledon, instead of being straight as it 
 
52 
 
 THE OPHIOGLOSSALES 
 
 is in Ophioglosstim and Botryiliiiim huioria, approaches the circinate form of that of 
 the Marattiaceae and the typical ferns. The young vascular bundle is now clearly 
 evident, lying somewhat toward the inner side of the leaf, and it can be readily 
 followed downward until it joins the root bundle, with which it is continuous. 
 
 In B. virginianutn, while the cotyledon is still quite small, the second leaf 
 appears close to the stem apex and nearly opposite the cotyledon, the sheathing 
 base of which surrounds the stem apex, together with the young second leaf. The 
 growth of the second leaf is also probably from a tetrahedral apical cell, and this 
 is the case with all of the later leaves. While the second leaf is still very small, 
 there begins the differentiation of the corresponding leaf trace, which joins the 
 bundle of the primary root close to the point at which it gives off another bundle 
 destined for the second root, the apical cell of which is cut out from the tissue near 
 the base of the second leaf, at a point almost directly opposite the cotyledon. 
 Indeed, the second root may be said to bear much the same relation to the second 
 leaf that the primary root does to the cotyledon (fig. 35, B). 
 
 Fin. 36. — Bolryclilum lunaria (after Bruchmann). 
 
 The cotyledon in B. virgtniaiium (fig. 7) is extraordinarily developed, reaching 
 a size that probably is unequaled by any other fern. Its slender petiole reaches 
 a length of 5 to 6 centimeters, and the lamina, unlike the simple oval primary leaf of 
 Ophioglossiim, is ternately divided and the segments deeply cut. The venation, 
 instead of being reticulate as in Ophioglossiim, is apparently pinnate, but this is 
 really the result of an unequal dichotomy, as it is in the ultimate branches of the 
 veins in most ferns. 
 
 The base of the petiole forms a stipular sheath like that found in the later 
 leaves, and it is traversed by two vascular bimdlcs, which unite into a single trace 
 before joining the bundle of the root. 
 
 While in B. hinaria all of the early leaves remain rudimentary, and the first 
 one to appear above ground is a fertile one, there are sometimes intermediate stages 
 which may show a small ternate lamina and traces of the fertile spike, although they 
 never appear above ground (see Bruchmann 2, fig. 50). 
 
 If horizontal sections of the young sporophyte are examined, at about the time 
 the cotyledon first emerges (fig. 41), the latter will be seen to have the form of a 
 
THE rMr.KVO 53 
 
 crescent, the horns of wliich unite at ahout the le\el of the stem apex, so that the 
 latter, together with the section of the seccjnd leaf, is completely inclosed by the 
 sheath formed at the base of the cotyledon. The leaf trace in both the cotyledon 
 and the second leaf is broadly oval in outline, increasing in breadth lower down, 
 so that the section there becomes crescent-shaped and the two together form an 
 imperfectly closed ring of small cells, the procambiiim of the future vascular cyl- 
 inder of the axis, which clearly is formed entirely from the united leaf traces. 
 
 The apical cell of the stem appears triangular in cross-section, and from it are 
 cut off regular segments, but, as we have seen from the study of the longitudinal 
 section, the tissues which are developed from these segments below the apical cell 
 remain undifferentiated and form merely a mass of large parenchyma cells with 
 thin walls, which fill the space within the ring formed b\- the confluent leaf traces, 
 and thus constitute the pith or medulla of the stem. 
 
 Jeffrey (Jeffrey 1, page 19) refers to the occurrence of short tracheids in the 
 prothallial tissue in one case which came under his observation. Theie was seme 
 evidence that an embryo had been present, but it had disappeared, and he thought 
 it possible that this was a case of apogamy, similar to that which has been repeatedly 
 observed in Ptens crettca and in various other leptosporangiate ferns. If apogamy 
 does really exist in B. virginuuuitti it is the only case which is known amon<^ the 
 Eusporangiatx> 
 
 Botrychnnii obliqiiiim (Lyon 1) differs strikingly in the foim of the eml)r\() 
 from either B. virguiiainini or B. Imuiria, and these differences are so marked that 
 Lyon proposes to separate B. obliquum and a number of allied species as a new 
 genus, Sceptridium. The ovum in Botrvcliium obliquum becomes very much 
 elongated after fertilization, and it is probable that the first wall separates the part 
 next the archegonium from the inner cell, the former developing into a suspensor, 
 very much as we shall see that it does in Dancea, and this suspensor becomes well 
 developed and is conspicuous in the later stages of the embryo. Unfortunately, 
 Lyon gives no details of the further history of the embryo until the definitive organs 
 are far advanced. He states that the suspensor becomes elongated, pushing down 
 the rest of the embryo, which he calls the "protocorm," believing that it is homol- 
 ogous with the so-called "protocorm" of Lvcopodiuni — a resemblance which would 
 probably disappear were a more thorough study of the subsequent history available. 
 This development of a suspensor is the first case described in the ferns, but, as we 
 shall see presently, a similar suspensor is regularly present in the genus Dancca, 
 which in other respects shows some interesting resemblances to B. obliquum in the 
 development of the embryo. Moreover, as we have seen, there is often an elonga- 
 tion of the upper cells of the \'oung embryo of B. virginianitm, which suggests a 
 rudimentary suspensor. 
 
 Lyon figures only a very advanced stage of the embryo, at about the time when 
 the root first emerges from the prothallium. At this time, aside from the presence 
 of the suspensor, the most marked difference between B. obliquum and B. virgin- 
 iaiium is the apparent absence of the foot in the former species, while this organ is 
 so remarkably developed in B.virginianum. The root \nB. obliquum penetrates the 
 lower surface of the gametophyte, the cotyledon breaking through on the upper 
 side, so that the relative position of the leaf and root is like that of Ophioglossum 
 moluccauum, or the Marattiace;e. In the orientation of the root and shoot, there- 
 fore, B. obliquum is much more like Dinuni or Opliioglossiim molucciiiium than it 
 is like B. vitgiuianum. 
 
 It is to be hoped that further investigations will lu- math- u|ion this extrenieh 
 interesting species. We anticipate that when the histor\ of the earlier stages is 
 
54 THE OPHIOGLOSSALES 
 
 known it will be found that a large foot is present in the embryo, as it is in the 
 Marattiaceae and the other Ophioglossaceae, and the apparent absence of this foot 
 in the older embryo is due to the root apex being formed deep in the central tissue 
 of the embryo, as it is in the Marattiaceae and in Ophioglossuni moluccanum, and 
 pushing its way downward through the foot, which thus becomes indistinguishable 
 from the outer tissue of the root. 
 
 Except for the absence of the foot the organs of the older embryo of 5. oblt(juum 
 show the same relative positions as those of B. virginiatnim, and, as in the latter, 
 the first leaf is apparently a functional green leaf which appears above the surface 
 of the earth. 
 
 THE EMBRYO OF HELMINTHOSTACH^'S. 
 
 The embryogeny of Helmtnthostachys, the third genus of the Ophioglossaceae, 
 is unfortunately very incompletely known. Lang (Lang I), to whom we owe the 
 only published account of the gametophyte and young sporophyte, was unable to 
 secure any young embrj^os and only very scanty material of the later stages. Lang 
 says of the youngest forms which he found (Lang 1, page 40): 
 
 "The large hemispherical foot is deeply inserted in the tissue of the prothallium. 
 The upper portion of the embryo had burst through the covering also of the latter; 
 in it can be distinguished the primary root — the first leaf— and (covered over by 
 the sheath of the first leaf) the depressed apex of the stem. The position of the 
 organs is thus essentially similar to what is found in Botrychiiim virgintanmn." 
 
 Lang's material was collected in the Barrawa Reserve Forest, in Ceylon, and 
 in February 1906 I visited the same locality and collected a large number of young 
 sporophytes, many of which were still connected with the gametophyte; but none 
 of the gametophytes were young enough to show young embryos. As we have 
 already suggested, it seems very probable that the gametophyte of HehnintJi- 
 ostachys, like that o{ Ophioglossum nioluccaiuim, is annual and produces normally a 
 single sporophyte, after which it perishes. The development of the gametophyte is 
 probably dependent upon the annual inundation of the forest, and this perhaps 
 accounts for the fact that all of the young sporophytes were of about the same age. 
 
 The youngest specimen which I found is shown in fig. 10. The petiole of the 
 leaf was about i centimeter in length, and at the tip the minute lamina, strongly 
 bent over like the young leaves oi Botrychiiim vtrgtnianurn, could be seen, showing 
 the three lobes which characterize the fully expanded leaf. The cotyledon is rudi- 
 mentary in Hehninthostachys, the "cotyledon" described by Lang being really the 
 second leaf. At the base, but separated from the root by a short internode, is 
 a swelling which marks the position of the apical bud, inclosed within the hollow 
 sheath at the base of the petiole. The root at this stage is still quite short, the first 
 green leaf seeming to be more precocious than in Botrychium and thus resembling 
 Ophioglossuni moluccanum. 
 
 As the leaf grows the lamina expands and is seen to be ternate in form. Very 
 often the two lateral lobes are of unequal size and show that the ternate form is due 
 to an unequal dichotomy, such as is common in many ferns in the early leaves, which 
 are transitional between the dichotomously divided cotyledon and the pinnate 
 leaves of the older sporophyte. The ultimate divisions of the veins, as in Botrych- 
 ium, are dichotomous.* 
 
 * In a recent note (Lang 2) it is stated that a suspensor also occurs in Helminthoitachys. 
 
55 
 
 1. rill'. ^ouNc; si'oRoi'iivi 1 
 
 TilK YOUNC; Sl'OROl'll^'lK OK Ol'i IIO(;LOSSUM. 
 
 As we have seen, the single vascular strand in the young sporophyte of Ophio- 
 glossiim rnoluccanum is continu<JUs through the cotyledon and root, and sections 
 at different heights show essentially the same structure throughout. The petiole 
 of the cotyledon (figure 27, A), which is traversed by two conspicuous lacunae, 
 shows that the axial bundle has a perfect collateral structure. The xylem consists 
 of a group of about a half dozen tracheids on the inner side of the bundle, of which 
 the endodermis, however, is not evident. If a section made in the mid-region of 
 the plant is examined, the only difference between its bundle and that within the 
 petiole of the leaf consists in the somewhat greater development of the tracheary 
 tissue. The section of the root bundle also almost exactly resembles that of the leaf". 
 Whether we call tln' root bundle monarch or collateral is merely a question of terms. 
 
 ng sporophyte of Ofiliio^hnum 
 developed; r, primary root of sporophyte; r', first root of bud; 
 D. The stem-apex more liighly magnified, showing the third leaf, /^. 
 
 In the mid-region of the sporophyte the endodermis is well marked, and this is even 
 more the case in the root. In the mid region the xylem is separated from the endo- 
 dermis by a single layer of pericycle cells. 
 
 About the time that the leaf emerges the first root of the bud begins to develop, 
 and, like the leaf, it seems to arise quite independently of the stem apex. The 
 tetrahedral apical cell is cut out from the cortical tissue of the root some distance 
 below the apex of the young bud, and as soon as the apical cell is established active 
 segmentation begins in the tissue below the bud and a strand of procambium is 
 developed, connecting the young root apex directly with the vascular bundle of the 
 primary root. Thus the vascular bundles of both the first leaf and the first root of 
 the young bud are directly connected with the primary root of the young sporo- 
 
 phyte and have nothing to di 
 which is still very small. 
 
 ith rhi 
 
 the 
 
 ipex 
 
 )t the stem ot tlu' Inu 
 
56 THE OPHIOGLOSSALES 
 
 The two bundles, belonging respectively to the leaf and the root of the bud, 
 join the bundle of the primary root very near together, and from near this point of 
 junction the development of the tracheary tissue proceeds toward the apices of the 
 young organs. The first tracheids, like those found m the median region of the 
 young sporophyte, are short reticulate ones. 
 
 The first leaf of the bud grows rapidly and closely resembles the cotyledon in 
 form and size, and at first it often looks as if two young sporophytes of equal size 
 were growing from the prothallium, but a very slight examination shows that the 
 second leaf belongs to the bud upon the primary root. The second leaf of the bud 
 also seems to be formed independently of the stem apex and arises nearly opposite 
 the first one (fig. 37, A). Its development is practically the same as that of the 
 first leaf, and it pushes through the outer tissue of the root at a point removed by an 
 appreciable distance from the base of the first leaf, from which it is separated by a 
 sheath about the base of the latter, formed by the outer root tissues. 
 
 The vascular bundle in the second leaf is soon developed and can be followed 
 down to the bundle of the first root of the bud, to which it bears somewhat the same 
 relation that the bundle of the first leaf of the bud does to the primary root of the 
 sporophyte. Fig. 37 will show clearly the arrangement of the vascular bundles in 
 a young bud in which the first leaf is fully developed, while the second and third are 
 pretty well advanced. The latter (/^) shows very plainly the conspicuous apical 
 cell and the beginning of the leaf trace. The apical cell of the stem at this time is 
 sufficiently conspicuous, but the amount of tissue surrounding it is quite limited 
 in extent. In fig. 37, D, it is probable that the fourth leaf has begun to develop. 
 Above the apex of the stem the cavity of the leaf sheath can be clearly seen. By 
 this time the first root of the bud has emerged below the first leaf and the tracheary 
 tissue is well developed in its basal region. 
 
 The second root of the bud originates below the apex of the stem in the vicinity 
 of the base of the second leaf. The vascular strand of the third leaf joins it, and the 
 third leaf seems to have somewhat the same relation to the second root that the 
 second leaf does to the primary root. 
 
 The study of the development of the young bud shows that, up to the time of the 
 production of the third leaf, the young organs arise quite independently of each 
 other from the tissue of the primary root. The young sporophyte is, so to speak, 
 made up by the union of several independent members. The third and fourth 
 leaves arise from the stem apex as all of the later ones do, but the first and second 
 leaves and the first root show no recognizable relation to the stem apex. Their 
 vascular bundles are directly connected with the bundle of the primary root of the 
 young sporophyte, and are in no way associated with the tissues which belong to 
 the stem region of the bud. No material was available for the study of the origin 
 of the leafy bud in 0. pendulum, but the first leaf evidently develops very much 
 later than it does in O. moliiccanum. Young leaves 10 centimeters or more in length 
 were found, which, there is some reason to suppose, were the primary leaves of the 
 young plant; but they were not recognized as such at the time they were collected 
 and it is therefore impossible to say whether they really were the first leaves devel- 
 oped. After the study had been made of the young sporophyte in O. mohiccanum, 
 and the secondary origin of the stem apex was made out, it was recognized that these 
 small leaves in (J. pnuhilntu also were probably the primary leaves of the plant, but 
 it was too late then to trace the connection of the roots from which they developed 
 to their connection with the primary root of the young sporophyte. 
 
 I was unable to determine just how soon the fertile leaves are developed from 
 the young sporophyte in O. moluccamnii. In all of the specimens that were sectioned, 
 
THK YOIINC SPOROIMIVTI 
 
 57 
 
 the Hist three or four leaves were sterile and ]iraetHaliv like the C()t\le(l()n, and it 
 is not certain which leaf, luulei oidinaiv circumstances, first gives rise to the spore- 
 bearing spike. In 0. moliiccanum and the other tropical species of 0/)hioglossum 
 growth is continuous, and it is evident that the development of the leaves does not 
 take the long period required in O. vulgatum and other species of temperate regions, 
 where growth is interrupted each year and where only one leaf is developed in the 
 
 12:L\'^ z...-{J ''A^\ 
 
 Sii of a series of transverse seclions of a young sporophyte of Ophioglossuin moliiccanum, still attached to priirarj 
 root,r. Section B passes through the stem apei. X35. 
 
 season. Mettenius states that O. pedunculosum, which, as we have seen, may per- 
 haps be identical with 0. moluccanum, develops three leaves each season. 1 his 
 was in the Botanical Garden at Leipzig, and it is highly probable, under the 
 much more favorable conditions of its native tropical habitat, that this number 
 would be exceeded. There was not time to make investigations in regard to this 
 ^ point, but it is very certain 
 
 that several leaves are de- 
 veloped in the course of each 
 year and that the develop- 
 ment of the individual leaf 
 does not require the long pe- 
 riod necessary in the species 
 of cold climates. 
 
 Figures 38 and 39 show 
 several transverse sections 
 forming a series taken from 
 a young sporophyte with 
 three fully developed leaves 
 and two younger ones. This 
 bud was developed from a 
 small root, but it was not 
 certain that this was the 
 primary root of the embryo. 
 The spirall)' arranged leaves 
 show the two-fifths diver- 
 gence. The seconti ami 
 third leaves of this Inul 
 were succes.sively larger than the first leaf, but the section of the petiole showed 
 a single centrally placed vascular bundle of the same type as that ol the piimarv 
 
 :ion of the sporophyte shown in 
 s the primary root. X about 60. 
 cular bundle of first leaf, more higlily 111 
 n apex, showing youngest leaf trace. /*; 
 
 3S, she 
 
 apical cell of ; 
 
58 THE OPHIOGLOSSALES 
 
 leaf and all of these leaves were sterile, nor could any indication of the develop- 
 ment of a central spike be seen in the development of the fourth leaf. The first 
 leaf of the bud, like the cotyledon, showed two large lacunae in the petiole, but in 
 the second leaf there was but a single one and this was interrupted at intervals, 
 while the petiole of the third leaf appeared almost solid, there being only small, 
 irregular, intercellular spaces, such as are always found in any loose parenchyma. 
 All of the vascular bundles are collateral. In the lower portion of the third leaf 
 eleven tracheids could be seen in the transverse section of the bundle, but higher 
 up the amount of tracheary tissue was reduced. 
 
 Figs. 38, A, and 39, A, were cut just above the stem apex and showed clearly 
 the arrangement of the first four leaves. The youngest leaf is still entirely surrounded 
 by the conical stipular sheath belonging to the third leaf. Fig. 38,/^, is taken imme- 
 diately above the stem apex and passes through the very young fifth leaf. This 
 section shows very satisfactorily the arrangement of the first five leaves. Section C 
 lies a short distance below the stem apex, r is the root from which the bud has 
 arisen, and the group of tracheids at its junction with the bud marks the point 
 of union of the bundles of the first leaf and the first root of the bud. In the center 
 of the section may be seen a mass of large-celled parenchyma, the central pith of the 
 stem, which is derived from the large-celled meristem of the stem apex. The broken 
 ring of procambium surrounding the pith is composed of obliquely cut traces of the 
 fourth and fifth leaves and the basal tissue of the second root of the bud. The leaf 
 traces from the second and third leaves are still free in the cortical tissue of the young 
 stem. E and F pass through the base of the bud. The stout bundle of the first 
 root is seen connected with the bundle of the root upon which the bud was developed, 
 and the second leaf trace (/'^) bends in to meet the bundle of the root but is cut 
 at a level above the point of junction. Close to the second leaf trace is the lower 
 part of the fourth leaf trace, which is connected with the trace of the second root 
 of the bud. The steles of both the first and second roots of the bud run horizon- 
 tally across its base before they emerge, and the leaf traces are almost perpendic- 
 ular to these. The second leaf trace at the base of the bud joins the stele of the 
 first root and the short tracheids curve out trom the opposite side of the bundle of 
 the first root to unite with the base of the third leaf trace, which does not, as might 
 have been expected, show any very obvious relation to the second root. At the 
 point where the first root bends downward to emerge from the overlying cortical 
 tissue the tracheary tissue of this, with the bases of the leat traces, forms a large 
 mass of irregular, broad, reticulate tracheids, occupying the center of the base of 
 the young bud. It was probably the presence of this mass or tracheary tissue at the 
 base of the bud which led Rostowzew to state that the bundle of the bud is at first 
 solid. It is possible that the buds formed on the old roots ot 0. vulgatum may be 
 somewhat different in structure from the early buds in O. inoluccanum and that 
 there really may be such a solid stele at its base. 
 
 According to Rostowzew (Rostowzew 1, 2) and Poirault (Poirault 2), the buds 
 upon the older roots ot O. vulgatum usually arise close to the apex, much as they 
 sometimes do in the formation of the bud upon the primary root of O. moluccanum. 
 Van Tieghem thought that the apex of the root itself became changed into the 
 leafy shoot, but both Poirault and Rostowzew demonstrated that the young bud 
 originated from a segment of the apical cell of the root and not from the apical 
 cell itself; the latter continues its growth and the root thus grows beyond the 
 point of insertion of the young bud. Rostowzew states that an outer cell of the 
 segment divides into two superimposed layers, of which the superior one, or the 
 one nearest the root apex, gives rise to the stem apex of the bud, while from the lower 
 
YOUNG SPOROI'HVTr: 
 
 59 
 
 one, i. e., the one turned toward the base ot the root, tlie first leaf arises. It thus 
 appears that in these adventitious buds in Ophioglossum viilgatiitn there is much 
 the same arrangement of the first organs of the young bud that we have seen to be 
 the case in O. moluccanum, the leaf and the stem apex being virtually independent 
 organs. From Rostowzew's figures it is evident that the bundle from the first leaf 
 of the bud in O. vulgatum is connected directly with the main root in very much 
 the same way as it is in the primary bud in (). moliiccaiium; but apparently the first 
 root of the young bud is inserted higher up and does not connect directly with the 
 primary root; this point, however, needs further elucidation. Moreover, according 
 to his statements, there is a good deal of difference shown as to the time of the emer- 
 gence of the leaves and roots, three or four roots sometimes being well developed 
 before the first leaf expands. This reminds one of the behavior of tlie primary 
 sporophyte, where Bruchmann found that sc\iral roots wvvv developed before the 
 first green leaf made its appearance. 
 
 40. — Vertical secti 
 right angles to the 
 
 young sporophyte of Botrychium virginumum. llie sections were c 
 the root. sA, stipular sheath of cotyledon; E, section of primary 
 in this case was triarch. X20. 
 
 THK YOUNG SPOROPHYTE OF BOTRYCHIUM. 
 
 Jeffrey (Jeffrey 1) has described at some length the structure of the young 
 sporophyte of B. virginiamnn, and Bruchmann (Bruchmann 2) has studied the 
 earlier stages of the sporophyte in B. lunarta. The present account is based mainly 
 upon my own studies of 5. vtrgiiiianum, made from material furnished me through 
 the kindness of Professor Jeffrey. Jeffrey found that from the beginning the con- 
 spicuous stele found in the axis of the young sporophyte was a hollow cylinder, 
 which he assumes to be a really cauline structure, and not made up of leaf traces. 
 The results of my own studies indicate that, as in the case of Opiitoglossuni, the 
 vascular system of the stem is made up exclusively of confluent leaf traces. 
 
 Fig. 35, /i, shows a longitudinal section of a young sporophyte in which the first 
 two leaves are evident and the continuation into these of the tissues of the vascular 
 bundle of the root is very clear. The large central pith continues up into the stem 
 apex, but the whole of the fibrovascular tissue is continued into the leaves, there 
 
60 
 
 THE OPHIOGLOSSALES 
 
 being no trace of procanibium inside of the leaf traces in the apical region of the 
 stem itself. Fig. 40 shows a sporophyte which was cut at right angles to the primary 
 root, which in this case had a triarch bundle. Near its base the three xylem masses 
 increase in size and form a nearly complete ring of tracheary tissue, which higher 
 up is continued into the leaf traces. The third leaf has already begun to form, but 
 is still very small. Below it, however, can be clearly seen the beginning of its large 
 leaf trace, which can be followed downward to its junction with the second root, 
 which is now just about ready to emerge. The stele of the second root is con- 
 nected with the central cylinder of the axis near its junction with this third leaf 
 trace. 1 he stem apex occupies a very small area, crowded between the base of the 
 second and third leaves. Its deep, narrow, apical cell is not as conspicuous as is 
 often the case. 
 
 s through 
 
 tern apex, 
 
 (, which 1 
 
 IS of a young sporophyte of Botrschium virginianum. X75. 
 i shown more enlarged in G; r~, second root; /^, second leaf. 
 
 A similar sporophyte is shown in cross-section in fig. 41. This was cut in 
 the plane of the first root, which makes a sharp angle with the base of the cotyledon. 
 The thick stele of the root is much expanded where it joins the leaf traces to form 
 the beginning of the tubular stele of the axis, and at the point of junction there is 
 a large nearly solid mass of short, irregularly disposed, reticulate tracheids. This 
 young sporophyte was evidently the further development of an embryo of the 
 type shown in fig. 35, A. The trace of the cotyledon makes a right angle with the 
 stele of the root and on the side opposite to its junction with the stele of the root 
 there can be seen the trace belonging to the second leaf, which is already well 
 developed. The apical cell is tetrahedral in BrArychiurn virginianum and the 
 cells are cut off in regular .segments from its three lateral faces. Segmentation is 
 
THK YOUNG Sl'OROPHYTE 
 
 61 
 
 evidently slow ami there is usually a good deal of difference in the stage of dexelop- 
 ment in the three segments composing a single cycle. The first division in each 
 of the lateral segments cuts off a small inner cell from a large outer one, and the 
 latter is then divided into two by an anticlinal wall. The divisions were not followed 
 beyond this stage and the limits of the segments comprising the .second cycle can 
 not be easily recognized, as by this time the youngest leaf begins to grow and the 
 stem apex is thus crowded into a ver\- small area between the bases of the two 
 youngest leaves. The shallow mass of meristematic tissue comprising the stem apex 
 merges gradually into the large-celled parenchyma which makes up the pith cylinder 
 inside the hollow stele of the axis. 
 
 Longitudinal .sections of a much older sporophyte are shown in fig. 42. 
 
 Except for the vascular bundles, the whole of the tissue of the young sporophyte 
 is composed of large-celled parenchyma which, especially in the region belonging 
 to the foot, contains large quantities of starch. 
 
 The form of the young leaves in B. virgiiiitnuuu is very different from that 
 in B. Innana. In the former the young leaf is bent forward over the stem apex and 
 the apical cell is at the tip of this bent-over portion. In B. Iiuitin'ti, however, the 
 
 Three longitudinal sections of a young spor )pliyte of Botrychtitm virt^ittiatiutti witli sever 
 /icr, absciss layer of peridirm, /, /, the \oungest leaves, sh, stipular shcitli. X20. 
 
 \u 
 
 bent-over jiortion becomes the stiimlai legioii and the apical cell of the 
 arises in the convex upper portion, so that the leaf grows straight u|iward instead 
 of being bent over as it is in B. virgiiiKuuttii. T!ie latter species in this respect 
 resembles the Marattiace:e. 
 
 The stipules in B. vtrgniianiim are lateral structures which exteml around the 
 next youngest leaf and the stem apex, but the sheath has a nariow cleft in front, so 
 that we may really speak of two stipules instead of a single stipular sheath (fig. 
 42, A^. In this respect also there is a strong resemblance to the Marattiaceap. 
 
 Fig. 43 shows a series of transverse sections of a young sporophyte of B. vir- 
 gintfiniitn in which tlie fourth leaf was just evident. Fig. 4;5, A and B, are near 
 the base of the fourth leaf and show the arrangement of the first four leaves. The 
 base of the cotyledon {rot) is a good deal flattened and the petiole is seen to be 
 traversed by very conspicuous lacunae, reminding one of those in the cotyledon of 
 Ophioglossiim mohiccanum. The conspicuous stipules can be seen extending 
 partly around the group of younger leaves, but not completeh' inclosing them. The 
 two vascular bundles of the petiole are beginning to unite to form the single leaf 
 trace which is found lower down. The xylem in each bundle forms a broad band 
 
62 
 
 THE OPHIOGLOSSALES 
 
 of wood, which is surrounded by the phloem, although the latter is much less 
 developed upon the inner side than upon the outer and the bundles thus approach 
 the collateral condition found in the trace lower down. Small protophloem ele- 
 ments occur in the outer part of the bundle and are continued part way around 
 toward the lower side, to the point where the two bundles are beginning to fuse, 
 but here they are no longer visible. 
 
 The second leaf shows no lacuna; and the base of the petiole is more nearly 
 cylindrical than is that of the cotyledon. Like the latter, the petiole contains two 
 vascular bundles which at the level of this section were still quite distinct, and the 
 permanent elements were just beginning to form. The stipules in the second leaf 
 completely inclose the third, but there is a very narrow slit in front where the two 
 stipules meet. The next section figured is taken just above the stem apex and cuts 
 through the youngest (fourth) leaf. The two bundles of the petiole of the third leaf 
 are evidently still quite separate, but closely approximated, and as yet show no per- 
 manent tissue. The stipules of the third leaf are just becoming evident. At the 
 level of the stem apex the bases of the third and fourth leaves are completely united, 
 
 Scries of transverse sections of a young sporophyte of Botryrlii 
 cylinder. D and E show the separated leaf traces whic 
 its junction with the stele of the primary root. 
 
 K shows the bundle of 
 
 and the base of the second leaf is also partially fused with these. The two bundles 
 h'om the second leaf are now beginning to unite to form the single leaf trace, while 
 in the third leaf the process is complete and a single oval mass of procambium cells 
 marks the position of the single trace of this leaf. On the opposite side of the 
 apical meristem is a similar but less developed mass of procambium, representing 
 the trace from the young fourth leaf. 
 
 For some distance below the stem apex the two bundles of the third and fourth 
 leaf respectively are quite distinct, and are separated from each other by a mass of 
 large-celled pith. Lower down the two bundles approach and coalesce, and on the 
 side of the second leaf trace (which is turned away from the third one) a group of 
 actively dividing cells can be seen which forms the beginning of the fifth leaf trace, 
 although the leaf itself can not be recognized at the summit. These three masses 
 of procambial tissue, viz, the young traces from the third, fourth, and fifth leaves, 
 
THK VOUNG SPOROl'HVTI 
 
 63 
 
 uiakea iieaily contiiiiiuuscrtscfnt-shapLd massot |ii()camhiuin. Fig. 44, .-V,/i, shows 
 the section of the central stele; and the space between the horns of the crescent, i. e., 
 the space between the third and fifth leaf traces, constitutes the so-called "foliar 
 gap" which faces the second leaf trace. The central cells of the young stele have 
 already begun to form cambium, which in the older stem constitutes such a char- 
 acteristic feature of the stele. 
 
 Before any tracheary tissue is formed in the young stele, a few thick-walled 
 cells can be seen in the outer part of the portion of the stele corresponding to the 
 third leaf trace. These are protophloem cells, and similar ones are developed later 
 in the younger parts of the stele. The first tracheary tissue appears in the region 
 ofthe steel section belonging to the third leaf trace. In the section shown (fig. 44, C) 
 
 A, B,C. The young vascula 
 
 D. Section of a voiing sporo 
 
 stele of axis. X75. 
 
 bundles of : 
 .hyie of B„tr 
 
 
 there are three groups of these primary tracheids, separated by a considerable interval 
 and placed at the extreme inner limit of the young stele. In the older portions of 
 the bundle the number of these increases and similar ones make their appearance 
 in the next younger section of the stele, that belonging to the fourth leaf. Last of 
 all, these primary tracheids are formed in the younger region of the stele and new 
 ones arise between these first-formed ones, so that there is developed a nearly 
 complete circle of tracheids marking the inner boundary ofthe tubular stele. 
 
 The second leaf trace approaches nearer and nearer the two horns of the 
 crescentic stele section as it is followed downward, and finally joins it, filling up 
 the gap between the regions belonging to the third and fifth leaf traces. Thus, for a 
 
64 THE OPHIOGLOSSALES 
 
 time the section of the stele appears perfectly circular and continues downward 
 until it reaches the next leaf trace, which marks the point of departure of the 
 trace of the cotyledon. Except for the greater development of the xylem, which is 
 composed of two to three concentric rows of tracheids, the section of the stele in this 
 region presents much the same appearance that it does in the younger parts nearer 
 the stem apex. The outer part of the stele at this point shows a somewhat broken 
 row of thick-walled bast fibers within which are large sieve tubes. The cambium, 
 lying just inside the phloem, is very much less developed in the basal region of the 
 stele than it is higher up and can scarcely be said to be present. The stele in this 
 region of the stem is quite similar to that of B. lunarta, figured by Poirault. The 
 endodermis is also much less definite than it is later, and I could not satisfy myself 
 that there was in B. virgiiiianum any trace of the inner endodermis which Poirault 
 states is present in the basal part of the stem in B. lunaria, but which disappears 
 later. The medullary rays also, which are conspicuous in the stele of the older 
 plant, are very imperfectly developed in this basal region. 
 
 In the intermediate region, between the base of the primary root and the trace 
 from the cotyledon, a section shows a thick irregular ring of tracheary tissue with 
 no clearly developed medullary rays (fig. 43, L). Indeed, at the lowest part of this 
 region the tracheary tissue forms an almost solid core with only a small amount of 
 parenchyma interspersed, thus forming an inconspicuous and irregular pith. While 
 the endodermis is not clearly delimited, some of the cells in the zone of tissue sur- 
 rounding the stele show the radial thickenings of the walls; but these cells are not 
 all of them in the layer immediately adjoining the stele, being irregularly disposed 
 throughout the three or four layers of cells surrounding the stele. As sections are 
 examined in succession upwards, the pith becomes better defined. Sections taken 
 downward show a ring of tracheary tissue gradually separating into the two or three 
 xylem masses of the primary root. 
 
 The primary root in the sporophyte which has just been described was diarch. 
 The stele shows a very evident endodermis, within which is the pericycle, for the 
 most part composed of two layers of cells. The xylems are composed of a large 
 group of tracheids, of which the two or three smaller ones next the pericycle repre- 
 sent the protoxylems. Small, deeply staining protophloem elements can be seen 
 toward the outside of the phloem, and within these are the other phloem elements, 
 thick-walled bast fibers, and sieve tubes. The rest of the stele is composed of small, 
 thin-walled parenchyma cells. The outer region of the root is made up of somewhat 
 compressed cells suggesting the periderm cells which are found in parts of the older 
 stem and which are also present in the cortex of the later roots. This outer layer of 
 cells, however, shows no evidence of active division. 
 
 A cross-section of the petiole of the cotyledon, from the same sporophyte, taken 
 at a point some distance above the stipules, shows within the epidermis about three 
 layers of parenchyma cells with no conspicuous intercellular spaces, but within 
 these are several very large lacunae, separated by narrow plates of tissue. In the 
 center of the section are seen the two concentric bundles that traverse the petiole. 
 As in the later leaves of this species, these bundles are truly concentric, thus resem- 
 bling those of the Marattiace:f? and differing from the collateral bundles of Ophio- 
 glossum and the smaller species of Botrychiuni. The phloem is better developed 
 upon the outer side and small protophloem elements can be seen at its outer limit, 
 but these are continued partially around the inner side of the bundle, where the 
 xylem is separated from the outside by about two layers of cells. No definite 
 endodermis can be made out. The two bundles are continued up to the base of the 
 lamina, where one of them passes into each lateral lobe of the ternate leaf. One of 
 
iiiK 'i()ii\(; si'oKoi'in 1 1 65 
 
 these divides and i;ives off a l.iaiuh whieli passes int.. ilu- teimmal lolu- . if the leat' 
 (see Jeffrey 1, paj^e 2^). 
 
 Ill a section of the older stiin, .dioiit where the tniiith leaf traei' heeoines joined 
 to the rtfth and sixth, the section of the stele appeals circular. In the portion hi- 
 longinif to the Hftli leaf trace there are present about half a dozen isolated tracheids, 
 arran<;ed in a low, while in the sixth trace only a single small tracheid is developed, 
 riu- two wlenis of the fourth leaf trace are now no longer distinguishable and the 
 uocuh pait of the trace has the form of an irregular row on the inner side of the 
 trace section, and is continuous with the xylems of the fourth and Hfth traces. 
 l\\en after tin fusion of rlu se three traces is practically complete, it is still evident 
 that the ring-shaped section is composed of three parts, the break between the 
 xylems sht)wing the limits of the three component leaf traces. With the increase in 
 the development of the .x\lem ring the limits of the individual leaf traces are finally 
 completely lost and the section of the stele shows a thick ring of the tracheary tissue 
 interrupted only at intervals by single rows of parenchyma cells in the medullary rays. 
 
 Outside the ring of tracheids there can be seen a /one of narrow cells arranged 
 in radially disposed rows. This /one constitutes a genuine cambium, precisely 
 similar in appearance to that found in the stems of Conifers and Dicotyledons. 
 Ibis cambium has long been known to occur in the older sporophytcs of the larger 
 species of Botryc/iiitm, and Jeffrey first noted it in the young sporophyte. Ihe 
 cambium is well developed in the leaf trace for a considerable distance above its 
 junction with the central stele. 
 
 The primary tracheary tissue is in immediate contact with the pith. Occa- 
 sionally the ring of tracheids is broken by a single parenchyma cell. Outside the 
 tracheary ring the cells are arranged in radial rows and in these the formation of 
 the walls is always periclinal, so that the cambium in these earlier stages presents a 
 very characteristict appearance. The rows of cambium cells continuous with the 
 primary parenchyma cells lying between the tracheids of the primary row do not 
 develop tracheary tissue, but remain parenchymatous and thus give rise to the 
 medullary rays, which later are so conspicuous. 
 
 The extreme outer portion of the vascular cylinder is composed of a ring of 
 thick-walled cells which are probably bast fibers, but the walls of these do not stain 
 strongly with safranine and are hence much less conspicuous than the tracheids. 
 These may possibly be considered as protophloem elements and evidently corre- 
 spond to the zone of thick-walled cells figured by Poirault for B. liinarui (Poirault 
 2, fig. 1 1). Just outside of this ring of thick-walled cells lies the endoderniis, which 
 has not as yet developed the characteristic radial thickenings which are so easily 
 seen in the older parts of the stele. These thick-walled phloem elements, as is the 
 case with the protoxylems, do not form an unbroken ring, but are interrupted at the 
 junctions of the confluent leaf traces (fig. 44, 6'). 
 
 In the older bundle, a zone of large cells, the \'oimg sieve tubes nia\- be recog- 
 nized l)ing inside the ring of bast fibers. Inside of the zone of sieve tubes lies 
 the cambium, the cells of which gradually pass into the xylem, to which constant 
 but slow additions are made from the cambium; indeed, Jeffrey thinks that we can 
 not speak of primary wood. He considers that all of the wood owes its origin 
 to the activity of the cambium ring. The radial rows of secondary wood are inter- 
 rupted at intervals by the medullary rays, and the resemblance of the stele at this 
 stage to the section of a young coniferous stem is really quite remarkable. The 
 endoderniis has now become exceedingly conspicuous, as the lignificd lateral walls 
 stain very strongly with safranine and stand out strongly in contrast with the cells 
 of the cortex lying outside the endoderniis. 1 he inner walls of the endodermal 
 5 
 
66 THK OPHIOGLOSSALES 
 
 cells also stain more or less strongly with the safranine, indicating that they also are 
 partially lignified. 
 
 The junction of the second leaf" trace with the central stele is shown in Hg. 43, 
 E, F, G, and similar stages in the fusion of the first leaf trace are shown in fig. 
 43, H, I. The junction of one of the early roots, perhaps the third root, with the 
 central stele is shown in fig. 43, /. At the level where the second leaf trace departs 
 from the central stele the .\ylem ring is composed of about three rows of cells. In 
 both this second leaf trace and the third the endodermis is very conspicuous upon the 
 outer side of the bundle, but is not developed upon its inner face. As the trace 
 approaches the central cylinder it can be seen that the endodermis is interrupted 
 across the leaf gap, but as the leaf trace approaches, endodermal cells are developed 
 between the endodermis of the central cylinder and that of the leaf trace. This 
 union takes place first upon one side, the gap being closed sooner on one side than 
 the other, but later a similar process takes place on the other side of the leaf trace 
 and the closing of the gap is complete, the endodermis now being continued without 
 interruption around the whole of the circular section. The leaf trace where it joins 
 the central cylinder is exactly like it in size and structure; indeed, it is nothing more 
 than a sector of the hollow cylindrical stele, which is evidently built up exclusively 
 of these fused leaf traces. There is no vestige of a«v tissue in the stele which is not 
 referable to the bases of the leaf traces themselves. This becomes very evident when 
 the structure of the leaf trace is examined before it closes the gap where it joins the 
 central stele. 
 
 An examination of the outer tissues of the older portions of the stem (fig. 42) 
 show that at certain points there is a conspicuous border of periderm (per), forming 
 a layer some four or five cells deep. This border of periderm occupies about a third 
 of the circumference of the section and marks the position of an old leaf base which 
 has fallen away. According to Jeffrey, this periderm is an "abciss" layer and 
 results in the cutting off of the old leaf bases. He suggests that the layer of cork 
 cells developed from the periderm closes the scar left by the dead leaf and may be 
 efficacious in preventing infection from fungi in the soil. The presence of a periderm 
 in the outer cortex of the Ophioglossaceae was first pointed out by Russow (Russow 
 1), and was afterward confirmed by Holle (Holle I). The periderm in B. vtrgin- 
 taniim is made up of the usual radially arranged rows of cells (fig. 42, per). 
 
 All of the tracheary tissue in the young sporophyte is composed of rather short, 
 reticulately marked tracheids. The secondary tracheids increase in size and the 
 thickened bars on their walls become broader, so that the larger elements have their 
 walls pitted rather than reticulately marked, and these pits, especially in the older 
 plant, are of the bordered type, not very unlike the round bordered pits found in 
 the wood of Conifers. 
 
 The elongated thick-walled elements in the outer part of the phloem are 
 probably bast fibers. In the older part of the bundle thesje stain quite strongly with 
 safranine and show irregular pits upon their walls. The sieve tubes do not show at 
 all clearly in sections mounted in balsam. They are evidently large, but the char- 
 acteristic dotted sieve areas can not be made out and no special study was made of 
 these. Between the large secondary tracheids and the phloem are the narrow, thin- 
 walled cambium cells. These probably contribute only to the wood, and there is 
 no secondary phloem. 
 
INI- \(»i \(; si'okonn I ; oi iii i.mi.n niosi \cins. 
 
 I lu' ()iil\ iKiDimt rli:ir li;is \\t Ihim piihlislucl of the \(iuni; s|)()rii|)hvfi' of" 
 Helnunthost(hh\s is rli:it of I.aiig {Lang I), wlio examined at sonic length the 
 vascular system in the \(iiinii spoiophyte. Lang states that the young sporophyte 
 of Hflniinthosttii /i\s develops its first leaf as a large teinate foliage leaf, and when I 
 first examined the young plants which were collected in Ceylon I reached the same 
 conclusion; hut on making sections of" the young sporophyte it was found that in 
 all of the young plants that were sectioned the supposed cotyledon was really the 
 second leaf, the cotyleilon itself hcing rudimentary and forming an inconspicuous 
 papilla at the base of the second leaf, which was the first to develop a functional 
 lamina. Further examination of a number of other specimens showed that this 
 rudimentary cotyledon was also present and it is probable that the cotyledon in 
 Hcluiinthostaihys, as in Botrychiuui huiaria and Ophioglossum vtilgatiinu is always 
 a rudimentary organ which never appears above the ground. 
 
 Fhe youngest sporophytes oi Helminthost(ich\s which were found alread\- had 
 the second leaf well developed and the first root had emerged. The ternate second 
 leaf (fig. lo, C) shows clearly that the ternate form is the result of an unequal dichot- 
 omy, such as is common in the early leaves of many ferns. The second leaf of 
 Helmnithostachys closely resembles the cotyledon of Rotrychnim vtrgmianutn and, 
 like that, usually possesses two vascular bundles extending through the long, slender 
 petiole. At the base of the leaf there is a conspicuous stipular sheath which 
 completely incloses the next younger leaf, as it does in Botrychium liinaria. The two 
 bundles of the petiole contribute to the two lateral lobes of the leaf as they do in the 
 cotyledon of Botrychium virgiuianiim, and one of these, forking again, gives rise 
 to the terminal lobe and smaller lateral lobe. The venation of the second leaf, 
 especially in its younger stages, is almost perfectly dichotomous and is more like 
 the venation in the leaves of Botrychium lunaria than that of 5. virginianum. With 
 the enlargement of the leaf segments the venation more nearly approaches the 
 pinnate arrangement found in the later leaves. 
 
 While the young sporophyte in Helminthostachys in its earliest stages shows a 
 general resemblance to that of Botrychium virginianum, a difference is very soon 
 noted. The young plant, very soon after the formation of the cotyledon, develops 
 a conspicuous internode between the cotyledon and the root, and the stem rapidly 
 elongates as the sporophyte grows, so that it soon shows the characteristic elongated 
 rhizome of the older sporophyte. 
 
 Unfortunately I was unable to secure any very young sporophytes, my youngest 
 specimens already having the second leaf nearly fully developed. According to 
 Lang, the arrangement of the primary organs of the young sporophyte is the same 
 as in B. virginianum, but I could make no study of the origin of the vascular system 
 in~the sporophyte before the first leaf and root had emerged from the prothallium. 
 
 To judge from longitudinal sections of the youngest plants obtained, the 
 relation of the first leaf and root is the same as in Botrychium virginianum and the 
 base of the young sporophyte closely resembles that of the latter. As in Botrxchium, 
 the foot is very large and conspicuous, forming a large, hemispherical body sur- 
 rounded by the prothallial tissue. The large primary root grows out at right angles 
 to the foot and is continued into the axis of the sporophyte, which grows up verti- 
 cally. None of the specimens were young enough to show the relation of the stem 
 apex to the first leaf and root, as in the \-oungest specinu-iis the axis hatl alriad\' 
 begun to elongate and an cvidcnr iiucrnodc was forimd between tiie hasi- of the 
 first leaf and the root. 
 
68 
 
 THE OPHIOGLOSSALEf 
 
 The longitudinal section of the young sporophyte at this time shows that the 
 thick stele of the root bends upward and is continued into the axis, which already 
 is a good deal elongated and has carried up the young leaves, which are separated 
 from each other by conspicuous internodes. The young sporophyte at this stage 
 thus differs strikingly from the corresponding stages in Ophioglossiim and Botrych- 
 iiim, with their very closely crowded leaves with no appieciable internodes between 
 them (fig. 46). 
 
 A single stele extends through the axis from its junction with the root stele and 
 forks below the insertion of the rudimentary first leaf, one branch passing into the 
 cotyledon, the other continuing upward and forming the trace for the second leaf, 
 at whose base is seated the apical bud. 
 
 Whether in the younger stages, before the development of the second leaf, the 
 vascular system of the young spoidph\tL- would show the same relations of the steles 
 of the root and cotyledon, as in /io/rvi /'///"/, n^nuiins to be seen; but from the general 
 similarity in the structure of the \()ung\sponiph\tes of the two this is quite likely; 
 
 yet it is possible that the stem apex is 
 more prominent in the early stages of 
 Helminthostachvs and part of the stele 
 which is found in the elongated stem 
 may be of cauline origin, although this 
 is hardly indicated by the condition of 
 things in the apical meristem of the 
 youngest sporophytes that were ex- 
 amined. 
 
 The foot is composed of large- 
 celled, thin-walled parenchyma, with 
 no noticeable differences between the 
 inner tissue and that of the periphery. 
 The stele of the root is continued up- 
 ward without interruption into that of 
 the shoot (fig. 46, B). Probably at an 
 earlier period there would have been, 
 as in Botrychium, simply the continu- 
 ous stele of the root and cotyledon. 
 At the junction of the root stele with 
 that of the axis there is a marked en- 
 largement, and the short, somewhat ir- 
 regular tracheids are decidedly broader 
 than those either in the root or in the portion of the stele above the junction. The 
 tracheids toward the outside of the central region are slender and have delicate 
 reticulations, which closely approximate the form of true spiral thickenings. These 
 are the protoxylem elements and can be traced upward and downward into the 
 stem and root, respectively. The secondary tracheary tissue is marked by broad, 
 reticulate thickenings, and these in the largest tracheids are replaced by conspicu- 
 ous oval or nearly round bordered pits, very much like those which are found in Bo- 
 trychium (fig. 46, £). 
 
 Above the base of the root there is a long internode below the cotyledon, and 
 through this runs the single axial stele. On the same side of the axis as the primary 
 root may be seen the cotyledon, which closely resembles in form the young cotyledon 
 of Botrychium virginianum, but instead of developing into a functional leaf it is 
 arrested in its growth before the lamina is fully developed. As in the cotyledon of 
 
 L Young sporophyte of Hehnitilhostachv 
 phyte, pr. ror, cotyledon; r',r^, first 
 i. Le.lf from an older sporophyte. Xl. 
 
 attached to gameto 
 
TIIK VOINC SPOKOI'inTI- 69 
 
 Botiycliiniii vivgnuanuiu, the lamina, wliicli althouyli \li)- j;ii.atl\- iliIullcI iilact- 
 theless is present, is bent over sharply against the stout petiole, which is strongly 
 convex on its adaxial side. In the specimen figured there was nf) trace of a stipular 
 sheath developed in the cotyledon, but sometimes this is very well developed, 
 although the ujiper part of the cotyledon remains quite rudimentary. The little 
 cotyledon has the rudimentary lamina distinctly ternate in outline and very much 
 resembles some of the rudimentary early leaves found in liotrychiuni lunaria, as 
 figured by Bruchmann. A single vascular bundle extends through the petiole of the 
 cotyledon, nearly to the base of the lamina, but except in the basal region the 
 development of the xylem is niL-.ch reduced and is entirely lacking before the lamina 
 is reached. The bundle of the cotyledon, where it bends at the base into the inter- 
 node, is nearly as well developed as the bundle of the second mttriiode which 
 separates from the former just below the insertion of the cotyledon. ( )n tin- ()|)|iosirf 
 side of the second internode from the C(ir\ledon, and exttiuling downwaiil tliinnuh 
 x\w Hrsr internode for a short distance below the level of the eon ledon. tlieie is seen 
 
 F.o. 46. 
 
 A, B. Two longitudinal sections of a young sporophyte of H^/mm/Ao5Mc/;vj; /)r, gametophyti'. X6. 
 C. Upper part of same, more enlarged; b, terminal bud; col, cotyledon; i, lacuna in first internode. 
 O. Apical region. /^, third leaf; sc, epidermal scales. X140. 
 E. Tracheids showing bordered pits and thickened bars. X320. 
 
 in the specimen figured a large lacuna, separated from the outside of the internode 
 by about half a dozen layers of cells and on its inner side from the endodermis of the 
 stele b\ two or three layers of cells. This lacuna, as Lang pointed out, is not always 
 present. At the upper end of the second internode can be seen the base of the second 
 leaf, the stipular sheath of which incloses the \()ung bud at the summit of the axis. 
 
 On its inner (dorsal) side the terminal bud shows the young third leaf, opposite 
 which is the section of the ridge surrounding the small free surface of the stem apex, 
 ver\- much like the condition in the \-oung bud of Opbioglossuni tnohucatuim. 
 The form of the young leaf at this stage is more like that of BotrYcli/iini liuuiriu than 
 that of B. v'lrg'niiantitn. The young leaf is a blunt cone with a marked projection 
 on its inner side, extending over the apex of the stem and marking the beginning 
 of the stipular sheath, which is entire as in B. hninnu, and not di\iileil into two fite 
 stipules as in B. virgnuniiiini (fig. 4'i, /)). .\ conspicuous strand of |ii(icambium 
 extends into the leaf and joins the trace of the second leaf. As in Bntixilninn and 
 Opliioglos.uim, the stele of the internode does not extend upwartl be\()nil the base of 
 the youngest leaf and there is no indication that it derives any of its tissues directly 
 
70 
 
 THE OPHIOGLOSSALES 
 
 from the apical meristem of the shoot. Fhe latter, as in the other Ophioglossaceae, 
 shows a large single initial cell, which in section closely resembles that of Ophio- 
 glossum mohiccaniim, having a broadly truncate base and being narrower above 
 (fig. 46, D). The apical meristem is of very limited extent and segmentation of the 
 apical cell is probably very slow. The second root, which in the case under con- 
 sideration arose near the base of the third leaf, is here very conspicuous and its stele 
 joins that of the third leaf near the base of the internode. The young leaf and root, 
 with the inconspicuous stem between them, have very much the appearance of a 
 bud at the base of the second leaf. The apex of the very young leaf grows from the 
 conspicuous single apical cell (fig. 46, D). This appears triangular in longitudinal 
 section and shows regular segmentation. The stele of the young leaf can be traced 
 up to the inner cells, cut off from the lateral segments of the apical cell. 
 
 The apical cell of the stem is quite large and, as we have seen, is truncate below, 
 as it is in Ophioglossum. Seen in its transverse section it is triangular, so that its 
 form is really that of a three-sided prism. Both lateral and basal segments are cut 
 off from it, but only the youngest segments can be clearly distinguished. Below the 
 apical cell there is only a very small amount of tissue between the apex of the shoot 
 and the point of the junction of the two leaf traces and the stele of the second root. 
 This junction is marked by the presence of very irregular reticulate tracheids (fig. 
 46, D, tr). The tissues on the upper side of the root stele extend to within a few cells 
 
 of the stem apex and may possibly 
 represent a portion of the internodal 
 stele that is not connected with the 
 leaf trace, but a study of later stages 
 does not support this view, and it 
 is almost certain that this youngest 
 tracheary tissue is the beginning of 
 the next leaf trace. 
 
 From the very first, as Lang 
 pointed out, the shoot is perfectly 
 dorsiventral, although at first it grows 
 vertically upward. The leaves are all 
 borne upon the dorsal side, while the 
 roots invariably arise upon the ven- 
 tral side. 
 
 The apex of the young root, 
 which in this case is on the point of 
 emerging, is occupied by a conspicu- 
 ous triangular apical cell, the segmen- 
 tation of which closely resembles that 
 of the root o( Botrychntni. At first the root cap is mainly derived directly from the 
 outer segments of the apical cell. These segments undergo periclinal as well as 
 anticlinal divisions and the inner layers of the root cap (which at this stage is rela- 
 tively thick) show a fairly regular stratification. The outer cells of the root cap are 
 a good deal elongated, so that the free end of the root is decidedly pointed. At this 
 time, except for the irregular conjunctive tracheids at the base, no permanent tissue 
 can be seen in the stele of the young root. 
 
 In the cross-sections of a young sporophyte of about the same age the third leaf 
 was rather further developed and the young fourth leaf could be seen, but the second 
 root was not so far advanced, or it may be that the root at the base of the third leaf 
 was the third root and not the second. The single apical cell, which was so prominent 
 
 ng sporophyte of Hetminthos- 
 ; Z'', second leaf. X65. 
 
 B. Section passing through I 
 
 C. Stem apex. X200. 
 
THE YOUNG SPOROI'HYTF 
 
 71 
 
 in the very young leaf, can no longer be recognized and the lamina is divided into 
 two nearly equal lobes, of which one is lateral with reference to the other, but these 
 two lobes probably arise as a result of the dichotomy of the original apex. A group 
 of narrow marginal cells somewhat to one side of the terminal lobe evidently consti- 
 tuted a meristem tissue, and this was preparing for the second dichotomy, by which 
 the ternate form of the young leaf is established. The branching of the lamina at 
 this early period therefore appears to be the result of an unequal dichotomy, as it is 
 in the first foliage leaf. The vascular bundles from each of the two primary leaf seg- 
 ments were just beginning to be recognizable and join into one at the base of the 
 lamina. Surrounding the leaf is the sheath developed from the base of the second 
 leaf. This sheath in the section appears as a closed ring some two or three cells in 
 thickness and looking very much like the sheath of Ophwglossiim. \ he top of the 
 conical sheath is closed and does not show the pore found in Ophioglossuni. The 
 sheath becomes thicker lower down, where four rows of cells show in the section. 
 Between the inclosed leaf and the surrounding sheath in the lower portion (fig. 49, B) 
 may be seen the section of flattened scales (sc) which grow from the base of the leaf 
 
 and from the tissue immediately about the stem apex. At its base the sheath of 
 the third leaf is already completely developed and surrounds the young fourth leaf 
 (/*), which in this section was cut horizontally and shows plainly its large apical cell. 
 The stem apex, which lies directly below the younger leaf, was cut obliquely in this 
 series and did not show the form of the apical cell plainly, but in a young plant of 
 about the same age (fig. 47, 6") the apical cell showed that it was regularl}- trian- 
 gular in form, seen in cross-section. 
 
 At the level of the stem apex the second leaf shows a single vascular bundle 
 which may be continued undivided upward, or it may divide into two bundles in the 
 petiole. In regard to the structure of the petiole, therefore, Helniintliostnchys is to 
 some degree intermediate between Botryrliium and Ophioglossum. The young trace 
 of the third leaf can be seen at this level as an oval mass of undifferentiated cam- 
 bium. Immediately below the stem apex the cells are smaller than is the case in 
 the young sporophyte of Botrychium, but one can not make out the beginning of a 
 central procambium cylinder. The smaller cells in the region immediately underlying 
 the stem apex may be explained by the fact that in these voung plants there is no 
 pith developed within the axial stele as there is later on. The youngest root in this 
 
72 
 
 THE OPHIOGLOSSALES 
 
 specimen was less developed than that in the specimen that we have just described. 
 This 3'oungest root was cut obhquely, but at its base there could be seen the undif- 
 ferentiated mass of procambium forming its stele, and the apex showed the tetra- 
 hedral apical cell, presenting much the same form as the longitudinal section. The 
 apical portion of the young root was somewhat shrunken away from the surrounding 
 tissue, so that there was ajjvery evident space surrounding it. Whether or not this 
 was the result of artificial shrinkage was not certain. 
 
 Below the terminal bud the third leaf trace proceeds downward into the inter- 
 node below, and the tracheary tissue, which is just beginning to develop in the bud 
 region, increases rapidly in amount and forms a central mass of wood which does 
 
 c 
 
 D. Shows the narrow pit, y^ bet 
 G-J. Show only the vascular bundles. 
 
 not appear solid, but is made up of two groups of tracheids, of which the one nearest 
 the second leaf trace has the larger tracheids. At this stage the bundle somewhat 
 resembles the section of a diarch root. Only a little further down the two xylem 
 masses are joined by the development of tracheids in the middle of the bundle, 
 which thus comes to have a solid xylem core. In the section figured there may be 
 seen a single small isolated tracheid on the side turned toward the second leaf trace, 
 which has now approached so close to the central bundle as to touch it (fig. 50, A). 
 The central bundle in the internode at this point is beyond question the leaf trace for 
 the third leaf. There is no difficulty in tracing its progress upward to where it enters 
 the leaf and there is no sign of any addition to it from tissue lying outside the 
 original leaf trace. The central stele of the upper part of the second internode is 
 throughout nothing more than the basal portion of the leaf trace from the third leaf. 
 At this point it is concentric in structure. Surrounding the solid .xylem core can be 
 seen about three rows of phloem cells, some of the outer ones being small protophloem 
 elements. The large ones lying inside are presumably sieve tubes. A single layer 
 of rather large pericycle cells separates the phloem from the large endodermal cells, 
 which begin to show very conspicuously lignified radial thickenings, especially on 
 the side of the bundle adjacent to the second leaf trace. The section of the second 
 leaf trace is somewhat extended transversely and the xylem band is composed of 
 about a dozen tracheids. In the mid region of the section the xylem is separated 
 from the inner side of the bundle by a single layer of cells, possibly sieve tubes, but 
 at one point the xylem is in contact with the pericycle, which is only imperfectly 
 developed and wanting at certain points. Upon the dorsal side the phloem may 
 be in direct contact with the endodermis, which is quite wanting at this point upon 
 the ventral side of the bundle. Higher up, where the trace is quite free, the endo- 
 dermis extends completely around it. While the second leaf trace, therefore, may 
 
73 
 
 be said to be concentric, it neveitbeless a|)[ii()acbe.s closel\' tbi collateiai tv|)e of" 
 the other ( )phioglos.sace:L'. 
 
 The fusion of the two leaf traces to form the sin<>;le axial binulle in the basal 
 part of the internode is very nuich like that found in the young sporophyte of /io- 
 trycluiini (fig. 49, D, F). l"he endoderniis of the two bundles becomes continuous 
 and there may be seen intermediate stages of a single bundle, oval in section, but 
 with two conspicuous xylems, separated by the mass of large, thin-walled tissue 
 which gradually disappears as the bundle is followed downward; and finally, in the 
 lower part of the internode, the section of the stele appears quite circular, the center 
 being occupied by the solid xylem formed by the complete coalescence of the xylems 
 of the fused leaf traces (fig. 49, J). The arrangement of the phloem, pericycle, and 
 endoderniis is the same 
 as in the single leaf trace '^ ')~^X^^ 
 
 in the upper region of the \/\1yh-( 
 
 niternocle. < \J'-0': {,.->(:%% 
 
 The base of the stip- 
 ular sheath can be fol- 
 lowed downward to about 
 the level where the two 
 leaf traces begin to coal- 
 esce. Its anterior free 
 portion, betwten which 
 and the internode is a 
 narrow slit, finally be- 
 comes quite free and its 
 section is visible as an 
 oblong mass of cells lying 
 quite separate from the 
 section of the internode. 
 
 Longitudinal sections 
 
 of a somewhat older sporophyte are shown in fig. 48, A, B; the third leaf is pretty 
 well developed, with the stipular sheath completely overarching the stem apex and 
 the fourth leaf, which already is conspicuous. The third leaf is still completely 
 covered by the large stipular sheath of the second leaf, which apparently forms a 
 closed cavity; but an examination of the adjoining sections showed that the ante- 
 rior margin of the sheath is quite free in front and that there is a narrow cleft 
 between it and the internode. Ihe real nature of the sheath and its relation to the 
 leaf base is very clearly shown in the nuilian section ot the terminal bud from an 
 older plant (fig. 48, Cy D), and the resemblance to the terminal bud of Botrvr/mitu 
 linifin'ri, except for the dorsiventral arrangement of the parts, is most striking. This 
 is equally marked in the adult sporoph\te, as Farmer has already pointed out. 
 
 rile specimen under consideration diffeied somewhat from the yoimger one 
 that was descrilnd in the relation of tht- \-oung organ.s. The cotyledon was better 
 developed and although the lamina was rudimentary the stipular sheath was large 
 and inclosed the base of the second leaf, while in the younger specimen discussed 
 no sheath was developed at the base of the cot\ledon. There was a diflVrent rela- 
 tion of the r(M)ts also, pidhahK associated with the greater (livelo|inuin of the 
 cot\'ledoii in rile specimen under consideration. The second root, instead of being 
 formed at the base of the third leaf in the terminal bud, was developed a short dis- 
 tance above the cotyledon, near the base of the second leaf, while in the \-oungei 
 specimen no root was developed above the primary root until the third leaf had 
 
74 THE OPHIOGLOSSALES 
 
 begun to develop; but this second root developed much earlier in its relation to the 
 development of the next leaf than was usually the case. 
 
 In the present plant the root corresponding to the third leaf, which was the 
 third root in this case, was in a very early stage of development and had not begun 
 to elongate at all. The third leaf in median section was already pretty well advanced 
 and the apical portion was very strongly incurved, as it is in the young leaves of 
 Botrychium virginiatium, which it resembles more than it does that of B. lutiaria, 
 and it also suggests the form of the young leaves in the Marattiaceae. The basal 
 part of the young leaf, however, is exactly like that of B. lunarta. The stipular 
 sheath forms a thick body, projecting forward and about equal in height to half the 
 total length of the young leaf. It is strongly concave below, and the next leaf is 
 fitted into the cavity (fig. 48, C). The forward margin extends downward as a sort 
 of lip which elongates rapidly and keeps pace with the growth of the young leaf 
 inclosed within it, and this is completely concealed until it has reached a large size. 
 The sheath finally forms a long, conical protuberance at the base of the leaf to which 
 it belongs. 
 
 The cavity below the sheath is extended backward, so that the base of the leaf 
 is hollowed out in front and in section appears much narrower than the part of the 
 petiole above the insertion of the sheath; this brings the stele of the leaf very close 
 to the epidermis at this point. The narrow cleft thus formed between the posterior 
 wall of the stipular cavity and the petiole of the leaf in front is the "canal," which 
 Gwynne-Vaughn described as occurring at the base of the petiole of the older plant 
 and opening above the insertion of the stipule. These canals are easily seen in longi- 
 tudinal sections of the young bud (fig. 48, B) and it is very clear that they are, in truth, 
 nothing but the narrow spaces between the bases of the adjacent leaves. Gwynne- 
 Vaughn's statement that they open above the insertion of the stipules is incorrect 
 unless the stipules of a given leaf are considered to be derived from the sheath sur- 
 rounding the base of the leaf, this sheath properly belonging to the next older leaf. 
 This canal can be seen as a narrow slit in the young leaf, extending below the inser- 
 tion of the next younger one, the young stele of the leaf being separated from the 
 epidermis at this point by only about three rows of cells. 
 
 The stele of the young leaf can be traced nearly to its tip, and then continues 
 downward toward the inner side of the leaf into the internode, below where it joins 
 the young stele belonging to the next younger leaf. 
 
 As in the younger sporophyte, the stem apex is of very limited extent and the 
 apical cell shows the same form. In the section figured (fig 48, C) the youngest leaf 
 is cut almost m a median plane and though externally no differentiation is visible the 
 young stele is already conspicuous and can be followed down without difficulty. 
 Passing on one side of the stem apex and above the leaf trace, there is no evidence of 
 procambium in the apical region of the stem. In the young sporophyte, therefore, it 
 is perfectly certain that no cauline stele is present. Farmer (Farmer 2) states that 
 in the older sporophyte the stele can be traced above the insertion of the youngest 
 leaves, but we believe a careful study of this point in the older plant would show, as 
 in the younger one, that the stele is developed in the young leaf at an exceedingly 
 early period and that the stele of the internode is composed entirely of leaf and root 
 traces. The procambium tissue developed on the ventral side of the stele near the 
 apex can be shown, by a study of its transverse sections, to be due entirely to tissue 
 derived from the leaf traces, which are extended downward until they meet on the 
 ventral side of the internode. 
 
 Near the stem apex are numerous small scales and hairs filling up the space 
 about it and probably associated with the prevention of drying up of the stem 
 
THF, YOUNG SPOROPHYTK 75 
 
 apex, which is doiibl) protected by these scales and its complete inclosuie in the 
 young leaf bases. It is possible that these scales secrete some mucilaginous sub- 
 stance, although there was no evidence of this in the stained sections. 
 
 In this plant the internodes were quite solid and there was no trace of the large 
 lacuna- which were described for the younger plant. Whether the presence of the 
 lacun;c is due to any differences in the environment remains to be seen. The speci- 
 mens were all collected under apparently the same conditions, but it is possible that 
 there may have been differences in the amount of water in the soil in which they were 
 growing, and this difference in the amount of water may have something to do with 
 the development of the lacunae. 
 
 The root at the base of the third leaf was cut transversely and was in a ver)- 
 early stage of development. The root was still completely buried in the cortex and 
 the stele still imperfectly developed, but showing its connection with the stele of 
 the fourth leaf above the junction of the third leaf trace with the stele of the inter- 
 node. No tracheary tissue was developed and the course of the leaf trace was bur 
 little disturbed by the formation of the root. 
 
 In longitudinal sections taken from a much older sporophyte (fig. 48, C, D) 
 the arrangement of the organs at the apex is seen to be exactly the same as in the 
 younger plants, but of course the parts were all larger. The youngest leaf (/ ' ) still 
 resembles in form and size the corresponding leaf in the younger stem, but the next 
 older leaf is relativel)' broader than the early leaves of the younger sporophyte, and 
 although the stipular sheath is now well developed the upper portion of the leaf 
 is still quite undifferentiated and the apex is scarcely bent forward at all. The 
 resemblance, therefore, to the young leaf of Botrychinni liiuario is even more 
 marked than it is in the younger sporophyte. The apical region of this young leaf 
 is occupied by an epithelium-like layer of columnar cells, and it is doubtful whether 
 any one of these can be certainly denominated the apical cell. Somewhat the same 
 doubt exists also as to the point of a single initial cell in the youngest leaf. In the 
 latter there was some shrinkage in the group of large meristem cells at the apex, 
 which made it still more difficult to decide this point, but a median section showed 
 somewhat indistinctly a single cell, apparently triangular in outline, which from its 
 form and the arrangement of the adjacent cells may very well have been the apical 
 cell. The small group of meristem cells forming the stem apex is crowded in between 
 the base of the youngest leaf and the elongated ridge which surrounds the stem apex 
 upon the ventral side. In consequence of this crowding the outer faces of the 
 apical cell and its youngest segments are very small. The base of the apical cell 
 is more than twice as broad as its free outer face. The ventral face of the apical 
 cell is convex and the young segments cut off from it are strongly curved. The inner 
 or basal wall is oblique, so that the axis of the young cells in the tissue below the 
 apical cell makes an angle with the long axis of the apical cell. 
 
 Below the apical region may be seen the section of the large central stele of the 
 internode. This is composed, apparently, of two strands of procambium separated 
 by a broad band of pith. As Farmer pointed out, in the older plant there is no 
 question that procambium tissue is developed upon the ventral side of the bundle 
 which extends into the apical region and which is, apparently, not connected with 
 the leaf traces. From this he believes that the stele is really a cauline structure. 
 
 A study of the cross-sections, however, as will be presently seen, shows that 
 this ventral tissue really does belong to the leaf traces, although it is possible that 
 the basal tissue of the roots may :idd to it in parr, bur I have not been able to 
 satisfy myself that any of this stelar tissue can be |ii(i|nrl\ assigned to the nctivit\- 
 of the stem apex. 
 
76 
 
 THE OPHIOGLOSSALES 
 
 As in Botrychium and Ophioglossuni, the tissue derived immediately from the 
 activity of the apical meristem, after the central pith is developed in the stele, 
 contributes only to this central region or pith, which is really not part of the stele 
 proper. On the dorsal side of the stele the whole of the procambial strand can 
 easily be followed into the leaf, and when this is followed downward it is seen to 
 run obliquely through the youngest internode, at the base of which it joins the next 
 leaf. The latter already has its lamina developed and strongly bent over, but is 
 still inclosed in the stipular sheath belonging to the next older leaf. 
 
 These older leaves have only a single leaf trace, but it divides into two strands 
 above the level of the stipular sheath. A median section of the thick central stele in 
 
 F.r..5,. 
 
 transverse sect 
 
 ions from a young sporophyt. 
 
 : of Helrr. 
 
 • hthoMclr 
 
 termi;i.nl bud i 
 
 nclosed by stipular sheath, j/;. 
 
 . In G, V 
 
 is pit bet^ 
 
 [, section of a yc 
 
 )ung leaf, showing the ternate 1 
 
 amina. 
 
 
 Xj;. /, petiole of leaf; 
 
 the upper region shows conspicuous protoxylem elements made up of much elongated 
 tracheids, whose thickenings are between annular and spiral in form, but very 
 different from the coarsely reticulate and the conspicuously pitted tracheids of the 
 meta.xylem developed in the older parts of the stele. In all cases where truly median 
 sections were seen the protoxylem elements occupied the innermost region of the 
 xylem and no metaxylem was seen within these, so that the bundle (at first, at any 
 rate) is endarch, as in the other Ophioglossaceae. It is impossible to see in this sec- 
 tion that metaxylem was developed inside of the protoxylem, as Farmer found to be 
 the case in the older sporophytes. 
 
THK voi'Nc; sroKoi'nii i: 77 
 
 In fig. 48, I\ iluic is sliowii .111 insntiDii oC ;i nnit ii|i(in ilic side .nul its vi-ry 
 broadly expaiuU-d li;isc- with rlu- iini;iil;n rnn|umti\i 1 1 ailuuls. I Ins \ii\ l)ici;i(l 
 base of insertion ami ilu dlnimis c<mtimiit\ ol tin- ii.u lu ai\ tissiu- ^A tin iciot with 
 that of the central stele of the stem siiggist tiiat tin \rntial jiait of the stele iiiiuht 
 be made up, in pait at least, of toot traces. 
 
 Cross-sections of oldei plants, such as that slioun m the seiies indieateil in 
 Hg. 51, show issentialh' the same arrangement di' the parts as d<i the Mummi ones. 
 .\t the ape.\ theie ma\- be seen in section the three \uungest lea\es, which wc will 
 tlesignate ies|iectively 1, 2, and i,. No. 1 being the oldest leaf shown. In this section 
 the oiliest leaf shows two vasculai bundles in the petiole, and this is tiiie also of the 
 secoml leaf, which is contained within the stipiilar sheath of the oldest leaf. I'he 
 lamina of leaf No. i ainath shows the characteristic ternate form. I'he \oung 
 leaves make an angle of ahoiif ^o'' with each other. Section H passes directl\ thioiigji 
 the stem a|)e.\ ami shows the sections of the basal part of the sheath of the two older 
 leaves. The stipular sheath of the youngest leaf paitiall}' incloses the stem apex, 
 and surrounding the apical tegion is a loose mass of cells, derived partly from the 
 tissue immediately around the stem apex and parth from the edges of the sheath 
 of the youngest leaves. These cells are realb sections of hairs and scales which 
 perhaps secrete mucilaginous matter, but there was no evidence of this in the sec- 
 tions. Turned towaitl the doisal side of the section may be seen the leaf tiace fVom 
 the second leaf and separated fVotn it by a considerable space on the ventral side 
 is a rather vagueb deHmil mass of young procambium, which marks the trace of 
 the youngest leaf". ( )n the ventral side of the latter is the stem apex itself. Further 
 down, as has already been described for the younger stem, the two leaf traces apjiroach. 
 At the level of the stem apex the trace of the second leaf is clearly defined and 
 nearly oval in outline, but more convex upon the dorsal side. I he bundle is inclined 
 toward the side of the stem opposite the insertion of the first leaf. Further down 
 the trace becomes much broader, this being mainly due to the development of the 
 tissue upon one side, which begins to bend downward toward the ventral side of the 
 stem. Fhe trace of the youngest leaf is now somewhat better defined and appears 
 somewhat kidney-shaped in section, the convexity, like that of the second leaf trace, 
 being turned toward the dorsal side. This trace also begins to show the extension 
 of tissue toward the ventral side of the stem and this is developed on the side opposite 
 to the extension in the next older leaf trace. Ultimately these ventral extensions of 
 the two bundles meet as the two leaf traces come nearer together, and the dorsal 
 ends also finally come into contact, so that the two bundles, the section of each of 
 which is approximately semicircular, form a complete ring inclosing a greater or 
 less amount of the ground tissue, which thus forms the pith of the hollow stele. The 
 stele resulting from the union of the two leaf traces is not at first perfecth- circular 
 in section, but shows plainly for a long time that it is composed of two separate 
 bundles (fig. 51, F). 
 
 Fhe first development of permanent tissue in the older leaf trace is evident 
 before it joins the younger one. The first elements are thick-w^alled cells (which 
 may be called protophloem) in the outer /one of the phloem, and a little later at the 
 inner limit of the bundle a small group of protoxylem elements appears. Very soon 
 after this a similar differentiation takes jilace in the younger leaf trace, so that when 
 the two leaf traces are completely fused the protoxylems have the appearance of 
 liaMns; arisen toward the central part of tin bundle, but in reality the component 
 bumlles, at least at their earlier stage of development, may be described as endarch 
 and agree, therefore, with the bundles of the other Ophioglossacea*. In the leaf 
 traces of the older part of the stem, however, theie may generally be found inside 
 
78 
 
 mOGLOSSALES 
 
 the protoxylem a few scattered, large tracheary elements, so that the bundle may be 
 said to be mesarch, as Farmer states is the case in the older rhizome. 
 
 During its earlier stages the stele shows no leaf gaps where the leaf traces 
 depart, and the leaf gaps are only gradually developed. Soon after the stele has 
 assumed the form of a hollow cylinder the leaf gaps are for some time absent, or 
 they are developed only in a very small degree and close almost immediateh upon 
 the departure of the leaf trace. 
 
 In none of the young plants that I examined could I detect any of the inner 
 endodermis which occurs in the older rhizomes. Farmer, however, states that the 
 inner endodermis is only imperfectly developed and concludes that it is the result 
 of the invagination of the outer endodermis through the leaf gap; or, to put it in 
 another way, it is the persistence of the inner endodermis of the leaf traces of which 
 the hollow stele is made up. The bundle at this stage most nearly resembles that of 
 Botryrhiuni Iiiiuiritu differing from that of B. vtrgimaiunn in the absence of a true 
 
 A. Section of the 
 
 B. Stele from lov 
 
 C. Part of centra 
 
 central stele of Helnitnlhostachvs, showing the two xylems. 
 t of an older sporophyte, showing junction of a leaf trace with t 
 more highly magnified. 
 
 cambium, the outer wood cells being directly in contact with the inner cells of the 
 phloem. Occasionally, however, there may be seen on the outer edge of the .xylem 
 ring a few imperfectly developed tracheids which probably represent a very rudi- 
 mentary development of secondary wood, but there is no other sign of the definite 
 cambium zone which is so conspicuous in the stele oi Botrychium virginianum. 
 
 From this study of the development of the leaf traces, following them from the 
 stem apex downward, it appears that the cylindrical stele in Helminthostachys 
 arises in precisely the same way as that of Botrychium, viz, by the union of the leaf 
 traces. The appearance of procambium upon the ventral side of the stele, which 
 in longitudinal section appears to be derived directly from the stem apex, can thus 
 be explained by the ventral extension of the broad leaf traces which meet on the 
 lower side of the stem as well as above, and the cylindrical stele is thus developed. 
 
79 
 
 It is not ini|)(issilil( th.it tlic kmh tiiuis m.i\ also cdntiihiitf to some (.xtciit to the 
 development ot this \ intr;il poi tioii ot tin- stilt . The xyleni masses belonging to the 
 two leaf trails lemain ijuite ilistimt tor a long time, the one belonging to the older 
 trace being better develo|ied than that trom the younger one (fig. 52, A). The 
 metaxylem consists of scattered tracheids arranged regularly in an arc outside the 
 protoxylem of the older bundle, but there are a few scattered smaller ones lying in 
 the tissue separating the two primar\- wiems. so that, as we ha\e seen, the bundle 
 may perhaps be described as mcsarch, aiul not iiidanh as it is m Ojilimoloisiini and 
 Botr\iliintii. 
 
 The two wlenis of the separate bundles are tinalh coniucted b\ intermediate 
 trachear)- tissue, so that a iuarl\- complete ring of wood is formed; but the xylem of 
 the older leaf trace is still clearly recognizable on account of its greater thickness. 
 In the section figured (rig. 52, 6") there is no tracheary tissue yet formed inside the 
 protoxylem and there is a well-marked pith occupying the center of the stele. The 
 endodermis is now well developed and inside it are about two rows of pericycle cells. 
 The protophloem is less conspicuous than in the younger stele and is best developed 
 in the outer portion of the older leaf trace. The rest of the phloem is made up of 
 
 A. Transverse section of apex of first root of Helminthostachys^ showing a 
 
 single tetrahedral apical cell. X200. 
 
 B. Apex of a later tetrarch root, with apparently no single initial cell. 
 
 thin-walled cells, the larger of which are presumably young sieve tubes. The section 
 is taken from a much older part of the stem and shows that the bundle has attained 
 practically the same condition as in the adult stem, except for the absence of the 
 inner endodermis. The central pith is surrounded by an unbroken ring of wood, 
 averaging about three cells in width. The outer part of the bundle is essentially 
 the same as in the one that has been described. 
 
 The canal described by Gwynne-Vaughn for the older sporophyte is plainly 
 evident here and extends backward from the sheath cavity, appearing in cross- 
 section as a narrow cleft between the inner side of the leafstalk and the internode 
 (rig. 49, D, v). It is evidently not a cavity in the cortex, but merely an invagination 
 of the epidermis, which is continuous with that of the inner surface of the stipular 
 sheath. 
 
 A peculiar feature of the ground tissue in the young stem and petiole in HcUnin- 
 thostachys is the pre.sence of cells containing roundish bodies which stain very 
 strongly with Bismarck brown, and these cells closely resemble the tannin cells of 
 the Marattiace;e, with which they are probabh' homologous (rig. 49, 6). I-ang 
 figures these cells (Lang I, fig. 65), but makes no mention of them in his text. 
 
80 
 
 THE OPHIOGLOSSALES 
 
 The trace for the second leaf, the first functional one, is usually provided with 
 a single undivided bundle passing through the petiole, but sometimes this divides 
 into two, as it does in the cotyledon of Botrychiiim virgtiiuinum. Where a single 
 bundle occurs in the petiole it occupies a nearly median position and is concentric 
 in structure, although the phloem is somewhat less developed upon the adaxial side. 
 The xylem consists of a large mass of tracheids surrounded by the phloem, which is 
 reduced to about two rows of cells upon the adaxial side. Whether the largest cells 
 in this portion were sieve tubes was not determined. Where two bundles are present 
 they are somewhat smaller and are more nearly circular in outline than the single 
 bundle, which they resemble in structure. In the ground tissue are large inter- 
 cellular spaces which disappear at the base of the leaf, where the ground tissue in 
 cross-section appears almost solid. This development of lacunas in the first foliage 
 leaf recalls the similar ones in the cotyledons ui~ Botrychiutu and Opiiioglossiim. 
 
 In the older part of the rhizome there is a slight development ot periderm on the 
 dorsal side, which Farmer found to be the case also in the older rhr/,ome. This 
 periderm is restricted to the dorsal region and is obviously associated with tiie leaf 
 bases, as it is in Botrvi/mun and probably also in Opliioglosiiini. It presumably 
 
 F,G. 54. 
 
 Df an older sforophyte of HelmimhosiMhys. > 
 sheath is absent from the third youngest leaf. 
 
 acts as an absciss layer, such as Jeffrey demonstrated in Botrycluum, and this not 
 only causes the separation of the old leaves but perhaps also acts as a protective 
 layer to the leaf scars. The absence of periderm from the ventral side is no doubt 
 connected with the strictly dorsal position of the leaves. 
 
 The later roots, like the first one, grow from a single tetrahedral apical cell, 
 very much like that of Ophioglossum and Botrychiiim. The root cap is not very 
 prominent and is usually somewhat pointed. It apparently owes its origin entirely 
 to the activity of the outer segments of the apical cell. The primary segments of 
 the root cap undergo periclinal divisions, but, as in Botrychiiim, the stratifications 
 of the older layers is much less marked than it is in the ordinary ferns. Each of 
 the lateral segments of the apical cell divides by a somewhat radially placed anti- 
 clinal, so that a transverse section of the three youngest segments shows six cells 
 having a nearly radial arrangement. Periclinal walls may arise and the inner cells of 
 the segments give rise to the central cylinder of the root, while from the outer ones 
 is developed the cortex. As in the other Ophioglossaceae, no root hairs are formed. 
 
Till': ■souNc; spoRonivri (SI 
 
 The piiniaiy root, as Lang pointed out, is usiiall) tiiauli, liiii occasional!) 
 iliarch loots occur, as they usually do in Botrychium and in Ophioglossum pciidii- 
 linii. In this case, however, one of the xylems was rather larger than the other. 
 The later loots are tetiarch and in the older sporophyte. as Farmer showed, the 
 roots {feneially lia\i si\ wiem masses, or occasionally seven. This type of root is 
 most like that of the Maiattiacex. The cndodermis is pretty well developed, but 
 not so conspicuous as it is in Botrychium. 
 
 In the cortical region of the first root there is a /one of cells in which occurs an 
 endophytic mvcorrhiza, such as is common in the roots of other ( )phioglossacea-, 
 and probablv the same as the endophytc which is found in the tissue of the prothal- 
 lium. Whether the infection of the primary root is direct from the prothallium oi 
 whether there is a new infection from the soil was not determined. Lang states 
 that the mycorrhiza is only developed in the first two or three roots, the fourth root 
 and those formed later not having the endophytc. In these later roots the cortical 
 cells are densely filled with starch. 
 
 The rhizome continues to grow upright for a good while and it was not deter- 
 mined at just what time it assumes the prostrate position which it has in the adult 
 form. At first there is usually one root formed for each leaf, but in the older plants 
 this regularity is lost, and Farmer states that there may be three or four roots 
 developed for a single leaf. On the other hand, the number of roots may be less 
 than that of the leaves, especially in the younger plants. 
 
 A curious abnormal form was seen in a young sporophyte, where for some reason 
 several of the earliest leaves had remained in the rudimentary condition of the cotyle- 
 don. Five of these rudimentary leaves could be seen formed in succession. The 
 sheaths were fully developed and there was a long internode between each pair of 
 leaves. Three of these had developed roots, but the others had failed to do so. This 
 rhizome was nearly 3 centimeters in length, but it had hardly increased at all in 
 thickness. A single ternate leaf had expanded at the summit, but whether this 
 was the first functional leaf that the plant had developed could not be determined, 
 as the rhizome was broken off below and there may have been one or more func- 
 tional leaves developed below the first of the rudimentary ones. 
 
 This repetition of vestigial leaves recalls the condition of things in Ophioglossum 
 vulgatum and Botrychiiun huuiria. For a good while the new leaves are of the same 
 ternate form as that of the first foliage leaf, but sooner or later, probably depending 
 on the vigor of the plant, the ternate form is gradually replaced by five foliate leaves, 
 the later divisions being the result of an unequal dichotomy of the lateral leaf seg- 
 ments, similar to that by which the second lateral segments of the primary leaf are 
 separated from the terminal leaflet. In these five foliate leaves the characteristic 
 "pecopterid" venation of the adult sporophyte is fully attained (fig. 45, B). Kach 
 lateral vein forks twice, the ultimate veinlets extending to the margin of the leaflet. 
 A section of the petiole of one of these leaves shows that it contains four vascular 
 bundles, arranged in pairs. The dorsal bundle of each pair is decidedly larger than 
 the ventral one. The base of the petiole is almost perfectly cylindrical, but further 
 up it becomes winged, so that a groove appears on its inner face extending for some 
 distance below the junction of the lamina and the petiole. 1 here is a very slightl}' 
 developed hypodermal tissue composed of two or three layers of cells, the walls of 
 which are colorless and considerably thickened, some of them showing thickened 
 corners like the collenchvma found in the leaves of the Marattiacex. 
 
82 THE OPHIOGLOSSALES 
 
 COMPARISON OF THE YOUNG SPOROPHYTES OF THE OPHIOGLOSSACE.E. 
 
 If, as we believe, the type of sporophyte found in Ophioglossuni moluccanum 
 is really primitive, we may assume that the sporophyte at first had a single axial 
 stele, collateral in structure and essentially the same in leaf and root. This primitive 
 sporophyte had no stem at all, but consisted simply of leaf and root. From the 
 primitive vascular skeleton, composed of a single unbranched strand, we can derive 
 the different types characteristic of the older sporophyte in the three genera. As 
 to the first origin of the stem apex, we can only conjecture; whether it originally 
 arose, as it does in Ophioglossum, as an endogenous structure repeating, as it were, 
 the origin of the primary root, we have no means of knowing, but this seems to be 
 the most probable explanation of the origin of the stem apex in the primitive sporo- 
 phyte from which are descended the different types of the Ophioglossaceae. After 
 the establishment of the stem apex the secondary leaves contributed their quota to 
 the developing skeleton of the sporophyte. In Ophioglossum these leaf traces remain 
 largely free and anastomose only to a limited extent, thus giving rise to the open 
 tubular dictyostele with very large meshes. The structure of the individual strands 
 of the dictyostele is essentially the same as that of the free leaf traces. 
 
 In Helmmthostachys the early leaf traces fuse completely and there is formed 
 a solid stele in the younger internodes with a central xylem core, composed of the 
 united xylems of the two leaf traces. These leaf traces are approximately collateral 
 in structure, although it may be that they have a small amount of phloem upon 
 their inner face. After entering the petiole of the young leaf, however, these assume 
 a distinctly concentric form. As the leaves increase in size their traces become 
 broader and in section appear more or less crescentic, so that when the leaf traces 
 come together there is left between them a certain amount of the ground tissue which 
 after they have united appears as a pith lying inside of the tubular stele. This 
 pith, however, it must be remembered, is not part of the stele proper, but is merely 
 an included portion of the ground tissue. With the complete fusion of these two 
 broad leaf traces the tubular form is established and the wood appears in section 
 as a continuous ring. In Botrychiitm, especially in the large forms like B. virgiti- 
 tanitm, the tubular condition which is secondary in Helmmthostachys is established 
 at once; this is probably to be explained by the fact that the vascular bundles of 
 the first leaves are much better developed, there being two strands in the cotyledon 
 and in the second leaf, and the leaf traces belonging to these are correspondingly 
 broad and on fusing include at once a certain amount of the ground tissue, so that 
 the stele appears tubular from the beginning. Botrychium virginuuiiim undoubtedly 
 represents the most specialized type of the Ophioglossaceae, and the development 
 of the cambium with a permanent secondary thickening of the wood is an evidence 
 of a higher degree of specialization in the vascular system than is found in any other 
 living Pteridophyte. While a very slight indication of this secondary thickening 
 has been found in Ophioglossum, and I have also noted some slight traces of it in 
 Helmmthostachys, there is never developed in these the genuine cambium ring, such 
 as we find in Botrychium virginiarnim. In the development of the spiral protoxylem 
 elements Helminthostachys differs from the other Ophioglossaceae and suggests the 
 true ferns. In the early development of its vascular system there rare strong sugges- 
 tions of some of the Marattiacea?, especially Kaulfussia and Dancea. The develop- 
 ment of concentric bundles in the petiole in Helminthostachys and Botrychium 
 also suggests the Marattiaceae. 
 
 Assuming that the collateral bundle, which is typical of the stem in all of the 
 Ophioglossaceae and occurs also throughout in Ophioglossum, is primary, the con- 
 
THE VOUNG SFOROI'Il^l !•: 83 
 
 centric bundles as they occur in the petiole ot Hcltnnithostacliys and liotrytliiutti 
 must be considered as secondary. Fhe monarch root of Ophioglossum moluc- 
 canum and the other members of the section Eiiophioglossum is to be considered 
 as a relic of the primary condition where the single axial stele, as in the young spo- 
 rophyte of O. moluccanum, had the single strand of practically uniform structure, 
 extending through the leaf and root. The development (jf diarch roots, such as 
 those of O. pt'iuliilinri and Botrvcliium, is probably also secondary and perhaps 
 associated with the early development of the second leaf in these forms. The diarch 
 root appears again in the young sporophyte of the Marattiaceae and is permanent in 
 most of the leptosporangiate ferns. The roots of the larger species of Botryclutini 
 and especially those of Helminfliostacliys, with their increased number of xylems, 
 are undoubtedly secondary developments, perhaps associated with the large size of 
 the roots; and we again find this same type developing in the Marattiaceae. In 
 regard to this point, Helminthostachys is the most aberrant of the Ophioglossaceae 
 and approaches nearest to the Marattiaceae. 
 
 The leaf in the smaller species of Ophioglossum is probably a very primitive 
 structure and the closed stipular sheath — which is not exclusively foliar in origin, 
 but at first owes part of its tissue to that of the cortex of the root, from which the 
 sporophyte arises — is probably an older structure than the strictly foliar, stipular 
 sheaths in Botrychium and Helminthostachys. The simpler types of Botrychium, 
 like B. simplex and B. hinaria, show a transitional condition between the closed 
 sheath of Ophioglossum and the open sheath of B. virginiatnim, which may really 
 be spoken of as composed of two stipules, in this respect recalling the Marattiaceae. 
 Helminthostachys, in the development of the stipular sheath, agrees exactly with 
 the simpler types of Botrychium. In these forms the sheath is hood-shaped, open- 
 ing by a transverse slit in front and below, and the upper portion of the sheath is 
 broken through when the inclosed leaf emerges, so that the two apparent stipules 
 in Helminthostachys are really secondary, caused by a tearing of this hood-shaped 
 sheath, and are not proper stipules as they are in Botrychium virginianum. 
 
 The simpler and probably more primitive species of Botrychium, like B. 
 simplex and B. luiiana, are obviously intermediate between Ophioglossum and 
 the larger and more specialized species of Botrvchiutri. This is shown in the form 
 and venation of the leaves, as well as in the character of their tissues. Instead of the 
 pinnate venation found in the leaves of B. virginianum, these more primitive species 
 show no midrib in the leaf, but the veins all radiate from the veins of the leaflet, 
 dividing dichotomously, so that they are arranged in a fan-like fashion. Were the 
 iiuls of the veins connected there would result a reticulate venation, exactly like that 
 of Ophioglossum. These leaves, moreover, are fleshy in consistence and have 
 stomata upon both sides, while in Botrychium virginianum and in Helminthostachys 
 the stomata are restricted to the lower surface of the leaf In all of these particulars 
 Ophioglossum and the simpler species of Botrvcliium are evidenrh' iimrc primitive 
 than the other genera. 
 
 In Botrychium virginianum, as we have seen, the c()t\ledon is extrauiilinarih 
 developed, more so than that of any other Pteridophyte. Ihe contrast between 
 this highly developed cotyledon and the very rudimentary one in Botrychium lunaria, 
 where the yt)ung sporophyte passes several years under ground before the first green 
 leaf emerges, is very striking. It is, however, to be assumed that the rudimentar\ 
 condition of the leaves in Ophioglossum vulgatum and Botrychium lunaria is a 
 secondary condition, connected with their long life under ground. 
 
 Helminthostachys, on the whole, approaches more nearly to B. lunaria in its 
 early stages of development than it does to B. viroinianum. This is shown in the 
 
84 THE OPHIOGLOSSALES 
 
 rudimentary condition of the cotyledon and in the venation of the first foHage leaf, 
 which, although it has a ternate lamina, is quite rudimentary. We have already 
 referred, however, to the rudimentary leaves of B. liiiiana, which also sometimes 
 develop a very small ternate lamina. The venation of the first leaf in Hchnnithos- 
 tachys also approaches the cyclopterid type found in the leaves oi B. hiiiana. The 
 structure of the stele in the young stem also, after attaining the tubular form, is 
 more like that of 5. lunaria than like that of 5. virgintanutti. 
 
 On comparing the early stages of development of the ( )phioglossace2' and 
 Marattiaceae we shall find that they have a good many points in common, which 
 will be discussed more at length after we have described the structures of the latter. 
 Ophioglossiim, as to the formation of the cotyledon, is very much like Kaiilfiissia, 
 while Helmtnthostachys is more suggestive of Dancca. The stipular structures of 
 the two families are unquestionably homologous and the entire conical stipular 
 sheath, found in the lower members of the Ophioglossace:e, is probably an older 
 type than the free stipules found in Botrychiiim virginianum and the Marattiaceae. 
 
 In both the Ophioglossaceae and Marattiaceae the young sporophyte at first is 
 made up almost entirely of leaf and root, and the whole vascular system is composed 
 of the leaf traces with no proper cauline vascular tissues, so that one might almost 
 say that the stem is made up entirely of the bases of the leaves, the dominance 
 of the leaf being the most noteworthy feature in the morphology of these plants. 
 On the whole, probably Helminthostachys most nearly resembles the Marattiaceae. 
 This is true both of the character of the venation of the leaves, the structure of the 
 root bundles, and the development of tannin cells in the young sporophyte. These 
 tannin cells seem to be quite absent from the tissues of both OpJuoglossuni and 
 Botrychium. 
 
THE ADULT SPOROPHYTE 85 
 
 IV. THK ADULT SPOROl'inTI,. 
 
 The essential characteristics of the sporophyte are established while it is still 
 quite small and the subsequent differences are mainly an increase in the size of the 
 parts and finally the development of the spore-bearing structures which constitute 
 the peculiar spike or sporangiophore so characteristic of these plants. As the 
 structure of the tissues has been repeatedly studied and described, no attempt will 
 be made here to take up a detailed study of these. 
 
 The Ophioglossacea- in general, except for the rather elaborate vascular bundles 
 ot the stem, especially in Botrychium, are marked by great simplicity in the structure 
 of the tissues. The surface is usually smooth except in the younger plants, where 
 there may be a development of scales and hairs, presumably for the protection 
 of the young tissues of the stem apex. The hard hypodermal tissues and bands 
 of sclerenchyma, so common in the leaves and stems of many ferns and developed 
 to a less degree in the Marattiacea?, are practically entirely absent from the Ophio- 
 glossacea;. 
 
 Except for the vascular system of the stem the bundles are usually less developed 
 than is common in the more highly differentiated ferns, this being especially the case 
 in Op/uuglossiini, where the very delicate bundles forming the veins of the leaf run 
 through the spongy green tissue of the leaf without causing any projecting veins at 
 the surface. This condition is true also in the simpler types o( Botrychium and in 
 the young leaves of Helminthostachys. The great bulk of the ground tissue in leaf, 
 root, and stem is parenchyma. The development of periderm in the outer tissues 
 of the stem is probably always associated with the leaf bases and serves, as we have 
 seen, both to separate the dead leaf bases and to protect the scar thus left after the 
 leaf has fallen away. 
 
 The roots are characterized by the complete absence of root hairs. The outer- 
 most layer of cells often has the walls much thickened and they mav show the reac- 
 tion of cork. The bulk of the cortex of the root, however, is composed of unmodified 
 parenchyma. The vascular cylinder in the root remains monarch in the section 
 Eiwphioglossiim, but is diarch in the smaller roots oi Ophioderma and in the smaller 
 species oi Botrychium, and ranges to hexarch in the larger roots oi Helminthostach\s. 
 The roots, especially in the larger species, are thick and fleshy and as a rule branch 
 sparingly and somewhat irregularly. In the section Euophioglossum no lateral 
 roots are formed and branching is rare. When it does take place it is the result of a 
 true dichotomy of the apex. The root, as in the youngest stages in the plant, grows 
 almost always from a single tetrahedral apical cell, which is much alike in all the 
 genera. 
 
 The most characteristic feature of the Ophioglossaceae is the peculiar sporangial 
 spike referred to. There is a certain correlation in the degree of development of this 
 spike and the sterile leaf segment with which it is associated. The fertile leaves 
 may arise very early in the history of the sporophyte. Bruchmann states that the 
 first leaf to appear above ground in Botr\chium lutuiria is already a fertile one and 
 in Ophioglossum vulgatum the second green leaf to be developed usually bears spores. 
 This early development of the fertile leaf is probably an indication oi" the primitive 
 nature ot these plants, as we must assume that the ancestral form must have at once 
 developed a sporangial structure on the first leaf, or what corresponded to that in 
 the embryo. In liatrychium luiuiria Bruchmann figures )oimg spore-bearing plants 
 which are still connected with the prothalliuni. and |effie\ states that in B.ingiiii- 
 anum he once found a fruiting plant with which the prothalliuni was still connected. 
 
86 
 
 THE OPHIOGLOSSALES 
 
 THE SPOROPHYTE OF OPHIOGLOSSUM. 
 
 Of the three existing genera of the Ophioglossaceae, the genus Ophioglossum 
 is much the largest and most widespread. The number of species, however, is very 
 imperfectly understood, as there is great confusion in the nomenclature, owing to 
 the inadequate study of the tropical species, of which 
 the number is undoubtedly much greater than has 
 generally been supposed. The smaller terrestrial species 
 look very much alike and it is evident that collectors have 
 often failed to discriminate among them. My attention 
 was especially called to this while I was collecting ma- 
 terial of Ophioglossum in Java, where there is evidently 
 a considerable number of species. Raciborski (Raci- 
 borski 1), who has described the Pteridophytes of this 
 region, mentions only a single species as occurring in the 
 neighborhood of Buitenzorg, where most of my collect- 
 ing was done. But in that immediate neighborhood 1 
 collected at least three very distinct species, and two 
 were collected at Tjibodas, some 4,000 feet above Buit- 
 enzorg. Raciborski, however, mentions only a single 
 species as occurring at Tjibodas. There is no doubt 
 that a critical study of these plants from other regions 
 where they abound will greatly increase the number of 
 species to be recorded. 
 
 The great majority of the species of Ophioglossum 
 belong to the subgenus Euophioglossum, which should 
 probably be separated completely from the very differ- 
 ent forms which comprise the other subgenera. The 
 species of Euophioglossum are all small or moderate- 
 sized plants with undivided leaves which are generally 
 lanceolate or broadly oval in outline and have the 
 sporangial spike borne on a very long stalk attached to 
 the sterile part of the leaf near the base of the lamina. 
 These plants are always terrestrial, the leaves growing 
 from an upright rhizome, which is sometimes a good 
 deal enlarged, sometimes more slender in form. From 
 the rhizome there extend numerous roots which are 
 sometimes developed one for each leaf, but this is by 
 no means always the case. The sporophyte oi Euophio- 
 glossum seldom exceeds a height of 30 to 40 centime- 
 ters and some of them are very much smaller (fig. 55, 
 also plate 3). 
 
 Except for their larger size the later leaves do not 
 differ essentially from the primary leaf of the young 
 sporophyte. The venation is always reticulate, usually 
 without any definite midrib, although sometimes there is a central vein which is 
 slightly stronger than the others. In many species, e. g., 0. moluccanum, O. vulga- 
 tum, small branches ending blindly are found within the large meshes. In other 
 species, e. g., O. lusitanicum, there are no free veins within the areoles. 
 
 In many of the species but one leaf is formed each year, but in others, especially 
 many of the tropical species, there may be as many as four or five formed in a 
 
 F,G.55. 
 Two specimens of Ophioglossum moluc- 
 canum (Schlecht), slightly reduced. 
 The larger one is the typical 0. 
 moluccanum ; the smaller one is 
 probably a second species. 
 
THE Ain I.T Sl'OROl'UVTE 87 
 
 single season. In O. vulgatum the development ot the leaf is very slow, the leaf 
 remaining three years inclosed in the bud and emerging in the fourth season. So 
 far as I know, no study has been made of the development of the leaves in those 
 species of the temperate climates where more than one leaf is developed in the season, 
 and it remains to be seen whether the leaves which unfold in the same season are 
 of the same age or not. In the rapidly growing tropical species it is exceedingly 
 unlikely that more than a few months are necessary for the complete development 
 of the leaf, which unfolds as soon as it is mature, but nf) data are available on this 
 subject. 
 
 While the spike m Euopliiogloisiirii is usuall) inserted at the base of the sterile 
 lamina it may be attached much further down. This is especially noticeable in the 
 small species O. hcrgtanum, from South America (see Bower 9, page 435). In O. 
 bergKinum the leaves do not show a clear separation into lamina and petiole and the 
 spike is inserted close to the base ot the linear leaf, so that it appears to be an en- 
 tirely distinct structure. 
 
 In nearly all of the species of Euophioglossiim the later roots, like the primary 
 ones, are monarch, but in O. bergunnim they are diarch. In most cases at least, 
 as Holle (Holle 1) showed, only one root is formed for each leaf. This seems to be 
 the case in O. moluccanum, which was investigated with some care. 
 
 Ophioglossum pendulum (plate 4, B) is the best-known representative oi Ophio- 
 derma, the second section of the genus. This differs a good deal in its general char- 
 acters from the small terrestrial species of Euophioglossum. It is not uncommon 
 throughout the eastern tropics, where it grows as an epiphyte upon the trunks and 
 branches of various trees. The rough stems of certain tree ferns and some palm- 
 like species of Phcenix form a favorite habitat for this fern in certain regions and, 
 as we have pointed out in the study of the gametophyte, it is frequently found rooting 
 in the masses of humus between the old leaf bases of some epiphytic ferns like 
 Asplentiim nidus. 1 he stem, instead of being upright, is markedly dorsiventral, 
 but is short, with the leaves crowded together near the end and all growing upon 
 the dorsal side of the rhizome, as they do in Helminthostachys (fig. 63). The leaves 
 are very much elongated and the ribbon-like lamina merges very gradually into the 
 extremely stout petiole. These leaves in large specimens may reach a length of 
 1.5 meters, or even more, and the long, strap-shaped leaves hanging down from the 
 boughs of the trees present a very characteristic appearance. In larger specimens 
 it is not uncommon to find the lamina dividing dichotomously (see plate 4, B) and 
 it is said that sometimes this dichotomy may extend to the petiole, in which case 
 each segment of the leaf bears a separate spike. 
 
 The venation of the leaf is reticulate, the veins inclosing long, narrow meshes 
 with no free veins, and the venation thus resembles the type represented in Euophio- 
 glossum by Ophioglossum lusitantcum. fhe larger plants always have several 
 leaves which are expanded at the same time and growth goes on uninterruptedly, 
 The leaves are crowded together and there is no evident internode between them, 
 but it is clearly seen that they are arranged in two rows on the fianks of the short 
 rhizome (fig. 63, A). The remains of the stipular sheaths surround the bases of 
 the leaves and these sometimes present the appearance of two small stipule-like 
 organs. As in Euophioglossum, adventitious buds are frequently formed upon the 
 roots. As a rule they are not terminal, but lateral structures and probably do not 
 differ essentially in their development from those of O. moluccanum. 
 
 The rhizome is more or less buried in the humus and the roots ramify through 
 this. I^ven in the very young plant the roots are branched and this branching 
 Ih-coiik-s very marked in the larger N]-)()roph\ fe. There seems to be some difference 
 
88 THE OPHIOGLOSSALES 
 
 in the number of roots, but in most cases two were found connected with each leaf, 
 instead of one as 1 stated in my earher description of the plant (Campbell 4). The 
 roots are very stout, sometimes attaining a diameter of over 3 millimeters. The smaller 
 roots are diarch, but in the large roots there may be three, four, or even five xylems. 
 
 Ophioglossum poululiim is much the largest member of the Ophioglossaceae 
 and the sporangial spike as well as the individual sporangia far exceeds in size those 
 of any other species (plate 4, B, 3-6). The spike is attached by a short peduncle and 
 hangs down from the pendent lamina. In large specimens the spike may reach a 
 length of 20 centimeters or more with a breadth of about a centimeter. Above the 
 insertion of the spike the leaf is thin, but the slender peduncle is continued downward 
 into a thick, flattened midrib which merges gradually into the petiole, so that the 
 spike has very much the appearance of being a terminal structure with the sterile 
 lamina adherent to it. 
 
 The young leaves in 0. pendulum emerge while they are still very young and 
 the sporangial spike is in an extremely rudimentary condition The general devel- 
 opment of the leaf is therefore very easy to follow in this species, as these young 
 leaves are entirely free. 
 
 The second species of the section Ophioderma is the rare O. intermedium 
 (fig. 69; plate 4, A). This for a long time was known only from one locality in 
 Borneo, but has lately been collected at other points in the Malayan region. The 
 specimens figured were collected by the writer near Buitenzorg in Java. In the 
 account of the Ophioglossaceae given in Engler and Prantl's "Natiirliche Pflanzen- 
 familien" (Bitter 2) it is suggested that 0. intermedium is only a terrestrial form 
 of O. pendulum; but there is no doubt at all that it is a very distinct species. The 
 plants described here were growing in masses of humus at the base of old clumps of 
 bamboo. The stem is usually very short, forming a small tuberous body, from which, 
 in most cases, only a single leaf was growing, although two were sometimes found. 
 A careful examination of this short rhizome showed that, like Ophioglossum pendu- 
 lum, it is dorsiventral. The roots were short and in all cases observed were without 
 branches. In one of the specimens a bud very similar indeed to the corresponding 
 buds in 0. pendulum was found growing from one of the roots. Ophioglossum 
 intermedium differs from O. pendulum in being rigidly upright. The peduncle is 
 longer and the lamina of the leaf much smaller and more clearly differentiated from 
 the petiole. As in O. pendulum, however, the petiole is prolonged into the peduncle 
 of the spike with the same midrib-like thickening, caused by the coherence of the 
 basal part of the peduncle with the lamina. 
 
 This plant is exceedingly variable. In the larger specimens, except for the much 
 shorter lamina, the plant a good deal resembles small specimens of O. pendulum. 
 In others, however, the lamina is almost completely suppressed and this condition 
 closely approaches the third member of this section, Ophioglossum simplex (fig. 71). 
 This latter species is at present known only from one locality in Sumatra, and is dis- 
 tinguished from all the other species of Ophioglossum by the practically complete 
 suppression of the sterile portion of the leaf (see Bower 8). The form of the rhizome 
 and the habit of the plant in O. simplex most nearly resemble 0. intermedium, to 
 which it is probably not very distantly related. Whether or not we regard the 
 absence of the lamina of the leaf as a case of reduction, the fertile leaf in 0. simplex 
 certainly very closely resembles what one would assume to have been the primitive 
 condition of the leaf in the ancestors of the Ophioglossaceae. 
 
 The third .section, Cheiroglossa, contains a single very peculiar species, O. 
 palniatum (plate 5). This occurs throughout tropical America, but does not seem 
 to be a common plant. It has been reported from Florida, where, however, it is 
 
THE ADULT SPOROPHYTE 89 
 
 very rare. It has been collected in various parts of South America and in the West 
 Indies, and apparently the same species is known also from the Island of Bourbon 
 and from the Seychelles in the Indian Ocean. The specimens from which the 
 figures and descriptions given here were made were collected in the summer of 1908 
 in the Ulue Mountains of Jamaica, and were found in one jihice only, growing upon 
 a rotten tree stump. 
 
 The rhizome in Ophioglossitrn palmatiim is an almost globular tuber, which is 
 radially constructed and in this respect resembles that of Kuopluoglossum more 
 than that of OphtoJcrma. The leaf bases are covered with long brown hairs, which 
 give a characteristic appearance to the rhizome. The roots are numerous and more 
 slender than those of Ophioglossum pendulum. They are sparingly branched and 
 the branching may have the appearance of a true dichotomy, but the material 
 available did not allow of a critical investigation of this point. However, since 
 in Euophtoglossum the branching of the roots is actually dichotomous, it is not 
 impossible that the same may be true in Cheiroglossa. The smaller roots, as in 
 O. prnduhim, are diarch and, as in all of the other forms, a niycorrhiza is developed 
 in the cortical region. 
 
 In sections of one of the largest roots available the bundle was triarch. 1 his 
 root was rather young and the trachcary tissue was only partially lignified. The 
 niycorrhiza was also less evident than in the smaller roots taken from a younger 
 plant. Two young plants were found which probably had arisen from buds upon 
 the roots of the older sporophyte. One of these is shown in figure 70, A. It had 
 developed a single nearly lanceolate leaf very much like the early leaves of O. 
 pe)idulum. A single root had developed near the base of the leaf and this was joined 
 to the small globular rhizome. A second root could be seen above this, but whether 
 this second root belonged to the leaf or was part of the root upon which the bud had 
 arisen could not be determined. 
 
 The older leaves may reach a length t)t 60 centimeters or more (plate 5). "1 hey 
 are more or less deeply palmately cleft into narrow segments, which give the leaf a 
 curious superficial resemblance to certain kelps, such as Laminana digitata. \ he 
 base of the fan-shaped lamina is abruptly narrowed into a nearly cylindrical petiole, 
 about ecjual in length to the lamina itself. The venation is rather of the type of 
 (). viilgdt 11)11, having numbers of free veins in addition to the elongated closed meshes. 
 
 In .some of the smaller leaves (plate 5, 3) there is present a single small median 
 spike, which, like that of O. prnduhim, has a short peduncle, but it is inserted below 
 the lamina, in this respect again being more like Euophioglossutn. In the larger leaves 
 there are several sporangial spikes arranged in a series on either side of the petiole, 
 usually below the insertion of the lamina, but sometimes some of them are inserted 
 above the junction of the lamina and the petiole. Bower (Bower 9, page 436) has 
 figured a series of specimens of this species showing the great variation in the number 
 and arrangement of the spikes. He emphasizes the fact that although these spikes 
 are apparently marginal, in reality they always ari.se from the adaxial side of the 
 leaf. 1 ie shows that in this species also there mav be a branching of the individual 
 spikes sinn'lar to that which occurs in Opliioglnssum prnduhim and less frecjuently 
 ill ccirain species of Kuopluoglossum. 
 
 IHK ANATOMY OF K UOPHIOGI.OSSIIM 
 
 The stem aj^ex of the adult plant has been repeatedK studied in the witlespiead 
 Opluoalussunt vulg/itum, hut thetc is little information in regard to this jioint in 
 other species. From a somewhat careful txamination of (). mohutunum it is eviilent 
 that this species agrees dosel}' with (). vulgntum, and a study by one of mj- students, 
 
90 
 
 THE OPHIOGLOSSALES 
 
 Mr. C. S. Morris, of this species and also ot a broad-leaved form from Ceylon, 
 probably 0. reticulatum, showed that the same characters prevail in the species he 
 examined as in 0. vulgatum. It therefore seems likely that all species of Euophio- 
 glossum agree in the main in the structure of the stem apex and the development 
 of their tissues. 
 
 We have seen that in the young sporophyte the stem apex is completely inclosed 
 in a small cavity formed by the hood-like stipular sheath of the youngest leaf. This 
 condition, once established, is permanent (fig. 56) and there is no essential difference 
 in the appearance of the stem apex between a young plant in which two or three 
 
 1 older sporophyte of phioglosium moluccanum, showing arrangement of 
 n apex; sp, young sporophylls. D shows apical region more enlarged. 
 
 leaves have been developed and the full-grown sporophyte. The stem apex is of 
 limited extent, crowded in between the bases of the young leaves, and in O. moluc- 
 canum the single apical cell is of the same form as that which was found in the early 
 condition of the young bud; i. e., in longitudinal section it is four-sided, with a 
 broad base and a narrower outer face. According to Holle (Holle 1), the apical 
 cell in 0. vulgatum is pointed below; but Rostowzew (Rostowzew 2) in his figures 
 shows the same form as the apical cell in O. moluccanum. It is not at all impossible, 
 however, that both forms may occur, as in cross-sections the apical cell in O. mol- 
 //rr///;//^« may be either three-sided or four-sided (figs. ^Q and 56, D). The apical 
 
THE ADULT SPOROPHYTE 
 
 91 
 
 cell in O. molurcanurti may rherefore be described as a three-sided or four-sided prism 
 or a truncate pyramid. 
 
 As in the young plant, the sripular sheath of each new leaf is formed mainly 
 from the basal tissue of the leaf itself, but includes also tissue from the margin of 
 the stem apex. The sheath is open above by a narrow pore through which it 
 communicates with the space between it and the next stipular sheath. In O. 
 moluccaniitn, as in (). vulgatiim, three leaves of different ages can generally be seen 
 inclosed in the bud and sometimes the first indication of the fourth young leaf can 
 
 Six of a scries of cross-sections from a sporophyte of Ophio^lossum molurcanum, j/>, young sporophy!!; x, stem apex. 
 
 also be distinguished (fig. 56, C,D). A cross-section of the youngest sheath shows 
 the pore as a minute opening, separated from the stem apex below by a very- 
 narrow space. 
 
 A careful study of the older stem gives no reason for assuming that the tissues 
 of the very open reticulum forming the fibrovascular system of the rhizome is in any 
 part due to additions from the apical tissue of the stem. Whether longitudinal 
 or transverse sections are examined it is perfectly clear that the strands of this 
 reticulum are made up entirely of leaf traces which can be readily followed into the 
 young leaves or the tissue immediately below. The immediate apical region of the 
 stem, as in the younger sporophyte, is unmodified parenchyma which adds onh' to 
 the large central pith, if pith this can be called, as the separate bundles of which the 
 vascular reticulum is composed run free for long distances through the ground tissue 
 
92 
 
 THE OPHIOGLOSSALES 
 
 of the Stem, which is identical in appearance in the cortical and the central regions. 
 The bundle from each young leaf can be traced to a junction with a root stele and 
 from this point of junction it extends through the ground tissue of the stem, running 
 almost horizontally until it joins the trace from the next older leaf. In this way is 
 built up the open, large-meshed vascular cylinder. So far as could be determined 
 in O. moliiccaiiiitii only one root was formed for each leaf. The tissues of the root 
 base are continued upward to connect with the young leaf and downward to join the 
 stele from an older one. 
 
 No endodermis can be detected in the bundles of the stem in O. vulgatuni and 
 the same is true for 0. moliiccamtm, but in O. bcrgianum, 0. capense, and 0. ellip- 
 iictim, according to Poirault (Poirault 2), both an inner and outer endodermis occur 
 in the older part of the rhizome; these however, disappear in the younger region 
 higher up. 
 
 A transverse section of the mature rhi- 
 zome (fig. 57) in Ophioglossum moluccanum 
 shows the widely separated sections of the 
 strands of the vascular cylinder as a circle 
 of small collateral bundles without any 
 endodermis, the mass of wood being in 
 immediate contact with the thin-walled 
 parenchyma of the ground tissue or sepa- 
 rated from it at most by a single row of 
 pericycle cells. In some of the smaller 
 species the leaf traces are relatively broader 
 and there is an approximation to the ring- 
 shaped section presented by the cylindrical 
 stele of Botrychium or Helminthostachys, 
 and sometimes the same appearance may be 
 approached in sections of the older rhizome 
 which happen to pass through a region 
 where there are numerous anastomoses of 
 the bundles forming the reticulum (fig. 57, E). 
 
 Kxcept for the vascular bundles, the 
 tissue of the stem is made up almost ex- 
 clusively of simple parenchyma. The de- 
 velopment of periderm, which takes place 
 to a limited extent in the outer region, is 
 doubtless associated with the old leaf bases, 
 as it is in Botr\chium. 
 
 The leaf structure of £i/o/j/z/o^/oj-x//m is 
 exceedingly simple. The mesophyll is made 
 up of thin-walled green cells, practically uni- 
 form throughout; and through this spongy mass of mesophyll the delicate veins pass 
 without forming any projections at the surface. Both sides of the leaf are provided 
 with a simple epidermis, stomata being developed on both sides of the leaf. In 
 those species in which the leaf lies more or less horizontally, as it does, for example, 
 in O. reticulatum , the stomata are less numerous upon the upper surface. 
 
 The arrangement of the bundles in the petiole has been already studied in the 
 commoner European species, O. vulgatuni and O. lusitmitcum. In all of the species 
 belonging to the section Euophioglossum there is given off from the vascular system 
 of the rhizome a single leaf trace, which divides at the ba.se of the leaf into two 
 
 section of a sporophyte of Opiiioglo 
 p;, older than the one shown in fig. 56. \ j 
 
■;r<)Rorm 
 
 93 
 
 stiaiuls. I'.ach of tlusr stiaiuls ma\- ilividi-, oi onl\ <.iu- ol tlKin ( ti<i. ho). In soim- 
 of the smaller species there are only three hiindles in the petiole, and in these foinis 
 
 hase. In the larger species these 
 
 ction is greater. Ihus, 
 
 each of the two ventral bundles may divide again in tin iipjier region of the petiole 
 that a cros.s-section at this point would show h\i hinulles, a large dorsal one 
 id two pairs of sinallei ventral ones. A similar condition ol things is touiul in tin 
 
 there are no anastomoses of the bundles above the ..«,i<.. 
 branches divide further and the number of bundles in tin 
 
 Vascular bundle from an adult rhizome of Opkiogloaum moluccanum. X 
 Section of an adult root showing monarch vascular cylinder. X35. The 
 is occupied by the mvcorrhiza. 
 
 small species. 0. californiciim. As sections are examined below the junction of the 
 lamina and the petiole, it can be seen that further forking of the bundles has taken 
 place preliminary to their entering the lamina itself. Some of the species have the 
 tissues of the petiole quite compact, others show a greater or less development of 
 lacunae or air-spaces in the petiole. 
 
 [HE ROOT IN EUOPHIOGLOSSUM. 
 
 The early roots in O. moluccanum grow from a large tetiahedral apical cell 
 showing a fairly regular segmentation, much like that of the typical ferns; but the 
 root cap is not so well developed nor does it show as definite a stratification. In the 
 
 early roots the whole of the root cap seems to be derived from the outer segments 
 of the apical cell. In the later and larger roots there seems to be somewhat less 
 regularity in the segmentation and no sections were found which showed the apical 
 
94 
 
 THE OPHIOGLOSSALES 
 
 cell and its segmentation as clearly as in the younger roots; however, only a small 
 number of these later roots were sectioned and it may be that they were not all 
 quite normal. 
 
 A section of an adult root is shown in fig. 59,5. The cortical region is made up 
 of simple parenchyma and the mycorrhizal zone is very conspicuous. The endo- 
 dermis is not very clearly defined, less so than in the earlier roots, and its limits are 
 
 Five sections of a full-grown sporophyll of Qphiogloisum molucranum. E, section of peduncle of spike; /, 
 
 not easily determined. The cortical cells inside the mycorrhizal zone are densely 
 filled with starch. More than half of the section of the bundle is occupied by the 
 large solid mass of wood. This is separated from the endodermis by a single layer 
 of pericycle cells. The single mass of phloem lying next the wood is not much more 
 than half as large. 
 
 ANATOMY OF OPHIODERMA. 
 
 Owing to the dorsiventral position of the rhizome and to the more rapid de- 
 velopment of the leaves, the section of the stem apex in Ophioglossiim pendulum 
 differs a good deal from a similar section oi Euophtoglossitm (fig. 64). The base of 
 the leaf is bent sharply upward, this being the case also in the very young leaf which 
 is inclosed within the bud. The leaves break through 
 the inclosing sheath in a very rudimentary condition, 
 although of such size that a section through the base 
 of the youngest visible leaf shows usually but a single 
 younger leaf inside the sheath, instead of the three or 
 four that are seen in a similar section of the bud in 
 EuQphioglossum. Within the sheath at the base of the 
 youngest visible leaf there is found a cavity in which is 
 the growing point of the stem and the youngest leaf 
 The apex of the stem forms a strongly inclined, nearly 
 plane surface and, as in Euophioglossum, the growth is 
 from a single apical cell which, so far as could be deter- 
 mined, has the form of a three-sided pyramid, whose 
 lateral faces are more or less strongly convex and whose apex may be truncate. 
 F"rom the few cases observed it is not possible to say definitely whether basal seg- 
 ments are regularly cut off as well as the lateral ones. The segments cut off from 
 the lateral faces are large and the divisions irregular. 
 
 Owing to the large size of the sporophyte there are practical difficulties con- 
 nected with the study of the vascular system of the rhizome from a series of micro- 
 tome sections. The numerous bundles in the leaf base do not unite into a single 
 
 E root of phioglossuti 
 1, showing dichotomy. 
 
a'OROI'HYTK 
 
 95 
 
 leaf trace, but pass downward througli the cortex separately, and even it it were 
 possible to make a satisfactory series of the sections of the large rhizome the task of 
 following the course of the very numerous individual bundles would be a difficult 
 one. From free-hand sections of the large base of the leaf, however, one can see 
 without difficulty the general plan of the vascular skeleton. The stout petiole is 
 slightly contracted at its base, but a section presents very much the same appear- 
 ance as one made higher up. 
 
 In specimens that have been preserved in alcohol so as to become decolored, 
 the tissues of the thick leaf base are sufficiently transparent to show quite clearly 
 the course of the bundles even without sectioning, by simply slitting the petiole 
 through the middle. The bundles are seen to anastomose freely at the extreme base 
 of the leaf, very much as thev do in the lamina. If such a section is examined where 
 the leaf base joins the rhizome (fig. 63, ])) the circle of bundles can be seen to become 
 
 A. Stem showing dorsivcntral form and branched roots. /, bases of leaves. 
 
 B. Buds attached to a root fragment. The older bud, 6', has developed a younj* sporophyll, j/j, and two sec<)nJary roots. 
 
 C. Base of a large leaf, showing arrangement of vascular bundles. 
 
 D. Section of same. 
 
 smaller as the bundles enter the cortex, but the bundles remain for the most part 
 quite separate, although there may be an occasional anastomosing of these, as occurs 
 in the petiole itself. This group of bundles composing the leaf trace — if such it 
 can be called — forms a circle about the opening in the vascular cylinder of the stem 
 where it joins the latter. A section of the vascular cylinder within the rhizome itself 
 presents the appearance of a nearly complete ring, which is probably formed by fu.s- 
 ing of the bundles derived from the leaf bases. The much greater development of 
 woody tissue in the stem of O. pendulum as compared with that of Euop/iioglossutri 
 is doubtless due to the very much larger size of the leaves and the correspondingly 
 greater number of vascular bundles contributed by them to the woody cylinder of 
 the rhizome. 
 
 The very young leaf, such as is shown in fig. 63, B, has a thick, fleshy leaf base, 
 terminating in a very small and pointed lamina strongly bent over as it is, c. g., 
 in the young leaf of Bolrvrhium virgiriinnum. This differs from the species of 
 Euopluoglossitm, in which the leaf is straight in the bud. llnder the arching hood 
 formed by this folded-over lamina is the young spike, almost equal in length at this 
 time to the lamina itself. A somewhat older stage is shown in fig. 82, B. The 
 leaf has now become somewhat flattened, but as yet there is no sharp distinction 
 between the lamina and the petiole. In this early stage of development the inter- 
 pretation of the spike as a terminal structure seems very plausible. An interesting 
 
96 THE OPHIOGLOSSALES 
 
 case is shown in fig. 8z, E, where the lamina is almost entirely siippiesscd and the 
 terminal character of the spike is very evident. 
 
 As the leaf develops, there is a very great increase in size in the lamina, which, 
 as we have seen, may reach a great length. The anatomical structure of the leaf 
 
 A. Section of stem apex. X3. 
 
 B. Young sporophyll. sp, sporangiopho 
 
 C. Stem apex, x, and youngest leaf, /. 
 
 -Ophioglossum pendulum. 
 
 D. Tlie stem apex, more enlarged. 
 
 E. Very young sporophyll. sp, sporangiopho 
 
 closely resembles that of Eiiophioglosstim, but the stomata are much larger and 
 around the stoma is a concentrically arranged series of cells, indicating that probably 
 the formation of the mother cell of the stoma is preceded by a series of preliminary 
 
 Fig. 66. 
 
 .A. Section of petiole of a sporophyll of (). pnidului 
 
 B. Section near base of lamina, sp, base of spike. 
 
 C-E. Sections of peduncle of spike. C and D arc 
 
 same leaf; E from a laiger one. 
 
 divisions in the epidermal cell. The tissues in the petiole are very like those in the 
 lamina of the leaf except that they are somewhat more compact. The walls of the 
 ground tissue in this region are very deeply pitted. 
 
 The structure of the vascular bundle of the petiole (fig. 65), is much like that in 
 Euophioglossum, except that there is a rather greater development of tissue upon the 
 
THK ADULT SPOROPHYTE 
 
 97 
 
 inin-r side of tlie wood, so that there is a slight suggestion, at least, ot a concentric- 
 structure of the bundle, such as is found in the petioles of Botrychium. The xylem is 
 composed entirely of tracheids with reticulate markings. The rest of the bundle is 
 made up of sieve tubes, mingled irregularly with smaller, thin-walled elements. It 
 , \ could not be certainly deter- 
 
 mined whetherany sieve tubes 
 were developed on the inner 
 side of the wood. 
 
 The number of bundles in 
 the petiole, especially in large 
 specimens, is much greater 
 than in any of the species of 
 Euophioglossum. In a spe- 
 cimen of medium size (fig. 66) 
 eighteen bundles could be 
 seen, of which probably seven 
 or eight are destined to sup- 
 ply the spike. In the fertile 
 leaves these form a complete 
 circle, but in sterile leaves 
 there are no bundles on the 
 adaxial side of the section. In the free portion of the peduncle, which is relatively 
 slender, the number of bundles is much less than in the broad basal portion. In 
 the smaller specimen figured there were but three bundles in the middle part of the 
 peduncle. The complete absence of the adaxial strands in the petiole of the sterile 
 leaf, even at its base, is a strong confirmation of the view suggested by both the older 
 leaf and the younger stages, that the peduncle is really an independent structure 
 whose bundles are joined directly to the rhizome. 
 
 Fic. 67. 
 1 of the root apex of Ophioglo: 
 
 ; niycorrliiial : 
 
 O. intertnciUum (fig. 69) closely resembles O. petidulum in its anatomy, like the 
 latter developing stomata upon both sides of the leaf, and these stomata are very 
 large and show accessory cells similar to those found about the stomata of O. 
 peiiJnJinn. Except for the difference in size, the arrangement of the vascular bundles 
 in the peduncle and their structure is closely similar to that in 0. pendulum. 
 
 7 
 
THE OPHIOGLOSSALES 
 
 The anatomy of O. simplex, so far as it has been investigated (see Bower 8) 
 corresponds closely with that of the other species of the section Ophioderma. 
 
 The roots of 0. pendulum also show the tetrahedral apical cell (fig. 67), but the 
 
 divisions are much less regular than in 
 O. moluccanum and probably the other 
 species of the Euophtoglosstim. The 
 segments divide slowly and increase a 
 good deal in size before the first divi- 
 sions take place, so that the youngest 
 segments may equal the apical cell in 
 size, and there is thus a certain resem- 
 blance to the apical meristem in the 
 root of the Marattiaceae. It is prob- 
 able that, as in the Marattiaceae, the 
 lateral segments of the apical cell also 
 contribute to the root cap, which is 
 rather better developed than it is in 
 O. moluccanum. 
 
 There is a good deal of difference 
 in the structure of the root bundle 
 (fig. 68). This is diarch in the primary 
 root, as we have seen, and also in the 
 smaller ones of older sporophytes; but 
 in the larger roots, which may reach 
 a diameter of over 3 millimeters, it is 
 triarch; tetrarch bundles are also com- 
 mon, and in one very large root which 
 I examined the bundle was pentarch, 
 but one of the xylem rays was shorter than the others, the xylem forming a solid 
 central mass having the form of an irregular five-pointed star. 
 
 intermedium Hk., slighlly 
 a very much reduced 1 
 
 fragment 
 
99 
 
 Material was lackiiiu tor a snuU' nf'rlu- anat(iin\- nC rlu- ilii/oim- in O. />/jl>niitiirri 
 (fig. 70), but tile structure dI rlic ])Lti()lL- is \ l 1 )• iiurIi like- that ot O. prnJuluNi and 
 probably the same relation of the leaf bundles to the bundles of the rhizome obtain 
 as in O. pendulum. The difference in the venation has already been pointed out 
 and the leaves differ also from those of 0. pcnduluru in the fact that stomata are 
 developed only upon the lower side. A feature of this species is the presence at the 
 base of the leaves of very conspicuous multicellular branching hairs, (juite different 
 from those found in any other species of Op/iioglossum (fig. 70, C). I find that the 
 cells are much more elongated in my specimens than those figured in the account 
 of the ( )phioglossace.-i' given in the Naturliche Pflaii/enfamilien.* 
 
 iiiK si'okoi'iniK OK Korm-ciiu.M. 
 
 There is much the same uncertaint\- as to the iiumlnr of s|Hcics iA' Butix, linnii 
 that there is in Opiuoglossum. Some of the species are e.xceedingl)- variable and 
 their limits are difficult to determine. Bitter, in his account of the Ophioglossace;c 
 in Engler and Prantl's "Naturliche Pflanzenfamilien," states that there are sixteen 
 species. Underwood (Underwood 1) says "about thirty," of which fifteen occur in 
 the United States. Christensen in his Inde.x Filicum recognizes thirty-four species 
 
 Many of the species are very widespread, being most abundant in the North 
 Temperate Zone. According to Christ (Christ I), B. lunarta is not only widespread 
 throughout the colder parts of the Northern Hemisphere, but occurs again in the 
 corresponding regions of the Southern Hemisphere, having been reported from Pat- 
 agonia, southern Australia, Tasmania, and New Zealand. A few species occur in 
 the mountain regions of the tropics; thus B. laiuiginosum is not uncommon in the 
 highlands of Ceylon, where I collected it at an altitude of about 7,000 feet. 
 
 The simplest member of the genus is Botrycluum simplex, a very variable 
 species occurring at various places in Noith America, northern Asia, and Kurope. 
 Its smaller forms closely resemble a small Opliioglossutii (fig. J2,A, B), except that 
 the sporangia are more distinct. From this primitive type, obviously not very 
 remote from Opiuoglossum, there is an interesting series of forms leading up to the 
 large species like B. virgiiiiatuim and B. silatfohum. This latter species, which 
 grows in the mountains of western North America, is perhaps the largest of the 
 genus (plate 7). The plant is sometimes 2 feet high, the leaf nearh' a foot in width. 
 and the very large panicle of sporangia 6 inches 01 more in length. 
 
 In the evolution of the leaf in Botrvrlinnu, xhv sjiorangiophort' shows a devel- 
 opment parallel with that of the sterile leaf. In the forms of B. simplex, lia\ing 
 an entirely undivided sterile lamina, the sporangiophore has the form of a simple 
 spike with a single row of large sessile sporangia on either side (fig. 72, J) and very 
 closely resembles a small Opiuoglossum. As the sterile leaf segment becomes more 
 and more dissected, there is a corresponding branching of the sporangiophore which, 
 
 • Prc.frssor Biiwrr. at the last meeting of the British .Association for the .Advanci-incnt of Science, read a paper on the Anatomy 
 of Ophiogloaum pulmaiuiii. This has not yet been published, but Professor Bower has kindly furnished me with an abstract of 
 his paper from which the following extract is quoted. 
 
 "It was thought probable that 0. palmaium (the only spiiies .>( Mction (./le/Vojj/ojju) would share with the species 
 named (i.e., the species of Ofiluoderma) the character of a d\\\A- 1 1 1 ii 1 1 i, , in,l material collected in Jamaica has shown 
 that it docs. The axis is much distended by parenchymatcn; i nili and cortex, and as a consequence the 
 
 meshes of the stele arc transversely widened. From their ni.ir l : i ; ■ m ' li ti, hut not quite simultaneously, arise two 
 strap-shaped strands, which are thus widely apart in their on;;ni. .\:ui ui .luibion into numerous smaller strands, these 
 range themselves into two fan-like semicircles, which spread till tlicir margins meet, forming the circle of strands of the 
 petiole. A remarkable feature of the stock is the intrusion of roots into the bulky pitli; this is especially obvious 
 towards the base, where they pass out as thick mycorrhizic roots." 
 This behavior of the roots recalls that of the Ma 
 
100 
 
 THE OPHIOGLOSSALES 
 
 in the large forms of B. virginianum and B. sUaifoliiim, forms a large panicle 15 
 centimeters or more in length and bears many hundred sporangia. Two types of 
 leaf division may be noted, a pinnate form represented by 5. lunaria and its allies, 
 which are sometimes separated as a special genus Eiihotrychiiim, and a ternate 
 division which is found in all of the larger species. These larger species form a 
 second section, Phyllotrichinm. A third genus, OsmunJopteris, has been proposed to 
 include Botrychium virginianum and its allies. In most of the species of Botrychium 
 
 Ophioglosmm simplex 
 Ridley (after Bower). 
 The fertile leaf has no 
 sterile lamina. 
 
 Fig. ^z.■ — (A-D, after Luerssen.) 
 A, B. Botrychium simplex Hitchcock. C,iS. ternatum S 
 D, B. lunaria (L.) Swz. E, B. virginianum (L.) Swz. 
 
PHE ADULT SPOROPHYTE 
 
 101 
 
 the tcxnuc (if the katis soft and Hcsh}-, like that of Opiiioglossum, hut in li.virgini- 
 niiutn it is thin and niemlManaceous, hkc that of many true ferns. In B. limartn 
 (fig. 72, D) the venation comes nearer to that of Ophinglossum than is the case in 
 the larger species. The fan-shaped leaflets have no midrib and the veins branch 
 dichotomously from the base of leaf, radiating from this point in the manner de- 
 scribed as "cyclopterid." In the larger species the leaflets have a distinct midrib 
 
 E.G. 73. 
 
 A. Rhizome of a strong plant of Botrychium virginianum. Xi. The base of the expanded le; 
 
 B. Section of the terminal bud. j(, stem apex, i, n, 111, the three youngest leaves. X2.66. 
 
 C. Section of petiole, enlarged. 
 
 D. Section of rhizome; p.pith; m, medullary rays; jr.xylem; f A, phloem; c, cambium; iA,endode 
 
 A-G. A series of sections of the sporophyll of 
 
 Botrychium la, 
 
 .ugSnosum \\z 
 
 position. Csi-cticmof pedunck-. 
 
 
 
 H. Section of the petiole of the cotyledon of B. 
 
 virg^inianum. i 
 
 , !, lacuna:. 
 
 and the forking veins are attached to this laterally (fig. 72, E). The type of B. 
 lunaria is with<nit question the more primitive, and were the ends ot the veins united 
 so as to inclose areoles there would be exactly the same type of venation as is found 
 in Ophioglossiini. 
 
 In Eiihotrychimn the anatomy of the leaf is almost exactly like that of Ophto- 
 glossum, stomata being developed upon both sides of the leaf; the mesophyll is 
 composed of absolutely similar cells throughout and the vascular bundles running 
 
102 
 
 THE OPHIOGLOSSALES 
 
 through the mesophyll do not project at the surface as veins. In the section Phyllo- 
 tric hi um, where the sterile lamina is horizontal, stomata are developed only upon the 
 lower surface. In these forms the mesophyll is more compact upon the upper side, 
 but does not develop a distinct palisade layer. The epidermal cells may be simply 
 elongated, e. g., B. lanuginosum, or they may be undulate, e. g., B. virgtnianum. 
 The development of the stomata was examined in B. lanuginosum and the stoma 
 was found to arise directly from an unmodified epidermal cell. The section Eubo- 
 trychium shows a further approach to Ophioglossum in the character of its stipular 
 sheath, which, as %ve have seen, forms a hood-like appendage at the base of the leaf, 
 while in B. virginianum the stipular sheath is open in front, so that there are really 
 two stipules. Moreover, the leaf in the latter is sharply bent over in the bud in a 
 way to suggest somewhat the circinate vernation of the Marattiaceae. For the most 
 part the surface of the plant is quite smooth, as in Ophioglossum, but the sheaths 
 
 Fig. 75. 
 
 A. Stem apex of young sporophyte of Botrychium virginianum. 
 
 B. Cross-section of stem apex. 
 
 C. Part of vascular cylinder of stem, pi, phloem; cam, cambiut 
 
 D. Some of the tracheids, more enlarged, showing bordered pits. 
 
 ;, apical cell; /, youngest leaf. X 
 i; A'v, xylem; m, medullary ray. 
 
 enveloping the young leaves, as well as the very young leaves themselves, are often 
 hairy, and in a very few species, like B. lanuginosum, long and slender unicellular 
 hairs are sparsely scattered over the surface of the adult leaf. The stem, as we have 
 seen, is always an upright rhizome and usually unbranched, but adventitious buds 
 may be developed (see Bruchmann 2). These adventitious buds may develop into 
 lateral branches. In one specimen of 5. lanuginosum examined by me there were 
 two equal branches, presenting the appearance of a dichotomy, but a careful exam- 
 ination indicated that the original apex had been destroyed and that these two 
 branches were lateral shoots, which had probably arisen as adventitious buds. In 
 Eubotrychium the leaves are arranged spirally upon the rhizome, as they are in 
 Euophtoglossum. This is also true of certain members of the section Phyllotrichium, 
 but others, e. g., B. ternatum, are arranged in two rows, thus suggesting the dorsi- 
 ventral arrangement found in Helminthostachys. 
 
 The structure of the adult stem does not differ, except in size, from that of the 
 younger plants already described. In Eubotrychium the cambium is absent and 
 apparently there is no secondary increase in diameter, but all of the larger species 
 oi Phyllotrichium show a greater or less developmentof the cambium, which we have 
 
TlIK ADL' 
 
 103 
 
 noted in the bundle of /i. virgniiaiunn. 1 lie position ot the sporangiophoie is some- 
 what different in different species. In Eulxjiryduutn it is attached, as a rule, close to 
 the junction of the sterile lamina and the petiole, very much as it is \n Eiiophioglossum. 
 In B. ternatum and B. obliquum the long peduncle is inserted very much further 
 down, sometimes almost at the level of the ground, so that at first sight the spor- 
 angiophore seems to be quite independent of the sterile leaf. In B. virgintanum 
 
 of the prniarv rrait. 
 
 B. Lmugniosuui it is inserted close to the base of the sterile lamina or even 
 above its base, and in the hitter species there may even be sometimes a second spo- 
 rangiophore developed on the fertile leaf above the insertion of the primary one. 
 
 Certain species show various monstrosities, this being particularly the case 
 with B. luuaria. In such forms certain of the segments of the sterile lamina may bear 
 sporangia. A number of rluse curious forms are figured by Goebel (Goebel 2). 
 
104 THE OPHIOGLOSSALES 
 
 In fig. 73, A, is shown the terminal bud of a strong plant of B. virginianum. 
 The base of the oldest leaf is cut away so as to show the young leaf for the next 
 year, which is seen to be strongly bent over, and the young leaf segments are rolled 
 inward in a way that very strongly reminds one of the young leaves of the typical 
 ferns. The young leaf is covered with hairs. The roots are thick and fleshy, prob- 
 ably one corresponding to each leaf, although this is not always easy to determine. 
 They are much contracted at the point of insertion. The roots branch freely, the 
 branching being monopodial, thus resembling Ophioglossum pendulum. An exam- 
 ination of the section of the bud shown in fig. 73 shows that the young leaf, which 
 is to unfold the next year, is already well advanced and the formation of the young 
 sporangia has just begun. The sporangiophore is also evident upon the leaf for the 
 following year, and the leaf which is to unfold in the third year can be recognized, 
 but its parts are not yet differentiated. 
 
 While the roots branch freely in B. virginianum, this is much less marked in 
 B. lunaria. In large plants of B. lainiginosum which were examined somewhat 
 carefully it was found that the roots appeared to be quite unbranched, although 
 equaling those of B. virginianum in diameter and showing, like them, a similar 
 tetrarch structure of the vascular cylinder. A more careful examination showed, 
 however, a few very short branches on some of the roots, but these were rare. It 
 may be that the rudiments of these lateral roots are formed, but that they do not 
 develop, and in this respect there is a certain resemblance to the condition which 
 obtains in Helminthostachys. 
 
 The anatomy of the roots does not differ from that of the younger sporophyte 
 already described, except that in the larger specimens the roots are regularly tetrarch, 
 while in Euhotrychium they are permanently diarch. 
 
 The earlier roots, as in Ophioglossum, have a tetrahedral apical cell which shows 
 a more regular segmentation than is found in the later roots and in these earlier 
 roots, at least in B. virginianum, there seems little question that the tissue of the 
 root cap is derived entirely from the outer segments of the apical cell. In the larger 
 roots, taken from the older sporophyte, this regularity is not so marked and it is 
 probable, although this could not be determined absolutely, that the lateral seg- 
 ments of the apical cell may also contribute to the tissues of the root cap. The 
 regularity of the segmentation is still less marked in the later roots of 5. obliquum. 
 A slight development of periderm may be seen in the older roots, but this does not 
 form a continuous layer, being developed only in irregular patches. The secondary 
 increase in diameter of the roots is apparently due almost entirely to simple enlarge- 
 ment of the cortical cells. 
 
 THE SPOROPHYTE OF HELMINTHOSTACHYS. 
 
 While Helminthostachys differs very much in habit from the other Ophio- 
 glossaceae, it nevertheless shows that it is unmistakably related to them and especially 
 to Botrychium. The dorsiventral prostrate rhizome with the ample palmately 
 divided leaves give it a very individual aspect, but, as is shown from a study of the 
 early history of the sporophyte and the gametophyte, there is no question that it 
 should be associated with the Ophioglossaceae. As yet but a single species is gener- 
 ally recognized, which grows throughout the Indo-Malayan region, extending from 
 the Himalayas through the Malay archipelago to northern Australia (plate 8). 
 
 The anatomical details of the adult plant have been very thoroughly investigated 
 by Farmer and Freeman (Farmer 2), and from a comparison of the structure of the 
 adult sporophyte with that of the younger stages it is clear that the essential 
 
THK ADULT SPOROPHYTE 
 
 105 
 
 characters are established at a comparatively early period. The rhizome Tfig. 79) 
 has the leaves arranged in two rows on the dorsal side, sf)mewhat as in Ophioglossum 
 pnidiiliim, but the leaves are less crowded and the internodes are evident, although 
 less marked than is often the case in the very young sporophyte. The youngest 
 leaves, still inclosed in the stipular sheath of the next older one, are conspicuous as 
 blunt projections, situated between the apex of the rhizome and the first expanded 
 leaf. As the leaf breaks through the stipular sheath of the next older one the 
 sheath is split into two stipule-like flaps, which much resemble superficially the 
 stipules of the Marattiace;e. From near the base of the leaves on the flanks of the 
 rhi/ome and belonging to the ventral portion of the stem there arise the thick fleshy 
 
 A. Sporophyll of Helnrinihostachysy somewhat redu 
 
 B. Base of sporangiophore, about natural size. 
 
 C. Stoma. X200. 
 
 Four sections of the sporophyll Fig. 3 passes 
 through the base of the spike, spi 4, section of 
 peduDcle of spike. 
 
 roots, which resemble those of the other Ophioglossaceae. According to Prantl 
 (Prantl 1) and Farmer (Farmer 2), the roots bear no definite relation to the leaf, 
 and the number as a rule somewhat exceeds those of the leaves. The roots branch 
 monopodially, as in the larger species of Botrychium, but Farmer states that the 
 lateral rudiments often prove abortive, or fall away at an early period, so that the 
 roots are apparently unbranched. This behavior recalls the condition found in 
 Botrychium latiuginosiiin. 
 
 The leaves are very different in general aspect from those of the other Ophio- 
 glossaceae, but are really not very different in their method of branching from the 
 ternate forms oi Botrychium. There are still traces of the three primary divisions 
 of the first leaves which undergo subsequent divisions, so that the leaf has a palmate 
 form (plate 8, B). To a certain extent, however, these divisions are the result of 
 
106 
 
 THE OPHIOGLOSSALES 
 
 unequal dichotomy, like that which results in the first divisions of the primary leaves. 
 In large plants the leaves may reach a length of 30 or 40 centimeters. The petiole is 
 nearly cylindrical at its base, but higher up becomes slightly winged at the margins, 
 these wings passing into the lamina. A section of the petiole taken a little way 
 below the insertion of the sterile lamina shows that the peduncle of the sporangio- 
 phore is recognizable for a long distance below the point at which it becomes free, 
 its adaxial surface projecting as a rounded ridge between the wings on the margin 
 of the petiole (fig. 78, D). In texture the leaves are firmer and more leathery than 
 in any of the other Ophioglossaceae, and in this respect, as well as in their venation, 
 they closely resemble such Marattiaceae as Dancea and Angiopteris. 
 
 The sporangiophore is to some extent intermediate in character between that of 
 Ophioglossum and Botrychium. It is inserted near the junction of the lamina and 
 petiole, but, as already stated, can be traced for some distance below the point where 
 it joins the petiole. The fertile portion of the sporangiophore, which may reach a 
 length of nearly 10 centimeters, is composed of crowded, short branches, upon 
 
 which are borne groups of nearly globular sporangia much like those oi Botrychium. 
 The ends of the branches which bear the sporangia may be expanded into small leaf- 
 like organs which, under exceptional conditions, look much like small leaves, upon 
 which the individual sporangia are borne (see Bower 9, fig. 239). The stem in 
 Helminthostachys almost always remains unbranched, except that there may be 
 formed adventitious buds somewhat as in Botrychium. The origin of these buds, 
 however, has not been investigated. The rhizome is rather fleshy in texture and 
 may reach a diameter of about a centimeter. Usually but one leaf is fully developed 
 at a time, but it is not uncommon to find two leaves produced the same season. 
 Farmer's description and figures of the structure of the terminal bud show that it is 
 essentially the same as we have found it to be in the younger plant, and, as has 
 already been pointed out, it shows strong resemblances to the terminal bud in 
 Botrychium lunaria. 
 
 The stomata (fig. 78, C) are confined to the lower surface of the leaf and are 
 surrounded by a series of concentrically placed cells like those around the stoma in 
 Ophioglossum pendulum. They are much smaller in size, however, and in this 
 
THK ADULT Sl'UROl'HVTE 107 
 
 resptct ni;i\- Intrn lie compaiiil to tlu- stonKita o{' J)(in,rtu oni.- of tlic Marattia- 
 cea?, whicli slmus a similar lonccntiic arrangement ot tlie accessor)- cells. A 
 cross-section ot the leaf shows that the mesophyll is better ditteientiatecl than in 
 Botrychntru. I here is a single compact layer of palisade cells below the epidermis, 
 and below these the mesophyll has very large intercellular spaces which communicate 
 with the stomata. In the anatomy of the leaf, therefore, as well as in its venation, 
 Helmintliostiuliys more nearly resembles the IVIarattiace;e than it does the other 
 Ophioglossace:e. The section of the jHtioie is not unlike riiat of Ophio gloss urn 
 pendulum (figure 78, D). A section nuule near the base of the petiole is almost 
 circular in outline and there is a ring of separate bundles. On the adaxial side, 
 however, two other bundles are seen inside the outer circle. Farmer and Freeman 
 (Farmer 2) state that the number of bundles is about seven or eight, but the 
 larger specimens which I have examined show a greater number. If a section of 
 the petiole is examined higher up it no longer appears circular, but is slightly 
 lobed, these lobes corresponding to the bases of the three divisions of the sterile 
 lamina, and on the adaxial side may be seen a fourth lobe which marks the position 
 of the peduncle of the sporangiophoie. In this upper region of the petiole the 
 bundles become more or less coalescent. but two adaxial bundles remain quite 
 distinct and are continued upward into the spike. The structure of the bundles is 
 concentric and the general appearance of the section is quite similar to that found 
 in the larger species of Botrychiutn. 
 
 The anatomy of the stem in the older sporophytc does not differ in an)- partic- 
 ular, except in the greater size of its parts, from that of the sporophytes already 
 described. The vascular cylinder forms a hollow tube with narrow leaf gaps above, 
 but quite solid on its ventral face. Farmer states that there is both an outer and an 
 inner endodermis and that wood is formed inside of the protoxylem, so that the 
 bundle is mesarch. 
 
 Reference has already been made to the invaginated canals communicating with 
 the stipular cavities and first described by Gwynne-Vaughn (Gwynne-Vaughn 1). 
 In connection with these he describes slight projections of tissue from the vascular 
 cylinder, which are surrounded by the endodermis. These projections he thinks 
 may perhaps be the vestiges of buds whifh formed at the axils of the leaves and 
 never developed. Whether this explanation is the correct one or not there is no 
 means of determining. Farmer found in the older plants that the roots were prevail- 
 ingly hexarch, ranging from pentarch to heptarch, and that tetrarch roots were rare. 
 
 The endodermis in the bundles of the later roots is much less distinct than it is 
 in the primary root, where it is very easily demonstrated. Within the endodermis 
 is a layer of pericycle cells, separating the endodermis from the radially arranged 
 masses of xylem and phloem. Farmer states that the apical cell of the root appears 
 triangular in section (presumably in longitudinal section), but that it is very- often 
 destroyed and as a result of the proliferation of the adjacent tissue an area of cells 
 results recalling the condition obtaining in the Marattiacea?. He also remarks that 
 most of the roots that he examined showed this abnormal condition. 
 
 I made sections of a number of roots taken from young sporophytes where the 
 root tips appeared to be (|uite normal. The result was rather interesting. In the 
 early roots there was usually present a tetrahedral apical cell, very much like that 
 of Botryrliiiini and showing a pretty regular segmentation (fig. 80, C). The seg- 
 ments cut ott from tlu apical cell are large, and even in the early roots it is pretty 
 certain that tin root cap owes its origin, in part at least, to the outer cells of the lateral 
 segments. Cross-sections of the apical cell show the same triangular form as longi- 
 tudinal sections and the divisions in the segments of the apical cell seem to be veiy 
 
108 
 
 THE OPHIOGLOSSALES 
 
 much as they are in Botrychtum. In one small root, however (fig. 80, D), the apex 
 was occupied by three cells which together formed a triangular group, but these 
 cells were of about equal size and it is not at all certain that the triangular cell x 
 was really the apical cell. 
 
 In longitudinal sections of the larger roots there is a close approximation of 
 the condition found in the Marattiaceae. There is probably still a single apical cell, 
 as there is in the early roots of the Marattiaceae, but this cell is somewhat truncate 
 below (fig. 53, B) and in a cross-section, taken from a tetrarch root, a single apical 
 cell was not clearly to be distinguished. The lateral segments are very large and 
 
 A. Section of pentarch root of Helminihoitachy 
 
 B. Part of the vascular bundle. X150. 
 
 C. Root of young sporophyte, with tetrahedral apical cell. 
 
 D. Root of an older .-sporophyte, with truncate apical cell. 
 
 there are cut off from them cells which contribute to the root cap. In short, the 
 apical growth in the later roots of Helminthostachys is evidently more like that of 
 DancBa, for example, than it is like Botrychtum. 
 
 THE SPORANGIOPHORE OF THE OPHIOGLOSS./\LES. 
 
 In all of the Ophioglossaceae the sporophyll, as we have seen, consists of two 
 parts, the fertile spike (or sporangiophore) and the sterile lamina. There is very 
 great diversity of opinion as to the real nature of the sporangiophore. The earlier 
 views were strongly influenced by the old theory that the sporangial parts were 
 
TH1-: ADULT SPOROPHYTE 109 
 
 secondary modiHcations of an originally sterile leaf. At present it is the belief of 
 most students of the ferns that the fertile structures of the sporophyll are older than 
 the sterile ones, and this has led to a quite different interpretation of the real nature 
 of the fertile segments. 
 
 Bower (Bower 9) has treated at length the different theories that have been 
 proposed to explain the nature of the fertile spike in the Ophioglossaceie. We can 
 only refer briefly to some of the more important of these. Mettenius (Mettenius 1) 
 considered that the two parts of the leaf were of equal value, but he gives no data 
 as to their origin. Later writers, e. g., Holle (Holle 1), Goebel (Goebel I), and very 
 recently Chrysler (Chrysler I), consider the fertile spike to be the equivalent of the 
 fertile pinnx' of such a fern as Aneimia. Hplle and Chrysler think that the single 
 median spike is the equivalent of two united pinn.ne. Goebel thinks that the spike 
 represents a single pinna which has arisen in a median position instead ot upon 
 the margin. Chrysler's results are based mainly upon a study of the distribution of 
 the vascular bundles in Botrychium. His results, however, do not take into account 
 the early development of the sporangiophore and its relation to the primordium 
 of the very young fertile leaf. 
 
 Bower, who has made the most complete study of the development of the spore- 
 bearing parts in the Ophioglossaceae that has yet been published, concluded in his 
 earlier work that the spike of Ophioglossum is morphologically equivalent to the 
 single sporangium of Lycopodium. In this view he was supported by Strasburger 
 and Celakovsky (Celakovsky 1). More recently, however, he has modified his view 
 (Bower 9) and concludes that the sporangial spike is a distinct organ, for which he 
 suggests the name "sporangiophore." This later interpretation we believe to be 
 the correct one. 
 
 In Ophioglossum the sporangiophore develops a peduncle which is often very 
 much longer than the sterile portion of the leaf and forms much the most conspicuous 
 portion of the sporophyll (fig. 55). The fertile portion of the sporangiophore is a 
 flattened spike along whose margins the two rows of large sporangia are borne. 
 These project very little, the cavity of the sporangium being deeply sunk in the tissue 
 of the spike and covered by several layers of cells forming the outer wall of the spor- 
 angium. The number of the sporangia in a spike may be only six or seven in some 
 of the small forms of Euophioglossiim, or it may be very large, as it is in the great 
 spikes of 0. pendulum. 
 
 In 0. pendulum and O. intermedium the spike is more flattened than it is in the 
 species of Euophtoglossum and the central sterile portion is wider in proportion. 
 Stomata are almost entirely absent from the epidermis of the spike in O. pendulum, 
 although occasionally a few are found in the central region. In O. intermedium 
 they are more numerous than in O. pendulum, but much less abundant than in the 
 species of Euophtoglossum. In the latter they may occur even upon the epidermis 
 of the wall of the sporangium. The most marked exception in the position of the 
 sporangiophore is seen in the section Cheiroglossa (plate 5). In this form there are 
 several sporangiophores developed upon the leaf. Although these are apparently mar- 
 gmal in position, it is claimed by Bower that they are really always developed upon 
 the adaxial surface of the leaf. Some of the smaller forms may show a single spike 
 which occupies a central position, very much like that of O. pendulum (plate 5, 3). 
 
 In some of the simpler forms of Botrvehinm simplex the sporangiophore is very 
 much like that of a small Ophioglossum, except that the individual sporangia are 
 more distinct, although the base of the sporangium is coherent with the margin of 
 the sporangial spike. In all of the other species of Botrychium, however, the spor- 
 angiophore branches, the degree of branching following very closely that of the 
 
no 
 
 THE OPHIOGLOSSALES 
 
 Sterile lamina. In the larger species, like B. virginianiim, many hundred sporangia 
 may be developed upon a single sporangiophore. In these larger and more special- 
 ized forms the sporangium is usually smaller and is better difFerentiated, always 
 having a more or less distinct pedicel. 
 
 The form of the sporangiophore in H ehnnithostach\s is to some extent inter- 
 mediate between that of Ophioglossum and Botrychium, but on the whole comes 
 nearer to Botrychium. The sporangia are densely crowded along the flanks of the 
 spike, thus forming two rows somewhat as in Ophioglossum, but they are very much 
 more crowded and are formed in groups or synangia which project free above the 
 surface of the young spike. The sporangia do not arise singly, but are borne upon 
 short branches or secondary sporangiophores which suggest the lateral branches of 
 the sporangiophore in Botrychium, but are far less regularly disposed. By the fur- 
 ther growth of the sporangia the axis of the spike is entirely concealed and a casual 
 examination would indicate that the sporangia were formed equally at all points of 
 its surface, but a study of the development shows that they are lateral in structure, 
 as they are in Ophioglossum. 
 
 THE DEVELOPMENT OF THE SPORANGIOPHORE. 
 
 The sporangiophore in both Ophioglossum and Botrychium becomes visible at 
 a very early period. In Ophioglossum moluccainnn I have found that the spor- 
 angiophore can be recognized even earlier than Bower has stated for 0. vulgatum 
 and O. rcticulatum. A median section of a very young fertile leaf in 0. moluccaiium 
 is shown in fig. 56, D. The apex of the leaf has scarcely become free and the apex, 
 
 as well as nearly the whole of the 
 adaxial surface, is made up of 
 large columnar meristem cells, of 
 which it is difficult to say that 
 any one is the single apical cell 
 of the leaf. The leaf trace is al- 
 ready evident, passing into the 
 base of the young leaf, which is 
 strongly inclined forward so that 
 the real apex of the leaf is not 
 directed vertically upward, but is 
 on the adaxial side of the leaf 
 rudiment. The apex of the young 
 sporangiophore arises from the 
 tissue immediately below the apex 
 of the sterile segment, and the 
 whole of the adaxial portion of 
 the leaf below the apex of the 
 young sporangiophore may be said 
 to belong to the latter and not 
 to the sterile portion of the leaf. 
 There is thus practically a dichot- 
 omy of the apex of the young 
 sporophyll, and the two branches (viz, the sterile segment and the sporangiophore) 
 are structures of equal rank. It would thus appear that the old view put forward 
 by Mettenius of two equal branches of the fertile leaf is probably the correct one. 
 It is probable that younger stages than those figured by Bower for O. vulgatum and 
 0. reticulatum would show the same condition of affairs that we pointed out in 
 
 A. Longitudinal section of young sporophyll of Ofihioglos. 
 
 still inclosed in the stipular sheath, sh X50. 
 
 B. Apex of sterile lamina, sp, sporangiophore. 
 
 C. .^pel of sporangiophore. X125. 
 
 D. Transverse section of a very young sporophyll, showing di< 
 
 apex. X125. 
 
THK ADULT 
 
 11] 
 
 O. tnoliiiirinuin and O. />riuli/li<ni, and flu- conclusion that the young sporangiophore 
 is an outgrowth of the sterile part of the leaf is the result of the subsequent rapid 
 growth of the upper sterile portion of the leaf, which thus carries its apex far beyond 
 the apex of the young spike. 
 
 The young sporophyll in O. />rn Julian (hg. 64, K) resembles that of 0. moluc- 
 canum, except that all the parts are much larger. Ihe apex of the young sporangio- 
 phore appears upon the adaxial side of the very broadly conical young sporophyll. 
 
 Bruchmann's studies on Botrxchium lunaria indicate that in this species the 
 separation of the fertile and sterile segments of the sporophyll is the result of a 
 diciiotomv of the apex of the sporophvli at a very early stage in its development. 
 
 in Opluoolosstim 
 ic ha\e not been 
 
 A. Young sporophyll of Ophiogloimm prmliilum. sp, llic sporangiopliorc 
 
 B. An older sporophyll. Xi. 
 
 C. A still older stage, showing venation of lamina. 
 
 D. Base of a full-grown sporangiophore. 
 K. Sporophyll with rudimentary lamina, /. 
 
 The sporoph}ll, however, is larger at this time than is the case 
 moluccamtni (Bruchmann 2, fig. 57). 
 
 The earliest stages in the development of the sporangioph( 
 studied in Hclninithostachys. Bower says (Bower 9, page 455): 
 
 "The origin 
 unlike. It appe; 
 frond, and it is 
 downward." 
 
 In the latter respect it might ver\- well be compared to the stronglv bent-ovi 
 sporangiophore of Botrychiiini virgiiiiaiiinn. 
 
 )f it is similiar (i. e., to Ophtoglossum) and its early stages not 
 rs at first as an outgrowth on the adaxial side of the sterile 
 urved over while youn<i so that the actual apex is pointed 
 
112 THE OPHIOGLOSSALES 
 
 After the separation ot the fertile and sterile portions of the sporophyll the 
 sterile segment at first grows much more rapidly than the sporangiophore and soon 
 extends beyond it, so that the sporangiophore has the appearance of" being merely 
 an adaxial appendage of the sporophyll. These changes in the relative importance 
 of the points can be very well shown by a study of the very large sporophylls of 0. 
 pendulum. In this species the young leaf emerges from the sheath while it is still in 
 a very early stage of development. These young leaves (fig. 82) have a very thick, 
 fleshy leaf base, ending in the small pointed lamina, bent over and almost hiding the 
 young sporangiophore which is attached to its inner surface. This curved-over 
 form of the young sporophyll differs from that oi Eiiophioghssum^ where the young 
 leaf is straight, and in this respect there is a suggestion of Botrychiiim or of the 
 true ferns. The young spike at this stage is almost equal in length to the sterile 
 lamina. 
 
 If the very young sporophyll of Ophioglossum mohiccatium is examined, it will 
 be seen that the vascular bundle leaving it passes downward and joins a young root 
 trace. Very soon after the bundle enters the young leaf it divides into two branches, 
 one of which passes into the base of the spike and the other and larger one into the 
 sterile section of the sporophyll. This latter bundle now divides again into two, 
 one of which passes to the dorsal side of the sterile segment while the other is turned 
 toward the ventral side of the leaf, which probably represents tissue belonging to the 
 sporangiophore. In the cross-sections of older leaves the petiole, which includes 
 the completely fused base or peduncle of the sporangiophore, shows three distinct 
 bundles, a large abaxial one and two smaller adaxial bundles. The former remains 
 undivided for some distance above the base of the insertion of the peduncle of the 
 sporangiophore. The two adaxial bundles divide into several branches, this begin- 
 ning just below the point where the lamina of the sterile leaf becomes differentiated 
 from the petiole. The lamina has its edges infolded so that it has a conical form, 
 open in front and inclosing the young spike (figs. 56, 57). The successive branching 
 of the two adaxial bundles follows in quick succession, and the outer bundles con- 
 tribute to the veins of the lamina; the inner ones pass upward into the peduncle 
 of the spike. 
 
 In a previous paper (Campbell 7) the conclusion was reached from the study of 
 a series of sections at different heights, taken from the full-grown sporophyll, that 
 the two adaxial bundles belonged exclusively to the spike. A study of the devel- 
 opment of the young sporophyll shows that this statement is not entirely correct 
 and that a considerable part of the vascular system of the sterile lamina owes its 
 origin also to branches from the two adaxial bundles before the latter enter the 
 peduncle of the sporangiophore. Of the two adaxial bundles, one represents the 
 original bundle of the sporangiophore derived from the primary forking of the leaf 
 trace; the other is derived secondarily from a subsequent division of^ the second 
 branch of the leaf trace, this also giving rise to the large abaxial bundle of the petiole. 
 
 THE DEVELOPMENT OF THE SPORANGIUM. 
 
 Much the most complete account of the development of the sporangium of the 
 Ophioglossace^e is that of Bower (Bower 5), who has made a very exhaustive exam- 
 ination of the development of the sporangium in all three genera. My own studies 
 have been for the most part confined to Ophioglossum pendulum, O. moluccanum, 
 and Botrychium vtrginiariurn. Burlingame (Burlingame 1) has more recently 
 examined the development of the sporangium in a species of Ophioglossum from 
 the Philippines, which was assumed to be O. reticulatum. Beers (Beers 1) has- 
 also published recently a somewhat incomplete account of the sporangium in Hel- 
 
THK ADULT SI'OROl'H "> ll-. 113 
 
 nnlwstaclixs. I'lin.iiuli the kiiuiiKss of Prof. 1,. L. Biulingame 1 havi- also had 
 soiiif of his |->ie|iaiations ot OpIiKJi^lossitm nin iil/iturn and 
 
 an opportunity to exanmu- 
 Helmtnthostach ys. 
 
 An examination of hotli lonijitiidinal and cross sections ot the \()un<i s|)ike in 
 (). ttiohiirtiinttn (and this is true also of O. piiululum) shows that they agree entirely 
 account of the de\elopnient in the species studied by him. There 
 
 ith H< 
 
 .A. 'I'ransvcrse i.cction of a very young sporangiophore of Ophioglos!um moluccnnum. X 1 50. 
 
 B. Section of an older sporangicphore. X75. The shaded areas mark the " sporangiogenit " hands. 
 
 C. Part of B, more highly magnified. The nucleated cells mark the sporangial region. 
 
 seems to be no question that, as Bower first pointed out, there is formed a continuous 
 band of tissue on each side of the sporangiophore, this band being the so-called 
 "sporangiogenic" band, from which the individual sporangia are differentiated 
 later. The sporangia arise from this sporangiogenic band at more or less definite 
 intervals,| these fertile areas being separated by bands of sterile tissue. In the 
 sporangial areas periclinal walls are formed by which an inner mass of tissue, the 
 
 Development of the sporangium in Opiiioglassum prnJu< 
 spt sporogenous tissue. A is a longitudinal section; 
 mature sporangia. X8. 
 
 archesporium, is separated from the outer cells which are to form the wall of the 
 future sporangium. The inner cells constituting the archesporium later give rise 
 to the masses of spores. Between the young sporangia there lie partitions or septa 
 formed from the intei-vening sterile cells of the sporangiogenic band. The cell 
 groups which form the archesporia and the sterile septa are derived from sister cells 
 
114 
 
 THE OPHIOGLOSSALES 
 
 of the original sporangiogenic bands. Bower thinks that all of the sporogenous 
 tissue can not be traced back to a single primary archesporial cell, and that spore- 
 bearing tissue is formed secondarily by periclinal divisions in the cells outside 
 of the original archesporium. In some of the very young sporangiophores of O. 
 pendulum there can sometimes be seen a single large epidermal cell which may 
 possibly be the mother cell of a young sporangium, but it is very difficult to determine 
 this point satisfactorily. 
 
 After the archesporium is differentiated there is a rapid division in its cells and 
 there is thus formed a very large mass of cells, whose limits, however, are not always 
 very clearly marked. Later on the contents of these cells become denser and are 
 more easily distinguished from the surrounding sterile tissue. 
 
 The cells lying outside the archesporium divide rapidly both by longitudinal 
 and transverse walls and give rise to the thick outer wall of the sporangium. In lon- 
 gitudinal sections through the sporangium two rows of cells may be seen extending 
 from the mass of archesporial cells to the outside of the sporangium. These two 
 rows of cells mark the point where the transverse cleft arises by which the spo- 
 rangium opens at maturity. The outer cells derived from the archesporial complex 
 
 A, B. Young sporangia; C, an older sporangium of Boirychwm vugummu 
 
 tissue is shaded; the nucleated cells adjoining form the tapetum. 
 
 do not develop into spores, but constitute the tapetum, which later becomes dis- 
 organized and forms a sort of plasmodium extending among the growing spore 
 mother cells, and is, no doubt, of great importance in the further development of 
 these. Bower thought that some of the inner cells of the archesporial tissue contrib- 
 uted to this Plasmodium, but a further study has led him to the conclusion that this 
 is not the case and that all of the inner cells of the archesporium develop into spores. 
 This view is confirmed by the recent paper of Burlingame (Burlingame I), who 
 investigated the development of the spores in O. reticulatum. 
 
 At maturity the sporangium opens by a transverse cleft whose position is already 
 evident in the younger stages of the sporangium. As the cells shrink with the drying 
 of the ripe sporangium the spores are crowded out through this cleft, but there is 
 no special mechanism like the annulus found in the sporangia of the higher ferns 
 which facilitate the dispersal of the spores. 
 
 Our knowledge of the development of the sporangium of Botrychium has been 
 based largely upon the study of 5. lunaria. I have investigated with some care the 
 development in B. virginianum, which differs mainly from B. lunaria in the smaller 
 
115 
 
 rh;ir it has a wull-nuirkfd thougli sliort indicil. 
 7/, siuh as B. siriiplf.x, rht- sporangia are nuich 
 
 ■lopnifiit ot the indiviclual 
 
 ingia hegin; 
 
 si/e of the spoiangiimi aiul the tan 
 In the simplest forms of Botryrhni 
 larger and are sessile. 
 
 In B. virginiatiurn (rig. 85) the 
 just about a year previous to their ripening, and if the plants are taken up about the 
 time the spores are shed the earliest stages of the sporangia will be found in the leaf 
 which is to expand the following season. At this time the sporangiophores in the 
 larger specimens are thrice pinnate, and the youngest recognizable sporangia are 
 borne at the tips of the branches. These young sporangia form slight elevations 
 which become smaller as they approach the tip of the segments, and if an exact 
 median section is made of one of these young sporangia it will be seen to have at its 
 apex a large pyramidal cell with a truncate base. Holtzman (Holtzman 1), states 
 that the whole sporangium may be traced back to a single cell and that the divisions 
 at rirst are like those of a three-sided apical cell. I have not been able to satisfy 
 myself as to the accuracy of this statement, but the youngest stages which I have 
 been able to find would not forbid such an interpretation, although there seems no 
 question that the basal part of the sporangium is derived, in part at least, from the 
 surrounding tissue. 
 
 Fig. S6. 
 Sporangiophores of Helmwihosiachys. The two lower figures seen from above. /, sterile appendages. X3. 
 Section of a young sporangium, showing sporogenous tissue, sp; tapetum, /; and wall. rr. (From preparation 
 made by Prof. L. L. Burltngame.) 
 
 . Section of a young spora__„ , ,__ 
 
 lade by Prof. L. L. Burltngame.) 
 
 The terminal cell of the sporangium is divided b\ a periclinal wall into a single 
 inner cell which forms the archesporium, and an outer one which contributes to the 
 wall of the sporangium. The outer cell rapidly divides and similar divisions occur 
 in the archesporium. H\ active growth in the basal part of the sporangium it pro- 
 jects more and more until it assumes the form of a projecting globular body with a 
 short stalk, thus diffViing stiikingK fiom the deeply sunken sporangia of Ophio- 
 olcissiirn. It is impossible to detect an\ difinire succession of divisions in the arche- 
 sporium, which ultimateh- becomes changed into a large globular mass of cells with 
 the usual dense contents, separated from the outside of the sporangium by half a 
 dozen or more layers of sterile cells. The cells immediately adjoining the sporo- 
 genous tissue constitute the tapetum. Bower states that the tapetum is exclusively 
 derived from this outer sterile tissue, but it must be said that, in Botrycluutn vir- 
 giniamim at least, the limits between the sporogenous tissue and the tapetum are by 
 no means easy to detect. The wall of the ripe sporangium has tVom four to six 
 layers of cells and the dehiscence is by a vertical cleft, whose position may sometimes 
 
116 THE OPHIOGLOSSALES 
 
 be recognized in the young sporangium by cells arranged very much as they are in 
 the sporangium oi Ophioglossiim. The stalk is traversed by a short vascular bundle 
 which can first be recognized v^^hen the sporogenous cells are pretty well advanced 
 in their development. 
 
 The sporangia of Helmintlwstachys on the whole more resemble those of 
 Botrvchium than those of Ophioglossurn. In the very young sporangiophore there 
 is developed a band of tissue on either side much like that in Ophioglossurn (see 
 Bower 9, page 455). Transverse sections of the fertile spike show that the spo- 
 rangiophores originate from certain groups of cells of this sporangiogenic band. 
 These young sporangiophores, however, project above the surface and later form 
 crowded projections which may undergo a greater or less degree of branching, so 
 that the number of sporangia borne upon a single one of these secondary sporangio- 
 phores may be considerable. These sporangiophores become irregularly lobed and 
 the final divisions of these develop into the sporangia in a manner which is very 
 similar indeed to that of Botrvchium, but the individual sporangia are usually less 
 distinct. Bower's account of the further development may be quoted as follows: 
 
 "It has already been noted that the position and number of the sporangia 
 which they bear are inconstant. In earl}' stages it is impossible to distinguish 
 the cells which will give rise to the sporangia, but from rather older stages it ap- 
 pears that the sporogenous groups, together with the superficial cells which cover 
 them are referable in origin to the segmentation of a single superficial cell. More- 
 over, the first periclinal division of the cells defines the whole of the sporogenous 
 tissue from the protective wall. As the sporangia grow older the}' project from 
 the surface of the sporangiophore. The sporogenous mass increases rapidly in 
 bulk, while the cells surrounding the sporogenous mass to the extent of several 
 layers assume the character of a tapetum which gradually becomes disorganized; 
 finall)' the sporogenous cells separate and divide into tetrads. As the sporangia 
 approach maturity the upper part of the sporangiophore may grow out into an 
 irregular rosette of laciniae of vegetative tissue." 
 
 A study of the older sporangia oi Hehninthostachys shows that the tapetum is 
 extraordinarily developed, constituting a very thick layer surrounding the mass of 
 spores (fig. 86). The clusters of fully developed sporangia are often very regular 
 in form (fig. 86, A), and remotely suggest the synangium of a Marattiaceous fern. 
 
Part II. THE MARATTIALES. 
 
 The members of the second order of the Eusporangiatae — the Marattiales — 
 resemble quite closely the typical Leptosporangiatae. Both in the form and venation 
 of the leaves and in the development of the sporangia upon their lower surface they 
 recall the common ferns. But the structure of the sporangium is very different, 
 and both in its structure and development the sporangium shows certain evident 
 resemblances to that of the Ophioglossaceae. 
 
 The Marattiaceous type is a very old one, and this fact lends a special interest 
 to these few survivors of the ancient Paleozoic fern flora. Like the Ophioglossaceae, 
 the leaves of the Marattiace;c are thick and fleshy in texture, but usually firmer than 
 is the case in the Ophioglossaccae. Some of the smaller species oi Dana-a are only 
 2 or 3 decimeters in height, but the larger species of Marattia and Angiopteris are 
 among the largest of the ferns. The latter, which is common in the eastern tropics 
 and e.xtends to Australasia, is a noble fern whose great caudex, covered with the 
 persistent leaf bases, is nearly as big as a barrel, and the gigantic leaves reach a 
 length off or 6 meters, with stalks almost as thick as a man's arm. In the dense, 
 wet forests of Tjibodas in Java, for example, the development of this fern is espe- 
 cially luxuriant, and groups of them form one of the most striking features of this 
 rich, tropical flora. 
 
 The Marattiacea? frequent, for the most part, the moist, shady forests of the 
 tropics, where they form a characteristic feature of the vegetation. Along the steep 
 banks of moist, shady ravines or streams they sometimes occur in profusion. 
 
 The number of living Marattiaceae'-' is small, some authorities recognizing 
 only about thirty species. There is, however, a good deal of difference of opinion 
 as to the number of species in some of the genera. This is especially true of 
 Angiopteris, which many authorities consider to have only one single, extremely 
 variable species, while others recognize many species. One of the genera, Marattia, 
 is widespread throughout the tropics of both hemispheres, where there are about 25 
 species. (See Bitter I, Christensen 1.) One of these, M. iloiiglasii, extends to the 
 Hawaiian Islands, where it is a common and conspicuous fern. Angtoptcris is 
 widespread throughout the tropics of the old world and reaches beyond the tropics 
 into Australia and southern japan. Archangiopteris, with a single species, A. 
 henryi, is a recently discovered form occurring in southwestern China. Kaidfussta, 
 also a monotypic species, occurs throughout the Indo-Malayan region and extends 
 as far as the Philippines. The genus Danwa is exclusively neo-tropical and com- 
 prises about 20 to 25 species, most of which occur in the West Indies, northern 
 South America, Central America, and Mexico. A sixth genus, Macroglossinii, from 
 Borneo, has recently been described (Copeland 1). 
 
 Of these, Kaiilfussia is the most aberrant, differing markedly from the other 
 genera in the form and venation of the leaves, as well as in the shape and position 
 of the synangia. Kaiilfussia and most species of Daiicra have dorsiventral rhizomes, 
 while the other genera have an upright, radially-constructed caudex. 
 
 The gametophyte, in all of the investigated species, is a large, dark-green, 
 fleshy thallus, much resembling superficially such a liverwort as Prllia. 1 he re- 
 productive organs are very much like those of Ophioglossum, and chlorophyll is 
 formed in abundance in the gametophyte. The large liverwort-like gametophyte 
 
 * Christensen (I) recogniics 6i species of Angiopteris, i of Archangiopteris, 26 of Danata, i of Kaulfussia, and z8 of Marattia. 
 
 117 
 
118 THK MARATTIALES 
 
 and the character of the reproductive organs are marked indications of the primitive 
 nature of these ancient ferns. The embryo also shows many points in common with 
 that of the Ophioglossaceae. As in the latter, the organs of the young sporophyte 
 are all of epibasal origin and the vascular system in the young sporophyte belongs 
 exclusively to the leaf and root, the stem being entirely destitute of a proper stele. 
 
 In Dancsa a short suspensor is usually developed, but the other genera have as 
 yet shown no examples of this. 
 
 The sporophyte is very much alike in its early development in all of the Marat- 
 tiaceae. The young plant consists at first mainly of the primary leaf and root, which 
 are traversed by a single axial vascular strand as in Oplnoglosstini mohiccanum, but 
 a stem apex is developed at an early period, although it remains relatively incon- 
 spicuous. 
 
 Kaiilfussia, Danaa, and ArcJiangtopteris are all ferns of moderate size and 
 comparatively simple structure ; but Marrattia, Angiopteris, and Macroglossum 
 are very large, and the arrangement of the vascular bundles in the stem becomes 
 exceedingly complex, corresponding to the numerous bundles in the enormously 
 developed leaves. 
 
 The form of the leaf in the Marattiace?e ranges from a perfectly simple leaf in 
 Dancea simpJicifoUa to the gigantic decompound leaves oi Angiopteris and Marattia. 
 Most of the species oi Datura (and this is true also oi Archangiopteris and Macro- 
 glossum) have simply pinnate leaves, while in Kaulfussia the leaves are palmately 
 divided into from three to seven leaflets, the larger leaves looking curiously like a 
 horse-chestnut leaf in outline, whence the specific name. The ternate form, which 
 is often found in Kaulfussia, recurs in the early leaves of the other genera and is 
 sometimes retained in quite large leaves in Marattia (plate 12, B). This recurrence 
 of the ternate leaf form suggests Helminthostachys and the ternate species of 
 Botrvtliiiim. The leaves are usually quite smooth, but in the earlier stages there is 
 sometimes a sparing development of hairs and scales. The latter are especially- 
 noticeable upon the leaves of the young sporophyte in Dancea. 
 
 The leaves are furnished at the base with very conspicuous fleshy stipules 
 which remain adhering to the stem after the leaves fall away, and these leaf bases, 
 with their attached stipules, more or less completely cover the surface of the stem. 
 As the leaves fall away they leave a characteristic scar marked by the remains of the 
 vascular bundles. The leaf base as well as the stalks of the leaflets shows a more or 
 less marked enlargement, recalling the pulvinus which occurs so commonly in the 
 Leguminosae. It is at this point that the leaf stalk separates, the smaller divisions 
 of the leaf often breaking away from the main or secondary rachis, in the same 
 fashion as the main leaf stalk falls. In the large species of Marattia and Angiop- 
 teris this enlarged leaf base with the two thick, fleshy stipules curiously resembles 
 in shape and size the hoof of a horse. 
 
 The leaves when young are coiled up in the same fashion as those of the typical 
 ferns and their venation in general is decidedly fern-like, the ultimate veins being 
 dichotomously branched and the venation very much like that of Helniinthostachys 
 or certain species of Botrychiuni. In Kaulfussia, however, the venation is reticulate 
 and strikingly like that of the typical Dicotyledons. In its earlier stages, however, 
 there is a marked resemblance to the venation of Ophioglossum. 
 
 The fleshy leaves do not, as a rule, show a very large development of mechanical 
 tissues; but there is developed, especially in the larger leaves, a thick layer of hypo- 
 dermal tissue, which is usually sclerenchyma, somewhat like that of the typical ferns, 
 but may be thick-angled tissue or collenchyma, such as is common in many flow- 
 ering plants. A marked anatomical feature is the presence of large mucilage ducts. 
 
THK c;amktophvtf 119 
 
 I. I UK GAMK'rOPHVri;. 
 
 The first account of the germination and development of the gametophyte 
 in the Marattiaceae was published by Luerssen (Luerssen 2), who studied the 
 germination in Marattia cicuta-foUa and in Angtopterts. Not long afterward the 
 Dutch botanist, Jonkmann, published an account of the development of the prothal- 
 lium in both Marattia and Atigioptens. His original paper (published in Dutch, 
 but afterward translated into French) gives an extremely satisfactory account of the 
 germination and development of the prothallium and reproductive organs (Jonk- 
 mann 1). Somewhat later he also published a preliminary account of the germina- 
 tion in Kaulfussia, but apparently the work was never published in full (Jonkmann 
 2). In 1892 Farmer (Farmer I) described the gametophyte and embryo in Angi- 
 opteris evecta and in 1894 the writer (Campbell 3) gave an account of the prothallium 
 and embryo in Marattia donglasii collected in Hawaii. Two years later Brebner 
 (Brebner 2) described the prothallium and embryo in Daticea simplicifolta. In 
 1908 the writer published an account of the prothallium and young sporophyte of 
 Kaulfussia (Campbell 9) and in 1909 a preliminary note was published in regard 
 to Daiia-a (Campbell, 10). 
 
 The following account of the germination is based mainly upon the work of 
 Jonkmann. I have, however, examined the early germination stages in Marattia 
 douglasii for comparison with Jonkmann's account, and, so far as my experiments 
 went, it agrees entirely with the species described by Jonkmann. 
 
 The ripe spores of the Marattiaceae are small and may be either of the bilateral 
 type or tetrahedral (radial). According to Jonkmann, the bilateral spores are much 
 more abundant in Marattia than the radial spores, but in Aiigiopteris the radial 
 spores predominate. The wall of the spore (see Jonkmann I, pp. 203, 204) shows a 
 differentiation into an inner membrane orendospore and a middle layer, the exospore, 
 which is often found divided into layers. There is a very thin external coat, the 
 perispore or epispore, which is generally thrown off in the early stages of germination 
 or even before germination begins. All of the membranes except the epispore are 
 colorless, while the latter is a more or less pronounced yellowish brown tint. The 
 surface of the spore is roughened by small papillae, which arise from the exospore. 
 The spores contain no chlorophyll, but there is a considerable amount of oil present, 
 which appears as drops of varying size, and there are also other granular contents — 
 starch and albuminous granules. The nucleus lies in the center of the spore and 
 is connected with the peripheral protoplasm by delicate protoplasmic filaments. 
 Germination begins quite promptly under favorable conditions and within about 
 a week the spores, which hitherto were quite colorless, begin to show a greenish 
 tint, due to the development of chlorophyll. Jonkmann states that the chlorophyll 
 appears first as flocculent masses near the nucleus, but these apparently amorphous 
 masses are really composed of very small, faintly tinted chromatophores, which lie 
 between the large oil drops and rapidly increase in size and depth of color as 
 germination proceeds, their number increasing by the usual division. The chloro- 
 plasts later become very conspicuous and are distributed in the periphery of thenow^ 
 very much enlarged spore, the outer membranes of which are ruptured so as to 
 expose the endospore, containing the nucleus and numerous large and conspicuous 
 chloroplasts. Starch granules are also to be seen in most cases. 
 
 The cell remains undivided until it has attained a size many times greater than 
 that of the spore. The first division wall, which is formed about a month after the 
 spores are sown, is transverse both in Angtopterts and Marattia, and, like that in 
 the germinating spore of Ophioglossum, divides the primary cell into two nearly 
 
120 THE MARATTIALES 
 
 equal parts (fig. 87, A). A rhizoid may be cut oft' either before or after this first 
 transverse wall is developed, but frequently no rhizoids are formed until a much 
 later period, as in Ophioglossum. The primary rhizoid, when present, is formed 
 much as in the typical ferns, the papilla from which it develops being cut off" from 
 the larger cell, and it contains little or no chlorophyll. F^ach of the primary prothal- 
 lial cells divides, in typical cases, by a longitudinal wall, so that the young gameto- 
 phyte consists of four cells, arising quadrant-wise (fig. 88, A), and closely resembling 
 corresponding stages in Ophioglossum, except for the absence of chlorophyll in the 
 latter. Where the young plants are crowded or light is deficient, as for instance 
 when the germination occurs within the sporangium, there is a tendency to the 
 development of a filament, a phenomenon often met with also in the typical ferns. 
 Usually one of the upper pair of cells in the four-celled stage assumes the role 
 of an apical cell, and for some time, as in the typical ferns, there is growth from a 
 two-sided apical cell (fig. 88, C). As soon as the apical cell is established, it grows 
 
 A. Two germinating spores of Mar. 
 
 B. Young gametophyte of same spe 
 
 C. Marattia sambucina Blume. X 
 D-H. M. douglmii Baker. D-F, 
 D and E represent the same prothall 
 
 i jraxinea Smith. X200. The rem 
 . X75. (.^, B, after Jonkmann). 
 
 very much as it does in such a liverwort as Aneiira and produces a thallus of the 
 same form and structure. But as the prothallium grows older a periclinal wall 
 is formed in the apical cell, and in the outermost of the two cells thus produced 
 there is a longitudinal wall dividipg it into two equal cells, and from this time on 
 it is impossible to recognize a single apical cell in the prothallium, the apex of which 
 is occupied by a group of apparently similar marginal initial cells. 
 
 At first the prothallium has a spatulate form, but before the single apical cell 
 is replaced by the group of marginal initials the outer cells of the younger segments 
 grow more rapidly than the inner ones, so that they project beyond the apical cell, 
 which thus comes to lie in a depression between the two lobes, and the familiar 
 heart-shaped form so commonly found in the prothallium of most ferns is established. 
 The marginal initial cells vary in number with the width of the depression in which 
 they lie. In a horizontal section they appear oblong in form, but in the vertical 
 sections made they have a semicircular outline (fig. 88, D, E). 
 
THE GAMETOI'HYTI 
 
 121 
 
 liasal segments are cutoff by a wall extending the wliolc depth of the piothalliuni 
 and the segment thus cut off is divided at once by a horizontal wall into a dorsal and 
 ventral cell of nearly equal size (fig. 88, E, d, v). Cell divisions are more active in the 
 ventral segments, more manifestly so at some distance back from the apex. It is due 
 to this more active cell division in the ventral segments that the strongly projecting 
 cushion-like mass of tissue is formed upon the ventral surface of the prothallium, 
 upon which the archegonia later make their appearance. The superficial cells of 
 both surfaces of the prothallium have a thick cuticle which often makes it difficult 
 to embed the prothallia without bad shrinkage. From the under side of the prothal- 
 lium nimierous rhizoids are developed, which in the case o{ Marattia and Angiopieris 
 are unicellular and thin-walled, but in DiuKra become divided into several cells. 
 Sometimes there seems to be no definite apical growth in the early stages, but on the 
 other hand Jonkmann states that both of the superior cells may function as apical 
 cells, thus inaugurating the early dichotomy of the young prothallium, and even a 
 third branch may arise from one of the inferior (]uadrants, which assumes the 
 character of a third apical cell. 
 
 Among the leptosporangiate ferns, the forms which most nearly resemble the 
 Marattiacese in the development of the gametophyte are the Osmundaces, especially 
 
 the genus Osmunda (Campbell, Mosses and Ferns, 2d ed., pp. 347, 348). In 
 Osmunda there is often a tendency also to the formation of a massive structure at an 
 early stage in the development, due to the formation of cell divisions in three planes, 
 so that the prothallium becomes from the first more than one cell thick. This 
 tendency in the Marattiaceae becomes much more pronounced as the gametophyte 
 grows and very soon there is evident a very thick midrib, which often becomes 
 exceedingly conspicuous in the older gametophyte, unlike the typical ferns, where 
 the thickening of the prothallium is confined to the region which bears the arche- 
 gonia. In the Marattiace;e this thickening extends almost to the margin of the 
 prothallium, so that it is only at the extreme edge that the prothallium shows but a 
 single layer of cells. The very old prothallia in Marattio branch dichotomously 
 (fig. 87, E, E), and the process is entirely similar to that found in manv thallose 
 liverworts. The original growing point becomes divided by the development of a 
 median lobe, thus inaugurating two growing points, and the midrib hack of tin- 
 growing point forks in exactly the same way as in many liverworts. 
 
 Besides this dichotomous branching, it is not at all uncommon to have adven- 
 titious buds developed upon the margin of the thallus. These form small secondary 
 
122 THE MARATTIALFS 
 
 piothallia, which may be detached and become independent plants; or reproduc- 
 tive organs may be developed upon them, usually only antheridia, while they are 
 still connected with the mother plant. The prothallia are very long-lived and 
 apparently may grow indefinitely so long as the archegonia are not fecundated. I 
 kept prothallia of Marattia doiiglasii for nearly two years, during which time they 
 grew vigorously and finally reached a length of over 2 centimeters. At the end of 
 two years there was no indication of the slightest decrease in their vigor. 
 
 The prothallia are monoecious, although it is not uncommon to find small 
 prothallia which bear only antheridia. They are very deep green in color, which 
 together with their more massive texture makes them easily distinguishable from the 
 prothallia of the ordinary ferns, and they very often look more like such a liverwort 
 as Pellta or Anthoceros. The prothallium of A ngtopteris very closely resembles that 
 of Marattia, but is often somewhat shorter. The difference in shape in the prothallia 
 
 of the form studied by [onkmann (//. 
 pruuiosa var. hypoleiica) and the Cey- 
 lonese form studied by Farmer and my- 
 self may be taken as an argument in favor 
 of recognizing a specific difference between 
 these two forms. 
 
 The archegonia in both Marattia and 
 _ Angiopteris, so far as my own observations 
 
 Fig. 89. extend, are confined to the lower surface 
 
 Two gametophytes of Angiopieris with young sporophyte of the midrib. Tonkmann, howcver, States 
 
 attached, em, embryo; cot, cotyledon. X^. , • ii i i i i 
 
 that occasionally he observed archegonia 
 developed upon the upper surface as well. The antheridia, while more abundant 
 upon the lower surface of the prothallium, are quite commonly met with also upon 
 the upper surface and are not restricted to the midrib, but may be found quite near 
 to the margin. The rhizoids in both Angiopteris and Marattia are probably always 
 unicellular, and their walls, which are quite strongly cutinized, are dark brown in 
 color. 
 
 THE PROTHALLIUM OF KAULFUSSIA.* 
 
 The prothallia of Kaulfiissia are usually much larger than those of either 
 Marattia or Angiopteris. They are very massive, strongly resembling a Pellia or 
 Aneura. The specimens described here were collected in Java in a small ravine 
 near the foot of the volcanic mountain Salak. The youngest prothallia found 
 were about 5 millimeters in length and, like the older ones, were decidedly elon- 
 gated, with a deep sinus in front (fig. 90, A, B). There were a few antheridia 
 occupying the forward part of the thick midrib, which is very largely developed 
 in Kaulfiissia, as it is in the other Marattiaceae. The older prothallia are relatively 
 somewhat broader and these larger ones usually bear archegonia. Only in a 
 few cases were young antheridia found upon the prothallia with the archegonia. 
 Whether this is always true could not be decided from the small number of prothallia 
 found. The antheridia in Kaulfiissia seem to be strictly confined to the lower sur- 
 face of the midrib and occupy much the same position that the archegonia do. 
 After the antheridia have matured and discharged the speimatozoids, archegonia 
 arise in the same position in regard to the apex. Unlike most ferns, the walls of the 
 empty antheridium do not become discolored, so that they are easily overlooked. 
 Careful examination of the sections of the older prothallia, however, will almost 
 always show the empty antheridia and it is probable that the prothallia are usually 
 
 ' Kauljussia Blume = Ch 
 
THE GAMF.TOPHYTF 
 
 123 
 
 pioteiandrous aiul not did'cioiis. as might be siippostd from a casual examination. 
 I'he margins of the large prothallium are more or less irregularly lobed and it is not 
 unlike that of Osnunuhi, but as in the other Marattiace;c the wings of the prothal- 
 lium are several cells in thickness near the midrib, and only at the extreme edge 
 are they reduced to a single cell in thickness. 
 
 rhe full-grown prothallia are a centimeter or more in length and nearly as 
 bioad. One very large specimen was found (fig. 90, /■ ). 1 his b;)re a young sporo- 
 phyte with two fully developed leaves, and the prothallium measured nearly 2.5 
 centimeters in length by 1.75 in extreme width, and was also very thick. A second 
 archegonial cushion was present, but whether this was due to a forking of the original 
 apex, such as not uncommonly occurs in Marattia and Angiopteris, was not deter- 
 mined, although this was very probably the case. 
 
 B. Apex of A, showing the anthcridia, f^ X20. 
 
 C. An older gametophyte, with archegonia, V. X4 
 
 D. Apical region of an older gametophyte; a-, apical cell. 
 
 E. An older gametophyte, with young sporophyte attached. 
 
 ityledon; 
 
 F. Two views of a large gametophyte, with attached sporophyte, natural size. 
 
 G. Rhizoid. X50. 
 
 H. Apex of rhizoid, more highly magnified. A single nucleus only is present. 
 
 The rhizoids are stout and they have thick but quite colorless walls. They 
 o'^ten show apparently transverse septa and the extremity is not infrequently forked 
 (fig. go, H). An examination of these rhizoids shows that these septa are not the 
 result of true cell division, as only one nucleus can be found in the whole rhizoid. 
 As the nucleus is large and conspicuous, and only one can be seen, it looks as if the 
 formation of septa is secondary and not connected with cell division. The structure 
 of the apex of the prothallium is exactly as in Marottia and Augioptens. The 
 apical meristem shows the usual row of marginal initial cells, of which one (fig. 90, 
 D, v) is often somewhat larger than the others and may possibly represent a single 
 apical cell. 
 
124 
 
 THE MARATTIALES 
 
 THE PROTHALLIUM OF DAN^.A. 
 
 In July, 1908, the prothallia of three species oiDancea were collected in Jamaica. 
 As the genus Damea has received comparatively little attention, especially as regards 
 the gametophyte, it seemed very desirable to secure as complete a series of the pro- 
 thallia and young sporophytes as could be done, and to this end the trip was made 
 to the West Indies. The genus Dancea is confined to the American tropics and 
 comprises, according to Christensen, 26 species of extremely characteristic ferns. 
 The type is evidently an old one, as some of the fossil Marattiaceae are closely allied 
 to the living genus Dana-a and may possibly be referred to it. The only account of 
 the prothallium hitherto published, so far as I am aware, is the paper of Brebner 
 (Brebner 2) on D. simplicifolia, a species from British Guiana. 
 
 In 1897 I collected a small number of specimens of prothallia oi Dantea, prob- 
 ably D. jenmani, and in the summer of 1 908 the same locality was visited, as well 
 as some others in the same district, and material of three species was secured. These 
 
 A. Three gametophytes of Danaa jenmani Underwood. X2. 
 
 B. Three young gametophytes of D.jamaicensis Underwood. X4. 
 C Four large gametophytes of D.jamaicensis. X2. 
 
 0,1, shows four groups of archegonia, ? . 
 
 collections were all made in the vicinity of Cinchona, a mountain station at an eleva- 
 tion of about 5,000 feet. As I have found in collecting other Marattiaceae, the most 
 favorable collecting ground for the prothallia is upon moist, clayey banks where the 
 fruiting plants are growing. In the shady crevices in such positions a careful search 
 will usually be rewarded by the discovery of numerous young plants, and with these 
 there are often associated prothallia in various stages of development. Brebner 
 described the prothallia of D. simplicifoUa as being nearly circular in outline, but 
 very few of the prothallia of the species collected by me showed this form; but they 
 were usually decidedly elongated, sometimes very strongly so, and were, as a rule, 
 very much larger than the specimens of Z). simplicifoUa described by Brebner. 
 
 The three species collected by me were D. jamaicensts Underwood, D. jenmani 
 Underwood, and D. elliptica Smith. In all three species the larger prothallia are 
 usually decidedly elongated and very irregular in outline, often showing conspicu- 
 ous leaf-like marginal lobes, like those occurring in the prothallia of 0.rmtiwi/a and 
 Gleichenia. 
 
THE GAMETOPHYTE 
 
 125 
 
 The germination of the spores and the early stages are quite unknown in 
 Daritta, but probably resemble those of the other genera which have been studied. 
 To judge from the younger stages collected, there is a good deal of variety of form, 
 as is the case also in the other Marattiacea;. The youngest specimens collected 
 belonged to D. jamaicensis. These were very broad in outline and unsymmetrical, 
 one wing of the prothallium being much better developed than the other (fig. 91, B). 
 
 Fig. 92 shows some older specimens of Z). clhptica which were greatly elongated, 
 the very much attenuated posterior region being quite thin and delicate in texture, 
 with no midrib and the archegonial cushion being confined to a small region just 
 back of the growing point. The marginal region in these young prothallia was 
 composed of a single layer of cells and a considerable portion was made up of but 
 
 liyte, probably developcil from 
 X20. rf antheridia. 
 
 B. Large qametophyte of D. flliptica, with two sporophytcs. 
 
 C. Multicellular rhizoid. Xioo. 
 
 D. A cell of the rhizoid, showinR nucleus, n. Xzio. 
 
 bud of Dntura janitiitftis 
 
 two layers, while the mid region, where the antheridia and later the archegonia are 
 developed, was not more than four to five cells in thickness. In general the prothallia 
 of Daticea are|lmore delicate in texture than those of the other Marattiace;e that 
 have been studied. 
 
126 
 
 THE MARATTIALES 
 
 A midrib usually begins to form at an early period and in the older prothallia 
 may become very conspicuous, sometimes reaching a thickness of eight to ten cells. 
 The margin, as we have seen, is always more or less irregular in outline and often 
 develops large leaf-like lobes which are particularly conspicuous in D. jamaicensis 
 (fig. 91, C), but are noticeable also in the other species. The large prothallia, which 
 are sometimes nearly 3 cm. in length, are often branched (fig. 92, D); the branching 
 
 B, antheridium of Maratlia douglaiii. tij neck ( 
 i>, ventral canal cell ; o, egg; m, mantle celli. 
 
 is sometimes unmistakably a true dichotomy, but again it seems to be adventi- 
 tious. In one large prothallium of £). jamaicensis of very irregular form, four groups 
 of archegonia were present, widely separated from each other. It could not be cer- 
 tainly determined whether these arose from a repeated forking of the original apex 
 or whether one or more of them was of adventitious origin (fig. 91, C). 
 
 The ape.x of the prothallium in the younger stages is usually uuknted by a 
 sinus, but in the older ones the heart shape is almost completely lost and the apex 
 may even project as a rounded protuberance, bearing the archegonia upon its lower 
 surface, or it may be fan-shaped, with little or no indentation at the growing point. 
 
 The prothallium is the usual dark-green color found in the other Marattiacea?, 
 but as we have seen is rather more delicate in texture. Of the three species 
 examined, Dana-a elUptica resembles more nearly the prothallia of the other Marat- 
 tiaceae in the thickness of the central portion, which may show from eight to ten 
 
12: 
 
 layers of cells, ihe form aiul division of" the apical cells is exactly the same as in 
 the other Marattiacea-. Stiff brown rhizoids are developed upon the lower surface 
 and are mainly confined to the midrib, over which they may be evenly distributed, 
 or there may be certain regions of considerable extent which are quite destitute of 
 them. The rhizoids in all the species which have yet been examined are truly 
 multicellular, as was correctly shown by Brebner to be the case in D. simpliclfolia. 
 The nuclei are not very large, but can be readily demonstrated. 
 
 The distribution of the reproductive organs is not always the same. Apparently 
 the usual course of development is present here, the antheridia appearing first upon 
 the lower side of the midrib and the adjacent points of the prothallium; later, 
 nearer the apex and upon the lower side, the archegonia arise. Sometimes, how- 
 ever, as in the other forms, there is a more or less marked tendency to dicecism and 
 some of the smaller prothallia seem to bear only archegonia or antheridia, and it is 
 common to find antheridia developing exclusively upon the lateral lobes, which 
 
 A. Section of a >(>ung prothallium, showing (\vi. 
 
 B-E. Longitudinal sections of antheridia. 
 
 F-H. Transverse sections. G, H are surface views showing the oper< nUr crli. in. iii.intlr trlU. \ iSo. 
 
 I. Cell from interior of prothallium, showing the endophvte. 
 
 apparently never produce archegonia at all. Antheridia may also occur upon the 
 upper surface, but this is not common, and in no cases were archegonia seen except 
 upon the lower surface of the midrib. The reproductive organs are sonutimes 
 found in great numbers upon the older prothallia, the whole lower surface of the 
 midrib being beset with archegonia. Quite as often, however, extensive areas along 
 the midrib are quite sterile and the archcgoni;i thus appear in |iatches, sepaiattd h\ 
 sterile intervals. 
 
 IHK KNDOl'HYIK () 
 rattiaceae that I havt 
 
 \I,\K \ 
 
 In all of the Marattiaceae that I have studied, an nulopln tic fungus \ei\ niucli 
 like that which occurs in the prothallium of the ()phioglossace;e has been found 
 occupying the central cells of the prothallium in nearly all cases. The endoph\ te 
 which infests the green prothallium of the Marattiaceae, when compared with that 
 found in the strictly saprophytic prothallia of the Ophioglossaceae, shows some 
 differences which are probably not without significance. The structure of the myce- 
 
128 
 
 THE MARATTIALES 
 
 Hum and its general behavior are so much hke the form which occurs in Ophio- 
 glossum as to leave little room for doubt that the two forms are either identical or 
 very closely related. The conidia (fig. g6, /), while occurring in the Marattiaceae, 
 are perhaps less frequent, but in form and structure are much like those of the 
 endophyte of Botrychnim. The most noticeable difference is the absence of the 
 digestive cells, i. e., those that contain the varicose swollen mycelium. No evidences 
 were found in the Marattiaceae of the destruction of the fungus by the cells of the 
 host and it is likely that the endophyte in these green prothallia is more nearly a 
 true parasite than is the case in the saprophytic gametophytes of the Ophio- 
 glossaceae. In the infested cells of the green gametophyte the starch and chroma- 
 tophores are destroyed by the action of the endophyte, but the nucleus of the cell 
 remains intact. 
 
 THE SEXUAL ORGANS. 
 
 THE ANTHERIDIUM. 
 
 The development of the antheridium, except for the details t)f spermatogenesis, 
 was correctly described by Luerssen and Jonkmann for Marattia and Jngioptcn's. 
 The other genera agree closely with these in the essential structure of the antherid- 
 ium. The development of the antheridium in the Marattiaceae agrees very closely 
 indeed with that of Ophioglossuni. 
 
 Fig. 97.— Developm 
 
 A. Section of prothallium, bearing antheridia on both surfaces. X8o. 
 
 B-H. Longitudinal sections. X300. E-H, D. elllplica; the others, D. jamaicemis. 
 
 The mother cell of the antheridium, as in Op/iioglossmn, divides first by a 
 periclinal wall into an outer cell, the primary cover cell, and an inner one, from which 
 the mass of spermatocytes is developed (figs. 96, 97). The mother cell of the anther- 
 idium shows much the same variation in form as that of Ophioglossum, sometimes 
 being relatively broad and shallow and at other times deeper and narrower (fig. 96, 
 B, C). The first division in the inner cell is usually transverse, but in the broader 
 type of antheridium this first wall may be longitudinal. The primary wall is followed 
 by a second one at right angles to it and the four cells thus formed are again divided 
 so that there result eight nearly equal cells. The first wall occasionally is somewhat 
 oblique, but even in such cases the regular quadrant and octant walls arise at right 
 
I III i.AMi.iorii^i I 12V 
 
 angles to till- |iiimai\ iluision wall. I lif ni-\r (livisii)iis an-, iisuall) at kast, anti- 
 clinals (fig. 97, (7), but before long the periclinal walls also are developed and sub- 
 sequent divisions do not show any recognizable regularity in their sequence; there 
 seems to be a good deal of variation in this respect, even in the same species. 1 he 
 number of cells ultimately formed varies a good deal, but the number of spermato- 
 cytes finally developed is probably never so great as that found in some of the 
 C)phioglossace;f. Ktiulfitssia, both in the si/.e of the antheridium and that of the 
 spermatozoids, approaches nearest to Ophioglossiuti. The number of sperinat(r/.oids 
 ma\ reach several hundred, fifty or more being visible in a single section of a large 
 antheridium, and nearly or (|uite as many may sometimes lie found in Dantra, 
 where, however, the spermatozoitls are much smaller than in K (iiilfussKi. 
 
 Ill the cover cill the divisions are all anticlinal and iKiri/oiual sections or 
 surface views show that the successive walls are arranged spiialK in a way suggesting 
 the segmentation of a time-sided apical cell. 'i"he last-foinud wall cuts out a small, 
 nearly triangular cell, the opercular cell (fig. 95 D, E. <>). In most cases, at least. 
 
 this opercular cell is thrown off when the antheridium opens, leaving a small 
 aperture through which the spermatozoids are ejected. Surrounding the mass of 
 spermatocytes is a layer of mantle cells cutoff from the adjacent cells of the prothal- 
 lium. These mantle cells, at the time of the opening of the antheridium. become 
 very much distended and project strongly into the cavity of the empty antheridium 
 (fig. 100, /?, til). They no doubt play an important part in the dehiscence of the ripe 
 antheridium. 
 
 Sl'KKM VroCENESIS. (Plate 2, figs 58-44.) 
 
 ()t the i\larattiace;e, Kaulfiissia is the most satisfactory for studying the details 
 ot spermatogenesis, owing to the much larger size of the spermatozoid. The 
 development of the spermatozoids agrees very closely in its details with that of 
 Opiiioglossuni. If the spermatocyte is examined just before the final division the 
 two blephaioplasts can be seen and the division of the cell into the two spermatoc\tes 
 proceeds very much as in Opliioglossiini. After the final division the nucleus of the 
 spermatocyte appears coarsely granular and, as in other cases, no nucleolus can be 
 seen. In favorable cases the blepharoplast is visible as a round body, stained 
 9 
 
130 
 
 IHE MARATTIALK! 
 
 rather stiongly and lying near the nucleus. 1 he blepharoplast soon takes on the 
 curved form and becomes much extended and the cilia begin to develop from it 
 before the nucleus has materially changed its shape. 
 
 The nucleus now becomes slightly pointed at one end and begins to stretch 
 out so as U) appear somewhat crescent-shaped, very much as in Ophioglossuni and 
 as it does in other ferns that have been described. With this change in the form of 
 the nucleus, the blepharoplast becomes still more elongated and strongly colored 
 and the cilia increase in length. The nucleus of the spermatozoid in Kaulfiissia 
 is less elongated than is usual in the ferns, and in this respect, as well as in its larger 
 size, it more nearly resembles Ophioglossuni than it does the other Marattiaceje. 
 The granular appearance of the nucleus is maintained until the spermatozoid is 
 almost fully developed, when there seems to be a fusion of the chromosomes so that 
 it appears almost homogeneous; this is accompanied by a noticeable diminution in 
 the size of the nucleus. The nucleus occupies only the large posterior coil of the 
 spermatozoid, while the anterior portion, which shows about two coils, is composed 
 of the blepharoplast with probably a certain amount of other cytoplasmic matter. 
 
 Fig. ioo. 
 
 A. Cross-section of a ripe anthcridium of D. jamaicensii. 
 
 B. Cross-section of an empty antheridium ; m, mantle cells. 
 C-D. Surface views, showing opercular cell. 
 
 All figures X 350. 
 
 In Marattia and Angiopteris (plate 2, fig. 44) the nucleus of the spermatocyte 
 becomes much more extended and the whole spermatozoid is more slender than in 
 Kaiilfussia. Indeed there is very little difference between the appearance of the 
 spermatozoids of Angiopteris and Marattia and those of the typical leptosporangiate 
 ferns. Some observations were made also upon Dancea, in which the spermatozoid 
 is somewhat intermediate in character between that o{ Angiopteris and Kaulfussta 
 (plate 2, figs. 42, 43). In size the spermatozoids are more like those ot Angiopteris, 
 but the nucleus is much less elongated and the general form of the spermatozoids 
 is more like that of Kaulfussia. 
 
 THE ARCHEGONU'M. 
 
 The archegonium in the Marattiaceae, like the antheridium, very much resembles 
 that of Ophioglossuni, but the neck of the archegonium is even less developed than 
 in the latter. Jonkmann ( Jonkmann I) has given a fairly complete account of the 
 development in Marattia and Angiopteris, and Farmer (Farmer 1) has described 
 and figured the archegonium of the latter genus. 
 
 Usually, at least, the archegonium is developed only upon the cushion of tissue 
 back of the apex, the young archegonia arising in acropetal succession. Jonkmann 
 
THE t.AMI 
 
 131 
 
 States, liowcvci, that lie lias tniintl tluiii a 
 Hum, but noiu- of the specimens 1 have 
 cell of the young archegoniimi is scarce! 
 antheiidium and like it is first divided In 
 and an inner cell, the latter usuall}, init 
 central and a hnsal cell (figs. lOi loj), as 
 
 Iso upon the uppii suit.ice ot the piothal- 
 ixamined have shown this. The mothei 
 V disringuishahle in form from the \oung 
 a jHiiclinal wall into an outer lovei- cell 
 nor always, divided suhsi(|ui nrly into a 
 in O/'/iioglossinn and in the t\ jiical ferns. 
 Sometimes the mothei cell of the archegonium is seen in tians\erse section to have 
 been cut out veiy nuich as the axial low of cells arises in the archegonium of the 
 Hepaticiv. Ihere seems no question that the so-called mother cill of the archego- 
 nium in all the ferns is really homologous onl\ \\uh rhi axial vow of cells of rhi- 
 
 'fi 
 
 rthe 
 
 opm 
 
 br\-ophyte archegonium, the four rows of neck cell 
 of the terminal cap cell of the liverwort archegoniu 
 
 The inner of the two primary cells, as we have alreach statid, ma\ havi' a basal 
 cell cut off from it before the further divisions arise, by which the egg cell and the 
 canal cells are divided. I'he neck canal cell is very broad and may become divided 
 into two cells, but usually the division is confined to the nucleus, which probabi)' 
 divides in all cases. 1 he ventral canal cell cut off in the usual fashion from the egg 
 is, with the exception of Datura, very large and conspicuous, thus differing from 
 
 \Tc\\r^on\A o{ Angiofiterii. X275. 6, ba 
 
 Ophioglossum, where the ventral canal cell is so difficult to demonstrate; bur in 
 Datura there is the same imperfect development ot the ventral canal cell that is 
 found in Ophioglossutri. 
 
 As the archegonium approaches maturity a layer ot mantle cells, much like 
 those which surround the antheridium, is cut off from the tissue surrounding the 
 venter of the archegonium. The archegonia of Marattia douglasii are confined to 
 the lower side of the midrib and begin to form at some distance back of the grow- 
 ing point; so tar as can be determined, any superficial cell of the apical meristem 
 can develop into an archegonium. The mother cell divides, as we have seen, 
 into three superimposed cells, of which the lowest, /;, usually divides later by 
 vertical walls, and forms the base of the archegonium. From the central one, b\- 
 transverse divisions, are formed the canal cells and egg, and from the uppermost 
 the neck. Compared with the typical ferns, the most striking differences are the 
 short neck and the very broad canal cells. The cover cell undergoes division into 
 four, by two intersecting vertical walls, as in Ophioglossutri, and each of these four 
 cells then undergoes division by nearly horizontal walls, but the cells remain short, 
 so that the neck projects very little and there are only three or four cells in each row; 
 occasionally there may be only two. Jonkmann states that, as a rule, two of the 
 rows of the neck contain three cells and two contain tour, but that there may occa- 
 sionally be as many as five. The neck canal cell often shows a trace of a division and 
 there ma\- he an :ictual division w:ill t'ormed (fig. loi, D), hut in Marattia douglasii 
 
132 
 
 THK MARATTIALES 
 
 this is not ordinarily the case. In one instance in this species I observed a division by 
 a vertical wall, so that two neck canal cells were formed, placed side by side, in a way 
 recalling very strongly the division of the neck canal cell described by Jeffrey in 
 certain species of Equisetnm (Jeffrey 2). The central cell is divided again by a hori- 
 zontal wall into two nearly equal cells, the lower one being the egg, the upper one the 
 ventral canal cell. The mature egg is nearly elliptical in form, the upper third being 
 almost homogeneous and quite colorless, forming the so-called receptive spot. The 
 nucleus is of moderate size and does not stain very strongly. The archegonium of 
 Augiopteris closely resembles that of Marattia, but is perhaps somewhat narrower 
 (fig- loi). 
 
 Jonkmann figures the archegonium ot M. ficittafolia, which shows that this 
 species also is, in the relative size of the canal and neck cells, very much like Angi- 
 opteris. Farmer thought that a basal cell was usually, if not always, present, and my 
 own studies tend to confirm this. 
 
 Archegonia of Kmill 
 
 , neck canal cell. 
 
 In Kaulfussia (fig. 102) the archegonia form, as a rule, only after the antheridia 
 have ceased to develop. Compared with the other Marattiaceae they are decidedly 
 large, and in this respect Kaulfussia approaches Ophioglossum. Like the mother 
 cell of the antheridium, there is a good deal of variation in the width of the young 
 archegonium in Kaulfussia. Some of the narrower types recall the archegonium 
 of Anthoceros and emphasize the resemblances between the archegonium of the 
 Anthocerotaceae and the eusporangiate ferns. The neck of the archegonium is very 
 short, each of the four original neck cells often dividing only once, so that there may 
 be but two cells to each of the four rows. More commonly, however, there is a 
 
 second division in some of the cells, so that each row consists of three cells. The 
 nucleus of the broad neck canal cell probably always divides, but the number of 
 available specimens was too small to decide whether or not there is ever formed a 
 division wall between these, although it is not at all unlikely that this may occur. 
 The ventral canal cell is conspicuous and equals its sister cell, the egg, in breadth. 
 In the peripheral portion of all the axial cells of the archegonium there are many 
 small granules of starch. Jonkmann figures similar starch granules in both Marattia 
 and Angiopteris. 
 
THE GAMETOPHYTE 
 
 133 
 
 Ihe archegonium in Danaa, while resembling that of the other Marattiaceae, 
 in its position and early development shows some marked differences, the signifi- 
 cance of which is not quite clear. The division of the mother cell into the primary 
 neck cell and the central cell follows in the same way as in the other forms, but the 
 inner cell, usually at least, does not have a basal cell separated from it, but develops 
 at once into the egg cell and canal cells. The absence of the basal cell is by no means 
 unknown, however, among the other Marattiaceic. The primary neck cell gives 
 rise to the usual four cells, each of which divides into three or four, and exception- 
 ally into five. There is a marked elongation of the inner cell before its separation 
 into the central cell and the neck canal cell (fig. 103, D). 
 
 Up to this point there is nothing peculiar in the development of the archegonium 
 in Danaa, but while in most of the other Marattiaceae a conspicuous ventral canal 
 cell is formed, its sister cell being the egg, in Datura the formation of a definite 
 ventral canal cell couKl nor be satisfactorily demonstrated. In a number of cases 
 (fig. 10^; phue i, fio. 4S) ;i siikiII nucleu.s-like body could be seen in a large, clear 
 
 space just above the somewhat contracted mass of protoplasm, with its laiue and 
 conspicuous nucleus, which constitutes the egg cell; but this ventral canal cell 
 nucleus, if such it is, is very different in appearance from the large and conspicuous 
 one found in the other Marattiaceae, and in the absence of anv division stages its 
 nuclear nature must for the present remain somewhat doubtful. 
 
 There is somewhat the same uncertainty in regard to the primary division of the 
 neck canal cell. This possesses a large and conspicuous nucleus which in large and 
 apparently mature archegonia was still undivided. In a few cases a division of the 
 protoplasm in the neck cells was observed; in other cases, without any such division. 
 
134 THE MARATTIAI.KS 
 
 a second body ot rather indefinite outline and staining much less strongly than the 
 lower nucleus could be made out; but up to the present time I have not been able 
 to satisfy myself that the division of the neck canal cell nucleus, which always 
 occurs in the other Marattiaceae, takes place here. The very uncertain nature of the 
 ventral canal cell recalls strongly the condition of affairs in Ophioglossum, where it 
 is equally difficult to demonstrate satisfactorily the presence of a true ventral canal 
 cell. 
 
 FERTILIZ.^TION. 
 
 It has not been possible to follow in detail the process of fertilization in the 
 Marattiaceae, but several stages were found in preparations of Marattta douglasii. 
 The entrance of the spermatozoid into the archegonium was not seen, but in a 
 number of cases the material had been killed immediately after, and twice spermato- 
 zoids were found which had penetrated into the egg. In these cases the spermato- 
 zoid was quite unchanged in form, but had not yet entered the nucleus itself. One 
 case was observed where there were apparently two nuclei in close contact, but the 
 egg nucleus was much contracted and it is doubtful whether this was really a normal 
 appearance. It is probable that the details of fertilization are quite similar to those 
 observed in other ferns. 
 
135 
 
 11. rilK LMHRYO. 
 
 The study of the embryogeny of the Marattiaceae offers many difficulties. 
 Fertilization does not seem to be of common occurrence and a very large number of 
 prothallia must be examined before even a small series of embryos can be secured; 
 moreover, the earlier stages of the embryo are peculiarly liable to shrinkage in 
 preparing them for sectioning and it is exceedingly difficult to secure really satis- 
 factory preparations of these early stages. 
 
 A marked peculiarity of the young sporophyte, which was first shown by Luers- 
 sen and Jonkmann for Marattia and Aiigiopteris, is the orientaticjn of the primary 
 organs of the young sporophyte with reference to the archegonium. The primary 
 or basal wall in the embryo is always transverse, as it usually is in Ophioglossum, 
 instead of being vertical as in most of the typical ferns. The first leaf, instead of 
 being formed from the portion of the embryo nearest the archegonium, as it is in 
 the common ferns, arises from the half of the embryo which is turned awa\' from the 
 archegonium, and grows straight upward, bursting through the prothallium u|)on 
 its upper surface, instead of appearing upon the lower side of the prothallium and 
 curving upward. Ihe external organs of the young sporophyte are dificrentiated 
 
 A. Two iongitudii 
 
 B. A similar sectif 
 
 C. Three transver 
 
 1 sections of a young embryo, t i, basal wall. X200. 
 
 of an older embryo, cor, cotyledori; r, root initial. 
 
 sections of a much older embryo, showing junction of the two first 
 
 much later than is the case in the Leptosporangiates, and the Marattiaceae in this 
 respect closely approach the Ophioglossaceae. 
 
 It has been generally assumed that, as in the leptosporangiate ferns, the cotyle- 
 don and stem are of epibasal origin, the root and foot hypobasal. This conclusion 
 was reached by both Jonkmann and Farmer, and my early studies on Marattia 
 douglasii led me to the same conclusion. A further study of this species, however, 
 as well as an examination of the embryos of Angiopteris, Kaidftissia, and Datura, 
 has shown that this is not the case, but that the whole of the hypobasal region is 
 devoted to the formation of the foot, and the root is developed secondarily from the 
 epibasal region, from which are also derived the stem apex and the cotyledon. 
 
 While in the typical ferns the young organs of the embryo at a very early stage 
 show a definite apical growth, the apical cells being readily traceable to the primary- 
 octants of the embryo, this is by no means so readily shown in the Marattiace;e. 
 In the later stages, such initial cells can be seen in the root and stem at least, but the 
 relation of thesi' initial ceils of rhe root and stem to the earl\- tli\isions of the embr\() 
 is ver\- difficult to (len-rminc In M.dniiclasn (Caniiihell .? ) I .tatiti rhat there 
 
136 THE MARATTIALES 
 
 was probably a single initial in the stem and in the primary root, and Farmer also 
 thought that a single root initial was always present in Angiopterts, but concluded 
 that such an initial cell was not present in the stem of the young sporophyte. Breb- 
 ner (Brebner 2) says that in Dancea simplictfolia such a single initial cell seemed 
 to be always present in the stem, and my studies on Angiopteris, Kaiilfussia, and 
 Dancsa indicate that a single initial is developed at an early period in the stem apex 
 and persists until the sporophyte has developed several leaves. 
 
 Before any division occurs in the embryo the fertilized cell increases markedh 
 in size, after which there is formed the horizontal basal wall (figs. io6, 109, Ij b). 
 This is probably followed at once by the median wall (except in Dancea), so that the 
 embryo at this stage is divided into four approximately equal quadrants (see Jonk- 
 mann 3, fig. 9). 
 
 The genus Daiicea, at least this is true for D. jamaiceiisis and D. eUiptica, 
 differs in the early divisions of the embryo from the other Marattiaceae. The egg 
 cell after fertilization elongates in a way which closely resembles that found in 
 Botrychium ohUquum (Lyon 1). The primary hypobasal cell either divides no further 
 or only once and forms a short suspensor, so that all the organs of the young embryo, 
 including the foot, are really of epibasal origin. As \'et no trace of such a suspensor 
 has been found in the other Marattiaceae. 
 
 THE EMBRYO OE MARATTLA. 
 
 Luerssen (Luerssen 2) found the one-celled embryo and young plants of 
 Marattia cirutwfoUa, but was not able to procure the older embryos. The writer 
 (Campbell 3) succeeded in procuring several stages of the embryo in Murtittia 
 douglasii, and somewhat later Luerssen described and figured some of the earlier 
 stages in M. fraxinea and M. wettimantiKrfolia. These, so far as 1 am aware, 
 complete the list of records upon the embryo o{ Marattia. 
 
 The fertilized ovum in Marattia doHglasii becomes much enlarged before it 
 divides and completely fills the venter of the archegonium. The granular contents 
 of the egg cell become evenly distributed without any apparent increase in quantity 
 as the fertilized ovum grows, so that the one-cell embryo contains comparativel\- 
 little granular contents, but the nucleus is very conspicuous. At the time of the 
 first division the young embryo is almost perfectly globular in form. I was unable 
 to find the stages immediately following this, but jonkmann has figured an eight- 
 cell embryo of M. fraxinea (see Jonkmann 3, fig. 9). The basal wall is transverse 
 and this is followed successively by the median and transverse walls, so that the 
 globular embryo is divided into approximately equal octants. From the hypobasal 
 half, which is nearest the archegonium, there is developed the foot alone, while all 
 of the other organs arise from the epibasal portion, which is turned away from the 
 archegonium. 
 
 The youngest embryo (except the one-celled stage) which I found in M. 
 douglasii, is shown in fig. 106, A. This is about the same stage as the one shown in 
 Jonkmann's figs. 11 and 13. Unfortunately this embryo was rather badly shrunken 
 in the process of embedding and the division walls were a good deal distorted, so 
 that it is rather hard to determine exactly their correct relation to each other; but 
 probably the wall h h represents the basal wall and m the median wall. The 
 embryo has now lost its original globular form and become oval, the long axis lying 
 transversely. The secondary divisions in the octants are mostly anticlinals, and the 
 first periclinals have just appeared in a few of the cells. Figure 106, A, shows two 
 consecutive median sections. It will be seen that cell division is more active in the 
 epibasal portion than in the hypobasal region. 
 
PHI 
 
 •MI1K'|( 
 
 13: 
 
 In my original snid\- <>{' Murattni I t'dncliidtd that the- primary divisions n-sulted 
 in the estabhshmcnt of the organs of the voiing sporophyte in a manner similar to 
 that in the l>ept()sporangiat:c, i. e., that the stem and leaf were derived from the 
 epibasal (]nadrant, the root and foot from the hypobasal ones. Farmer thought that 
 this was true for Angiopteris and Jonkmann assumes that it is the case also in 
 Marattia. A fuither study of my preparations, however, has led me to believe that 
 the whole of the hypobasal region is devoted to the foot, while the root, together with 
 the leaf and stem apex, are of epibasal origin. A large cell (hg \ob,A) occupies 
 approximately the same position as the stem apex in the older embryo and it is 
 possible that this may be the apical cell of the stem, but in the absence of the inter- 
 mediate stages this can not be positively asserted. For a long time the embryo 
 retains the oval form and there is scarcely any sign of the young organs of the 
 sporophyte which in the leptosporangiate ferns are so early manifest; indeed, 
 Jonkmann states that there is no differentiation at all, but a careful study of the 
 older embryo shows unmistakable indications of the definitive organs. The devel- 
 opment of the tissue in the epibasal region is not uniform, but somewhat to one 
 side of the center (fig. io6, B) there may be seen a group of columnar superficial 
 cells, which mark the position of the future growing point of the stem. Wlurher 
 one of these supeificial cells is the 
 definitive apical cell it is impossible 
 to determine, but from a compari.son 
 with the older stages, as well as with 
 )rresponding stages of the embryo ) 
 
 in the oti 
 
 hi ■, at least, tha 
 
 and perhajis this 
 
 traceil back to tl 
 
 cell noted in the \()imt£e 
 
 genera. 
 
 t seems pro 
 
 this is the c; 
 
 nitial cell ma\ 
 
 centralh' pla 
 
 v<). 
 
 ba- 
 
 in b 
 
 considerable portion of the epibasal 
 tissue is not uicUuleil ui the men- 
 stematic region, winch is tleiued 
 mainly, at least, from only one of the 
 original epibasal quadrants. The 
 tissue adjacent to this meristematic region is made up of large cells with less dense 
 contents and smaller nuclei and differs in no way from the large cells of the foot, into 
 which this tissue insensibly merges and of which it may be said to constitute a part. 
 In the embryo figured, it was not quite certain whether the root apex had 
 begun to form or not. It seems probable that the cell r is really the apical cell 
 of the young root, but except for its position it was not noticeably different from the 
 cells adjacent; however, as in Angiopteris and Datura there is no question that the 
 root originates in this position, it is probable that this cell is really the initial for the 
 young root and is cut out from the base of the epibasal tissue near its junction with 
 the foot. This marked endogenous formation of the root is very much like that in 
 the embryo of Opliiogloisiini nioliiicauiini, but differs entirely fVom the superficial 
 origin of the apical cell of the root in the embryo of the Leptosporangiates. 
 
 The stem apex in the older emhr\(> is of very limited extent, consisting of onl\- 
 a few cells, of which one, as we have said, is probably the definite apical cell, although 
 it must be said that this point is very difficult to decide, as all of the central cells 
 of this apical group look a good deal alike, but a careful study of both transverse 
 and longitudinal sections seems to point to one of these as the apical cell, which in 
 shape is a good deal like that in the stem apex of Ophioglossum. In longitudinal 
 
138 
 
 THE MARATTIALE5 
 
 sections this is somewhat wedge-shaped, the narrow end turned outward and the 
 broad base below. In cross-section this cell is nearly square in shape (fig. 107, A, x) 
 and there may be seen a fairly regular series of four segments cut from the lateral 
 faces. From the broad truncate base of the cell, segments are also cut off which 
 contribute to the inner tissue of the stem.* 
 
 Much the greater part of the epibasal meristem contributes to the cotyledon, 
 which is soon evident as a broadly conical protuberance, somewhat flattened on the 
 side adjacent to the stem apex and merging gradually on the outer side into the large- 
 celled tissue which adjoins the foot. To judge from the limited number of young 
 embryos which I could examine, it seems that the growth of the cotyledon is not due 
 to the activity of a single apical cell, but this point was somewhat uncertain. 
 
 By the time that the cotyledon is established, growth has progressed in the young 
 root, which now is seen to have a conspicuous apical cell, which divides rapidly 
 so that the root quickly elongates in a direction opposite to that of the cotyledon 
 (fig. 108). The apical cell of the primary root in M. douglasii is not triangular in 
 outline, but more or less quadrilateral, whether seen in longitudinal or cross 
 sections (fig. 107, B). In form and segmentation it perhaps more nearly resembles 
 that of Daiura than it does the tetrahedral apical cell which occurs in the young 
 
 primary root oi Angwpteris. 
 Active cell divisions take 
 place also in the tissue of 
 the foot, which completely 
 incloses the young root and 
 becomes practically merged 
 with it, so that it is quite 
 impossible to say, at the time 
 the root emerges, just how 
 much of the tissue ot its 
 outer portion really belongs 
 to the root itself and how 
 much is derived from the 
 original tissue of the foot. 
 The latter is now no longer 
 recognizable as such, the 
 young sporophyte apparently at this time being composed almost entirely of the 
 cotyledon and the very large root, with the stem apex, lying near their junction. 
 From a comparison with younger stages, however, it is perfectly evident that the 
 enlarged central region of the embryo at this time is composed mainly of tissue 
 belonging to the foot, which is, so to speak, perforated by the root in its downward 
 growth. 
 
 The development of the vascular bundles at a very early period is first evident 
 in the cotyledon, which almost as soon as it can be recognized at all is seen to have 
 a strand of procambium extending below it. If this procambium strand is traced 
 downward, it is seen to continue without interruption into the base of the root, exactly 
 as it does in Ophtoghssum moluccanum. No trace of a procambial cylinder can be 
 found extending into the stem apex, which gives rise only to the parenchyma of the 
 central pith. This early development of the vascular bundles in the cotyledon and 
 root was correctly observed by Jonkmann and Farmer, both for Marattia and 
 
 A. Section of an advanced 
 
 emergence of cotyled 
 
 B. The cotvledon, more er 
 
 * In a very recent paper 
 tem in Marattia alata is at first 
 group of initial cells. 
 
 l), Miss Charles states tha 
 r sporophvtes the single api( 
 
 the apical cell of the 
 I cell was replaced by 
 
THE EMBRYO 
 
 139 
 
 Angiopteris, but they seem to rakr it for granted that a vascular strand is also 
 developed in the stem; at any rate the\- make no reference to the absence of such 
 a vascular strand from the stem region. In my earlier study of Marattia, I supposed 
 that a strand was developed which belonged to the stem itself, but a further exam- 
 ination of many plants, after a study of this point in the other genera, has made it 
 pretty clear that this supposed stem bundle really belongs to the second leaf, the 
 rudiment of which appears at a very early period. 
 
 rilK KMHRYO OF ANCIOI'TI- RIS. 
 
 In Angiopteris the embryo very early becomes more greatly elongated trans- 
 versely than is the case in Marattia, so that in a longitudinal section it appears as a 
 very much depressed oval (fig. 109); the cpibasal region is larger than the hypobasal 
 
 A. Two sections of a young embryo of Angiopteris. 
 
 B. Diagrams showing arrangement of cells of same 
 
 C. An oljer embryo, b h, basal wall; si, stem apex 
 
 and the quadrant divisions are often still evident. Whether octant divisions are 
 formed in all the quadrants could not positively be determined, but Jonkmann 
 states that such is the case. There can usually be found in the young embryo a 
 nearly centrally placed large cell (fig. 109, C. st), which probably is the initial cell of 
 the young stem. The position of this cell is not unlike that which occurs in the 
 
 nhryooi Angiopteris. X200. The yovmg root, 
 :x; col, cotyledon; /, foot. 
 
 embryo of Eqtiisettini, and at this stage there is also a certain resemblance to the em- 
 bryo of Botr\chium virginuinum. As the embryo grows it tends to assume a more 
 nearly globular form (fig. i lo). The basal wall can still be imperfectly followed, the 
 hypobasal portion of the embryo being made up of the large cells forming the foot, 
 while above the basal wall the cell divisions are more active and the rudiments of 
 
140 
 
 THE MARATTIALES 
 
 the Stem and cotyledon can be recognized, although the embryo still retains its oval 
 outline. As in the embryo of Marattia, growth is most active in the central region 
 of the epibasal part of the embryo and there can generally be recognized a large 
 central cell, which is presumably the single initial of the stem apex, although, as 
 in the case oi Marattia, it is not absolutely certain that a single initial cell is always 
 present. Cell division is especially active on one side of the stem apex, and this 
 area marks the position of the young cotyledon. The limits of this growth area 
 
 A. Two longitudinal sections of 
 
 B. Stem region of same embryo. Xi6o. 
 
 C. Root apex. Xl6o. 
 
 not very sharply defined and it is difficult to say whether it can be traced back to 
 a single quadrant or octant cell, but it is probably not always constant in its 
 position; indeed, the stem looks as if the stem apex and the cotyledon both arose 
 from the same quadrant, the second epibasal quadrant contributing, at least in 
 part, to the foot. 
 
 As seems to be true in all the Marattiaceae, the root makes its appearance at a 
 comparatively late period and it is evident that the root, as in the embryo of Maratlia, 
 is a strictly endogenous structure. The first indication of the root is the development 
 
 A. Longitudinal section of an advanced embryo of An 
 
 B. Cotyledon of same embryo, showing dichotomy of 
 
 C. Root apex of same. X220. 
 
 of the cotyledon. X50. 
 
 of a group of actively dividing cells, almost in the center of the embryo below the 
 stem apex. This meristematic region is probably always of epibasal origin, but it is 
 close to the basal wall and it is possible that sometimes it may arise below it. The 
 apical cell, which in the early stages almost always appears triangular in section, 
 becomes conspicuous and was recognized by Farmer in his study of the embryo of 
 
Tin: IMI'.R'iO 141 
 
 Jiigioptcns {Fainifi I). Tin- loor initial now divitlcs by icj^Lilar SLiinicntatidn and 
 the root apex pushes rapidly down tlirough the underlying toot and ultimately 
 emerges on the lower side of the prothallium. In the meantime the cotyledon grows 
 actively and there is a rapid elongation of the whole embryo in a vertical direction. 
 As in the case of Mcirattia, it is difficult to prove that the cotyledon grows from a 
 single initial cell. The cotyledon has very much the form of that in Marattia, 
 growth being more active on the outer side, so that it curves over the stem ape.x 
 very much as the cotyledon does in the embryo of Hotrycliiiini virgiiiiatium. About 
 the time that the cotyledon is ready to emerge, the apex becomes flattened out and 
 (sometimes, at any rate) there is a true dichotomy of this apex (see fig. 112, B). 
 
 In the primary root there seems to be no question of the presence of the single 
 initial cell which, at first at least, has the tetrahedral form, but later on is apt to 
 have the base truncate, although it usually has three series of lateral segments. 
 I have not been able to confirm Farmer's statements that the single apical cell is 
 later replaced by a group of similar initials, as in the later roots of the sporophyte, 
 although it is not at all impossible that such may sometimes be the case. The 
 development of the embryo of the vascular bundles in the young sporophyte is 
 exactly the same as in the coiiesponding stages of the embrvo ni Mmattin. 
 
 rilK KMIUnO ()1- KAUI.I TSSIA. 
 
 So far as I am aware, no account has been published of the embijo ui kuul- 
 fiissiii, except one of my own (Campbell II). Only two young embryos were found, 
 so that it was impossible to follow in detail the early history of the young sporophyte. 
 The basal wall, as in the other Marattiaceae, is transverse and, to judge from a 
 comparison of similar stages of the embryo in Marattia and Aiigiopten's, all of the 
 organs of the young sporophyte in Kaulfussia, except the foot, are also of epibasal 
 origin. Figure 113, A, shows a nearly median longitudinal section of the young 
 embryo. This is very much elongated transversely, and to judge from the position 
 of the cells the basal wall is probably followed by the median walls, forming nearly 
 equal quadrants. The large cell {st) in one of the epibasal quadrants corresponds 
 in position to the similar cell found in the embryos of the other genera and very 
 likely may represent the primary initial cell of the stem. The embryos, however. 
 
 Fig. 1 13.— Young embryos of Kimljuniit. X275. 
 
 a similar embryo, hh, ba^al wall; n, t]iiadrant wall. 
 
 were too young to make clear the relation of the cotyledon and primary root to the 
 stem. Three nearly transverse sections of an embryo of about the same age are 
 shown in fig. 113, B. To judge from the structure of the older sporophyte at the 
 time when it first emerges from the prothallium, Kaulfussia agrees in the main with 
 Marattia and Angiopteris in the origin of the young organs of the sporophAte. 
 
142 
 
 THE MARATTIALE; 
 
 THE EMBRYO OF DAN.+'.A. 
 
 liiebiKT (Brebnci 2) has described the older enibiyt) of Duiura simpliiifolm, 
 but did not secure the earlier stages. The youngest specimens he figures closely 
 resemble in form a corresponding stage in the species studied by me, and the slightly 
 pointed basal region suggests the possibility of the presence of a short suspensor 
 in D. simpUcifolia like that which I have found in D. jamaicensis and D. elliptua. 
 While my own collection of young embryos is not as complete as might be wished, 
 still enough stages were secured to show that at least D. jamaiceusis and D. rlli pticu 
 
 A. Archegoniuni containing a one-celled embryo 
 
 B. Outline of next section of same embryo. 
 
 C. Three-celled embryo, showing suspensor, ius. 
 
 D. Outline of next section of the same embryo. 
 
 E. Three longitudinal sections of a four-celled 
 
 F. .An older embryo, which was shrunken. 
 
 G. Diagram of F, showing probable arrangen 
 
 nbryo. 
 It of cells. 
 
 differ remarkably from the other Marattiacea? that have been studied in the develop- 
 ment of a short suspensor, thus showing an interesting analogy with the embryo of 
 Botrychium obliquum, described by Lyon (Lyon 1). Whether D. jenmani shows 
 the same peculiarity was not determined, on account of the failure to obtain the 
 young embryo of this species; but as in its later development it corresponds very 
 closely to the other species, it is highly probable that a suspensor is developed. 
 
 section of an older embryo of D. jamaicensis. The su 
 B. Three sections of a young embryo of D. elliptica. sus, suspensor. X200. 
 
 Before the first division takes place in the embryo, the fertilized ovum enlarges 
 to several times its original size and becomes decidedly elongated. The first division 
 wall, as in the other Marattiaceae, is transverse, but of the two primary cells thus 
 formed the hypobasal one divides no further, or only once, and forms a short sus- 
 pensor which pushes up into the neck of the archegonium whose cells become more 
 or less completely disorganized (fig. 114, A). The next division wall is a nearly 
 median one in the epibasal cell, and this is soon followed by a second wall in each 
 
IMHKVi 
 
 143 
 
 .f tlic 
 
 basal iills, so ihat this portiini of tin cmhrvo is dunK-.l into foui luarly 
 equal tiuadrants. I lure is sonic cvickncc that these are tollowed by hoii/ontal 
 octant walls, so that the epibasal region is thus divided into octants in much the same 
 fashion as obtains in the whole embryo in other Marattiacea- and in the lepto- 
 sporangiate ferns. How far these divisions are constant can only be conjectured, 
 owing to the small number of young embryos which were available. The elongated 
 pear-shaped embryo in Datura appears very different indeed, in these early stages, 
 from the broadly elliptical and much depressed embryos of corresponding stages 
 in the other Marattiace;e (fig. i 14. f , /_), E). 
 
 older embryo. X2cx3. Section B is a nearly median ; 
 A is the next section in the series; C shows the suspensor, sus. 
 
 A further study of the embryos shows that all of the lower halt of the epibasal 
 region, probably that derived from the four lower octants — i. e., the four octants 
 that were in contact with the .suspensor — develops into the foot, while all of the 
 other organs of the embryo (leaf, stem, and root) arise from the four terminal octants. 
 
 Quite early in the development of the embryo, there appears the same centrally 
 placed large cell which we have observed in the embryos of the other genera and 
 which, as we have seen, probably represents the primary stem initial. Brebner 
 concluded that a single initial was present in the stem of Z). simplicifolia and this is 
 true of the three species e.xamined by me. The initial of the stem in Daiura becomes 
 
 very easily seen in the older stages and is perhaps more clearly defined than it is in 
 any other of the Marattiace;e. Whether or not this cell can be traced back to one of 
 the original terminal octant cells is difficult to say, but it is quite possible. 
 
 No single initial could be made out for the cotyledon, and it seems quite likel}' 
 that the position of the cotyledon is not always exactly the same, being determined 
 perhaps by the position of the embryo with reference to the light or to some other 
 factor. The cotyledon, as in Marattia and Jngiopteiis, first appears as a slight 
 prominence close to the nearly centrally placed stem apex, and there is soon visible 
 
144 THE MARATTIALES 
 
 a group of actively dividing cells which constitute its growing point, but no one of 
 these can be certainly named as the apical cell (fig. 117, 3). The embryo now 
 rapidly increases in breadth until the apex is almost flat, except for the slight pro- 
 jection in the center, where are situated the stem apex and the young cotyledon. 
 The whole embryo at this stage may be described as "top-shaped." The stem and 
 cotyledon occupy a comparatively small portion of this broad terminal area and 
 are surrounded by a ring of large absorbent cells, physiologically, at least, belonging 
 to the foot and merging insensibly into the similar cells which make up the lower 
 half of the embryo. 
 
 As in the other Marattiaces, the very young embryo shows no trace at all of 
 the root and this appears only after the embryo has reached a comparatively large 
 size. The origin of the root is exactly the same as in the other genera and it is a 
 strictly endogenous structure. The single initial cell arises deep down in the tissue 
 of the embryo, usually below the cotyledon and probably from the same octant as 
 that from which the cotyledon is formed, but owing to the displacement of the 
 original division walls this can not be determined positively and it may be that the 
 position of the root is not always exactly the same. The apical cell of the root soon 
 
 i-iG. 119. — u.jamatcetisis. 
 A. Two sections of an advanced embryo. Xioo. B. Stem apex of the same embryo. X220. 
 
 becomes conspicuous, but it does not usually show the triangular form as seen in a 
 longitudinal section, but appears more nearly square, as in Marattia. This root 
 initial cell is quite variable in form, but more commonly it appears in longitudinal 
 section with a truncate base (fig. 117, r). In transverse .section it approaches the 
 triangular form, but is more or less irregular in outline. The lateral segments cut 
 off from the initial cell are large and contribute later to the root cap as well as to the 
 inner root tissues, and it is not impossible (as Farmer believes to be the case also 
 
III. IMHK^I 
 
 145 
 
 in :lugi()[^trni) that sdiiutmu's HKire tliaii out- ot tlicsf tcimin.il iills tuiution as 
 tlic apical cell. 
 
 As soon as the toot apex is established its growth is veiy ia|iiil and, as in the 
 other genera, there is a rapid elongation of the whole embryo, whose vcitical dianutci 
 very soon becomes greater than the transverse diameter, and ultimatel> the whole 
 embrvo becomes veiy much elongated. In the a.xial region of the embiyo just below 
 the stem ape.x there is a marked elongation of the central cells which might at first 
 sight be taken to represent the central vascular cylinder, but these elongated cells do 
 not give rise to vascular elements, but remain as elongated parenchyma and belong 
 really to the pith. Ihe part of the foot which lies below the growing root apex 
 acts as a very massive root cap and is pushed down with the growth of the toot until 
 the latter emerges from the lower side of the piothalliiim. The cor\'Iedon in the 
 meantime grows actively ii|nvar(l and finally penetrates tlu' piorhalliinn, emerging 
 
 on the upper side. The growth of the stem is slight and the cotyledon and root form 
 nearly a straight line, as in the other Marattiaceae and in Opluoglossum, so that the 
 young sporophyte may be described as bipolar. The equatorial region is surrounded 
 by the large absorbent cells in contact with the prothallium and ail of these may be 
 said to function as the foot, although it is impossible to sa}' how much of this tissue 
 is derived from the original foot. 
 
 THE ANATO\n .AM) HISTOLOGY OF THE YOUNG Sl'OROl'H^' IK. 
 
 The young sporophyte, at the time the root emerges from the prothalliinii, is 
 very much alike in all of the genera (figs. io8, iii, I2i). The cotyledon at this 
 stage is a thick, conical protuberance, strongly curved inward over the stem apex, 
 owing to the more rapid growth upon its outer side. Close to its base lies the 
 very limited growing region of the stem, adjacent to which is the rudiment of the 
 second leaf, which at this time projects very slightly above the level of the stein apex. 
 The cells of the latter, and also those of the young leaves, are evidentl)- actively 
 growing cells with conspicuous nuclei and are quite different in appearance from the 
 large, transparent cells that compose the enlarged mid-region of the young sporo- 
 phyte. This is largely made up of the original tissue of the foot, but this tissue merges 
 insensibly into the basal region of the cotyledon and the upper part of the root, 
 these two organs forming in their growth almost a straight line. \\'hether or not 
 we term this middle region of the young plant the "stem," it must be borne in 
 mind that it does not arise from the activity of the extremely limited meristem, 
 forming the true stem apex. The young sporophyte is traversed by a single conspic- 
 uous vascular bundle which extends through the cotyledon and root without intei- 
 
 10 
 
146 
 
 THE MARATTIALES 
 
 riiption, and it is quite impossible to say just exactly where the point of junction is. 
 In Angiopteris elongated tannin cells, which stain very strongly, accompany the 
 vascular bundle, both in the root and cotyledon; but these are either entirely wanting 
 or but slightly developed in Marattia and are quite absent from the young sporophyte 
 at this stage in the other genera. In the relation of the primary root and leaf, there- 
 fore, the embryo of the young sporophyte in the Marattiaceae shows a most striking 
 resemblance to the condition found in Ophtoglossum moluccaiium. The cotyledon, 
 as in Ophioglossum, is not to be looked upon as an appendage of the stem, but as an 
 organ siti generis. 
 
 Almost as soon as the second leaf is recognizable, there is evident, connecting 
 it with the primary vascular strand, a short group of procambium cells, and the 
 stem apex is seen to occupy the space between these two leaf traces. No procambium 
 IS developed in the stem above the junction of the leaf traces, hut the inner cells, derived 
 from the apical meristem, contribute solclv to the medullary tissue of the sporophyte 
 
 (hg.izuJ). _ ' ' 
 
 ..h 
 
 
 
 Fig. ill.—Dim.cajamaicens 
 
 15. 
 
 
 Large embryo, cot 
 Median section of ( 
 
 , cotylcdo: 
 ;otyIedon; 
 
 n; I', second leaf. X50. 
 h, epidermal scale. X95. 
 
 C. Second leaf. 
 
 D. Trichomes f 
 
 X,5o. ^ 
 rom apical region. 
 
 The first tracheary tissue arises in the mid-region of the young sporophyte and 
 consists of a group of short reticulate tracheids. From this point the development 
 of the tracheary tissue proceeds upward into the cotyledon and downward into the 
 root. The single strand of tracheary tissue in the cotyledon is seen to be continuous 
 with one of the xylem masses of the diarch root. The second xylem of the root arises 
 somewhat later and is connected with the xylem of the second leaf. This second 
 xylem mass is decidedly smaller than the first one, at least in Dancea, where I have 
 studied this point carefully, and an examination of the other genera points to a 
 similar inequality in the xylems of these as well, but this may not always be the case. 
 
 THE COTYLEDON. 
 
 The cotyledon at a very early period bends strongly over the stem apex, very 
 much as it does in Botrychium virginianum, and very soon afterwards begins to 
 flatten out, so as to indicate the separation of the lamina from the petiole. The 
 flattening of the apex is followed, in many cases at least, by a true dichotomy, which 
 is soon repeated, so that the lamina becomes fan-shaped, with a strictly dichotomous 
 venation (fig. 87, G, H; fig. 127). This shows especially well in Marattia douglasii. 
 In M. sambucina (fig. 87, C) the cotyledon is more nearly orbicular than in M. 
 douglasii, but the venation indicates a similar early dichotomy of the apex. 
 
 In Angiopteris, the form of the cotyledon is extremely variable (fig. 124). 
 Farmer states that it has a distinct midrib extending to the apex of the cotyledon 
 
THF YOUNG .SPOROPHYTF 147 
 
 ami that the seionclaix' \tiiis aii' |iinnattly aiiangtcl with leti-ieiuc: to the midrib. 
 These later veins, however, may show clichotonious branching. I made an e.xamina- 
 tion of a considerable number of young sporophytcs collected at the same place where 
 Professor Farmer secured his specimens, but very few of the plants that I collected 
 siioweci this pinnate \enation in the cotvUdon. although it is usualh' conspicuous 
 
 11^ sporophyte of .1 
 to tlie gametopl 
 
 Young sporophyte of Kaulju 
 
 gametophyte,/ir. X2. 
 Cotyledon from another plant . 
 A later leaf; ir, stipule. / i . 
 
 in the second leaf. Figure 124, 6", shows drawings of a number of the forms from 
 my collection, showing the variation from a strictly pinnate to a perfectly dichoto- 
 mous venation. Many of them show an intermediate venation, but with very few 
 exceptions, even where a midrib was present, this was forked at the apex, indicating 
 
 F.r.. ,24. 
 
 A. Young sporophyte of .In^lopicrisj with three leaves. 
 
 B. Young cotyledon. X20. 
 
 C. Three cotvh-dons, showine the variation in form. ; 
 
 I his primar\ 
 ung leaf (fig'. 
 
 B) wl 
 
 .t the 
 re thei 
 
 ipe.K 
 .■ wa,^ 
 
 thai llun iiad been an eai I\- ilichotonn . Ibis primar\ tli 
 well shown in a section thiough a vei 
 (juestion that such a dichotomy was taking phi 
 
 In Daiiira there is much the same variation in the form of the cot\leilon as in 
 A)igioptcris. Brebner's figures of D. stmplicifoUa show that the cotyledon in this 
 species has a midrib with two lateral veins near the base, but sometimes the midrib 
 forks at the summit, indicating again the early dichotomy of the cotyledon. In one 
 
148 
 
 THK MARATTIALES 
 
 of his figures, where there is no forking of the mid-vein, the hiteral veins seem to 
 follow the margin of the leaf and rejoin the central vein, thus inclosing two large 
 areoles like those found in the cotyledon of Kaiilfussi a; no mention of this, however, 
 is made in the text. D. jaNiaimisis (fig. 125, J) shows a very similar form and vena- 
 
 Four young bporophytes still M 
 Young cotyledon of D. elliptka. 
 An older cotyledon of D. elliptii. 
 
 tion, but the cotyledon is larger and there is sometimes a forking of the lateral veins. 
 In this species also there may sometimes be found a fan-shaped lamina with true 
 dichotomous venation like that of Marattia, but as I had httle material of this 
 species I can not say how common this form is. It is probable that an examination 
 
 lung sporophyte of D. jamtiicen 
 lightly enlarged, i, the cotyledon 
 ise of the fourth leaf, showing st 
 cale, K. X20. 
 
 ith four leavi 
 . </. and pelta 
 
 of a large series of young plants would show examples of the same type ot cotyledon 
 as in the other species. The cotyledon of Z). elliptica (fig. 125, B, C) is of much the 
 same form as that of D. jamaicensis, but the cotyledon is larger as a rule and the 
 venation may be more complex, although conforming to the same general type. 
 
II F YOL'NC srOKOI'HYTI-: 
 
 149 
 
 The cotyledon in Ka/il/nsMti (lig. \z^) is s|iatiilatL- in loim and not unlike- that 
 of DciiKva in shape and also closely resembles in outline the broader leaf forms of 
 Ophioglossiirn tuoliicctinurri. I he resemblance to the latter is much increased by 
 the venation, which is reticulate and extraordinarily like that of the cotyledon of 
 Ophioglossum. As in Ophioglossuni, there is usually a mid-vein, but this is not, 
 however, noticeably thicker than the laterals. The latter, instead of extending 
 free to the edge of the leaf, are connected at their distal ends with the mid-vein so 
 as to inclose more or less elongated arcoles. The central vein does not extend to the 
 tip but, as in Dancea, divides into equal branches which join above so as to inclose 
 
 ityledon of Marattia dou^lasli. X260. 
 of cotyledon. X300. 
 
 onchti\ veins shows that theii 
 'ins of the other Marattiace;e 
 rgin of the leaf, or more rareh 
 
 a terminal areole. A careful examination of the s 
 divisions are really dichotomous, as in the lateral 
 and some of the ultimate branches ma\' reach the n 
 they may end freely within the areoles. 
 
 As the cotyledon develops there is more or less elongation of the petiole, which, 
 however, very seldom exceeds the lamina in length. The petiole is usually more or 
 less deeply channeled upon its inner face. In Marattia doiiglasii (fig. 133) the base 
 of the cot\ ledon is much enlarged just above the le\el of the stem apex, which is 
 
 A. Section of la 
 
 B. A young 
 
 C. An older one. 
 
 thus almost inclosed in a sort of cleft in the base of the cotyledon, very much as it is 
 in the case oi Botrychiitm or H elmijitliostachys. In Anginpteris there is but slight 
 contraction at the base and the stem apex forms a strongly oblique surface, hardl}' 
 covered at all by the base of the cotyledon. The base of the cotyledon is still more 
 conspicuously hollowed out in Dantrn and especially in Kaiilfiissia (figs. 136. 137), 
 and in the latter there is a circular ridge completely surrounding the stem apex, 
 exactly as is the case in the young bud in Ophioglossuni tnoliircarnnn. Traces ot 
 this ridge can be seen also in Damra and Marattia, but this is much less conspicuous 
 than in Kaiilfiissia. In the latter this cavity is especially marked, .so that one might 
 almost speak of a stipular sheath, and this resemblance to Opiiioglossiint in the 
 
150 
 
 THE MARATTIALES 
 
 relation of the leaf base and stem apex, especially in connection with the Ophio- 
 glossiim-WkQ venation in Kaulfussia, is not without significance in its bearing upon 
 a possible relationship between the Marattiaceae and the Ophioglossaceae. 
 
 In all of the genera the lower part of the cotyledon and sometimes the stem 
 apex are more or less covered with hairs and scales. I hese are less abundant in 
 Marattia and Angiopteris, where they are, for the most part, short hairs whose 
 terminal cells contain a great deal of tannin and stain very strongly. In Kaulfussia, 
 and especially in Dancea, conspicuous scales of peculiar form are abundant. 
 
 ^"^^"^^^^ 
 
 A. Section near base of lamina of cotyledon. X75. D. An older stoma. X200. 
 
 B. Part of the lamina, jf, stoma. X200. E. Cross-section of a peltate scale, from young cotyledon. X200. 
 
 C. Young stoma. 
 
 In Marattia dougJasii the nearly cylindrical petiole is somewhat channeled on its 
 inner side and the single and nearly centrally placed vascular bundle appears almost 
 circular in section. The xylem forms a somewhat crescent-shaped body, completely 
 inclosed by the phloem, but the phloem is less developed upon the inner side and 
 the bundle closely approaches the collateral type. Tannin cells occur, but are not 
 so conspicuous as they are in Angiopteris. The lamina of the cotyledon is similar 
 in structure to that of the latter, the leaves differing mainly in the much smaller 
 development of the mesophyll, which is often reduced to a single layer of cells (fig. 
 
 127,5). The smaller veins have 
 the xylem reduced to a few — 
 one to three — small tracheids 
 situated upon the upper side of 
 the bundles, so that the bundles 
 are strictly collateral. Stomata 
 of the ordinary form occur upon 
 the lower surface of the leaf and 
 these sometimes show a series 
 of concentrically arranged cells 
 about them. 
 
 The petiole of the cotyledon 
 in Angiopteris resembles in sec- 
 tion that of Marattia, but it is slightly winged (fig. 130), and the tannin cells adja- 
 cent to the vascular bundle are more conspicuous than in Marattia. 
 
 In Kaulfussia (Rg. 137, C), where the petiole is stouter than in the other genera, 
 the section is more nearly circular than in Angiopteris, but it is also slightly winged, 
 these wings being developed almost to the proportions of stipules at the contracted 
 
HK VOUNG 
 
 151 
 
 basal portion, llie vascular bundle lias a greater development of xyleni than in the 
 other genera and the xyleni forms a band of considerable extent in the oval section of 
 the bundle. The structure of the bundle is concentric, but, as in other cases, the 
 phloem is better developed upon the outer side. Below the free base of the cotyledon, 
 where the trace enters the stem, there is a reduction of the inner phloem, or the 
 inner phloem entirely disappears, and the bundle there becomes truly collateral. No 
 definite endodermis can be recognized and the tannin cells found in the cor\le(lon 
 of Angiopteris are quite absent. 
 
 A. Section of petiole oi 
 
 B. The vascular bundle. X200. 
 
 C. Section of a peltate scale. X75. 
 
 The section of the petiole of the cotyledon in Danaa (fig. 131) is nearly cir- 
 cular, and the wings found in the other forms are almost entirely obliterated. The 
 ground tissue is composed of undifferentiated parenchyma, with no trace of tan- 
 nin cells. The vascular bundle is circular in outline, with a small group of three 
 or four tracheids placed somewhat toward the inner side of the bundle, but sepa- 
 rated from the outside by about two rows of cells, probably phloem, but not 
 showing any differentiation. Small protophloem elements, staining darkly, can 
 be seen upon the dorsal side of the bundle and an occasional similar cell can be 
 
 A. Two longitudinal sections of a somewhat older sporophyte of A/, doiigjaiii. 
 
 B. Apical region of stem, showing apical cell and second leaf. /". X2oa 
 
 seen toward the ventral side, but most of the ventral surface is destitute of these. 
 1 he petiole broadens at its base, becoming flattened, and in this region the inner 
 phloem disappears and the bundle becomes collateral, as it is in Ophioglossum. The 
 tracheary tissue of the cotyledon in all of the Marattiacerr is composed of slender, 
 reticulately marked tracheids, like those of Ophioglosstim. Sieve tubes are prob- 
 
152 THE MARATTIALES 
 
 ably present in the phloem, but in material mounted in Canada balsam the sieve 
 tubes are very poorly differentiated. 
 
 Upon the cotyledon as well as upon the later leaves in Dnna-a very character- 
 istic peltate scales are developed. These arise as single club-shaped cells, the 
 swollen end of" which is cut off by a cross wall. In the upper cell, which broadens 
 rapidly, a series of vertical walls arises, sometimes arranged very regularly (fig. 
 129, E). In the case figured there had evidently been two median, intersecting walls, 
 so that the cells were arranged in a somewhat quadrant fashion; more commonly 
 the divisions are less regular and the cells increase considerably in size by 
 further divisions. A section of an older scale may be seen in fig. 131, C. Stomata 
 are developed upon the lower surface of the cotyledon in all of the Marattiaceae. 
 These have already been referred to in the case of Marattia, and in Angiopteris 
 they are very similar to those found in Marattia. In Dancea, previous to the 
 formation of the stoma, there are several preliminary divisions, the cells being cut 
 
 off in a spiral fashion, the last division resulting in the formation of the mother eel 
 of the stoma (fig. 129, C). The complete stoma appears, therefore, surrounded by 
 a series of somewhat concentrically arranged accessory cells (fig. 129, D). Similar 
 accessory cells, but much less developed, can be recognized in Marattta also. 
 
 The stomata of Kaiilfussia have long been known for their great size and the 
 fact that they form permanently open pores, a peculiarity already developed in the 
 stomata of the cotyledon. These are very much larger than those of the other genera 
 and can be readily recognized by the naked eye as little dots. The guard cells 
 become very large (fig. 128) and strongly curved, so that the stoma appears almost 
 circular in outline, with a very large permanently open central pore. This large 
 pore is surrounded by concentrically arranged series of cells, suggesting that the 
 mother cell of the stoma is cut off very much as in Dana-a, but material was want- 
 ing for a thorough investigation of this point. 
 
THE YOUNG Sl'OROPHVTI 
 
 153 
 
 THE STKM or THE VOINt; SPOKDPHVTE. 
 
 The stem apex projects very little or nor at all, and occupies but a limited area 
 close to the base of the cotyledon. Its outer surface is nearly horizontal, except in 
 Angtoptcris, where it is generally strongly inclined (fig. 134). I here probably is 
 always a single apical cell. Segments are cut off from the sides, and very often also 
 inner segments, where the base of the apical ceil is truncate. 'Ihe lateral segments 
 also have, cut off from their inner faces, cells which contribute to the growth in length 
 of the stem, which remains, however, very short in the young sporophyte, the leaves 
 being very much crowded together. As we have already stated, no central cylinder 
 properly belonging to the stem can be demonstrated at this time, all of the procambial 
 
 , Stem apex, showing single apical cell. 
 . Second leaf, /''.of same sporophyle. 
 
 Nearlv median sect 
 /-, second le.if; m, r 
 Inphyte. X about ; 
 
 of a young sporophyte of Ktiulfuss 
 ilage ducts; at, stem apex; fit 
 
 tissue belonging either to the cotyledon or to the root. In the older embryo of 
 Marattia douglasii the apex of the stem (fig. 139. B) is occupied by a group of rela- 
 tively large cells, which at first sight seem to be pretty much alike; but a careful 
 examination of these makes it probable that one of them may be considered as the 
 real initial cell. This, in cross-section, appears almost square and shows a fairly 
 regular segmentation. While it must be said that the segmentation is not so regular 
 that one can assert without (juestion that all of the meristem is really due to the 
 activity of a single cell, I believe that this is probably the case. 
 
 My own investigations of the stem apex in Angtopteris have led to the conclusion 
 that there is a single initial cell here also in the young sporophyte, although Farmer 
 
154 
 
 THE MARATTIALES 
 
 in his earlier studies of this plant failed to find a single initial cell; however, in a 
 more recent paper (Farmer 3) he has decided that a single apical cell is present in 
 the young sporophyte of Angiopteris, and he gives figures of this. In longitudinal 
 section (fig. 134) the apical cell appears oblong, with a markedly truncate base, 
 while in transverse section it is imperfectly triangular in outline. The apical cell 
 of the stem in Kaiilfussia (fig. 136, C) is usually somewhat broader than it is in 
 Angiopteris, and closely resembles that of Ophioglossum nioluccanum. It has a 
 broadly truncate base and is somewhat narrowed above. In transverse section it 
 approaches the triangular outline of the apical cell in Ophioglossum, but the seg- 
 mentation is less regular and in this respect it is not unlike that oi Angiopteris. The 
 stem apex was examined in three species of Datia-a, all of which seemed to agree 
 
 A, B. Two transverse sections of a young sporophyte, above 
 level of stem apex, showing closed sheath formed by 
 base of cotyledon. /^, second leaf. X50. 
 
 Transverse section of petiole of cotyledon. X 
 Vascular bundle of cotyledon, more enlarged. 
 
 closely in the form of the apical cell, which is much more definite in Danaa than it 
 is in any of the other genera. Brebner's account and figure of the apical cell in 
 D. simplicifolia agree closely with my own preparations of Z). eUiptica, D. jerimani, 
 and D. jamaicensis. 
 
 The apical cell (fig. 140, A), seen in longitudinal section, is very deep and, 
 unlike that of the other genera, is not usually, at least, truncate below, but pointed, 
 so that it closely resembles in form that of Botrychinm or HeJmintbostachys. In 
 transverse section it also usually appears triangular, so that it is tetrahedral like 
 that of Botrychinm. While the segmentation is somewhat less regular than in 
 Botrychinm, nevertheless the limits of the younger segments can be followed without 
 much difficulty, especially in longitudinal sections, although the exact sequence of 
 the divisions in the segments themselves did not seem to be absolutely regular. 
 
 In all of the species the tissue derived from the inner cells of the segments of 
 the apical cell, or cut off from the base of the apical cell itself, remains as undif- 
 ferentiated parenchyma and contributes only to the central pith of the stem. I^'xcept 
 
THE YOUNG SPOROPHYTF. 
 
 155 
 
 that the leaves arise from the segments of the apical cell of the stem, and their 
 tissues are therefore indirectly derived from the stem apex, we may say that the stem 
 apex takes no part in the differentiation of the fibro-vascular system, which in the 
 young sporophyte is composed entirely of the leaf and root traces. 
 
 Only a small number of sections of the primary root were obtained in Marattia 
 donglasii, and these showed some variation, so that it is not possible to state posi- 
 tively what is the typical form of its apical cell, but there is no question that the 
 primary root grows from a single initial cell, as in the other genera. In longitudinal 
 section it appears oblong, with a broadly truncate base from which segments arise, 
 as well as from the lateral faces. There seem to be regularly four sets of lateral 
 
 .\. Nearly median section of a young sporophyte of DatnEa jatt 
 
 B. Another section of same, passing through second leaf, /" 
 
 C. Section of another similar sporophyte cut at right angles to that shown 
 
 of bundles of cotyledon and primary root. /)r, gametophyte; tc, scales. 
 
 segments, although cross-sections of the apical cell sometimes appear almost trian- 
 gular. The root cap is formed in partffrom segments derived from the outer face 
 of the apical cell, in part from similarjsegments cut off from the outer portion of 
 the youngest lateral segments of the apical cell. The central cylinder of the root 
 is formed in part from the basal segments of the apical cell, but the lateral seg- 
 ments also contribute to it (fig. 141, D). 
 
 In Angiopteris (fig. ill, C) the apical cill in rhe earlier stages of the root may 
 appear triangular in longitudinal section, bur in rht- later stages it is usualh' moie or 
 less truncate. 
 
 The primary root in Kuulfussni, in the few cases where satisfactory sections 
 were made of it, did not show the tetrahedral apical cell, but this was truncate at the 
 base and in cross-section appeared four-sided (fig. 142). In this respect it seems to 
 approach M/irnttia. The primary root in all of the Maratriace:f is ordinarily diarch; 
 
156 THE MARATTIALES 
 
 the Statement (Campbell 3, page 14) that in Morattia douglasii the primary root 
 is tetrarch I have found to be based upon a mistake, the root which was so de- 
 scribed not being the primary root, but evidently one of the later ones. I have 
 
 occasionally found very strong plant- 
 lets in Dancea, where the primary 
 root was triarch, as it sometimes is 
 in Botrychium, and very commonly is 
 in Helminthostachys; but such spe- 
 cimens are not common. I he root 
 bundle is bounded by a conspicuous 
 endodermis, whose cells show very 
 plainly the characteristic radially 
 thickened bands (figs. 143, en; 144). 
 The endodermis gradually loses its 
 definite appearance at the junction 
 of the root and leaf bundles, and I 
 have not been able to recognize it 
 above the level of the base of the 
 dousia^ii. cotyledon, although it is not at all 
 impossible that with proper staining 
 it might be traced further up. 
 
 THE SECOND LEAF. 
 
 The second leaf begins to develop while the cotyledon is still small, and by the 
 time the latter ruptures the overlying prothallium tissue the second leaf is clearU' 
 evident as a flattened cone, closely resembling the early stage 
 of the cotyledon. In KaulfussKi the second leaf arises quite 
 close to the cotyledon and on the same side of the stem apex, 
 so that the dorsiventral character of the stem is already indi- 
 cated at this early stage of development. In the other genera 
 the second leaf arises nearly opposite the cotyledon, but not 
 
 Fig. 139. 
 Transverse section of a young sporophv 
 
 sr, stem apex. X150. 
 The stem apes. X360. 
 
 exactly so, and the subsequent leaves form a spiral, and for a time at least, except 
 in Kaiilfiissia, the anatomy of the shoot is radial. This radial arrangement is 
 retained permanently in Mnrattia and Atigiopteris and also in some species of 
 
Till-; voi'N(; spoKoi'inri: 
 
 157 
 
 Ddihiti, i:. ^., I), rlli/'t/tii; Inir in D. ifinnuin ;mcl in /). jininii ,i)isi s, ws wvW ;is in 
 most of the other species, the stem hii;ill\- heeonus dorsiventral. ;is it is from the 
 first in Kaiilfiissia. 
 
 '\he development of tlie seeond leaf is marked h) an increased activit) in the 
 apical meristem, (jiiite close to the apical cell, and it is clear that the leaf arises from 
 the younijer sejjments, hut it is impossihie to state exactly what the relation of the 
 xoimij leaf is to the yoimj^er seijments of thi apical cell, or to say whether the whole 
 
 A-C. Three transverse sections o( primary root of Marattia douglwii. 
 D Longitudinal section of apex of primary root. 
 
 of the loaf is the derivative of a single segment. When the young leaf can first be 
 distinguished it is elevated very slightly above the level of the stem apex and its 
 growing point is composed of a group of columnar cells. Sometimes, especially in 
 DaiKra, there may be recognized near the center of this group of cells a single, 
 somewhat larger one, which may be interpreted as a single apical cell, but this is not 
 always recognizable. 
 
 A, longitudinal, B, 
 C. Apex of second 
 D Transverse sect 
 
 Kven before there is any noticeable projection of the leaf rudiment, there is 
 developed beneath it the beginning of the short strand of piocambium which is to 
 form its vascular bundle, and this can be followed downward to a point below the 
 stem apex, where it joins the trace from the cotyledon. Above this junction there is 
 no evidence of procambium tissue extending into the stem apex, but the whole 
 tissue lying between the junction of these two primary vascular bundles and flu- 
 apical meristem of the stem is made up of undifferentiated parench\ma. 
 
 The outer surface of the stem apex may be nearly horizontal or (especially in 
 Angi(jpteris) it may be strongly inclined and a single fairly conspicuous initial cell 
 
158 
 
 THE MARATTIALES 
 
 may usually bt recognized without difficulty. Iliis ceil is usually quite deep and 
 in all of the genera except Dantea has a truncate base in longitudinal section. In 
 Daticea, however, the base is usually pointed and the cell closely resembles the corre- 
 sponding cell in Botrychtum, or the typical leptosporangiate ferns. The growth of 
 the stem is usually rather slow and the leaves of the young plant are much crowded. 
 
 In all of the forms e.xamined, the tissue derived from the inner cells of the 
 segments and from the basal segments of the apical cell, where such segments were 
 formed, remained as undifferentiated parenchyma and contributed only to the 
 internal pith of the stem. Except that the leaves arise from the segments of the 
 apical cell of the stem and all of their tissues are therefore indirectly derived from it, 
 the stem apex of the young sporophyte takes no part at all in the development of 
 the fibrovascular system of the stem, which, except for the bundles of the roots, is 
 entirely composed of the leaf traces. 
 
 The second leaf in Angiopterts usually shows a pretty well-defined but slender 
 midvein, which may, however, fork at the apex, and this is true also of the third leaf. 
 It is in the third leaf, both in Marattia and Angiopteris, that the characteristic 
 stipules are first developed. These appear as lateral wing-shaped bodies close to the 
 
 A. SectioD of primary root of Dai 
 
 B. The vascular bundle. X175. 
 
 C. Triarch bundle from a primary 
 
 jainatcensis. 
 cndodermis. 
 )t of D. elliflii 
 
 base of the petiole, partially inclosing the next youngest leaf and the stem apex. 
 The third leaf in Angiopteris and several others succeeding it assume a more and 
 more pronounced lanceolate form with a prominent midrib and pinnately arranged 
 lateral veins which usually fork once. The margin of the leaf is serrate. After 
 several of these lanceolate leaves have been developed, leaves are formed in which 
 the base is provided with auricles, and these gradually pass into the next type of leaf, 
 which is ternate, the terminal leaflet being very much larger than the lateral ones 
 (plate 13, figs. 4, 5). These ternate leaves then pass by intermediate stages into 
 the pinnate form of the leaves of the adult plant. 
 
 In Kaulfussia the second and third leaves closely resemble the primary one, 
 except that they are somewhat larger (fig. 123). The second leaf develops well- 
 marked stipules which are found in all of the later leaves. The earlier leaves do not 
 show a definite midrib, though there is usually a delicate midvein at the base of the 
 lamina, but this vein forks usually about half-way up. In the older leaves (fig. 123, 
 C) there is developed a stout midrib, from which extend distinct pinnately arranged 
 lateral veins, much as in Angiopteris, but these lateral veins are connected by a 
 system of anastomosing veins, so that the leaf most strikingly resembles that of a 
 typical Dicotyledon. It is not until a very late period that the leaves in Kaulfussia 
 assume the compound form of the older plant (plate 11, fig. i). 
 
THF YOUNG SPOROPHVTF 
 
 159 
 
 In Diuhi-a a dozen or more kavcs may be tormeil before there is any brancliinij 
 of tile lamina. While the cotyledon may show an approach to the pinnate venation 
 of the older leaves, there is almost always a dichotomy of the apex of the slender 
 midvein, and this is generally true also for a number of the subse(]uent leaves, 
 although in these later leaves the midrib becomes well marked and the lateral veins 
 are developed apparently monopodially. In D. siniplicifolid the leaves remain 
 permanently unbranched, but in all of the other species the later leaves are once 
 pinnate. The intermediate leaves have the ternate form found in Marattia and 
 Aiigiopteris, and after a few of these, about two or three, have been formed, five 
 foliate leaves appear and the number of leaflets gradually increases, as new leaves 
 are formed, until the full number of leaflets is developed. The early ternate leaves 
 often have the terminal leaflet much larger than the lateral ones, and the leaf 
 
 closely resembles the corresponding stages in Augiopteris. The petiole in most 
 species is more or less obviously winged. In Daua-a the stipules are (sometimes, at 
 least) not formed until the fourth leaf. In the fourth leaf (fig. 126, B), however, 
 they are well developed, appearing as two conspicuous wing-like organs with 
 serrate edges. 
 
 In all of the genera there is a root formed for each leaf in tlu' carh\i stages of 
 the sporophyte, but how late this continues was not determined. In the older 
 sporophyte more than one root may be formed for each leaf. In Kaulfitssia the 
 second root is developed earlier than in the other genera, beginning to develop while 
 the second leaf is still very small, and in general the roots in K aid fuss in seem to 
 develop earlier than in the other genera. The second root is especially late in 
 developing in Dmura, the very first rudiment of the root making its appearance 
 between the first and second leaves, after the latter is aliead\ (uiire will advanced. 
 
160 THU MARATTIALKS 
 
 111. IHK OLDER SPOROPHVIK. 
 
 Ihe vasculai s) stem ot the adult sporophyte in the Marattiace;e is extremely 
 complex, especially in the larger species of Marattui and Angioptnis. In Dauini 
 and Kniil/iissi/i it is simpler, but in these also the arrangement of the bundles in the 
 adult stem is by no means easy to trace. A thorough study of the condition in the 
 early stages of the sporophyte is therefore specially important for understanding 
 the much more complex arrangement of the bundles in the older plant. 
 
 In view of the conditions that we have found to exist in the ( )phioglossace;e, 
 and particularly the nature of the vascular system in the young sporophyte of 
 Ophioglossiim nioluccaniini, a careful reinvestigation of the developmental history 
 of the vascular system in the young sporophyte of the Marattiaceae, especially the 
 simpler and presumably more primitive genera Dancta and Kaulfussia, was very 
 much needed. To this end series of sections were made of young sporophytes of 
 both of the latter genera, ranging from individuals in which the second leaf was 
 still undeveloped to those in which a dozen or more leaves had been formed. These 
 series were fairly complete and made it possible to trace quite satisfactorily the evo- 
 lution of the vascular system in the young sporophyte of both Dancra and Kaulfussia. 
 Brebner (Brebner 3) has given an accurate account of the distribution of the bundles 
 in the young sporophyte of Damea simplictfolia, which agrees quite closely with the 
 arrangement of the bundles as I have found them in the species studied by me. 
 Jeffrey ( Jeffrey 3) has described and figured a few specimens of D. alata ( .?), but 
 his material was too incomplete to make the study at all satisfactory. 
 
 The available material of Marattia was much less complete, although a number 
 of stages were secured. Farmer has made an exhaustive study of the vascular system 
 in the young sporophyte of Angiopterts (Farmer 3), and an extended study of Aii- 
 giopteris did not seem necessary, therefore, except for a study of the early stages to 
 determine the origin of the vascular bundles. All of the writers mentioned con- 
 cerned themselves almost exclusively with the fully developed bundles, beginning 
 their study at the base of the stem and working upward. No attempt was made to 
 follow out the development of the vascular bundles from the meristematic region 
 of the apex and proceeding downward,- a method of study which we think would 
 have materially modified the conclusions drawn by these observers in regard to the 
 relation of the different parts of the vascular system. 
 
 In my studies as to the origin of the vascular system in the young sporophyte 
 of the Marattiaceae the course of the bundles has been carefully followed in series 
 of microtome sections, passing downward from the apical meristem to the base of 
 the sporophyte, where the axial primary bundle passes into the primary root. These 
 series of transverse sections were then compared carefully with series of longitudinal 
 sections of young plants of approximately the same age, and in this way a clear 
 conception was obtained of the arrangement of the vascular system within the stem 
 of the young sporophyte. 
 
 THE DEVELOPMENT OF THE VASCULAR SYSTEM IN DAN^,A. 
 
 Especial attention has been given to the development of the vascular system in 
 Dana-a and Kaulfussia, as these have received less attention than Marattia and 
 Angiopteris, and moreover are probably more primitive forms whose structures have 
 departed less from the ancestral type than those of either Marattia or Angiopteris. 
 
 A longitudinal section of the young sporophyte of Dancca jamaicensis at the 
 period when the root is just about to emerge is shown in fig. I2i, A. The root in 
 this species, as in D. simplicifolia, described by Brebner, is the first organ to appear 
 
161 
 
 oiitsiik- tin- |ii(itlialliiim; the (.-()t\ Iidoii ;it this time is still very small ami thi' lamina 
 is scarcely developed. The root apex does not show in the section figuied and all 
 that can be seen is the upper portion of its stele, which joins the two young vascular 
 bundles that extend respectively into the cotyledon and the second leaf, which can 
 already be recognized. The greater part of the young sporophyte is made up of 
 large thin-walled cells. Figure I2i, li, shows a median section of the young cotyledon 
 from this same embryo, which is strongly bent over the stem apex. It is possible 
 that the square terminal cell, which in the drawing has the nucleus indicated, is the 
 apical cell of the leaf, but this is not at all certain. The outer part of the cotyledon 
 is made up of large thin-walled cells like those which compose the bulk of the 
 sporophyte, but at its apex there is active cell division, and traversing the young 
 cotyledon and placed somewhat toward its inner side, can be seen the primary 
 vascular bundle, made up of elongated procambium cells. This bundle can be 
 easily followed down into the body of the sporophyte, where, as we have seen, it is 
 
 sections of a very young sporophyte of Danxa jamaicensii. ) 
 A shows cotyledon apex; C, apices of the stem and second leaf; G shows 
 
 without interruption into the stek 
 
 the 
 
 lb 
 
 ipex, 
 
 which does not show in the figure, is in a slight depression close tt) the base of the 
 cotyledon and shows, as usual, a single deep apical cell. Nearly opposite the cotyledon 
 is the beginning of the second leaf (/ "); this is shown in median section in fig. 121. B. 
 Ihe group of narrow columnar cells that compose its apex does not show a recog- 
 nizable single initial cell. A short distance below its apex, the procambiimi of its 
 vascular bimdle becomes differentiated and is continued downward until it joins 
 the bundle from the cotyledon. The tissue arising above the junction of these two 
 bundles and continuing upward into the apex of the stem is composed of absolutely 
 undifferentiated parenchyma and there is no evidence at all of an\- stelar tissue in 
 the stem above the junction of these two primary leaf traces; indeed, the stem apex 
 lies decidedly on one side of the plane which traverses the vascular bundles of the 
 two leaves. Fig. 145 shows a number of cross-sections from a series taken from a 
 large embryo of Daneta jamaicensis, before the emergence of the cotyledon (//), 
 11 
 
162 
 
 THE MARATTIALEt 
 
 whose apex is still undivided. C shows the apices of the stem and second leaf, the 
 former being shown on a larger scale in G. The section of the cotyledonary bundle, 
 lying to the right of the stem apex, is plainly evident in section C. D shows the 
 separated bundles of the cotyledon and very young second leaf, which lower down 
 unite into a single strand. 
 
 Whether we can properly speak of a stem at all in connection with the tissues 
 composing the central region of the young sporophyte is questionable. It is per- 
 fectly evident that this central region is composed partly of tissue belonging to the 
 base of the cotyledon, which is in no sense to be looked upon as an adjunct of the 
 
 -C. Three sections near the stem apex, which is shown 
 and more highly magnified in D. Fig. C is revers 
 position with reference to A. 
 E. The central region of C. 
 
 F-J. Show the condition of the Vi 
 
 region of the sporophyte. 
 
 H, I. Are in the transition region. 
 
 J. Shows the root stele. X200. 
 
 stem, and partly of the root base; the two structures, i. e., the bases of the root and 
 the cotyledon, merge insensibly into each other. To this tissue, derived from the 
 root and the cotyledon, must be added a large amount of superficial tissue derived 
 originally from the foot of the embryo. In speaking of the axis of the young sporo- 
 phyte as a "stem" it must therefore be borne in mind that this stem is really a 
 composite structure, partly leaf, partly root, and partly foot. 
 
 Fig- ^3^y ^ '>"<i ^' shows two longitudinal sections of an older stage after the 
 emergence of the cotyledon. A passes through the stem apex and it is evident 
 that there is no vascular bundle extending into the stem. The tracheary tissue is 
 now well developed in the leaf trace. B shows the trace from the second leaf 
 at its point of junction with the bundle from the cotyledon. In fig. 138, C is a similar 
 
Till- OLDKR SI'OROPH^rr. 153 
 
 stasif to ihat shown in ./, hut tut at rit^lit angles to it, so as to sliow tlic continuity 
 oftliL- Inuullcs ot the cotyledon and ]m imary root, which together constitute the prinia- 
 ly axial vascular bundle of the sporophyte, as they do in Op/iioglossum nioluccaiuitii. 
 ihe similarity of such a section as that figured to the similar stage in O. rnoliirr/uiimi 
 is suHicicntly obvious. The stem apex is shown in fig. 140, .7, and the lutliment of 
 the second leaf is shown in fig. 140, B. 
 
 F'igure 146 shows ten transverse sections from a series made from a young sporo- 
 phyte ofi). jainaiccnsis of almost exactly the same age as that shown in fig. 138, B. 
 A is cut above the level of the stem apex and passes through the cotyledon and apex 
 of the second leaf. The arrangement of the cells at the apex of the latter suggests 
 the possibility of a single apical cell. The base of the petiole of the cotyledon is 
 much flattened and is made up of large, thin-w^illed parenchyma with no traces of 
 tannin cells or mucilage ducts. The vascular bundle is oval in outline and decidedly 
 collateral in structure. There is no recognizable endodermis, and I shall not 
 attempt to say whether the cells immediately surrounding the bundle should be 
 spoken of as endodermis or pericycle. Except for two small tracheids on the inner 
 side of the bundle the vascular bundle is composed of delicate tissue, apparently 
 parenchyma, but some of the larger elements undoubtedly represent sieve tubes. 
 Brebner states that the sieve tubes in D. stmplicifolia (Brebner 3, page 524) do not 
 show clearly in sections mounted in Canada balsam, but when examined in other 
 media they show thick, glistening walls and pitted areas. This point was not further 
 examined in the species under consideration, but would probably agree with the 
 observations made by Brebner upon D. simplicifolia. 
 
 Section B passes through the stem apex, whose apical cell is shown on a larger 
 scale in D. The cotyledon trace in this section shows three tracheids. In D the 
 section is made a short distance below the stem apex and shows the second leaf trace 
 bending in to join the trace from the cotyledon. The central part of the section is 
 shown more enlarged in E, and it is perfectly clear that the region immediately 
 below the apex and lying between two leaf bases shows nothing which can be 
 interpreted as a cauline stele. F and G show sections taken still further down, 
 illustrating the coalescence of the two leaf traces to form the single axial stele. In 
 G there is seen the first tracheid belonging to the leaf trace from the second leaf, 
 and in H the central bundle is complete and there are seen the two xylem masses 
 corresponding respectively to the two leaf traces. One of these, belonging to the 
 primary leaf trace, is more developed than the other one. The larger xylem mass 
 in this case had five tracheids and the smaller one three. / and / are transitional 
 stages between the bundle of the cotyledon and the root, and, proceeding down- 
 ward, the bundle becomes somewhat smaller and there is a reduction in the number 
 of tracheids, which, however, are plainly seen to be continuous in their development 
 with the two .xylem masses in the central bundle of the stem, formed by the junction 
 of the two primary leaf traces, which maintain their identity until they merge into 
 the two xylem masses of the diarch root. Within the root itself the endodermis 
 surrounding the bundle is very conspicuous. The cells making up the endodermis 
 are of large si/e and the characteristic markings upon their radial walls are very 
 easily .seen. The limits of the endodermis become much less clear in the transi- 
 tional region, and higher up it is difficult to make out a clearly defined endodermis. 
 Brebner figures a very irregular endodermis in the stem of D. siriipliiifolia, and 
 perhaps we may assume that a proper endodermis is present also in the young stem 
 of D. jamatccnsis, but it is certainly very inconspicuous, to say the least, in the central 
 region of the young sporophyte, in strong contrast to the very conspicuous endo- 
 dermis in the root. Farmer (Farmer 3) was able by treatment with sulphuric acid 
 
164 
 
 THE MARATTIALES 
 
 to demonstrate an external endoderniis in the stem bundles oi Angiopteris, Marattia, 
 and Kaiilfussia, and presumably similar treatment would show its presence in the 
 young s em of Daiicea. 
 
 It is evident that we can not speak of the young sporophyte in Dancea jamai- 
 censis as possessing in its early stage a "protostele," or "haplostele," to use Brebner's 
 terminology, unless we choose to call the single leaf trace of the cotyledon a protostele. 
 Just so soon as the two first leaf traces unite, the axial vascular bundle of the young 
 sporophyte has two quite independent xylems. 
 
 Figure 147 is a longitudinal section of a sporophyte of D. cUiptica, slightly 
 older than that of D. jamaicensis shown in fig. 138. The second leaf (/^) had already 
 begun to develop the lamina, whose apex was forked so that there were two veins 
 present. A longitudinal section of one of the two leaf lobes is shown in D. The 
 stem apex closely resembles that of D. jamaicensis, the apical cell having much the 
 same form and size. The stem apex does not lie in the same plane as the leaf traces 
 
 A. Nearly medi 
 
 origin of second 
 
 B. Central region. > 
 
 ¥u,. n-j.— Dait^a elliptii 
 section of a young sporophyte, showing 
 
 C. Median section of second leaf. 
 
 D. Apex of second leaf. X150. 
 
 E. Stem apex. X150. 
 
 and hence does not show in the figure, which is a section passing directly through 
 the leaf traces. This plant showed very clearly the very earliest stage of the second 
 root (fig. 147,//, r). The apical cell of the young root is cut out from one of the cells 
 of the parenchyma below the stem apex, lying almost exactly half-way between the 
 two primary leaf traces. Only a single segment had been cut off from the apical 
 cell in this case and both the apical cell and this primary segment were easily dis- 
 tinguished from the surrounding tissue by their denser contents. There is nothing 
 to indicate that the cell which had assumed the function of the root initial was in 
 any way essentially different from the neighboring cells. If it is an endodermal cell 
 there is no way of recognizing this, as the endoderniis could not be recognized in 
 this part of the sporophyte. Figure 147, B, shows an enlarged view of this young 
 root initial, lying above the junction of the two primary leaf traces. Conspicuous 
 reticulate tracheids with pointed ends are present in the trace from the cotyledon, 
 but as yet no permanent tissue has been developed in the bundle from the second leaf. 
 
THE OI.nr-R SPOROPHYTE 165 
 
 A somewhat older spoiophyte from D. janiaicensis is shown in fig. 148. In 
 this plant the third leaf was plainly visible and the second root had developed into 
 a short conical body which was pushing its way through the tissues between the 
 
 insertion of the first and second leaves. By this time tracheary tissue haS developed 
 in the bundle from the second leaf, and now for the first time we note the presence 
 of large tannin cells in the neighborhood of the vascular bundles. Three of these 
 are shown in fig. 149, lying next the bundle of the second leaf. The irregular ring 
 
 of the sporophyte shown in fiq. 148. 
 
 of narrow cells shown in the- figure and hing 
 point of insertion of the second root. Figure 
 of the apex of this second root. 'I"he apical 
 
 below tile a]iical meiistem marks the 
 149, /i, shows a nearh' median section 
 cell is nearly oblong in form and the 
 
 lateral segments are very large and periclinal segments which contribute to the root 
 
166 
 
 THE MARATTIALES 
 
 cap are cut off from them. The outer cells of the root cap are very much enlarged 
 and there is an evident space between the root and the surrounding tissues. The 
 trace from the third leaf joins that of the second one close to its union with the 
 bundle from the cotyledon (fig. 148, B). 
 
 In fig. 150 are shown nine cross-sections from a series made from a plant of D. 
 jamaicensis, in which four leaves were evident. A microscopic examination ot the 
 stem apex showed, however, that a fifth leaf was also present, which stood nearly 
 directly opposite the cotyledon. A shows sections only of leaves 3 and 4, the latter 
 having the stipules conspicuous, while these are absent or scarcely developed at all 
 on the first three leaves, at least in the specimen in question. Between the leaves 
 are seen sections of the numerous scales which beset the surface of the young leaves. 
 B is taken somewhat lower down and includes a section of leaf 5. C passes through 
 the stem apex and shows clearly the spiral arrangements of the first five leaves, 
 each of which has in its petiole a single concentric bundle which becomes larger in 
 
 ig sporophyte, with t1 
 apex. Roman figures 
 
 D. jamaicensis. 
 
 fully-developed leaves and two younger t 
 
 each successive leaf, with a corresponding increase in the development of the xylem. 
 As the sections are examined, farther and farther down in the stem {D to F), one 
 can see very clearly the way in which the single bundle in the lower part of the 
 stem owes its origin to the coalescence of the leaf traces. Proceeding downward 
 the traces of leaves 4 and 5 are seen to approach gradually and finally to become 
 completely coalescent; and still further down {F) the leaf trace from 3 joins that 
 from 4, and a single bundle results, crescent-form in section, but showing ckarl\ its 
 compound nature. 1 he three xylems never become entirely confluent. 
 
THE OLDKR 
 
 167 
 
 For a long time sections of the stem show this single central Inmclle of crescentic 
 form, at Hrst tleiived from the coalescence of the third, fourth, and Hfth leaf traces, 
 but continued upward in the same form and added to by the addition of the traces 
 from the subsequent leaves. I'his crescentic stele, which, for convenience may 
 be spoken of as the stele of the stem, is entirely of foliar origin. The crescent never 
 becomes completely closed and its opening in the earlier stages of development can 
 not be properly called a foliar gap. The parenchyma which is inclosed within its 
 curve belongs from the first to the ground tissue and is not part of the stele. Some 
 of the surrounding cells show traces of the typical endodermal markings and it is 
 perhaps safe to say that the stele is bounded by an endodermis, as Farmer states is 
 the case in Angtopteris and Brebner in D. simplicifolia. The limits of the endo- 
 dermis, however, especially upon the concave side of the stele, are very vague. The 
 stele, after the complete fusion of the three leaf traces, may perhaps best be described 
 
 Fig. 151. 
 
 Four longitudinal sectic 
 
 rth leaf, /*, has the stipules well dc 
 
 as concentric in structure, with phloem developed all around the xylem, but there 
 are probably traces of phloem also between the three xylems which represent the 
 three confluent leaf traces. 
 
 In the older portions of the stem the bundles become still more completely 
 fused and the compound bundle is oval in outline, but still shows plainlv the three 
 xylems of its constituent leaf traces (fig. 150, G). In this region the endodermis 
 is rather better developed than it is higher up, but its limits are still rather vague. 
 At this level the traces of the leaves, i and 2, are still free, but have approached 
 nearer to the central bundle than is the case higher up. 
 
 Still lower down, the trace from the second leaf joins the Iniiulle formed from 
 
 the later leaf traces, 
 proaches the conditi 
 using Brebner's teir 
 
 which no longer clearly shows the sepa: 
 11 whiili has been desciilnd as " pi dtusleli 
 iin()li)"\-. I he wiem tknunts. liu\\i\ei. 
 
 but 
 
 osteli 
 
168 
 
 THE MARATTIALES 
 
 core, but are more or less scattered, with thin-walled elements between. After 
 fusion of the second leaf trace with this is complete, the section of the stem shows 
 only two bundles, representing the two first leaf traces, and finally (fig. 150, H and /) 
 the section of the stem shows the condition already described for the sporophyte 
 with but two leaves. These finally merge into a single primary axial bundle (/), 
 which, followed downward, merges imperceptibly into the stele of the primary root. 
 Near the middle of these lower sections can be seen a section of the second root, 
 which pursues a downward course through the cortical tissue for a very long distance, 
 but finally emerges and grows downward, side by side with the primary root. The 
 third root (fig. 150, G, r") emerges much higher up and breaks through the cortex 
 at about the level of the junction of the three youngest leaf traces. 
 
 Fig. 153 shows the details of the central tissue from the same sporophyte as 
 that just described. J passes through the stem apex and shows the apical cell cut 
 somewhat obliquely, and near it the section of the trace from the fifth leaf. D is z 
 section of the fourth leaf trace from the same level. This shows but two tracheids 
 at a point near the inner limits of the bundle, while on the outer side there is a con- 
 spicuous curved line of protophloem cells. 5 is a section taken somewhat lower 
 down and shows the bending in of the leaf trace as it descends into the stem. This 
 is still more marked in the fifth leaf trace, which, at this level, has a crescentic form, 
 
 A. Apex of fifth leaf from sporophyte shown in fig. 
 
 B. Stem apex of same. 
 
 C. Base of third root, showing triarch bundle and r 
 
 but no permanent elements yet developed. It is probable that a portion of this 
 crescentic mass of procambium represents the sixth leaf trace. The fourth leaf 
 trace at this point shows three tracheids which are decidedly larger than those 
 higher up, and still lower down the tracheids increase still more in size and several 
 tannin cells can be seen in contact with them. The section of the fourth leaf trace 
 is connected with the large but undifferentiated trace of the fifth leaf by an isthmus of 
 procambium cells, so that the sections f)f the two bundles form the crescent-shaped 
 section, similar to that seen lower down (fig. 153, C). The tracheary elements of the 
 fused portion of the fourth leaf trace are noticeably larger again than those in the 
 free portion of the leaf trace. We now notice, for the first time, traces of the mucilage 
 ducts which afterwards become so conspicuous in the stem. These first appear in 
 section as rounded cells (m), and it is evident, as Farmer has shown in Arigiopteris, 
 that the earliest mucilage canals are of lysigenous origin, i. e., they are formed by a 
 
169 
 
 fusion of ctlls and not 
 
 splirtinii apa 
 ibout this intercellular spac 
 
 )t" the- (.(.lis and tin- dcvilopnicnt of" 
 In liis taiiicr paper Hrehner states 
 that the mucilage ducts are schizogenous, but in his later paper he admits that they 
 may he of lysigenous oiigin, which certainly is the case in the species of Datia'a 
 described here. 
 
 Ihe details of the older central bundles of the sections /• and (/', Hg. 150, are 
 shown in fig. 153, E, F: the shaded cells in 6' are tannin cells. 
 
 Figure 151 shows longitudinal sections of a young sporophyte of /^. elltptica, 
 of about the same age as the sporophyte which has just been described. This al.so 
 shows three fully developed leaves, while the fourth and fifth leaves are well advanced. 
 
 -Ql^i P 
 
 153.— Details of th< 
 
 lar system of sporophyte : 
 
 A. Passes through the stem apex; B and C, lower down 
 D. A single leaf trace. X150. E, F. Vascular bundles frc 
 
 ,mfig..50 
 showing the fourth and fifth le; 
 1 the lower part of the stem, m, h 
 
 young mucilage ducts 
 
 The central bundle of the stem formed by the junction of the three first leaf traces 
 is now conspicuous and forms a stout central strand, continuing downward into the 
 primary root (;'). // is a nearly median section, passing through the stem ape.x, 
 whose large and conspicuous apical cell is shown in fig. 152, B. The ape.\ of the 
 fifth leaf is seen just above this and a more enlarged figure of this is shown in fig. 
 152, A. The apex of the fifth leaf (which is here cut transversely) is occupied bv a 
 small group of large cells, of which the central one may perhaps represent a single 
 initial cell. Below the stem apex, but separated from it by a considerable amount 
 of tissue, can be seen the bases of the leaf traces from the fourth and fifth leaves 
 respectively. These are joining the central bundle of the stem near the junction 
 
170 
 
 THE MARATTIALES 
 
 of the third leaf trace, and at this point there may be seen the base of the third root 
 (r"). Between the fourth and fifth leaf traces is present a section of the first mucilage 
 duct. 5 is a section some distance to one side of the apex, also passing through the 
 primary mucilage canal and the fourth root, which arises between the fourth and 
 fifth leaf traces. Figures C and D are sections on opposite sides of the stem apex 
 from B. The fourth leaf, with its conspicuous stipules (st), showsjn these sections, 
 and the third root can also be seen. Fig. 152, C, is a more enlarged view of the 
 third root shown in fig. 151, D. The root is triarch and in the peripheral part of the 
 cortex there is visible a row of very conspicuous cells whose walls stain very strongly 
 with safranine and in section closely resemble large tracheary elements. A longitudinal 
 section of these cells, which are also conspicuous in the later roots (fig. 155), shows 
 that they are elongated sclerenchyma cells whose thick walls are conspicuously 
 pitted. Farmer (Farmer 3) observed similar cells in the roots of Jngiopteris. 
 
 Figures 154 and 155 show three longitudinal sections of a pretty well advanced 
 young sporophyte of D. jenmatii. The root (shown in the figure emerging at the 
 base) is probably the second root. The stem has already begun to assume a dorsi- 
 ventral form, and the basal part of the root, which is strongly cui-ved, is cut away 
 and so does not show in these sections. The earliest leaves are no longer recognizable 
 and it is impossible to determine just how many leaves have been formed. The 
 leaves first formed have the single trace characteristic of the first leaves in all of 
 the species, but the youngest leaves show that the leaf trace is double. In this sec- 
 tion (fig. 155, A) is shown a very young leaf, cut through parallel to its surface 
 and showing clearly the stipules at its base and also the young vascular bundles. 
 Within the leaf are two bundles, which are seen to join at its base, but separate 
 again lower down, the two bundles thus forming two distinct traces in the stem. In 
 fig. 154, B, the section passes through the youngest root, which is seen to have its 
 stele joined to one of the leaf traces from the youngest leaf. Several conspicuous 
 mucilage ducts are now present in the stem near the vascular bundles; these are 
 
THE OLDER SPOROPHYTE 
 
 171 
 
 formed of rows of large cells which become more or less fused tf)gether and their 
 contents, composed of a dense granular mucilage, stain very strongly, either with 
 safranine or with IJismarck brown, which were the principal stains used in preparing 
 the sections. Numerous tannin cells are also present, scatrcncl irregulari\- through 
 the cortical tissue and also in the vascular bundles. 
 
 Figure 155, B, is a median section through the base of one of the later roots, 
 showing its connection with the corresponding bundle in the stem. The elongated 
 cells {sc) are the cortical sclerenchyma cells already referred to. These are shown 
 on a larger scale in C and D, where the conspicuous pitted walls can be seen. 
 The shaded cells in the figure are tannin ceils. 
 
 Figure 156 shows cros.s-sections taken at different heights from a sporophyte of 
 D.jenmani, a plant of about the same age as that which was shown in the preceding 
 series of longitudinal sections. Sections made near the apical region, A, B, and C, 
 
 A. Another section of specimen siiown in fig. 154. 
 
 B. Base of a root showing sclerenchyma, sc, in cor 
 
 C. Sclerenchyma cells, showing pitted walls. X5: 
 
 D. Surface view, showing pits. 
 
 show two pretty well advanced leaves and the first trace of a third, the apex of which 
 is indicated in C and E. Unlike the earlier leaves with their single leaf tiace, these 
 later leaves, as we have seen, have the petiole traversed by two vascular bundles 
 whose sections are plainly visible. These, as we have remarked from a study of the 
 longitudinal sections, fuse into one near the base of the leaf (//), and two small 
 bundles are also present on either side, belonging to the stipules. As the sections 
 are followed downward, the main bundle is found to divide into the double leaf trace 
 entering the stem, and the two small stipide bimdles become uniteil with these, so 
 that within the stem, at about the level of the stem ape.x, two distinct bundles are 
 seen, corresponding to each leaf. Ihe leaves are now arranged in two series, 
 indicating that the dorsiventral character of the stem in this species is established 
 and the youngest leaf lies almost opposite the next xoungest oni\ whose leaf traces 
 can be seen. E shows the stem :i]iex with the apical cell. \. ami the \(umgest leaf 
 rudiment, /. 
 
172 
 
 THE MARATTIALES 
 
 Below the level of the stem apex, the stele of the axis closely resembles that 
 already described from the somewhat younger plant and forms a somewhat irregular 
 crescent, evidently composed of two portions which are joined on the side facing the 
 two leaf traces from the oldest leaf which shows in this section. The opening of the 
 crescent at this level is turned away from the leaf. No permanent elements are 
 present in the central stele as yet. Following this downward, the crescent is seen 
 to open on the side next the two oldest leaf traces and there is thus formed a foliar 
 gap, while on the opposite side the space becomes closed. In section D, which 
 is taken lower down, the position of the section of the central stele is reversed. At 
 
 A-D. Four transverse sections of a young sporophyt 
 
 E. The apical meristem with apical cell of stem, 
 
 F-I. Four sections from the same series, but taken 
 
 stele. »;, mucilage ducts. Between the i 
 
 • oi D.jenmani. X20. Section C shows stei 
 r, and a leaf initial, /. X180. 
 lower down. X20. Figs. F, G, show fusion 
 lucilage ducts in G can be seen the small i 
 
 this point the section of the central stele forms a crescent ot tissue, most of which 
 has not passed beyond the procambial stage. This crescent-shaped mass is evi- 
 dently composed of two parts and represents mamly the basal undivided bundles 
 of the two youngest leaves. Tracheary tissue has begun to develop in the older of 
 these two masses. Occupying the space within this crescentic stele are three young 
 mucilage canals, but none are yet developed in the cortical tissue of the stem. 
 Within this central space there can also be seen the section of a small isolated group 
 
THK OLDKR SI'OKOrinTr, 
 
 173 
 
 ;i\lal stele in loiiimissural j 
 itii Hvc leaves. The older k 
 Ininclle, e-resceiit-shaped in 
 
 laiul whuh hac 
 t base, shown ii 
 iLitlinc and eon 
 bundles belong 
 
 of procanibial cells. This is the singk' axial stele oi eonmiissural strand which had 
 
 not yet begun to develop in the plant \ 
 
 this section, has a single large vascula 
 
 centric in structure, and there are also seen the two small concentru 
 
 ing to the stipules. 
 
 All the leaf traces in the older leaves are concentric in structure, with well- 
 developed internal phloem, and thus differ from the collateral traces of the earliest 
 leaves. This concentric structure of the bundle is already well marked in the fourth 
 leaf, where, however, the protophloem is mainly developed upon the outer side. /• 
 shows the arrangement of the bundles in a section somewhat further down. I he 
 two strands from the oldest of the three leaves seen in the apex have joined the horns 
 of the crescentic central bundle, which still shows clearly its dual nature. In the 
 younger of the two portions of these sections the first tracheids have just appeared 
 near the free end, while in the older half a line of tracheids extends nearly through 
 the whole of the center. The mucilage ducts at this level are much larger and their 
 cells have become confluent. Close to the large mucilage canal can be seen the 
 
 
 A 
 
 Fig. 157. 
 
 A. Section of an older stem of Danaa jenmaitiy showing cortical iniicilaj;c ducts, w, and tcntral or 
 , commissural strand, c. //, leaf trace, r, root. X20. 
 
 B. Shows a section higher up, with free leaf traces and commissural strand. 
 
 section of the axial vascular strand, which somewhat higher up turns outward and 
 becomes fused with the inner face of the crescentic central stele. The leaf traces 
 soon become completely merged with the broad crescentic stele and their limits 
 can no longer be recognized. The foliar gap gradually closes up and the two por- 
 tions of the crescent separate at a point opposite so as to form a leaf gap, nearly- 
 opposite the first one; but I have seen no cases where a section of the central stele 
 appears as an unbroken ring, although it is possible that this might occur for a 
 brief period. Much the same condition of things is seen lower down and it is clear 
 that in all cases, after the double leaf trace has been developed, one bundle fuses 
 with an older and one with a younger leaf trace. The axial bundle shows a central 
 mass of tracheary tissue, and in somewhat older stages an endodermis is pretty well 
 developed. Finally this bundle can be traced downward to where it connects with 
 the main stele, composed of the united leaf traces. The main stele gradually passes 
 down into the base of the stem, where it shows the structure already described for 
 the younger sporophyte. 
 
 In older sporophytes the crescentic central stele becomes brokt n up perniaiu iul\ 
 into two elongated masses when seen in section (fig. 157), but the junction of the 
 
174 
 
 THE MARAITIALES 
 
 leaf traces with these takes phice exactly as in the younger plant. Fig. 157, B, shows 
 the separate leaf traces and the free axial strand, while in A the leaf traces are 
 becoming fused with the broad central strands and the axial strand is also anastomos- 
 ing with one of the latter. This has taken place at the point of junction with a 
 root. While in the younger sporophyte the mucilage ducts are confined to the central 
 region, in the older plant they occur also in the peripheral region of the stem (fig. 
 157, >n). The details of the vascular bundles are shown in figs. 158 and 159. 
 
 To recapitulate: The vascular system in the young sporophyte of Daiura 
 begins as a single axial strand, which is continuous through the cotyledon and root. 
 At a very early period a second vascular bundle or stele is formed in the second leaf 
 connecting with the primary strand, and this is followed by a similar single strand 
 or stele in each succeeding leaf, up to about the seventh. Up to this time, except 
 for the steles of the secondary roots, the whole vascular system is built up of united 
 
 leaf traces and there is no cauline bundle in the strict sense of the word, although 
 we may speak of the bundle, or stele of the stem, as soon as there is a solid central 
 strand formed below the junction of the earlier leaf traces. This primary stele 
 never has the form of a true protostele, however, as the xylems belonging to the 
 separate leaf traces can be recognized and the compound nature of this central 
 bundle is unmistakable. 
 
TIIK OLDKK Sl'OKOl'H'lTl' 175 
 
 At a later stage-, piohablv about the time that the seventli leaf is toinieii, theie 
 arises the single axial (coiiiniissuial) stiaiul. which is caiiline in its oiigiii aiul which, 
 as lirebner has stated foi- D. snti/>licifalia, is the only strictly caiiline portion of the 
 vascular system ot the stem. Ihis can first be seen at about the same time that the 
 tlrst double leaf trace appears. 
 
 The development of the vascular system was not traced fuither, as it has already- 
 been adequately described by Kiihn ( Kiihn 2), and except tor the increased number 
 of leaf trace and commissural strands the vascular system is essentially the same 
 in a young plant like that shown in fig. 157 as it is in the adult sporophyte. 
 
 Brebner's account of the development of the young sporophyte in Dan, in 
 simplicifolia agrees essentially with my own studies of D. eUiptica, D. jetmnnii, and 
 D. jamaicensts. He failed, however, to get successful sections of the transition 
 region between the root and the cotyledon, or, as he puts it, "between the root and 
 stem," and his statements that the central bundle begins as a "haplostele," i. e., 
 a protostele, is not borne out by my studies on the other species. Moreover, his 
 statement that "in this way the somewhat irregular adelosiphonic dictyostele of 
 Dantra is established now by a process which is, to all intents and purposes, a 
 branching of the haplostele, due to the departure of the leaf traces of closely set, 
 spirally arranged leaves," might better be stated, "the dictyostele arises by a process 
 which is essentially the fusion of the leaf traces of closely set, not necessarily spirally- 
 arranged leaves." He fully recognized the important part which the leaf traces 
 play in the building up of the dictyostele of the stem, but he did not apparently 
 recognize these as the sole factors in the development of the dictyostele of the older 
 stem. He does,^however, specifically state that the axial t)r commissural strand is 
 the only part of the vascular system which is strictly of cauline origin. 
 
 THI<; ADULT Sl^OROPHYTE OF DAN.ilA. 
 The anatomy of what was supposed to be the sporophyte of Dancea was first 
 described by Holle (Holle 2), but the later investigations of Kiihn showed that the 
 plant investigated by Holle could not have been a Dancea, nor could it even have 
 belonged to the Marattiaceae, since Holle describes the rhizome as developing a 
 sheath of sclerenchyma, a condition of things which has not been found to exist 
 in any Marattiaceae. Kiihn's account is very far from complete and the description 
 of the sporophyte given by Brebner (Brebner 3) fori), simplicifolia does not deal 
 with the adult sporophyte. The account given here is based mainly upon a study 
 of D. jamaicensts and D. elhptica, but an investigation of certain points was also 
 made in D. jenmani. The rhizome in D. jamaicensts and D. jenmani is a markedly 
 dorsiventral one. These two species are a good deal alike, but are readily dis- 
 tinguished by the difterence in the form of the leaves. In the former the leaves are 
 rather larger and have more numerous pinnae, while the terminal pinna is developed. 
 D. jctimant has somewhat smaller leaves with from seven to nine pairs of leaflets, 
 and the leaves are abrujitly ])innate, tlie terminal pinna not being developed ( plate q. 
 
 A. fig^ ,). 
 1 he 
 
 leaves are arranged very much as they are in llchntiittiostdilivs 01 in 
 Ophtoglossum penJtilnm, the fleshy leaf base being provided with two vei\ conspic- 
 uous stipules which are developed very early in the life of the sporoph\ tc. I'lic 
 exact number of roots corresponding to a leaf is not easv to make out, but there are 
 at least two, and it is (juite possible that the number may sometimes be greater. 
 These roots branch freely, especially toward the tips. 
 
 D. clliptica (plate lo) is a good deal larger than the other species and the stout 
 rhizome, which, including the persistent leaf bases, has a diameter of 3 centimeters 
 
176 
 
 THE MARATTIALES 
 
 or more, is upright, the leaves being spirally arranged as they are in Marattia and 
 Angiopteris. The petioles in this species are characterized by curious enlargements 
 or nodes, which may occur in the other species also, but are much less prominent. 
 The leaves of D. elliptica reach a length of about 50 centimeters and the leaflets, 
 which are much larger than those in the other species, are less numerous, there being 
 usually about eleven. The conspicuous nodose swellings seen in the petiole are 
 repeated on a smaller scale in the rachis at the base of each pair of leaflets. Sections 
 of the stems were examined in D. jamaiceiuis and D. flliptica, which, e.xcept for a 
 difference in size, agree closely in their structure. The ground tissue, as was first 
 pointed out by Kiahn, consists entirely of parenchyma, through which are scattered 
 many conspicuous tannin sacs which, in preserved material, appear to the naked eye 
 as numerous black specks. The section of the vascular cylinder looks very much 
 like that of the younger sporophytes already described, but the number of bundles 
 seen in the section is larger, and there is a group of .several bundles instead of the 
 single medullary strand seen in the section of the younger stem. In a section from 
 a. large specimen of D. elliptica in which the rhizome, including the old leaf bases. 
 
 was about 3 centimeters in diameter, the outer ring ot bundles containid thirteen, 
 while within this was a smaller circle composed ot five medullary bundles, three 
 large ones and two smaller ones. In a similar section of a medium-sized section of 
 D. jamaicensis, in "which the rhizome was about half the size of that in D. elliptica, 
 there were nine bundles in the outer ring and four in the middle, two large ones and 
 two small ones. In this section the junction of a root was cut through and the root 
 stele was seen to be applied by its broad base to one of the larger of the medullary 
 bundles (fig. 161, J). 
 
 The relation of this central group of bundles in the stem to the prinuu)- medullary 
 strand was not investigated. Brebner has shown that in the later stages of the 3'oung 
 sporophyte in D. simplicifolia the original medullary strand — which, as we have seen, 
 is really the only part of the vascular system which is strictly of cauline origin — 
 fuses at certain points with the bundles of the outer ring, and at this point of fusion 
 there may be a liranching so that a section above this shows two of these medullary 
 bundles. Whether all of the central group of bundles seen in the stem of the adult 
 
/;. 
 
 177 
 
 (1 D. rlliptuii aif line to iuitlui l)i aiuliinj; of tlic 
 
 uiline 
 
 primary incdiillary strand, or whether some of thtin may be new bundles 
 oiigin, remains to be investigated. 
 
 The arrangement of the characteristic mucilage ducts is the same in the two species 
 studied. There is a ring of these in the outer region of the cortex and a second group 
 in the central region, these central ones usually being close to the vascular bundles. 
 
 lizoinc of Dtifuea jamtiUcui 
 :oi D. el 1 1 /yi lea. X i .5. 
 lagc ducts arc not shown. 
 
 THE AN.ATOMY OF THE LE.'Vl-. 
 
 A section of the petiole of the adult leaf appears very much like a similar section 
 of the stem, except that there is developed in the cortical region, separated by several 
 rows of cells from the epidermis, a band of sclerenchyma, yellowish or brownish 
 in color, and very much indeed like that common in the typical ferns. Holle's 
 statement that a similar mantle of sclerenchyma occurs in the stem was shown to 
 be an error, the result of his having investi- 
 gated a fern which was not a Daticea. 
 There certainly was no trace of scleren- 
 chyma in any of the specimens I have 
 examined. 
 
 The distribution of the vascular bun- 
 dles in the petiole is very similar to that in 
 Hclnunthostachxs, except that inside the 
 ring ot bundles there is a single larger 
 central one, instead of the two small ones 
 which occur in Helminthostachys. All the 
 bundles are continued into the stem as 
 separate strands or there may be a certain 
 amount ot anastomosing among them, as 
 there is in Ophioglosstim pendulum, which 
 Daneea resembles more in the character of its leaf traces than it does Helmin- 
 thostachys, where the bundles of the petiole are united into a single trace within 
 the cortex of the stem. 
 
 The distribution of the mucilage ducts in the petiole is very much like that of 
 the stem. There is a definite ring of these inside the zone of sclerenchyma and several 
 larger ones in the central region, near the vascular bundles. 1 he form and venation 
 ot the leaf is essentially the same as in the younger sporoph\tes which have alread\' 
 been described, and stomata are developed only upon tiie lower side, the epidermal 
 cells showing a strongly undulate outline, except tor the series ot narrow cells which 
 12 
 
178 
 
 THE MARATTIALES 
 
 immediately surround the stoma. A section of the lamina presents an appearance 
 very much like that of Hclninithnstachys. Below the upper epidermis is a well- 
 developed palisade layer, while below this the spongy mesophyll has the characteristic 
 intercellular spaces found in the leaves of most of the higher plants. Dancea tricho- 
 manoides, a very small species from Peru, has delicate membranaceous leaves, which 
 probably would show a much simpler structure than that of the coriaceous leaves 
 of the larger species. D. siiitciisis, a species from Porto Rico, which is in the herba- 
 rium of the British Museum, has adventitious buds developed at the leaf tip. 
 
 A, B. Two lont^itudinal 
 
 and third leav 
 
 C. Stem apex. XiSi 
 
 Tig. 163. 
 :tions of a young sporophyte of Kaulju 
 ; r', second root. X20. 
 D. Apex of third leaf. XiSo. 
 
 THE APICAL GROWTH OF THE ROOTS. 
 
 The earlier roots grow from a single initial cell, but this is later replaced by a 
 group of similar initials which, in the large roots, are very much like those described 
 for the root apex of Jngioptcris (Koch 1). Figure 162 shows a section of a root 
 from the young sporophvte which is somewhat transitional in character between the 
 form with a single detinue mitial cell and the larger root with its group of initials. 
 The cell x, which is very much like the apical cell of the primary root, may perhaps 
 still be considered as the single apical cell, but the adjacent segments are almost 
 equal in size to this, and it is clear that they contribute to the plerome cylinder of 
 the root as well as to the cortex and the root cap. 
 
 In sections of large roots taken from the adult sporophyte, there is found at the 
 growing point not a single initial cell, but a group of apparently similar initial cells 
 somewhat wedge-shaped in longitudinal section and arranged in a radiating fashion. 
 No single one of these can be clearly recognized as a primary initial cell. These 
 cells divide at intervals by longitudinal walls and from the bases additions are made 
 
Till- OLDI-K Sl'OKOl'IIVTE 179 
 
 to the plerome cylinder, while from the outer cells of the meristem group arise the 
 new cells for the root cap. The outer segments cut oft' from the peripheral cells of 
 the apical group of cells undergo more rapid periclinal divisions than those nearer 
 the center and the cells thus formed contribute to the cortex of the root. 
 
 Unlike the roots of most of the Ophioglossaces, those of Daitam branch 
 freely in a monopodial fashion. This is paralleled among the Ophioglossacex- by 
 some of the larger species ot Botryc/iiuni, Helninithostachys, and (Jphioglossum 
 pcnJuhini. Kiihn's statement that the roots are unbranched is incorrect. The 
 branching of the roots was perhaps somewhat more frequent in D. jamaicensis 
 than in D. elliptica, but the latter species also shows numerous short lateral rootlets. 
 The root hairs are not conspicuous on these older rf)ots, but are readily seen upon 
 the younger ones and are then found to be multicellular, as was first shown by 
 Brebner for D. simplicifolia. Sections ot the larger roots were examined in these 
 two species and, while they agreed in the main, there were some slight difterences. 
 In a large root of/), elliptica, about 2 millimeters in diameter, the outer portion was 
 occupied by about four rows of brownish thin-walled cells, very probably more or 
 less suberized. Within this was a ring of brownish sclerenchyma cells about three 
 cells in thickness, and between this and the exceedingly conspicuous endodermis was 
 a broad zone of parenchyma, the cells containing much starch. A short distance 
 outside of the endodermis was a ring of conspicuous mucilage ducts. There were 
 twelve xylem rays alternating with as many rays of phloem. These did not extend 
 to the middle of the stele, which was occupied by a pith of considerable size. 
 The roots of Z). jamaicensis were somewhat smaller and the main diff"erences were: 
 first, the thicker layer of tissue outside the ring of sclerenchyma, which was much 
 less developed than in D. elliptica (in many places this ring was but one cell in 
 thickness and very seldom more than two cells); second, the mucilage ducts were 
 relatively very much larger; third, there were but nine xylem masses instead of the 
 twelve in the root of D. elliptica. 
 
 The endophyte, which is usually found in the primary roots of all the Marat- 
 tiaceie is generally wanting from the larger roots of the adult sporophyte. 
 
 THE SPOROPHYTE OF KAULFUSSIA. 
 
 The deveK)pment of the young sporophyte in Kaulfussia resembles that of 
 Dancea in most respects, but from the first it is markedly dorsiventral, like the 
 sporophyte of Helninithostachys. The cotyledon, while not provided with the char- 
 acteristic free stipules of the later leaves, still has its base connected with a ridge 
 which extends around the stem apex and youngest leaves, which thus lie in a cavity 
 at the base of the cotyledon in much the same way that the stem apex is situated in the 
 young sporophyte of Ophioglossum; but the sheath in Kaulfussia forms a shallow 
 open cup, instead of the closed conical sheath found in Ophioglossum. This stipular 
 sheath seems to be in all respects similar in its origin to that found in Ophioglossum. 
 
 The second leaf lies nearly opposite the first and thi- thirtl next to the fiist, 
 and so on, the distichous arrangement being evident from the beginning. \\'e have 
 already pointed out that in the very young sporophyte the vascular bundle from the 
 cotyledon is continued directly into that of the root and that the bundle from the 
 second leaf joins this common bundle, as shown in fig. 163. These primary bundles 
 run on one side of the apical meristem, which, as in Damra, is of very limited extent 
 and does not contribute at all to these bundles. 
 
 The apical cell of the young sporophyte, up to the oldest stages that were in- 
 vestigated, is very much like that of the very young plant and is broader than that 
 of Daiicca, more resembling both in cross and longitudinal sections the apical cell 
 
180 
 
 THE MARATTIALES 
 
 oi Ophioglossum. The young leaves arise right and left in succession on either side 
 of the stem apex, and from a very early period are strongly bent over, showing the 
 characteristic circinate vernation of the Marattiaceie. Sections made through the 
 apex of the young leaf show that it has at first a definite apical cell a good deal like 
 that which is found in the stem apex, and although no satisfactory cross-sections of 
 this apical cell were found, there seems to be no question that there is present a 
 single definite initial cell which probably has much the same form as that of the stem 
 apex. The young leaf, at this stage, has the dorsal region strongly convex and 
 composed of large parenchyma cells. The apex is bent over forward and downward. 
 
 Fig. 164.— Scries ol 
 leaf 
 D. Stem apex more enlarged. 
 F-H. Central region of the 
 
 The tissues of the apex and the adjacent region are composed of small, actively 
 dividing cells. The young vascular bundle can be clearly seen extending nearly to 
 the apex and continuing downward into the stem until it joins the bundle from the 
 next older leaf. 
 
 The second root arises at about the same time that the third leaf is recognizable 
 and apparently its origin is exactly the same as it is in Daiura. Its stele joins the 
 central bundle of the stem near its junction with the third leaf trace. The young 
 leaves, like those in Daiia-a, are sparingly covered with hairs and scales, but these 
 scales do not have the peltate form found in Datura, being usually short rows of 
 stout cells; or the hair may have its base composed of an oblong mass of cells 
 attached by a narrow stalk and having its apex drawn out into a slender filament 
 composed of two or three elongated cells. The terminal cells of these hairs usually 
 stain very strongly, indicating that they contain tannin. 
 
 The young leaves are relatively stouter than in Dnnera, but otherwise resemble 
 them closely. In the young sporophyte, up to the time of the formation of the third 
 leaf, there may be no trace of the mucilage ducts, although these may be developed 
 
Till-: OLDKR SPOROl'IiyTI- 181 
 
 at an earlier ixriod. I'lie mucilage ducts throughout the hte of the sporophyte 
 are less conspicuous in K/iiiI/iismh than in Daiuca. 
 
 A section taken at the level of the stem apex in the viiy \nung s|)<)rophyte 
 (fig. 137) presents an appearance not unlike that found in the bud in Oplnuglossutri. 
 The base of the cotyledon extends around the stem apex, so that in the section the 
 stem apex and the second leaf are inciosetl in this cavity formed by the stipular 
 sheath of the cotyledon. Higher up, the base of the cotyledon becomes free from the 
 second leaf, but this takes place earlier on one side than on the other. 
 
 The development of the vascular system in the stem is exactly as it is in Daiicca. 
 The stem apex in the young sporophyte takes no part in the development of the 
 vascular bundles, the stele in the stem being made up entirely of the united leaf 
 traces. In the specimen figured (fig. 164), in which the second leaf was still quite 
 young, the bundle from the cotyledon passes downward into the stem and continues 
 its downward course until it joins the second leaf trace. The resulting solid stele 
 at first shows the two separated xylems of the component leaf traces which lower 
 down merge into a single thick band of xylem completely surrounded by the phloem 
 
 (fig. 164, E-H). This stage corresponds to Brebner's "haplostele" in DancFa sini- 
 pUcifoUa. This condition merges insensibly into the primary root with its diarch 
 bundle, the two xylems in the younger part of the root being quite separate, but 
 forming a single band at the base of the transitional region, between the root and 
 the bundle in the stem. The bundle of the root shows a conspicuous endodermis 
 which becomes less and less evident in the transitional region, although it prob- 
 ably never quite disappears. The primary root usually shows the presence of an 
 endophytic fungus like that occurring in the prothalh'um and in the priniarv root 
 of most of the Ophioglossaceae. 
 
 Fig. 166 shows cross-sections of a young sporophyte in which the fifth leaf is still 
 quite small, but with its trace showing the first tracheary tissue. This fifth leaf trace 
 unites below with the trace from the fourth leaf and forms a single bundle, with 
 the xylems separate. The xylem from the fourth leaf above its junction with the fifth 
 leaf trace forms at first a continuous band which divides into two parts, one of which 
 unites with the xylem from the fifth leaf trace, while the other remains distinct. 
 This separation of the xylem in the fourth leaf trace is the first indication of the for- 
 mation of the double leaf trace, such as we shall find occurs in the later leaves. 
 
182 
 
 THE MARATTIALES 
 
 The separated xylems become joined lower down into a single transverse plate, just 
 before the junction of the strand with the trace from the third leaf (fig. i66, H), 
 with which it joins in the same way and is continued downward to the junction with 
 the strand arising from the fusion of the second leaf trace with the cotyledon. Above 
 the junction of the first and second leaves, the stele of the second root, which grows 
 vertically downward, joins the stele formed by the junction of the three youngest 
 leaf traces. 
 
 Above the junction the section of the bundle is nearly circular in outline, the 
 large tracheids forming a broken group in the center of the bundle, quite like a 
 
 A-C. Serif 
 
 D. Stem apex. 
 H-J. Three sections lower down. 
 
 K. Central vascular cylinder still lower down. 
 
 corresponding stage in Dancrn. Some of the cells adjacent to the bundle show the 
 typical thickenings on the radial walls, indicating that they belong to the endodermis, 
 but the limits of the endodermis are e.xtremely vague. In this plant two mucilage 
 ducts had developed in the fourth leaf, one on each side of the vascular bundle 
 (fig. 1 66, B, m). 
 
 The second root is stouter than the primary one and its bundle is triarch. The 
 mycorrhiza present in the first root could not be detected in the second root, but 
 possibly may be developed at a later stage. All of the roots are provided with root 
 
THT OLDPR SrOROPFn'Tr 
 
 183 
 
 hairs, wliicli rcseniblf tliost- in Diuin-n in bLiiii^ iiuiltinlliilai. In rlic laif^er roots 
 there may sometimes he five or six cells, each with a conspicuous nucleus, makinj; 
 up these root hairs, and sometimes they show signs of branching at the apex. 
 
 It is evident that in Kaulfiissui, up to the time that five leaves have developed, 
 the vascular system ot the stem consists of a single axial strand, a sort of sympodium 
 formed by the completely united traces of" the young leaves. The xylem may appear 
 as a sin-gle mass at the points where fusion of the leaf traces is complete, but at most 
 points the individuality of the component strands is maintained, so far as the xylem 
 is concerned. 
 
 Figure i68 shows transverse sections from a iiuuh older sporophyte. The 
 section of the petiole still shows the single vascular bundle having near its inner face 
 a single large mucilage duct, which follows the leaf trace downward into the stem. 
 As the bundle from the leaf enters the stem it broadens, and there is a separation 
 of the xylems into two masses, first seen in the fourth leaf. In the specimen shown 
 here the earliest leaves w-ere not sectioned, and, as it can not be stated exactly how 
 many leaves the young sporophyte had developed, it must therefore remain uncer- 
 tain which leaf first shows the completely divided leaf trace. In the specimen in 
 question the oldest leaf that showed had a single trace with two separate xylems, 
 
 while the next leaf had the leaf trace completely separated and resembling a corre- 
 sponding stage in Datia-a (fig. i68, /'). As the traces from the two leaves approach 
 preliminary to their fusion, the double leaf trace has its parts united again so that 
 the traces become single before they join. The leaf trace now forms a single bundle, 
 crescentic in section and very much like the single broad leaf trace from the earlier 
 leaf. The two traces come nearer and nearer and finally fuse on one side, so that 
 a section shows a single bundle almost circular in form, but broken on one side 
 (fig. 1 68, H). This break is finally closed up, and the section of the bundle is 
 completely circular and resembles very closely that oi Botrxchiutn or HfUniutlios- 
 taclixs, except that phloem is developed inside the ring of xylem (fig. 168, A"). A 
 similar condition may sometimes be f^ound for a very short rime in the voung plant 
 of /3rt,/,jv/ also. 
 
 No certain evidence of an internal endodermis coidd be made out ami tlu' 
 external endodermis is also somewhat vague, although there is pn>babl\- no question 
 of its presence. The outer phloem is bounded by a fairly well-defined peric\cle, 
 between which and the starch-filled cortical cells of the stem arc two or three layeis 
 of transparent cells, some of which presumably constitute the endodermis. The 
 endodermis is certainly present in the free leaf traces, although less conspicuous 
 
184 
 
 THE MARATTIALES 
 
 than it is in the roots. In the upper region of the stem the leaf traces are further 
 apart, and as the number increases a cross-section of the stem shows a circle 
 of separate bundles closely resembling the arrangement in Ophioglossum. The 
 central "commissural" strand, which is found in the adult stem, had not developed 
 in any of the young sporophytes I examined and evidently arises much later in the 
 history of the sporophyte than it does in the other Marattiaceae. In this respect, 
 as well as in the arrangement of the bundles themselves, Kaulfiissia is more like 
 
 A, B. Two sections above level of stem a 
 
 C-J. Sections taken lower down from 
 
 K. Central bundle of section J. XI, 
 
 L. Part of the bundle of section H, 
 
 of a young sporophyte of Kaulju 
 series, m, mucilage ducts. 
 
 enlarged. G, F, I, J, show only the 
 
 I of about same age ; 
 
 Ophioglossum than it is like the other Marattiaceae. In the intermediate condition 
 between the single solid stele found in the very young stem and the entirely separated 
 leaf traces of the older rhizome, there is a transitional condition with a "siphono- 
 stele," very much as in Helminthostachys (fig. i68, K). 
 
 In one very young sporophyte there was found, in the cortex of the primary 
 root near its base, a large lacuna which recalled the similar lacuna found in the 
 first internode o{ Helminthostachys. 
 
HK OLDER SPOROPHYTE 
 
 185 
 
 Tlu- youiig s[)on)\^h\tv (>{' Kaiilfussid is almost completely destitute of the tannin 
 cells which are so conspicuous a feature in Datura, and these tannin cells are prac- 
 tically entirely absent from the sporophyte throughout its whole existence. In this 
 respect there is a marked resemblance to Opliioglossuni. The mucilage ducts are 
 also somewhat less developed than in the other Marattiace:e and are mainly con- 
 lined to the central part of the stem in proximity to the vascular bundles. At a late 
 period they may also appear in the cortical region. Their lysigenous origin is less 
 evident than in Dancea and it is not impossible that they may sometimes be of 
 schizogenous origin, as Brebner states is often the case in Dancea. 
 
 The structure of the vascular strands is very much like that of Dancva, as a 
 reference to the figures will show. There are developed slender, spirally-marked 
 protoxylem elements like those in Dana-a and which we have already seen occur 
 also among the Ophioglossace^, in Helminthostachys. 
 
 The stipules in the young plants of Kaulfiissta are much less definite than they 
 are in Daiicca. A cross-section through the bases of the young leaves shows that 
 instead of the bases appearing free, each with its distinct stipules, the leaf is confluent 
 with the next one, one side being slightly extended and free. The leaf base thus 
 incloses the next younger leaf completely on one side, but leaves it free on the 
 
 A. Bundle from intermediate region of .i \ 
 
 B. Section of primary root. X150. 
 
 C. Section of bundle of second root, fn, endo 
 
 Other (fig. 168, 5). In the later leaves the stipules become more conspicuous, but 
 they are always connected by a very conspicuous commissure in front (fig. 171, (oni) 
 so that a section of the young leaf base shows a space between the front of the petiole 
 and the stipular sheath. In short, the structure is very similar to the open stipular 
 sheath of Botrychium virginianum and apparently arises in much the same way. 
 
 The structure of the adult sporophyte has been carefully investigated by Kiihn, 
 who pointed out that it is the simplest of the Marattiace:e in the structure of the 
 stem (fig. 171, B). As in the young sporophyte, the stem shows in section a single 
 circle of vascular bundles, but there is in addition a single median strand within the 
 circle of bundles belonging to the dictyostele, and this bundle is presuniahl\- of 
 cauline origin, like the corresponding one in Daiiera. 
 
 The whole vascular skeleton of the stem, when removed by maceration, was 
 found by Kiihn to be a hollow cylinder with large open meshes. These are more 
 elongated on the ventral side and from these ventral strands alone, according to 
 Kuhn, the roots are developed. In this respect Kaiilfiissia offers a certain analog\- 
 to Helmnithostachys and perhaps to Ophioglossiini pendulum. 
 
 As in the other Marattiaceic, the ground tissue of the rhi/ome is composed of 
 simple parenchyma. The outer layers have their walls brown in color and show the 
 reaction of cork. Except for the presence of the central strand, a .section of the 
 
186 
 
 THE MARATTIALES 
 
 rhizome oi Knulfiissin presents almost the exact appearance of a similar section of 
 Ophioglossiini. 
 
 In a rhizome having a diameter of about a centimeter twelve bundles showed 
 in a cross-section taken through the internode, one of these being the medullary 
 strand. The section of a petiole of a leaf taken from the same plant is shown in 
 fig. 171, C, and the structure of the section closely resembles that of the rhizome. 
 There were eight bundles arranged in a circle, within which were two medullary 
 strands. The cortical region was composed of several rows of brownish cells, prob- 
 ably similar to those in the outer part of the rhizome, and within these, separated 
 
 bundle from a young sporophyte of Kutil- 
 ftissia, showing two xylems. X150. 
 
 Fig. 171. 
 
 Rhizome of a large sporophyte of Kaulju 
 ules; com, commissure. X]f. 
 
 B. Section of rhizome. X2. 
 
 C. Section of petiole. X2. 
 
 from them by three or four rows of cells, was a conspicuous band of collenchyma, 
 a tissue characteristic of the leaves of most of the Marattiaceae. 
 
 According to Kiihn, the bundles of the leaves are continued separately into the 
 cortex of the stem and do not unite into a single leaf trace. There are occasional 
 anastomoses of the vascular bundles of the petiole, quite like those in the stem itself, 
 and this becomes especially marked at the base of the leaf, so that the leaf trace (if 
 such it can be called) forms a hollow reticulate cone, quite equaling in diameter the 
 vascular cylinder of the stem itself, and attached to this by the open base, the sepa- 
 rate strands forming a crescent-shaped group when seen in section. The arrange- 
 ment of the bundles within the petiole itself and their method of junction with the 
 vascular system of the rhizome are very much like tho.se in Ophioglossiim pendiiluju. 
 
 The vascular bundles, both in the petiole and rhizome, are concentric. The 
 endodermis is not recognizable and the central xylem is completely inclosed by the 
 phloem, sieve tubes being developed throughout. 
 
TIIF 
 
 OI.nivK SI'OROPMVTi: 187 
 
 \\ hik' iiuicilafi,i- diKts aii' pre 
 
 ■sint in hdth leaf and sti-m, they aie less numerous 
 
 and conspiiuous than in Daiiua. 
 
 There is a ring of" them in the outer cortex and 
 
 several larger ones in the central : 
 
 region. Kidin states that, although these mucilage 
 
 ducts sometimes look as if they 
 
 were of schi/.ogenoirs origin, in reality the)' are 
 
 always lysigenous, i. e., tiiey are 
 
 lornu-d h\ tlie fusion of several mucdage-secreting 
 
 In the cotyledon and the earlier leaves a section of the lamina shows, hetween 
 the large-celled epidermis of the upper and lower sides, a mesophyll composed of 
 about three layers of (juite inditierent parenchyma, exactly as it is in Opiitoglossuni. 
 In the adult leaves the upper part of the mesophyll is compact, but there is no 
 proper palisade tissue developed. Toward the lower surface the mesophyll becomes 
 looser. 
 
 Until the plant has reached a considerable si/e the leaves remain undividetl 
 and, except for the much greater size, closely resemble the early leaves of the )-oung 
 sporophyte (see plate 1 1, fig. i). 'The simple leaves are succeeded later by trifoliate 
 ones, and finally five leaflets are developed. The long, fleshy petiole may reach a 
 diameter of a centimeter or more and a length of some 50 centimeters in large 
 specimens. The venation strikingly resembles that of a dicotyledonous leaf, the 
 pinnately arranged, secondary veins being connected by a plexus of small veins, 
 inclosing nearly square areoles within which a few free terminations are occasionally 
 found. The large pores upon the lower side of the leaf, formed by the greatly 
 enlarged stomata already referred to in the younger leaves, are plainly visible to the 
 naked eye, and sections of these look curiously like the pores upon the thallus of 
 certain Marchantiaceae (fig. 128). 
 
 The stipules of the older leaves are very conspicuous and show much the same 
 structure as that noted for the earlier leaves. The young leaf is included within the 
 stipular sheath of the next older one, and near the apex the unbroken stipular sheath 
 of the youngest expanded leaf shows very plainly that the large stipules are joined 
 in front by a broad commissure, which extends entirely around the apex of the shoot, 
 exactly as it does in Helminthostachys, from which the sheath differs mainl}^ in the 
 fact that it is divided above into the two stipules as it is in Botrychium virginiatiiim, 
 and there seems no reason to doubt that we have to do with entirely homologous 
 structures (fig. 171, A). The stipules of the young leaves are covered with hairs 
 and scales similar to those noted in the young sporophyte. 
 
 The roots arise from the ventral surface and flanks of the rhizome, and out- 
 number the leaves, to which they apparently bear no definite relation. 
 
 The apical growth was not studied in the root of the older sporophyte, but it 
 is probable that in the large roots of the older plant the single apical cell found in 
 the early roots is replaced by a group of initials such as are found in the other 
 Marattiaceae, although it is barely possible that the single apical cell may be retained 
 here as it is in the Ophioglossaceae. Unfortunately, material was not available for 
 a study of this point. 1 he roots in most cases are quite unbranched, as they are in 
 huof^luoglossiitn. 1 he only cases where branches were seen, were young sporoph\tes 
 in which the end of the root had been destroyed and two branches had arisen on 
 each side of the destroyed apex. This looks somewhat like a dichotomy ami it is 
 barely possible that, as in Opiuoglossum, there may be a real dichotomy of the root, 
 but all of the cases that were found had much more the appearance of the formation 
 of two lateral roots. The young roots of the sporoph\te are provided with multicel- 
 lular root hairs, like those of Dniitiii. but these disappear as the root becomes oliler. 
 The ci nti;d \;iscid:ir c\liiuKr of the root is surrounded by a circle of ver)- large 
 nuicilage ducts. TIk riulocU 1 mis is clearly evident and the routs which I examined 
 
188 THE MARATTIALES 
 
 irom a medium-sized plant were tetraich. Kiihn states that pentaich and hex- 
 arch roots also are formed. The reduced number of xylems in the root of Kaul- 
 fussia, as compared with the other Marattiaceae, is another indication of its probable 
 nearer relationship to the Ophioglossaceae, the structure of the root being very similar 
 indeed to that of Hehninthostachys or Ophioglossum pciulidum. 
 
 TliF. Sl'UROPIlYTE OF MARATTIA. 
 
 The published observations upon the young sporophyte of Marattia are far 
 from complete. Kiihn (Kuhn I) has described the stem structure in young plants 
 of M . fraxniea, but it is evident that these plants were already too far advanced 
 to show the early arrangement of the bundles, as the stems he described had 
 reached a length of nearly 2 centimeters. Farmer and Hill (Farmer 3) have also 
 given some details as to the early stem structure in the same species. These young 
 plants, especially those described by Kuhn, have the stem relatively longer than is 
 the case either in M. doiiglasii, which I have studied somewhat in detail, or M. 
 alata, or M . sambucina, which I have also examined. In all of these species the stem 
 of the very young sporophyte is still quite short and very soon assumes the compact 
 globular form with the crowded leaves that it has in the adult sporophyte. It may 
 be said, however, that in M. alata the buds which develop upon the old leaf bases 
 have the young stem somewhat more elongated, but not nearly so much as Kiihn's 
 figures would indicate to be the case in M. fraxinea."^'' 
 
 In the youngest specimens examined by Kuhn the cross-section of the stem 
 showed a ring of bundles corresponding to the leaf traces and a central medullary 
 strand, the whole arrangement being very similar to that which is found in the 
 adult rhizome of Kaiilfiissia. This stage is also very similar to the condition found 
 in the young sporophyte of Dancea. 
 
 Farmer and Hill describe the vascular skeleton of the very young plant of M. 
 fraxinea as a "siphonostele" with much larger foliar gaps than those found in 
 Angiopteris, and thus more nearly resembling Kaidfussia or Dancea. The leaf 
 traces are at first single, but the later leaves have double leaf traces, such as we have 
 described in Dancea and Kaidfussia. 
 
 The writer has already published some details in regard to the young sporo- 
 phyte of Marattia douglasii (Campbell 3, 4), and some additional facts are here 
 added to those that have already been published; but as the series of specimens 
 available for study was not at all complete, further investigation is desirable to 
 complete the history of the development of the vascular system in Marattia. The 
 only material available for a study of the young sporophyte in M. douglasii was a 
 series of slides made a good many years ago from material collected on the island 
 of Kauai in the Hawaiian Islands. This material was supplemented by a small 
 number of very young plants of Af. sambucina, collected in Java. Material of older 
 sporophytes of M. alata was collected in Jamaica in the summer of 1908. Un- 
 fortunately, all the preparations of M. douglasii, except the very youngest stages, 
 were longitudinal sections, so that it was diflScult to follow out satisfactorily the 
 course of the vascular bundles in the later stages. 
 
 Longitudinal sections of a young sporophyte before the cotyledon was com- 
 pletely expanded are shown in fig. 133. The section was cut nearly in the plane of 
 the cotyledon and the bent-over apex of the latter was cut so as to show plainly the 
 two lobes arising from the first dichotomy of its apex. The second leaf is already 
 formed and differs in no essential particular from the corresponding leaf in Kaul- 
 
 * Since the above was written a paper has appeared (Charles l) describing the vascular system of the young sporo- 
 phvtr of M. alula. The stele of the very young plant is described as a protosteic which passes abruptly into a solenostele. 
 
nil. ()LI)i:k si'droi'iivti; 189 
 
 fussiti or Diiihrii. It was not (|iiit(.' ctitain whctlur a siiiiilf apical clII was |Misent 
 at this stage, but a triangular cell which could be seen at the apex was probably 
 the apical cell of the young leaf. The root apex of this specimen had been injured 
 so that the form of the apical cell could not be clearly made out, but it was probably 
 the same as we have already described iov the young sporophyte. I he apical cell of 
 the stem meristem was broader in outline than that of Daiuva and truncate below . 
 The vascular system at this stage, as we have already indicated earlier, consists 
 of the common bundle of the root and the cotyledon, which is joined at a point quite 
 close to the stem apex by the second leaf trace. The first tracheaiy tissue is visible 
 at this point in the form of short, reticulately marked tracheids such as we have 
 already seen in the young sporophyte of Dcuuva. At this stage no mucilage ducts 
 or tannin cells had develojud. About the base of the young leaves are short iiaiis 
 
 A, B. Two longitudinal sections of 
 sporophyte of Marattm ilouglo 
 second leaf; r' , second root. X 
 
 C. Part of primary root, showing 
 fected by endophytic fungus. 
 
 
 Ho. 173. 
 a young sporophyte of Maraitia 
 . 5/, stem apex; /, young leaves; 
 
 X2S. 
 
 and scales like those we have seen in the other genera. They are more like those of 
 Kaulfussia than like the peltate scales found in Dtuuca. 
 
 Sections of a somewhat older sporophyte are shown in figure 172, tiiis section 
 being made at right angles to the one just described. The cotyledon in this specimen 
 was fully expanded and the primary root had penetrated into the earth. I'hc 
 arrangement of the bundles was the same as in the younger sporophyte described, 
 except that the third leaf was now visible and the second root was already well 
 advanced. The section passed exactly through this root, the stele of which is seen 
 to join that of the second leaf and is practically continuous with it. The young 
 trace fVom the third leaf joins the second leaf trace near its junction with the second 
 root. The apical cell of the latter was readily seen and appears in longitudinal 
 section of nearly triangular form, but with the base somewhat truncate (fig. 177, C). 
 The tracheaiy tissue in the middle of the sporophyte is pretty well advanced 
 and the formation of the tracheaiy tissue has extended for some distance into the 
 primary root and the cotyledon. In the former a single elongated tannin sac could 
 
190 
 
 THE MARATTIALI' 
 
 be seen, like those which are much more developed in the primary root o{ Jrigiop- 
 teris, and it was also seen that certain of the cells in the cortex of the primary roots 
 were invaded by the endophytic fungus. In the cotyledon there was a single large 
 mucilage duct. 
 
 A series of transverse sections was made of a young sporophyte of M. sam- 
 bucina and these agreed in all essential particulars with similar sections of Daticea 
 or Kaidfiissia. 
 
 In a young sporophyte of M. douglasti, in which two roots were developed in 
 addition to the primary one, there could be seen at the apex two young leaves, prob- 
 ably the fourth and fifth, but as only the median sections of this series had been 
 kept the exact number of leaves could not be determined. 
 
 The central part of the stem in this plant (fig. 173) was occupied by a single 
 thick bundle, but whether this was solid or open on one side, as Farmer states is 
 
 the case in M. fraxinca, could not be satisfactorily determined. At this stage the 
 section of the young plant almost exactly resembles a similar one in Danira, and 
 presumably the single central strand is composed of the confluent leaf traces from 
 the three first leaves, as it is in Daiiara. 
 
 The .xylem of the bundle is still composed exclusively of short, reticulate tra- 
 cheids. Large mucilage ducts and numerous tannin sacs are developed in the 
 petioles of the older leaves and the elongated tannin sacs, like those in the roots of 
 Daria-a and Angiopteris, are sparingly developed in the young roots; but both tannin 
 sacs and mucilage ducts were absent from the stem tissues at this period. 
 
 In a still older plant (fig. 174) the single central bundle of the basal region is 
 replaced by the separated steles of the single leaf traces, which are beginning to form 
 the open dictyostele characteristic of the adult sporophyte. The exact nature of its 
 origin could not be followed in the material at my disposal, but it is presumably 
 
THK OLDER SI'OKOI'HVTr: 191 
 
 mucli the same as that (.lescrihtil for Dmiati, since tlie ap|)earaiue of the Idii^itiidinal 
 sections in Marattia is exactly like that of coriespondinj; sta<;es in Duiuni. The 
 apex of the stem, however, is much broader than in Datia-a. There is in the middle 
 of the apical region a cell which from its size and position may be pretty certainly 
 denominated the apical cell (fig. 174, C). There is present a strand of procam- 
 bium which ends abruptly a short distance below the stem apex. This strand 
 |irobably represents the primary commissural strand, which, as in Datia-t:, is in 
 all probability a truly cauline bundle and has no direct connection with the leaf 
 traces. In the central region of the stem there are now seveial large mucilage 
 ducts, but tannin cells are still absent. In the roots, however, the tannin cells are 
 abundantly developed. 
 
 Farmer and Hill's brief account of the development of Maratttti fiaxinea agrees 
 with my own observations so far as they have gone, except for the interpretation 
 of the vascular bundles. The "protostele" found in the lower part of the stem is 
 undoubtedly the common bundle of the primary root and the cotyledon, and the 
 open "siphonostele" is really made up of separate leaf traces, which anastomose 
 at certain points to form the large meshes of the very open dictyostele. Farmer and 
 Hill call attention to the fact that the "foliar gaps" are much wider than in Aiigi- 
 optcn's, and in consequence the separate strands, seen in section, form a circle of 
 apparently quite separate bundles, evidently closely approximating the condition 
 found in Datuea and Kaiilfussia. Farmer and Hill do not make it quite clear that 
 the commissures which they found developed later, connecting the strands of the 
 dictyostele, were really parts of the central cauline strand, but they presumably 
 assume that such was the case, as this is explicitly stated in the case of Jngiopteris, 
 which they also studied. The structure of the bundle in the latest stage described 
 by Farmer and Hill agrees pretty closely with the condition described by Kiihn in 
 the youngest specimens which he studied, where the stem was about a centimeter 
 in diameter. 
 
 The early leaf traces, as in the other Marattiace;e, are single, but later on double 
 traces are formed. According to Farmer and Hill, the two bundles of the leaf trace 
 unite before joining the vascular cylinder of the stem. No material was available 
 for a further study of the development of the bundle in M. cloiiglnsn, bur it |irobablv 
 agrees with that of M. fraxinca. 
 
 Several young plants of M. alata were examined for comparison with the young 
 germ plants of M . douglasii. The specimens in question (fig. 175) had arisen as 
 adventitious buds upon the old leaf bases which had become separated from the 
 stem. This manner of formation of the young plants is very common in this species 
 (see plate 12, A). In one of these yt)ung plants, in which the stem was about 4 
 centimeters long and 1.5 centimeters in diameter, exclusive of the leaf bases, the 
 oldest expanded leaf had a petiole of about 8 millimeters in diameter. In these 
 young plants there is usually only one leaf expanded, so that they have a very different 
 appearance from the fully developed sporophyte with its crowded circle of leaves, 
 and these monophyllous plants reminil one very much of sterile specimens of some 
 of the larger species of Botrychiutn (plate 12, B). A section of the petiole of this 
 specimen showed ten vascular bundles arranged in a circle, within which, on the 
 ventral side, were two other bundles. The section resembled almost exactly that 
 of the petiole of a fully developed lenf of Fffhninf/iostarlivs. At the base of the leaf 
 are the two very conspicuous stipules, one of which overlaps the other. These 
 stipules are connected by a commissure which joins the two stipules near the base 
 and is entirely concealed within the cavity formed by the overlapping stipules, 
 but can be seen on raising these, as a hood-like membrane, overarching the next 
 
192 
 
 THE MARATTIALES 
 
 younger leaf. Fig. 175, E, shows a young leaf from another plant, in which one 
 side has been cut away so as to show the relation of the stipules to the leaf base. 
 This leaf was still coiled up and its apex was quite concealed within the large 
 overlapping stipules. Both as to its position and its relation to the stipules, the 
 commissure exactly resembles the lip-like basal extension of the stipular sheath in 
 Botrychiitm and Hclminthostachys, and there is no reason to suppose that it is not 
 exactly homologous with this. 
 
 Partially inclosed by the stipular sheath of the expanded leaf is the next younger 
 leaf, which is cut in a plane nearly at right angles with that of the oldest leaf, so that 
 the overlapping of the stipules, which entirely conceal the rest of the leaf, is very 
 plainly seen (fig. 175, B, P). Within the stipular sheath of this leaf is a still younger 
 one, which is entirely concealed from view from the outside, but shows plainly in 
 longitudinal sections, taken next to the center of the bud. The stem apex is entirely 
 concealed within the sheath of the younger leaf. The resemblance of this section of 
 
 Fig. 175. — Marattia alata Smith. 
 
 The outer tissue has been { 
 
 A, B. Yuunp plant ilcvclopcd as a bud upon an old I 
 of the thick central section are shown. XI. 
 
 C. Section of petiole of leaf. 
 
 D. Three sections of a rhizome of a small plant. X1.33 
 
 E. Young leaf. The lamina is coiled up within the large stipules, st. 
 
 to show the commissure, com. Xl.'^3 
 
 of the stipules has been i 
 
 the young plant to a similar section of the bud in one ot the larger species of Botry- 
 chtum or of Helminthostachys is sufficiently striking. 
 
 The bundles forming the ring in the petiole anastomose freely in the leaf base, 
 as they do in Kaidfiusia or Ophioglossum poidulum, so that the number of strands 
 in the leaf trace is smaller than the number of bundles within the petiole. 
 
 The vascular bundles in the stem form a single very open mesh-work, with 
 which the leaf traces join and which is, with little question, composed entirely of 
 these leaf traces, as it is in the earlier stage. Indeed, except for the larger size of 
 these plants, there is no essential difference between them and the young germ 
 plants, described in M. douglasii. Several roots are usually developed before the 
 first leaf of these young bud-plants unfolds. Probably one root develops for each 
 leaf, but it is not certain that this is always the case. Owing to the thickness of the 
 cortical region which has to be traversed by the root before it emerges, the roots 
 have already reached a large size before they appear upon the outside. This deep- 
 
Till- OLDKR SPOROPHYTi; 193 
 
 scatfd origin of tlic roots of" tlu- Marattiacex* has often liti-n Hgiinil and tk-sriilud. 
 The roots from the base of this plant showed the hexarch stiuctuie, the six xylem 
 masses being united in the middle of the stele. About half-way between the stele 
 and the epidermis of the root is a circle of large mucilage ducts, but these were quite 
 absent from the outer cortex. No mycorrhiza could be found, except that in the 
 outermost layers of cells, which constitute a iiidimentary periderm, occasional fungus 
 filaments can be detected, but these are very different in appearance from the 
 typical mycorrhi/.a found in the primary root of the germ plant. No root hairs could 
 be found, but occasionalU' short stumps wire seen which looked as if a root hair 
 had been broken off. 
 
 Fig. 175, 1), shows three consicurivi' free-hand sections of a stem from a 
 plant of about the same size as the one we have just described, but evidently much 
 older, as there were the remains of many leaves and the whole caudex, except for 
 its smaller size, was very much like that of the adult plant. The broad vascular 
 strands made up of the confluent leaf traces formed an irregular circle which in 
 fig- '75' ^' ■^' ii shows the free central strand which lower down forms the com- 
 missure across the central parenchyma. The arrangement of the bundles, therefore, 
 in the young plant of Marattiaceae is the same as in the adult rhizome of Kaulfiissia. 
 In these sections two roots are shown, one of which is cut through the point of 
 junction with a strand of the dictyostele (fig. 175, D, i). As in the younger germ 
 plants, tannin cells are quite absent from the stem, although in the roots they are 
 conspicuous. A few large mucilage ducts, however, occur in the ground tissue of 
 the stem. 
 
 Farmer and Hill give one figure of a cross-section of the stele from the young 
 stem of Maratti (I which shows that it has essentially the same structure as that of 
 DatKsa. There is a fairly well-marked endodermis, within which lies a broad zone 
 of phloem, entirely surrounding the central mass of xylem. 
 
 According to Kiihn, there is in the older plant a second circle of bundles within 
 the first, but no satisfactory account is given as to the relation of the bundles making 
 up this second circle with those of the outer dictyostele. However, it is probably 
 composed also of elements derived from the leaf traces, but it is possible that some 
 of the strands may be of cauline origin. 
 
 THE ADULT SPOROPHYTE OF MARATTIA. 
 
 The genus Marattin includes about 25 extremely variable species, some of 
 which, e. g., M. fraxiuea, closely resemble Angioptcris in their general habit (see 
 Bitter 1, page 441; Christensen 1). They occur in the moist, tropical regions of 
 both the Old and New World, and one species, M. salicifolia, extends as far south 
 as the Cape Region of South Africa. M. douglasii is a conspicuous fern of the 
 Hawaiian Islands. This species I have studied somewhat in detail, as well as the 
 West Indian M. alata, which is abundant in the mountain forests of Jamaica. These 
 two species closely resemble each other in general habit and have the leaves very 
 much more divided than is the case in M.fraxitiea and its allies (see plate 12). 
 
 The stem in the adult plant is an almost globular, upright caudex, a foot or 
 more in diameter in large plants. The closely set, spirally arranged leaves have 
 very stout petioles, 5 or 6 centimeters in diameter at their base, which is enlarged 
 and provided with two very large wing-like fleshy stipules, which, with the base of 
 the leaf, remain attached and completely cover the caudex after the petioles have 
 fallen away. There is at the base of the leaf, as in the other Maraftiacent% a pulvinus- 
 likc enlargement, where the leaf breaks oft", leaving a clean scar marked by the broken 
 ends of the vascular bundles. The leaves may reach a length of 2 to 3 meters rr 
 13 
 
194 
 
 THE MARATTIALES 
 
 even more. The thick, fleshy texture of the leaflets, much Hke that of Botrychium 
 tcrnatiim or B. silaifolium, at once distinguishes Marattia from any of the large 
 leptosporangiate ferns with which it may be associated. The resemblance to the 
 leaves of Botrychium is especially marked in the young plants growing from the 
 stipular buds. These leaves show a marked triangular outline, curiously similar to 
 that of the larger species of Botrychium {ste plate 12, B). The venation of the leaf- 
 lets is also very like, resembling that oi Botrychium more than it does that oi Dancea 
 or Angiopteris. In the species with large linear leaflets, e. g., M. fraxinea, the form 
 of the leaf is quite similar to that of Jngtoptcris or Dainra, but the veins are not so 
 closely approximated. 
 
 The structure of the leaf was particularly studied in M. alata, which closely 
 resembles the Hawaiian M. douglasii. In well-grown specimens the leaves measure 
 2 or 3 meters in length, with a petiole which is 5 or 6 centimeters in diameter above 
 the insertion of the stipules and is somewhat larger lower down. Where it joins 
 the stem there is a marked constriction and the leaves often become broken off at 
 this point, the fleshy leaf base remaining alive and often giving rise to adventitious 
 buds (plate 12, A, 3). A section of the leaf base above the stipules shows the vascular 
 
 A. Section of ultimate rachis of a leallet of Marattia alata. The shaded ; 
 
 B. Collcnchyina cells, more highly magnified. 
 
 C. Part of vascular bundle. The shaded cells are tannin sacs. 
 
 ullcnchvina; m, mucilage du 
 
 bundles to be arranged in two large concentric circles, within which is a smaller 
 third ring corresponding to the two inner bundles found in the section of the younger 
 leaf. Below this point there are numerous anastomoses of the bundles before they 
 enter the cortex of the stem, and the leaf trace consists of a much smaller number 
 of bundles, eight in the specimen examined, which are arranged in a single circle, 
 open on the adaxial side. The distribution of the bundles to the stipules was not 
 studied, but it no doubt corresponds to that in other species, where the veins form 
 an extensively branched system, connected with the bundles of the petiole. 
 
 The large leaves of M. alata, when fully developed, are five-pinnate, the ultimate 
 leaflets being about 2 centimeters long. From the midrib extend lateral veins which 
 usually fork once but may remain undivided (plate 12, A, i ). The bundles of the 
 rachis in the later divisions of the leaf diminish in number and show the horseshoe- 
 like arrangement seen in the leaf trace after it leaves the petiole. In the final divisions 
 the crescent of bundles seen in the larger rachis is always completely united and 
 in section appears as a single horseshoe-shaped bundle (fig. 176, A). In the main 
 stipe and secondary and tertiary rachis there is a conspicuous hypodermal sheath 
 of sclerenchyma. This sclerenchyma passes into collenchyma near the base of the 
 
IHK 0LD1;R Sl'OKO 
 
 195 
 
 stipe and Hnally disappears entirely before the stem is reached. 1 he scieienchynia 
 is also replaced by collenchyma in the rachis of the terminal divisions of the leaf. 
 Tannin sacs are quite absent except from the immediate vicinity of the bundles, 
 where they occur upon the inner concave side, either in direct contact with. the bundle 
 or actually within it. 1 he leaf lamina is only about half as thick as that of the leath- 
 ery Ics^f of Ddiuva clliptica, but shows much the same structure. 1 he palisade cells 
 are not (juite so well developed and there is a complete absence of sclerenchyma, but 
 otherwise the structure is very similar. The cells of the epidermis are undulate in 
 outline and the stomata, like those in Dana-a, are surrounded by a series of accessoiy 
 cells. 
 
 The hui^e roots may attain a diameter of 6 to 7 millimeters. In the absence 
 of the hypodermal sheath of sclerenchyma they more nearly resemble Kaulfussia 
 than Daiicea, but they are more like Dancca in the greater number of xylem rays, 
 which number twelve in the largest roots examined. In these later roots the central 
 part of the stele is occupied by the pith, the xylem rays not being united as they 
 
 f a small root of Marattiti dau^lasii 
 
 B. I'ranivirse section of apex of a similar root. 
 
 C. Longitudinal section of the apex of the second r 
 
 D. Apex of a large root of M. alata. x .v, initial cells 
 
 are in the roots from the younger plant, but in some cases a few scattereti tiacheids 
 appear in the central region. There is usually a ring of conspicuous mucilage 
 ducts in the cortex, but these were sometimes not very well developed. Numerous 
 tannin sacs are scattered through the cortex and also occur within the stele. 
 
 THE SPOI^OPIIYTI'; OK ANCIOffKRIS. 
 
 AtigiofHcris is the largest and most specialized of the Marattiace;c. As we have 
 indicated before, there is much difference of opinion as to the number of species 
 which should be recognized and many botanists consider that all of the forms belong 
 to a single extremely variable species. Bitter, however, in his account of the Marat- 
 tiace;e in the "Natiirliche Pflanzenfamilien," thinks that from 20 to 30 species 
 should be recognized, following in this respect the classification given by Presl and 
 l)e Vries; while Christensen (Christensen 1) recognizes 62. 1 have examined ma- 
 terial from Australia furnished by Mr. j. H. Maiden, director of the Botanical 
 Gardens in Sydney, as'well as material collected by myself in Ceylon and Java. 
 Some of the specimens growing in the Botanical Garden at Buitenzorg are the 
 largest that I have seen. Leaves 6 meters and upwards in length were measured and 
 the upright caudex was almost as big as a barrel. The specimens seen in Ceylon 
 were somewhat smaller. 
 
196 
 
 THE MARATTIALES 
 
 Forms oi Angiopteris occur from Polynesia to Madagascar and northward as 
 far as the Himalayas. The genus also occurs in northern Australia. 
 
 The anatomy of the adult plant has been investigated more thoroughly than 
 that of any other member of the Marattiaceae. The most recent account of the 
 anatomy is that given by Miss Shove (Shove 1), while Farmer and Hill have care- 
 fully investigated the vascular system of the young sporophyte. A number of prep- 
 arations were made by the writer for the purposes of comparison with the other 
 genera, but no attempt was made to follow out in detail the extremely complicated 
 vascular skeleton of the adult sporophyte. Farmer and Hill (Farmer 3) have given 
 a detailed account of the development of the vascular system in the young sporo- 
 phyte ot Jngioptcns, so that it was not thought necessary to make a large number 
 of preparations of this species, the material of which was collected at the same place 
 where Farmer obtained his plants. However, as these authors did not trace the 
 development of the young bundles from the apical meristem, it seemed worth while 
 to examine this point for the purpose of supplementing their account based upon 
 the study of the fully developed vascular skeleton. 
 
 young sporophyt 
 
 The vascular system begins, as was recognized by Farmer in an earlier paper 
 (Farmer 1), as a single strand connecting the root and cotyledon, exactly as is the case 
 in the other Marattiaceae. In the further study of the development of the vascular skel- 
 eton Farmer and Hill employed the method of constructing a model of the vascular 
 system, built up of superimposed sheets of wax, corresponding with the outline of 
 the vascular bundles seen in the serial sections. The figures which they give, 
 drawn from such models, show very clearly the relation of the vascular strands 
 which make up the complicated skeleton in Angiopteris. The assumption is made 
 that the single stele found in the young plant is really a cauline structure, the leaf 
 traces being subsidiary. The early history of the vascular system of the young 
 sporophyte is given by them as follows: 
 
 "The vascular skeleton in the young plant oi Angiopteris consists of an axile 
 rod of tissue, from which strands are given off to the roots and leaves respectively. 
 The first lateral root is given off at a point not quite opposite the formation of the 
 first leaf trace. It is separated from it by about 130°. The regular relation be- 
 tween the leaf and the corresponding root is, however, soon lost. The gaps produced 
 by the early leaf traces are very small and are immediately made good above. The 
 
IHE OLDKR 
 
 197 
 
 first deeply depressed aperture or gap occurs at about tlie sixth or seventh leaf. 
 The leaf traces still continue to issue from the stele as single strands till a varying 
 number have been formed, but they begin to bifurcate while still within the cortex 
 of the stem. 
 
 "As the stem increases the leaf traces become more numerous and crowded and 
 they take away a larger portion of the vascular tissue from the axile strand. The 
 result is that the leaf gaps become less rapidly repaired. Ihe stele is already hollow 
 in this region, that is, it consists of a cylindrical vascular mass with perforations 
 corresponding to foliar gaps and inclosing a core of parenchyma. Sooner or later, 
 the gap above one leaf fails to be repaired until after the exit of the trace of the next 
 leaf and then the original vascular cylinder becomes broken up and assumes a 
 condition, in transverse section, conforming with that of polystely or dialystely." 
 
 There is thus a gradual transition from the solid stele found in the earliest stages 
 to a hollow cylinder or siphonostele with a core of pith, and by the formation of larger 
 and larger leaf gaps there is a transition to an open dictyostele like that found in the 
 other genera. Finally there appears the commissural strand which forms a thick 
 
 A-D. Four transverse 
 
 the fusion of the 
 
 E. Centr.il region of C. 
 
 F. Central region of D 
 
 stem apex. B, C, s 
 
 strnd traversing the pith and undoubtedh' of cauline origin. This central strand 
 becomes more and more important as the plant develops and from its central 
 position might be mistaken for the original axial strand of the young sporophyte. 
 The vascular system at this stage is described as follows: 
 
 "The leaf traces become more complex and anastomoses take place at irregular 
 intervals with the strands which can still be recognized as the relics of the original 
 siphonostele, as well as with one another. Irregularities also commence to become 
 apparent as to the relative heights at which the two members of the leaf traces be- 
 come freed from the plexus of tissue, and a stage is thus reached at which the vascular 
 skeleton appears to consist of a stout axile strand, surrounded by upwardly diverging 
 zones of steles, which ultimately pass out above to the leaves." 
 
 It is evident that at this stage there is a condition which is \er\' much like that 
 figured by Kuhn for Kaidfussia, and it is pretty clear that the whole of the dictvostele, 
 exclusive of the axial or commissural strand, is made up of the "upwardl\- diverging 
 zones of steles which pass out above into the leaves," i. e., in other words, the 
 dictyostele is composed of a union of multiple leaf traces. 
 
198 
 
 THE MARATTIALES 
 
 Transverse sections of a very young sporophyte of Angiopteris show the same 
 relation of the parts as in the other genera. The junction of the^traces of the first, 
 second, and third leaves takes place at a very short distance below the stem apex, 
 which, as Farmer pointed out, shows an unmistakable apical cell of somewhat irregu- 
 lar form. Compared with the other genera, the primary leaf trace in Angiopteris has 
 the xylem better developed, and composed of about half a dozen tracheids, where 
 the trace enters the stem. This primary leaf trace is concentric, as it is in the later 
 leaves, and not collateral, as it is in Daucra (fig. 179). 
 
 Near the point of the junction of the three primary traces, there may be seen 
 in the young traces from the second and third leaves the first tracheary tissue, 
 consisting of one or two tracheids. As the sections are examined lower down the 
 tracheary tissue of the second and third leaf traces increases in amount until it 
 forms a solid band, separated from the corresponding xylem of the primary leaf 
 trace by a band of parenchyma. The three leaf traces are now completely fused, but 
 the sections of the two xylems are perfectly evident, and this band of tissue between 
 
 A-D. Four sections of an older sporophvte tha 
 
 E. Stem apex. X130. 
 
 F. Central bundle from transitional region. 
 O. Bundle of primary root. 
 
 the two xylems (which really belongs to the ground tissue of the stem and not to the 
 stele itself) corresponds to the "pith" figured by Farmer and Hill for the transi- 
 tional region between the root and stem (fig. i8o, C). Sometimes, in still lower 
 sections, the two xylems are connected, but I have found no cases where the pith 
 was entirely surrounded by the xylem, and the two xylems of this axial strand merge 
 gradually into the two xylems of the diarch primary root (fig. 179, G). The endo- 
 dermis is clearly evident in the intermediate region and is recognizable also at higher 
 levels, but is less easy to distinguish. 
 
 In a series of sections, from a plant in which the fifth leaf was just recognizable, 
 a section at the level of the stem apex shows the apical meristem to be composed 
 of several large cells, one of which is probably the single apical cell, but this is not 
 easy to determine. The base of the fifth leaf is close to the apical group and the 
 section of its stele is indicated by a group of small cells close to the stem apex. The 
 stele of the third leaf is clearly seen, separated from the fourth leaf by about one- 
 third the circumference of the stem. Below the stem apex these two leaf traces 
 
THK OI.nFR SPOROPHVTR 
 
 199 
 
 ;ippro;uli, l)iit an- still srpaiatccl In a cdiisicki al)l(.- mass of tissiit-, l\ins^ directly 
 under the stem apex. Ihe two bundles finally join and between them there seems 
 to be a sort of connective tissue which may perhaps belong to the stem itself, so 
 that this central stele of the stem may possibly have a certain amount of cauline 
 tissue in addition to that deiived from the leaf traces; but it is quite as likely that 
 this connective tissue between the bundles is nothing more than a lateral extension of 
 the leaf trace itself, very much as is the case in J / rhuiiit/i'>.^t/ir/iv and liatt \chium. 
 
 . sporophytc of Angioptcr 
 r", r', second and third roots. X25. B, traverses the stem apei. 
 
 When the fusion is complete, the section of the stele appears circular in outline 
 and the whole of the tissue is apparently (piite uniform, with the exception of a single 
 tracheid marking the jiosition of the wleni of the oUler leaf trace. Ihere follows 
 almost immediateh- tin- oUKr K :if trace which joins the solid stele, formed by the 
 fusion of the traces tiom tin- third and fouith leaf, witiinut causing an\' break. At 
 this level the first trachear\ tissue can be seen in xhv fourth leaf trace also. The 
 structure of the stem below the junction of the second and third leaf traces is the 
 same as that already described for the younger sporophyte. A very young loot was 
 
200 THE MARATTIALES 
 
 formed just opposite where the second leaf trace joined the stele. The apical cell, 
 which had apparently developed from an endodermal cell, had only undergone the 
 first division. This root was probably not the first lateral root, or if it was, it had 
 formed very much later than is usual. The base of the plant had been somewhat 
 injured and there was what looked like the remains of another root, which was 
 probably the first lateral root. 
 
 The oldest specimen of which sections were made had four fully developed 
 leaves and the rudiments of the fifth and sixth. An examination of the lower series 
 of sections showed the remains of one or two other leaves and it was probable that 
 eight leaves all together had been formed. At the level of the stem apex (fig. i8i, 5) 
 six separate leaf traces could be seen and the relation of the younger ones to the apex 
 was exactly the same as in the younger sporophyte. Below the apex, the central 
 stele of the stem shows a broken ring of procambium inclosing larger parenchyma, 
 representing the pith. The procambium ring is made up evidently of two portions, 
 one of which is certainly referable to the sixth leaf trace, while the opposite one 
 perhaps represents the beginning of the next leaf trace, although the leaf to which 
 it belongs is not yet evident above (fig. i8i, D). The junction of the leaf traces so 
 close to the stem apex makes this point very difficult to decide. 
 
 The fusion of the fifth and fourth traces follows quickly and the resulting stele 
 shows plainly the three separate xylems of its component bundles. The older 
 portion of the stele shows that it is still solid, with no evident leaf gaps (fig. i8i, 
 EG), but with several groups of tracheids probably corresponding to as many leaf 
 traces, though less distinct than in the younger plant. This is complicated by the 
 fusion of the root traces, which disturbs the arrangement of the xylems from the 
 leaf traces, but the xylem is less compact than is figured by Farmer and Hill and 
 the "pith" much less definite; indeed, one can hardly speak of a pith in this connec- 
 tion. No trace could be seen of the central commissural stiand which later makes 
 its appearance and which presumably arises in the same way that it does in Datura. 
 
 In the early stages Angiopteris appears to agree closely with the other Marat- 
 tiaceaein the development of its vascular system, but the single central stele without 
 leaf gaps is retained much longer than in the other genera and it also becomes much 
 larger and has a better-developed xylem, and the open dictyostele, formed from the 
 anastomosing of the early single leaf traces, characteristic oi' Dana-a and Marattia, 
 is not present. It is not quite clear whether the "siphonostele" with its small leaf 
 gaps, which is the next stage in the development, is made up entirely of leaf traces, 
 and it is possible that the stelar tissue, connecting the adjacent leaf traces, may be 
 composed in part of cauline tissue. On the other hand, it is quite as likely that the 
 connection of the leaf traces is brought about merely by a broadening at the point 
 of contact, such as occurs in Helminthostachys. Indeed, at this stage, the stele of 
 Angiopteris is more like that oi Helminthostachys than like that of the other Mafat- 
 tiacea?. 
 
 THE ADULT SPOROPHYTE OF ANGIOPTERIS. 
 
 In habit Angiopteris closely resembles the larger species of Marattia. The 
 enormous leaves are arranged spirally about the thick upright caudex, which is 
 covered completely by the persistent bases of the old leaves. The leaves are usually 
 twice-pinnate; the linear leaflets have more or less conspicuously serrate margins 
 with a venation very much like that of Danaa, the veins being more closely set than 
 in Marattia. In addition to the ordinary subterranean roots, large aerial roots are 
 not infrequent, but it is doubtful if there is any essential difference between the two 
 kinds of roots. 
 
THF OLDER SPOROPHYTK 201 
 
 The anatomy of the stem closely resembles that of Marattia, Inir the hiiiulks 
 seen in section are much more numerous and instead of being arranged in two 
 circles, exclusive of the central commissural strand, are usually arranged in four or 
 five. The first complete study of the arrangement of the bundles in the stem was 
 made by Mettenius (Mettenius 2). A summary of his results is given in the paper 
 by Miss Shove. Briefly stated, Mettenius describes the distribution of the bundles 
 as follows (Shove I, p. 498): 
 
 "The vascular bundles form funnel-shaped zones, with their lower ends in the 
 axis of the stem and their upper portions continued out into the leaves as leaf traces. 
 It is the transverse sections of these concentrically arranged funnels which appear 
 as the rings of separate bundles in the section of the stem. Segments from the outer 
 zone pass into the leaves as the leaf traces and the gaps thus left are filled up by 
 corresponding segments from the next inner zone." 
 
 The detailed account made by Miss Shove was based upon a moderate-sized 
 specimen from Ceylon. This stem is described as being somewhat dorsiventral in 
 structure, which is certainly not usual in Atigiopteris and may perhaps be explained 
 by the conditions under which the plant was growing. I found Angiopteris growing 
 frequently upon steep banks and it is quite possible that plants growing in such a 
 position might be obliged to bend upward, in which case a somewhat dorsiventral 
 structure would be developed. Plants growing upon level ground, so far as my 
 observations go, are always strictly radial in structure. 
 
 Miss Shove found essentially the same structure as that described by Mettenius, 
 except that she states that the leaf traces are developed exclusively from the strands 
 of the outer zone, while Mettenius states that strands are also contributed to the 
 leaf trace from the second zone. The following is taken from her paper (Shove I): 
 
 "The general scheme of the arrangement of the vascular tissue in Angiopteris 
 is almost clearly conceived by considering it in connection with the insertion of the 
 leaves. The leaf bases, which are set in a rough spiral on the stem, shcnv in their 
 lower parts a meshed segment of vascular tissue having the form of part of the sur- 
 face of a cylinder. This segment passes from the leaf ba.se into the outermost zone 
 of the stem, uniting right and left with the strands of this zone. Then, continuing 
 in an obliquely descending direction, it passes on into the second zone and so on 
 until it reaches the longitudinal axis of the stem, where it unites with other leaf- 
 trace bundles and loses all individuality." 
 
 Miss Shove found that the steles in the stem were both mesarch and endarch 
 in structure. The proto.xylem is found in groups of two or more spiral tracheids, 
 some of them at the periphery and some in the center of the stele. The number of 
 protoxylems varies with the size of the stele, the larger ones usually containing five 
 or six such groups. The earliest protoxylem appears at the periphery of the stele. 
 Protophloem is developed upon the outer side of the stele in the form of discontinu- 
 ous, small, thick-walled elements, which Miss Shove regards as sieve tubes. This 
 protophloem does not occur upon the inner side of the xylem. A remarkable pecu- 
 liarity of the bundle is the fact that the secondary sieve tubes are formed outside the 
 protophloem instead of within it. The phloem is of greater breadth on the outer 
 side of the stele than on the inner. No endodermis could be detected about the 
 steles in the stem. The usual absence of sclerenchyma w^as noted for the stem. Miss 
 Shove was unable to determine the nature of the apical meristem and the question 
 still remains open, whether the stem grows from a single apical cell orfVom a group, 
 such as is found in the larger roots. 
 
 It appears from the account of both the older and more recent investigators 
 that the complicated system of concentric meshed zones in the adult stem oi Angiop- 
 
202 
 
 THE MARATTIALES 
 
 teris is really built up of leaf traces, the so-called "compensating segment" being 
 nothing more than the lower part of a leaf trace which higher up emerges as a meshed 
 segment from the outer zone and passes into the base of the leaf. 
 
 The structure of the petiole (fig. 182, /f ) is like that of Maratti a, except that 
 the vascular bundles are more numerous and are arranged in several circles, inclos- 
 ing one or two small bundles in the center. A cross-section of a leaflet (fig. 182, B) 
 shows the vascular bundle which traverses the midrib to be horseshoe-shape in 
 section, having a central mass of tracheids, with reticulate or scalariform markings, 
 and surrounded by the phloem, largely made up of large sieve tubes, but having 
 also protophloem cells and bast. The ground tissue is composed largely of paren- 
 chyma, but on both sides below the epidermis is a conspicuous band of coUenchyma. 
 In the larger divisions of the leaf the collenchyma, as in Marattia, is replaced by 
 sclerenchyma. The structure of the lamina is very much like that of Dancra, but 
 the palisade parenchyma is even better developed. 
 
 A comparison was made of the leaves of two forms (species t) of Angiopteris, 
 one from Australia, the other from Ceylon. These showed several notable differ- 
 ences. The leaflets of the Ceylonese specimens were thinner and sharply serrate. 
 
 Fig. i%i.—Angiopte\ 
 
 A. Section of petiole from an adult sporophyte, somewhat reduced. 
 
 B. Section of a leaflet, co/, collenchyma; p, palisade tissue. X14. 
 
 C. Part of sporophvll, showing the sori. X4. 
 
 while in the Australian specimens the serrations were almost wanting, except at the 
 tapering apex of the leaflet. There were also marked anatomical diff^erences. In 
 the Australian form the palisade cells are very much elongated and are separated 
 from the epidermis by a layer of colorless hypodermal cells, and the spongy meso- 
 phyll of the lower part of the leaf is decidedly more compact than in the form from 
 Ceylon. In the latter, the palisade cells are noticeably shorter and abut directly 
 upon the epidermis. Undoubtedly the diiTerences in the anatomy of the two forms 
 are associated with the diff^erence between the moist, tropical climate of Ceylon and 
 the drier and cooler climate of Australia, and perhaps do not necessarily imply that 
 the two species are distinct, although it is highly probable that such is the case. 
 
 The roots originate, for the most part, in the inner zones, but a few may arise 
 in connection with the bundles of the outer zones and must necessarily traverse a 
 very large amount of tissue before they finally emerge. 
 
 Russow (Russow 1) described two sorts of roots — earth roots which were 
 branched and had but five xylem rays completely lignified to the center of the bun- 
 dles, and aerial roots which were much larger and unbranchcd and had twelve to 
 
POROPUVTF 
 
 203 
 
 twenty xylein rays, <)nl\' the outer traelieids lieiiii^ lii^iiitiecl. In tin- |)lant which was 
 examined by Miss Shove only the earth roots were found, and these, instead of" 
 having only five xylenis as Russow asserts, had from ten to thirteen and the xylem 
 elements were entirely lignified. The pioiiahilities are that there is no sharp line 
 to be drawn between the aerial roots and the eaith roots. Figure 183 shows a sec- 
 tion of a root from the base of a small plant from Australia, in which there were 
 fourteen xylem masses. As will be seen, there is the same circle of large mucilage 
 ducts and the tannin cells that are found in the root of Mmattia, which it very 
 closely resembles. Root hairs are nearly or quite absent, in which respect these 
 resemble the older roots of other Marattiacea^. in two cases Miss Shove found a 
 dichotomous branching of the root. This, while probably anomalous, is interesting, 
 as it recalls the method of branching in the roots of Op/iioglossiim. 
 
 ARCHANGIOPTKRIS. 
 The genus Archangioptnis. with a single species, A. henryi Christ and Giesen- 
 hagen, is at present known onlv from a single locality in .southern China. Arrli- 
 ano'iDl'tcris in habit resembles a large Dmur/i. Vhv leaves, which re.ich a length of 
 
 Fig. ii-i,.—Angiopteris. 
 A. Section of a large root, m, mucilage ducts. X14. 
 R. Part of the central cylinder of the root, fw, cndodcrmis. X70. 
 
 about a meter, have from seven to twelve leaflets, much like those o\' Dtnuvn cUipticd, 
 both in form and venation. The stalk of each leaflet is swollen in a manner that 
 suggests the nodular swellings in the petiole of Dan a- a elliptica (fig. 184). 
 
 The only account of the structure of the plant is that ot Gwynne-Vaughn 
 (Gwynne-Vaughn 2), but as he had only a fragment of a stem he was not able to 
 make a complete study of the vascular system. To judge from the fragment of the 
 stem which he examined, the leaves seem to be arranged spirally and the stem 
 is probably radial in structure. The general structure of the leaf base seems to be 
 most like that oi' Kaiilfussia. The leaf trace consists of only two vascular strands, 
 which divide up later into several (eight or nine) separate strands, arranged at the 
 base of the leaf in a horseshoe curve, quite like that of Kaulfussia. The arrange- 
 ment of the vascular system of the stipules is probably the same as in the other 
 Marattiacea'. The structure of the stem is also apparently very much like that 
 of Kniilfiissia, a section showing only one circle of bundles with a single central 
 strand. The root, which is liki' that of the other M.iiattiaciii'. h;is from seven to 
 ten xylem rays. 
 
204 
 
 THE MARATTIALES 
 
 MACROGLOSSUM. 
 
 This represents a new genus of Marattiacese, the one species of which, M. altdce 
 Copeland, has recently been described (Copeland 1). It is a large fern from Borneo, 
 and is evidently related to Angtopteris, from which it differs mainly in its simply 
 pinnate leaves, which may reach a length of about 3 meters (see Copeland 1, plate V). 
 
 TISSUES OF THE MARATTIACE^. 
 
 There are often present upon the leaf bases of the Marattiaceae peculiar lenticel- 
 like structures to which German writers have given the name "Staubgrubchen." 
 These arise beneath the stomata, and form small cavities whose peripheral cells 
 become detached and dried up, forming a dust-like powder. 
 
 Rod-shaped bodies, mainly composed of cal- 
 cium pectate, are of common occurrence in the 
 intercellular spaces of the tissues of the Marat- 
 tiaceae. Siliceous deposits and crystals of calcium 
 oxalate have also been observed in Aiigiopteris 
 (see Bitter 1). 
 
 The elements of the vascular bundles are much 
 like those in the bundles of the ordinary ferns. The 
 tracheary tissue is composed mainly of scalariform 
 elements, and the sieve tubes have numerous lateral 
 sieve plates, like those of the other ferns. Hill and 
 Farmer have noted a slight secondary formation 
 of wood in Aiigiopteris, somewhat similar to but 
 much less marked than that in the stem of Botryr/i- 
 iiivi vngitiianiiDi (see Hill I, Farmer 3). 
 
 THE SIH)R01'HYLL OF THE MARATTIACET,. 
 
 The Marattiaceae differ most strikingly from 
 the Ophioglossaceae in the character of the sporo- 
 phylls. In all of the living Marattiaceae the sporan- 
 gia or synangia are borne upon the lower surface I'lc \'&^.--Arch^„gwi,u, 
 of the leaves, which usually are not at all different 
 from the sterile ones. In Dancea (fig. 189) the sporophylls are decidedly contracted 
 and the very large synangia almost completely cover the lower surface of the leaflets, 
 but in the other genera the sporophyll is not at all contracted and much the greater 
 part of the leaf surface is free (figs. 182, 184, 185, 188). 
 
 Kaulfiissia (fig. 1 88) differs most from the other genera, the very peculiar cir- 
 cular or rarely oval synangia being scattered apparently without any definite order 
 over the whole lower surface of the leaf. In Dancea the elongated synangia, which 
 lie over the veins, extending from the midrib nearly to the margin of the leaflet, are 
 crowded so that they leave very little of the surface free. In Marattia and Arigiop- 
 teris the synangia lie over the veins as they do in Datura, but they are very much 
 shorter and are formed near the margin of the leaflet, within which they form a 
 single line which occupies only a very small part of the leaf. Archangiopteris, except 
 that the synangia are very much longer, agrees closely with Angioptens. 
 
 No proper indusium is present in the Marattiaceae, although very commonly 
 about the base of the synangium there are found a few hairs or scales which have 
 sometimes been considered to represent an indusium. In Dancra there is a growth 
 of tissue between the elongated synangia, which grows up into a sort of ridge, the 
 
 rvi (after Chri 
 nd Giesenhagen). 
 
205 
 
 top ot which is txjiaiuliil so rliar ir o\iraiclits the s\iiany;ia. I his ii(l<:;i- in section 
 appears T-shape. Whether this growth of leaf tissue between the synangia in 
 Daiicea is to be considered as an indusium may be questioned. It appears to be 
 very simihir in structure to the oblong cavir\- or fovea, within which is sunk the nor 
 very ilissiniihir s\nangiiini u[' I unirs (Hg. i S9, B). 
 
 rHi; Sl'OKANCIl'M OV THE MAKATIIACi; l- 
 
 In .Iiii^iiiptnis and Archangioptcris there are formed sejiarate sjioiangia not 
 very unhke those found in Bntrychiitm or Hclmnitltost(uli\s, but ui all the otiier 
 genera the sporangia are fused into a synangium which might be compared with 
 the spike of Ophioglossnui. The synangium in Danaa sirn/^licifaha may reach a 
 length of over 3 centimeters and contain upwaid of 100 loculi. I he first study of 
 the development was made upon Mardttia h\ Lueissen (Luersseii 3). According 
 to his statement, the differenti- 
 ation of the synangium begins 
 while the leaf is still very small 
 and rolled up between the stip- 
 ules. The tissue about a vein 
 begins to develop into an elevated 
 cushion following the vein. Upon 
 this receptacle there are devel- 
 oped two parallel ridges of tissue, 
 separated by a cleft. These two 
 ridges later grow upward and 
 meet above, so that their edges 
 completely close up the cleft, 
 which no longer shows from the 
 outside. In each half of this syn- 
 angium, very much as is the case 
 in the sporangial spike of Op/iio- 
 glossum, there are developed sep- 
 arate archesporial groups corre- 
 sponding to the separate cham- 
 bers that are found in the fully 
 developed synangium. Luerssen 
 states that the whole process takes 
 about six months for its completion (fig. 186). 
 
 In Angiopteris (fig. 187) the development of the sporangium begins at a much 
 later period, when the leaf is almost completely developed. Ihe sporangia here are 
 arranged in an oblong group or sorus, which corresponds to the synangium in Marat- 
 tia. The formation of the sorus begins, according to Goebel, as an oblong depression 
 above a young vein, and about this depression there is formed a circle of short hairs 
 which are sometimes supposed to represent an indusium. Two ridges corresponding 
 to those found in the young synangium of Marattia are formed, but upon these the 
 young sporangia develop separately, very much as they do in Botr\cliium. 
 
 Goebel states that the archesporium in Aiigioptcris can be traced back to a 
 single hypodermal cell. This cell divides repeatedly but apparently without any 
 definite order, and there is finally formed a large mass of sporogenous tissue, each 
 cell of which gives rise in the usual way to four spores. The cells about the arche- 
 sporium develop the tapetum. Goebel states that these tapetal cells are destroyed 
 before the division of the spore mother cells, but I have found that this is not the 
 
 Fig. 185.— A/araH/a douglmii. 
 
 A. Leaf from a young sporophyte. stj stipules. 
 
 B. Leaflet with synangia. X4. 
 
 C. Horizontal section of a synangium. Xio. 
 
206 
 
 THE MARATTIALES 
 
 case and that the tapctal cells persist until the division of the spores is complete. 
 Bower has confirmed this statement and found that the same condition of things 
 obtains in the other genera. 
 
 The most complete account of the development of the sporangium is that of 
 Bower (Bower 6), who has studied the development in all of the genera except 
 Archangiopteris. Except for the difference in form, there is no essential difference 
 in the development of the sporangia in Kaidfussia and Dana'a from that found in 
 Marattta. In Datuea, however, the individuality of the loculi of the synangium 
 is less clear than in the other genera. Very often Bower found that the arche- 
 sporium became divided more or less completely by parts of the sterile tissue, some- 
 what in the same fashion that the so-called "trabeculae" are formed in I sokes. 
 In DancEci, moreover, the synangium from the first is solid, and the cleft which is 
 present in the young synangium of Marattia is absent. In Kaulfussia the develop- 
 ment of the synangium differs in that a single circular ridge is formed instead of the 
 
 ; synangium of Marattia fn 
 jlder synangium. x, the 
 
 two parallel ones found in Marattia, and the loculi or chambers of the synangium 
 are thus arranged in a circle around a central pit-like depression. Bower states 
 that the sporogenous tissue of each loculus in all the forms he examined can usually 
 be traced to a single mother cell. He also found that the tapetum always arises 
 from the cells adjacent to the archesporium, and that normally all of the sporogenous 
 tissue develops into spores. In these respects the Marattiacea? closely resemble 
 H elmi uthostachys and Botrychiiim. 
 
 In Daiicea and Kaulfussia there is no mechanical tissue representing the annulus 
 found in the more specialized ferns. The dehiscence of the sporangium in these 
 forms is brought about merely by a shrinking of the cells on either side of the slit by 
 means of which each loculus opens. This slit is very short in Daua-a and may finally 
 appear as a circular pore, but it is not essentially different from the more elongated 
 slit found in Kaulfussia and Marattia. In the latter there is developed, in the outer 
 tissues of the synangium, mechanical tissue which causes the two halves of the 
 synangium to separate, very much like the two covers of a book, and the elongated 
 slit from each loculus opens into this space between the widely separated halves of 
 the synangium. 
 
OLDliR SrOKUl'llV 
 
 207 
 
 Ihe ckliisccruu of tlic iiidiv idiial loculi is affcctc-d by the contractiiin of the 
 tliiiiiici-wallcil cells siii loumled by firmer tissue. The spoian^iuiii (A ."ingiopteris is 
 undoLibtedly a more speciahzed structure than the synaiifjium of tlie other Maratti- 
 ace;e. In Angtoptcns each individual sporangium has the wall on the outer side 
 much thicker than that on the inner one, and the superficial cells have their walls 
 much thickened. The inner wall is sometimes composed of but one layer of cells 
 outside of the tapetum, but more commonly there are one or two layers of cells 
 between the tapetum and the epidermis. Near the top of the sporangium on its 
 outer side there is a transverse band of cells with thicker walls, and these constitute 
 a rudimentary annulus very much like that found in the Osmundaceae. By the 
 contraction of this thickened annulus the longitudinal slit on the inner face of the 
 sporangium is made to open widely at maturity. 1 he number of s|iorcs produced 
 in each loculus, according to Bower, is approximately 1,750 for Dmian, 7.500 foi 
 KauljnssKi, 2,500 for Marattia, and 1,450 for Angiopteris. 
 
 Which type of sporangium in the Marattiacea; is the more primitive is very 
 difficult to say, as both the free sporangium like that of Angioptrris and the compact 
 
 synangium like that of Manittia 
 and Daiicca are of about equal 
 antiquity, so far as the geological 
 record goes. It must be remem- 
 bered that the living Marattiace^e 
 are almost certainly merely a few 
 isolated fragments of a once large 
 group, and it is by no means 
 necessary to assume that the spo- 
 rangia of the living forms must 
 necessarily all conform to a com- 
 mon primitive type. It seems 
 quite as likely that the fVee spo- 
 rangia, like those of Angtoptcns, 
 may have originated directly from 
 some ancient prototype which 
 resembled, perhaps, forms like 
 , older sporangiu,,,. BoUychi um OY H Ami nthostculivs, 
 
 r. the annulus; Mhcrorsistent tapetum. X75. -' -i i l \ 
 
 while the genera with the solid 
 synangium like Dtirura may have come from forms with completely united spo- 
 rangia like Opiiioglossum. 
 
 It appears from a study of the most ancient ferns that these bail dinioipliic 
 leaves, the sporophyll probably resembling the fertile spikes of Botryihimii or the 
 fertile leaf segments of Osmuiuia. How the modern Marattiace;e originated fVom 
 forms of this type is not by any means cieai. i'he development in Hclminthostaihys 
 of sterile leaf-like lobes associated with the sporangia may perhaps afford a clew to 
 the method of sterilization by which the sporangiophores of some ty|ie allied to the 
 (^phioglossace;e may gradually have developed sterile green leaf segments, bearing 
 upon their lower surface sporangia or synangia like those of the modern Marattiace;c. 
 These sterile segments found in Hchniutlmstachys are occasionally strikingly leaf- 
 like, and as the groups of sporangia in Helvitnthostachys are sometimes imited into 
 small synangia, the development of similar .synangia upon the lower suitaii ol siu h 
 green sterile segments of the sporangiophore is quite C()ncei\ abli-. 
 
 Fio. \%T.—Angiopterii. 
 .A, B. sections of young sporangia (after Goebel), 
 
 Another explanation might be the coalescence of tht 
 of a sporophyll like that found in the Ophioglossace;e, 
 
 fertile a: 
 bur the 
 
 the 
 
 that thi 
 
 spo 
 
208 
 
 THE MARATTIALES 
 
 langiophoie in all of the living Ophioglossaceae is adaxial, while in the Marattiaceae 
 the synangia are abaxial, is a serious objection to such a hypothesis; and the most 
 plausible explanation at present seems to be that of a progressive sterilization of 
 part of the sporophyll itself and the expansion of the sterile tissue into the broad 
 lamina of the leaf bearing the separated sporangia or synangia upon its lower surface. 
 It must be admitted that the difference between the sporophylls of the existing 
 Marattiaceae and those of the Ophioglossaceae is very great, and the leaves of the 
 former, superficially at least, are much more like those of the leptosporangiate ferns; 
 but the sporangia themselves are very similar in their development to those of the 
 
 Ophioglossaceae, and offer no obstacle to the assumption of a fairly close relationship 
 between the Ophioglossaceae and the Marattiaceae, which is indicated by the history 
 of the development of the gametophyte, and the vegetative organs and tissues of the 
 sporophyte. Perhaps the most marked difference between the tissues of the Marat- 
 tiaceae and the Ophioglossaceae is the prevalence of the mucilage ducts in the tissues 
 of Marattiaceae, but these arise rather late in the history of the sporophyte, and it is 
 
 Fig. ii^.—Danxajamakensh. 
 • . A. Base of a fertile leaflet, showing synangia, up. X2. 
 
 B. Transverse sections of three synangia. X12. 
 
 C. Horizontal section of a synangium. 
 
 significant that they are less developed in the presumably more primitive Kaulfiissia 
 than they are in the more specialized types like Angiopteris. Another difference is 
 the development of sclerenchyma, which is quite absent from the Ophioglossaceae; 
 but this again is also quite absent from Kaulfussia, which, on the whole, must be 
 considered to be the most primitive of the living Marattiaceae and also the form 
 which most closely approaches the Ophioglossaceae. 
 
PARr III. THl^ ORIGIN AND RI:LA llONSI 1 1 PS ()!< 
 1HI<; PlUSl^ORANGIAT/i:. 
 
 IIk luspoiangiate ferns as they now exist represent l)ut a fVu isolatKl hagnu nts 
 (ifwliat in earlier geologic time was presumably a very nuicli larger and more (.(in- 
 nected gr()U|>. As the geological history of these forms is very far from clear (this 
 being especially true of the Ophioglossaces), we are, perforce, dependent niainl\- 
 upon a comparative study of the few existing types for information concerning their 
 relationships. In the foregoing pages an endeavor has been made to trace the de- 
 velopment of these forms, both in their gametophytic and early sporophytic stages, 
 and the result of these studies has been to confirm the belief that a real genetic 
 relationship exists between the Marattiace^e and the Ophioglossacex'. 
 
 While the three genera of the Ophioglossace;e differ in certain particulars from 
 each other, there is no question as to their being comparatively closely related, and 
 the same is true of the different genera of the Marattiace;e, although perhaps the 
 differences here are somewhat greater than in the Ophioglossaceae. 
 
 It has been assumed in these studies that the sporophyte of the ferns is the 
 result of a progressive specialization of the sporogonium of some form allied to the 
 Bryophytes, though it is highly improbable that any of the existing Bryophytes are 
 directly related to this progenitor of the primitive ferns. No attempt will be made 
 here to discuss the reasons for accepting the "antithetic" theory of the alternation 
 of generations, rather than the homologous theory. These reasons have been set 
 forth at length elsewhere. 
 
 It is hopeless to expect that any satisfactory fossil traces will be found of these 
 predecessors of the true ferns. The fern type is exceedingly ancient, but it must 
 have been preceded by simpler forms connecting it with some bryophytic type. 
 These forms, as well as the earlier true Pteridophytes, were almost certainly plants of 
 small size and delicate texture, probably not very unlike some of the small and 
 delicate species of Ophioglossum. Such plants, composed entirely of soft, pciish- 
 able tissues, could hardly be expected to leave recognizable fossil traces, and their 
 absence from the ancient Paleozoic rocks is not to be wondered at. 
 
 There still exist, however, among the Bryophytes, certain forms which, if they 
 are not directly related to the eusporangiate ferns, nevertheless show many striking 
 similarities in structure, which help to explain at least what may very well have been 
 the character of the liverwort-like ancestors from which the ferns are descended. 
 Among the living Bryophytes, as is well known, an interesting series of types may be 
 traced, showing the gradual increase in the importance of the neutral generation — 
 the sporophyte — starting as little more than a mass of spores, finally by a progressive 
 sterilization of what wasoriginalh' sporogenous tissue, and an accompanying special- 
 ization of the sterile tissues thus formed, attaining a condition of almost complete 
 independence. The importance of this process of sterilization of potentially sporo- 
 genous tissue, has been especially clearly expounded in the works of Professor Bower. 
 
 In two classes, the true mosses or Musci and the horned liverworts or Antho- 
 cerotes, the sporophyte continues its growth for several months and develops an 
 elaborate system of green, assimilative tissue, quite comparable to that found in 
 the vascular plants. The spore-producing function is correspondingly subordinated 
 and the spore formation is delayed until a late period in the life of the sporophyte. 
 14 2oy 
 
210 
 
 ORIGIN AND RELATIONSHIPS OF THE EUSPORANGIATAE 
 
 The large, green sporophyte, however (fig. 190, A, sp), never attains complete inde- 
 pendence, as no proper root is developed and it is dependent for its water supply 
 upon the gametophyte. 
 
 In all of the PZusporangiates, the embryo is much later in developing its organs 
 than in the Leptosporangiates, and thus resembles the Bryophytes; moreover, the 
 embryo reaches a very much larger size before it attains its independence. Even 
 after the young sporophyte has developed several roots and leaves, it may still 
 
 Fig. 190. 
 A. Gametophyte of Anihotcros sp. with the large 
 sporophyte, 5/>, attached to it. Upper part 
 of spnropliyte is split into valves which allow 
 the escape of spores. 
 B, C. Two gametophvtes of Marattta douglasii, with 
 attached ^jiorophytc. Sporophyte emerges 
 upon tipper side of gametophyte, very much 
 as in Anthoicro^, but a root, r, is developed 
 which penetrates lower side of gametophyte 
 and connects the young sporophyte with the 
 earth, g, gametophyte; jp, sporophyte; 
 /, first leaf; r, primary root. 
 
 maintain its connection with the gametophyte and be to a certain extent dependent 
 upon it in a way which forcibly recalls the relation of gametophyte and sporophyte 
 in the mosses. This condition, which furnishes a very strong argument in favor of 
 the primitive nature of the Eusporangiates when compared with the Leptosporan- 
 giates, is especially pronounced among the Ophioglossace^e, where it may even go so 
 far that spores are developed before the connection of gametophyte and sporophyte 
 is finally sundered. 
 
 In the true mosses the sporophyte is exceedingly specialized and shows but a 
 very remote resemblance to that of the Pteridophytes. In Aiithoccros, however, the 
 
 Fig. 191. 
 
 Diagrams which show the similarity in form of the very 
 young sporophyte in AnthoceroStA^ and Ophioglossiini^ 
 B. In the former the upper portion becomes an elong- 
 ated capsule within which spores are produced. The 
 shaded portion shows layer of sporogenous tissue sur- 
 rounding central sterile tissue or columella. The 
 zone, m, between capsule and foot, /, is composed of 
 meristematic tissue. In Ophioghsmm the upper part 
 of the young sporophyte forms the primary leaf, or 
 cotyledon, cot, which is not sporogenous. The colum- 
 ella of Anihoccros is replaced by the axial vascular 
 bundle, v.b. At the junction of the cotyledon and 
 foot arises the primary root, r, which finally penetrates 
 the foot and enters the ground. 
 
 mosl highly developed of the horned liverworts, there are a number of very significant 
 structural details that are very strongly reminiscent of the young sporophyte of 
 Oph'toghssum moluccanum. In both of these forms the young embryo has a very 
 large foot and a conical upper region (fig. 191). This upper portion in Ophioglossum 
 develops into the cotyledon; in Anthoceros, into the spore-bearing part of the sporo- 
 phyte. Above the foot in Anthoceros there is developed a zone of meristematic 
 tissue by means of which new cells are added to the base of the growing spore- 
 
ORIGIN AND RELATIONSHIl'S OK THE EUSPORANGIATAE 
 
 211 
 
 phyte, and in Op/iioglossum much the same thing occurs in the cotyledon, although 
 at first this grows from a single apical cell. It is probable that in the ancestors of the 
 Ophioglossace.ne this upper conical portion of the embryo was developed directly 
 into a spore-bearing organ. There is, however, a marked difference which soon 
 appears, viz, the development of a true root in Op/iioglossutii. This root, pushing 
 down through the foot, penetrates the lower side of the gametophyte and places the 
 young sporo|)hyte directly into communication with the water supply from the earth, 
 and thus the latter becomes entirely independent. Were the large sporophyte of 
 Antlioceros to develop a similar root from the basal meristem, it also would become 
 entirely self-supporting. Indeed, so marked are the resemblances in the early stage of 
 development that they make the inference almost irresistible that the Ophioglossaceae 
 must have descended from some simpler forms whose sporophyte bore a strong 
 resemblance to Antlioceros. This "pro-Ophioglossum," if we may call it so, pre- 
 sumably produced spores upon the first leaf, instead of its being a sterile cotyle- 
 don as in the existing forms. From a study of the development of the sporophyte 
 
 192.- 
 
 Comparison of sexual organs in Anthoccrotacca; and Marallia. 
 
 of the thallus of Anihoccros pearsoni Howe, showing an anthcridium mother cell, 
 
 A. Longiludin; 
 
 witli the superimposed cover cell {d). 
 
 B. An older anthcridium within a cavity covered by a double layer of cover cells. 
 
 C. Young archegonium of Xtegaceros IjiboiUnsii Campbell, a form closely related to Anihoceios. 
 
 D. Sections of young antheridia of Marattia tiouglasii Baker. The inner cell divides at once into the sperm 
 
 cells, and the cover remains single-layered, dy the cover. 
 
 E. Young archegonia of Mara/f/fl rfoHjj/ajiV. I, the basal cells of the archegonium; </, neck. 
 
 in O. moluccaniim, it may be inferred that this ancestral form had no stem, but 
 consisted simply of this spore-bearing leaf and a root. The sporangia were presum- 
 ably simpler than in the existing forms of Ophioglossiiui, and perhaps intermediate 
 in character between such sporangia as those of the smaller species of Opiiioglossutn 
 and the imperfectly segregated spore groups which occur in the Anthocerotes. 
 
 Not only does the sporophyte of Antlioceros show important resemblances to 
 that yif Opliioglossnni, but the development of the sexual organs also shows striking 
 analogies. The archegonium, unlike that of other Bryophytes, is sunk below the 
 surface in much the same fashion as that of the eusporangiate ferns, and the short 
 neck of the archegonium in the latter is probably to be compared to the four terminal 
 neck cells which may occur in the Anthocerotes (fig. 192). The endogenous anther- 
 idia of the Anthocerotes also may perhaps be compared to *he sunken antheridia of 
 the eusporangiate ferns, and in their early stages the resemblances are very close. 
 
212 ORIGIN AND RELATIONSHIPS OF THE EUSPORANGIATAE 
 
 Ihe most marked difference in the character of the reproductive organs is the 
 spermatozoid. The spermatozoids in the luisporangiates, especially in Ophio- 
 glossum, are very large and possess numerous cilia, while in Anthoceros they are 
 minute and have but two cilia, like the other Bryophytes. This is perhaps the strong- 
 est reason for assuming that there is not a direct connection between the Anthoce- 
 rotes and the Ophioglossaceae; but both gametophyte and sporophyte have so many 
 points in common that it may be pretty safely assumed that the progenitors of the 
 Ophioglossaceae were not very different, in appearance at least, from the living //«- 
 thoceros. Whether the differences in the spermatozoid are secondary remains to be 
 seen, but in view of the extraordinary constancy of the form of the spermatozoids 
 in all of the main groups of the Archegoniates, one would certainly expect that large 
 multiciliate spermatozoids would be found in the ancestors of the Kusporangiates, 
 supposing these should ever be discovered. 
 
 There is no question that the subterranean prothallium of the Ophioglossaceae 
 is a secondary condition, derived from some green gametophyte, probably very much 
 like that of the Marattiaceae. Not only has the chlorophyll been lost, but in Op/iio- 
 glossum and Helminthostachys the dorsiventral form of the gametophyte has been 
 replaced by a radially symmetrical thallus. While this may be partially explained 
 as the result of the absence of light, it must be remembered that the equally sub- 
 terranean prothallium of Botrychiiim is dorsiventral, although the reproductive 
 organs are borne upon the dorsal surface and not upon the ventral one, as they are 
 in most ferns. Whether this position of the reproductive organs is the result of the 
 conditions of growth or whether it indicates that the green gametophyte from which 
 this saprophytic underground form is descended also bore the reproductive organs 
 dorsally, as most liverworts do, can only be conjectured. In the Marattiaceae 
 antheridia are not uncommon upon the dorsal surface, and according to Jonkmann 
 archegonia also may occur dorsally. Whether this is normal or is the result of unusual 
 light conditions was apparently not investigated. The formation of archegonia in the 
 ordinary ferns may be induced upon the dorsal side of the prothallium, provided 
 it is illuminated from below. If the illumination is equal on both sides, archegonia 
 will develop both on the ventral and dorsal surface. 
 
 Of the investigated species of Ophioglossiim, 0. moluccanum and the nearly 
 related and perhaps identical O. penduncnlositm probably represent the most primi- 
 tive type. In these species there may be a feeble development of chlorophyll in the 
 gametophyte under certain conditions and the saprophytic habit is much less pro- 
 nounced than it is in the other species of Ophioglossiim and Botrychium. 
 
 The development of an endotropic mycorrhiza in the prothallium of the 
 Marattiaceae is an interesting suggestion as to the probable beginning of the sapro- 
 phytic habit which characterizes the gametophyte of the Ophioglossaceae. There 
 seems to be good reason to suppose that the peculiar type of symbiosis, which is 
 developed so highly in the Ophioglossaceae, began by the development of a fungus 
 associate in some green gametophyte like that of the Marattiaceae. 
 
 The reproductive organs of the C^phioglossaceas and Marattiaceae are very 
 similar indeed. It is probable that the short-necked archegonium found in the 
 Marattiaceae and in Ophioglossiim is more primitive than the long-necked arche- 
 gonium of Botrychium or Helminthostachys. This conclusion is based upon the 
 assumption that the four-necked archegonium of the ferns is a development of the 
 four terminal cells in the neck of the Bryophyte archegonium, Anthoceros most 
 nearly approaching this hypothetical ancestral type. This being the case, the forms 
 with the shortest neck would most nearly resemble this assumed ancestral type. 
 
ORIGIN AND RELATIONSHII'S Ol- THK KUSPORANGIATAE 213 
 
 The question of tlif canal cells is a |ni/./ling one. In the Maiattiacea-, with the 
 exception of D/ui.ni, two neck canal cills, or at any rate two nuclei, are present and 
 the ventral canal cell is consjiicuoiis. In Dainra, however, the canal cells are very 
 much less perfectly developed, the ventral canal cells especially heing exceedingly 
 difHcult to demonstrate. In the latter respect Daiuva shows a remarkable resem- 
 blance to the ()phioglossace:e, especially to Ophioglossuni, where the ventral canal 
 cell is equally difficult to demonstrate. Ihis apparent degeneration of the ventral 
 canal cell is probably secondary, as all of the Bryophytes and the other Pteridophytes 
 have the ventral canal cell ccjnspicuous. 
 
 The antheridium in all of the Marattiace;e is almost exactly like that of Op hi o- 
 glossum, while Botrychnim and Helminthostachys differ in having the outer wall of 
 the antheridium more or less completely two-layered. If the antheridium is derived 
 from a type like that of Anthoceros (fig. 192), where the antheridial cavity is covered 
 by a double layer of cells, it would seem that Botrychiiim and Helminthostachys are 
 more primitive than Ophioglossuni or the Marattiace;e in this respect. 
 
 Of the Eusporangiates, there is no question that the type of embryo found in 
 Ophioglossum moliiccaniim comes the nearest to the assumed ancestral form. This 
 hypothetical pro-fern may be assumed to have developed simply a spore-bearing 
 organ or sporangiophore, perhaps more or less leaf-like, and a root developed from 
 the inner part of the massive foot in the same way that the root actually does develop 
 in the embryo of the Marattiaceae and Ophioglossum mohucanum. Except that the 
 cotyledon of O. moluccanitm is sterile, the young sporophyte in this species is actually 
 like this hypothetical type, viz, it is composed simply of root and leaf. It is probable 
 that the fertile spike of Ophioglossum is not unlike the primitive sporangiophore, 
 and the ancestral form presumably developed such a spike-like sporangiophore from 
 the first, the sterile lamina being a secondary structure arising perhaps from a basal 
 meristem, like that found in the sporophyte of Anthoceros. The early development 
 of the spores in the Ophioglossaceae is a further indication of their primitive nature. 
 
 The monophyllous condition which prevails among the Ophioglossacea:' must 
 also be regarded not as a secondary condition but as a primitive one, and a study of 
 the development of the sporophyte lends no support to the theory that we have to 
 do with a reduced strobilus. Of the Marattiaceae, Kaulfussia normally approaches 
 the monophyllous condition more closely than any other genus. The older plants, 
 as a rule, have only two or three leaves expanded at one time and there may fre- 
 quently be but one. In strong contrast with this is the dense crown of leaves found 
 in Marattia and Angioptciis, although even in these forms the number of leaves is 
 less than it is in the majority of leptosporangiate ferns. The tendency to the mono- 
 phyllous condition is shown in the younger plants of Marattia, developed from the 
 buds on the old leaf bases, where, as a rule, only one leaf is expanded at a time. 
 
 The young sporophyte of O. molurranutti has ?io stem. Without exception, for 
 a considerable time at least, the vascular system of the sporophyte is composed 
 exclusively of tissue belonging to the leaves and roots, the stem apex playing no 
 part in the building up of the vascular skeleton. The theory that there is a special 
 stele belonging to the stem, of which the leaf traces and the root traces are sub- 
 sidiary structures, is not borne out by a detailed study of the evolution of the vascular 
 system in either the Marattiace.e or Ophioglossacc;c. In all of these that I have 
 examined the vascular system begins as a single strand common to the primary 
 leaf (or cotyledon) and the primary root, and the more complicated vascular system 
 of the older stem for a very consitlerable period is built up exclusiveh' b\- additions 
 (it new liaf traces or root steles. 1 his condition is permanent 111 rlu- ( )phioglos- 
 
214 ORIGIN AND RELATIONSHIPS OF THE EUSPORANGIATAE 
 
 saceae. In the Marattiaceae small cauline bundles (commissural bundles) may be 
 developed at later stages, but these are insignificant compared with the complex of 
 united leaf traces. 
 
 The antithetic theory of alternation implies that the sporophylls are older 
 structures than the sterile leaves, and those ferns in which the sporogenous function 
 of the sporophylls is most pronounced may be assumed, other things being equal, 
 to be the most primitive. In this respect, as well as in the usually monophyllous 
 condition, the Ophioglossaceae are distinctly more primitive than the Marattiaceae. 
 The sporangiophore or fertile leaf segment is a practically independent structure, 
 being differentiated at an extremely early period in the development of the leaf. 
 In the Marattiaceae the sporangia are subordinate and the leaves upon which they 
 are borne usually differ in no respect from the sterile leaves. The most marked 
 exception to this is Da?icra, in which the sporophylls differ decidedly from the sterile 
 leaves and the synangia are much larger than those of the other genera, although 
 the leaves themselves are usually smaller. 
 
 Bower thinks that the circular synangium of Kaulfiissia is the most primitive 
 type among the existing Marattiaceae, but there is some reason to doubt the sound- 
 ness of this conclusion; at any rate, on theoretical grounds, the large synangia and 
 contracted sporophylls of Dancca must be assumed to be more primitive than the 
 type of leaf found in Kaiilfussia. In regard to the character of the sporophylls, 
 therefore, none of the living Marattiaceae can be regarded as being very primitive. 
 All of the oldest fern fossils show the sporangia to be borne upon special leaves or 
 leaf segments, in which the lamina is nearly or quite absent (Scott 1, Bower 9). 
 In this respect, therefore, these ancient fossil ferns were more like the Ophiogloss- 
 aceae than like the Marattiaceae, although their sporangia came nearer to the type 
 of Marattiaceae. It is by no means impossible that these oldest ferns, e. g., the 
 Botryopterideae, were related to the Ophioglossaceae. 
 
 The single apical cell found in the stem apex throughout the life of the sporo- 
 phyte in Ophioglossum is probably a more primitive condition than the group of 
 cells found in Angwpteris. It still remains to be seen whether the single apical cell 
 found in the other Marattiaceae in their early stages is persistent throughout life or 
 is replaced, as in the roots, by a group of initial cells. 
 
 The single axial strand of collateral structure throughout cotyledon and loot 
 may be taken as the starting point for two types of vascular skeleton which have 
 arisen from it. The first is that of Ophioglossum, where, with the development of 
 the new leaves, there is built up the wide-meshed cylindrical network or dictyostele, 
 composed of single collateral leaf traces. A further development of the same type 
 results in the much more complicated dictyosteles of the Marattiaceae, where, more- 
 over, the collateral bundles are replaced by concentric ones. This greater com- 
 plexity is due primarily to the much larger leaves of the Marattiaceae, in which the 
 leaf traces are compound. In the earlier stages of these forms, however, we have 
 seen that the young leaf traces are single and the structure of the vascular skeleton 
 exactly as it is in Ophioglossum. Of the Marattiaceae, Kaulfussia more nearly 
 retains this primitive condition than the other genera and next to this comes Dancea, 
 while further removed from the primitive type is the exceedingly complex skeleton 
 found in the massive stem of Angiopteris. The concentric bundles, characteristic 
 of the Marattiaceae, are presumably of secondary origin and we find that in the 
 early bundles, especially of Danaa, a true collateral structure is present. 
 
 The second type of skeleton is that found in Botrychium and Helmmthostachys. 
 This is a solid, hollow cylinder with inconspicuous leaf gaps, resulting from the 
 union of the broad leaf traces, which fuse completely to form this hollow stele. That 
 
ORIGIN AND RELATIONSHIPS OF THE EUSPORANGIATAi: 215 
 
 the cyliinlriiiil Inindlc, or siphonostele, is not due to the formation of" :i pitli within 
 a protostele is clearly shown hy the study of the developmcnt|^of the hundle in the 
 young sporophyte of lioirycliiiim, where it can easily be seen that the component 
 bundles are separate at first, and that the pith, so-called, of the siphonostele is merely 
 a portion of the ground tissue that is included between them, and which later be- 
 comes entirely separated from the cortical tissue. A similar condition of things may 
 be found in tracing the development of the vascular cylinder in the young stem of 
 Helminthostachys. The development of the young bundles in Angioptcns is more 
 like Botrychimn than Ophioglossuni, and this, in connection with the separate 
 sporangia that occur in Augiopteris, suggests that perhaps the type ot the Marat- 
 tiaceae represented by Augiopteris may have originated independently from forms 
 like Helminthostachys, instead of having had the same origin as tiiose forms in 
 which true synangia are developed. 
 
 The essential similarity of the leaf structures throughout the Ophioglossacea' 
 is sufficiently clear, antl the steps in the increasing complexity of the sporophyll can 
 be easily traced in existing forms, ranging from the undivided fertile and sterile 
 segments of Ophioglossiim to the decompound leaves and much-branched panicle 
 oi Botrychium. Helminthostachys is much the most aberrant of the Ophioglossaceae 
 and in many respects shows a marked resemblance to the Marattiaceie. The form 
 and venation of the leaves strongly recall Dana-a or Angiopteris, except that the 
 leaves are not pinnately divided, in which respect they closely resemble Botrychium. 
 It must be remembered, however, that the ternate form of leaf characteristic of 
 Botrychium and Helminthostachys is very generally met with in the early stages of 
 the sporophyte in the Marattiaceae. The anatomy of the leaf in Helminthostachys 
 very closely resembles that of the Marattiacere, there being a well-developed palisade 
 layer in the mesophyll, a character which is either wanting entirely or very imper- 
 fectly developed in Botrychium. 
 
 The leaves of the Marattiaceae, except in Kaulfussia, in their form and in the 
 circinate coiling of the young leaves, suggest a relationship with the leptosporangi- 
 ate ferns rather than with the Ophioglossaceae. However, the larger species of 
 Botrychium and Helminthostachys show an approach to this circinate form of the 
 young leaf, this being especially conspicuous in the young sporophylls oi Botrychium 
 virginianum. With the exception of Kaulfussia, the general form and venation 
 are alike in all the Marattiace;e and have their nearest analogy in Helminthostachys. 
 There seems no reason to assume that the stipular sheath in the Marattiaceae is dif- 
 ferent in its nature from that of the Ophioglossaceae, especially Helminthostachys 
 and Botrychium. The ternate form of the leaf may be persistent in Kaulfussia, 
 where, however, the venation is more like that of Ophioglossum. Whether this 
 occurrence of both reticulate and pinnate venation in the two families is a case of 
 parallel development, or whether it indicates the connection of the families at dif- 
 ferent points, it is at present impossible to say. 
 
 In the structure of the roots the Marattiaceae find their nearest ally in Hel- 
 minthostachys, where the number of the xylem rays, four to seven, equals that of 
 Kaulfussia, which has the simplest structure among the Marattiacea% and, as we 
 have seen, probably on the whole is most nearly related to them. Helminthostachys 
 also resembles the Marattiacex in the character of the apical growth of the root, 
 where, although there is a single apical cell, such as occurs in the young roots of the 
 Marattiacea-, its form is more like that of the Marattiaceae than it is like that of 
 Botrychium. The endophytc which occurs in the older roots of the Ophioglossacea', as 
 well as in the primary one, I have not fouml in the older roots of the Marattiaceae, at 
 least notas a regularthing, although it seems to be always present in the primary root. 
 
216 ORIGIN AND RELATIONSHIPS OF THE EUSPORANGIATAE 
 
 The Ophioglossaceae and Marattiaceae are alike in the absence of any mechan- 
 ical tissues in the cortex of the stem. Sclerenchyma, which is well developed in 
 the leaves of most of the Marattiaceae, seems to be entirely wanting in all of the 
 Ophioglossaceae, but the coUenchyma which replaces sclerenchyma in certain 
 organs of the Marattiaceae, and is the only form of mechanical tissue in Kaidfttssia, 
 is occasionally found in the Ophioglossaceae. A notable case of this is the broad and 
 very conspicuous zone of coUenchyma surrounding the root bundle in Ophioglossiim 
 palmatum. The mucilage ducts so characteristic of Marattiaceae are apparently 
 quite wanting in all of the Ophioglossaceae, and the tannin sacs which are a nearly 
 constant feature in the Marattiaceae are absent in Botrychium and Ophioglossum, 
 but occur in Helmmthostachys. 
 
 In the general structure of the vascular bundles there are important resem- 
 blances. In both families the first-formed tracheids in the young sporophyte are 
 all of the short, reticulate form, and this is retained permanently in Ophioglossum, 
 but is replaced later by pitted tracheids in Helminthostachys and Botrychium. 
 Hchninthostachys alone among the Ophioglossaceae shows spiral protoxylem ele- 
 ments like those found in the later bundles of the Marattiaceae. 
 
 It is probable that the collateral bundles of Ophioglossum represent the prim- 
 itive type from which the concentric type found in the Marattiaceae has been derived. 
 In Botrychium and Helminthostachys, while the collateral type of bundle is found 
 in the stem, in the petioles the bundles are concentric, as they are in the Marattiaceae. 
 The collateral type of bundle reappears again in the early stages of the sporophyte 
 in Dancea. 
 
 Angiopteris, which is probably the most specialized of the living Marattiaceae, 
 has concentric bundles only. 
 
C()N(-LUSION. 
 
 W'liilf tin- nioiv specialized Maiattiacea', like Muruttui and Angioptcris, 
 apparently differ very much from the Ophioglossacea-, it appears from a study of 
 the development of these forms, as well as that of the simpler and presumably more 
 primitive genera Dana:a and Kaulfussiti, that the most conspicuous differences are 
 of secondary rather than of primary importance, and the conclusion is justified that 
 the two families of the I'^usporangiata- really arose from the same primitive stock. 
 Of the Ophioglossaceae, Helmmthostachys is undoubtedly the form which, nn rlu- 
 whole, comes nearest to the Marattiacea;, although Ophioglossinn is more like them 
 in the character of the reproductive organs and in the vascular skeleton of the 
 sporophyte. Op/iioglossiirn also recalls Kaiilfiissm in the form and venation of 
 its leaves. 
 
 While the green gametophyte in the Marattiacea' is undoubtedly a more 
 primitive structure than the saprophytic prothallium of the Ophioglossaceze, the 
 sporophyte of the latter is certainly a more primitive one than that of the Marat- 
 tiacea. The monophyllous condition, the early development of the sporangia, and 
 the fact that the whole of the sporangiophore is spore-bearing, all point to this. 
 
 The way in which the sporophyll of the Marattiacea has been derived from a 
 type like that of the Ophioglossacea is not clear, and it is by no means certain that 
 all of the Marattiacea have been developed in the same way. The occurrence of 
 fossil forms with sporangia of the marattiaceous type grouped together — much like 
 the sporangia of Botrychiiim orOsmimda — suggests that such a condition might have 
 been preliminary to the separation of the sporangia by the development of sterile 
 green tissue in the sporangiophore, such as sometimes occurs in Helminthostarhys. 
 'fhe converse is not tenable unless we accept the view that the sterile leaf is an 
 older structure than the sporangiophore, a view which we believe is not warranted 
 by the facts. It has been suggested in a previous chapter that the sporophyll of the 
 Marattiacea might be accounted for by a theory of concrescence of the sporangiophore 
 and sterile lamina in the Ophioglossaceae. Ophtoglossum palmatum, with the two 
 series of small sporangiophores and broad leaf lamina, is the form which comes 
 nearest to such a condition. A very serious objection to this view was mentioned, 
 however — the fact that the sporangiophore is an adaxial structure, whereas the 
 synangia of the Marattiacea are always abaxial in their position. 
 
 Comparing the eusporangiate ferns with the leptosporangiate, it is generally 
 conceded that the Osmundacea are the ferns which most closely approach the 
 Euspoiangiata. The Osmundacea show points of resemblance to both the Ophio- 
 glossacea and Marattiacea, this being true both of the vegetative tissues and the 
 sporangia. In Osmunda the arrangement of the sporangia suggests Botrycliium, 
 while in Todea the sporangia are borne upon the backs of the leaves, much as they 
 are in the Marattiacea. Bower has suggested that the Gleicheniacea, which are, 
 however, certainly allied to the Osmundacea also, show certain suggestions of a 
 marattiaceous affinity, especially in the arrangement of the sporangia. 
 
 We may then briefly summarize our conclusion as follows: From some form, 
 allied to the simpler existing species of Ophioglossiirn, the whole fern series is 
 descended. In this whole series the leaf is the predominant organ, the stem at first 
 being quite subordinate in importance. This ancestral fern was monophyllous and 
 the leaf at first was a sporophyll, perhaps without any definite sterile segment. 
 
 217 
 
218 CONCLUSION 
 
 From this central type presumably several lines diverged, of which only a few frag- 
 ments persist. One of these is seen in the different forms of Botrychium, to which 
 Helminthostachys is probably not very remotely allied. The whole of this series is 
 characterized by a subterranean gametophyte, and a more or less saprophytic habit 
 of the sporophyte is indicated by the development of a mycorrhiza in the roots. 
 
 The Marattiaceae, as they now exist, probably do not represent a single un- 
 broken line of descent, and show strong evidences of a possible multiple derivation 
 from the earlier stock. The point of contact with the Ophioglossacea; is probably 
 in the neighborhood of Helminthostachys, which, on the whole, is more like the 
 Marattiaceae than are the other Ophioglossaceae; but it is improbable that the solid 
 synangium, such as characterizes most of the Marattiaceae, was derived from a 
 group of separated sporangia like those in Botrychium or Helminthostachys, and it is 
 more likely that they arose from a structure which resembled the spike of Ophio- 
 glossum. Angiopteris is, with little question, the most specialized of the living 
 Marattiaceae and has probably departed further from the ancestral type than any 
 of the other forms, while Kaulfiissia, on the other hand, is probably the most 
 primitive. On the whole the Marattiaceae are nearer to the leptosporangiate ferns 
 than the Ophioglossaceae are, and it is likely that the Leptosporangiates are directly 
 descended from some ancient fern forms, allied to the Marattiaceae, but differing 
 from any of the existing types. 
 
BIBLIOGRAPHY. 
 
 Arber, E. a. Newell. 
 
 I. On the Past History of the Ferns. Ann. I?ot., 1906, 20, 215. 
 Atkinson, G. F. 
 
 I. Symbiosis in the Ophioglossacea-. Bull, lorrey Hot. Club, 1893, 20, 356. 
 Bary, a. de 
 
 I. Comparative Anatomy of the Vcj^etalive Organs of the I'lianerogams and Ferns. Oxford, 
 1884. 
 Beer, R. 
 
 I. On the Development of the .Spores of Ilrlminthostachys zeyltinica. Ann. Bot., 1906, 20, 177. 
 Belajekf, W. 
 
 1. L'eber Bau und Entwickelung der Spermatozoiden der (Jefasskryptogainen. Bet. d. dcutsch. 
 
 bot. Gesell., 1888, 7, 122. 
 
 2. Uebcr Bau und Entwickelung der Spermatozoiden der I'flanzen. Flora, 1894, 79, Erganzsb., 
 
 1-48. 
 
 3. Ueber den Nebenkern in spermatogenen Zellen und die Spermatogencsc bei den Farnkrauter. 
 
 Ber. d. deutsch. bot. Gesell., 1897, 15, 339. 
 
 4. Ueber die Cilienbildner in den spermatogenen Zellen, Ber. d. deutsch. bot. Gesell., 1898, 16, 
 
 140. 
 
 5. Ueber die Centrosomcn in den spermatogenen Zellen. Ber. d. deutsch. bot. (lesell., 1899, 
 
 17, 199-205. 
 Bitter, G. 
 
 1. Marattiace.T. Engler and Branrl. Die natiirlichen Fflanzenfamilicn, 1900, Th. I, Abt. IV, 
 
 422-444. 
 
 2. Ophioglossaceae. Ibid., 449-472. 
 Bower, F. O. 
 
 1. Comparative Morphology of the Leaf of the Vascular Cryptogams and Gymnosperms. Proc. 
 
 Roy. Soc, 1884, x.xxvii, 61; Phil. Trans. Royal Soc, 1884, 175, 565. 
 
 2. The Comparative Examination of the Merisrems of Ferns as a Phylogenetic Study. Ann. 
 
 Bot., 1889, 3, 305. 
 
 3. Is the Eusporangiate or the Leptosporangiate the more Primitive Type of Fern? Ann. Bot., 
 
 1891, s, 109. 
 
 4. Studies in the Morphology of the Spore-producing Members: Equisetineoe and Lycopodineae. 
 
 Phil. Trans. Roy. Soc, 1894, ser. B, 185, 473-572. 
 
 5. Studies in the Morphology of Spore-producing Members, 11: Ophioglossaceae. London, 1896. 
 
 6. Studies in the Morphology of Spore-producing Members, in: Marattiace.Te. Phil. Trans. 
 
 Roy. Soc, 1897, ser. B, 189, 35-81. 
 
 7. Studies in the Morphology of Spore-producing Members, v. General Comparisons and Con- 
 
 clusions. Phil. Trans. Roy. Soc, 1903, ser. B, 196, 191-257. 
 
 8. Ophtoglossum simplex Ridley. Ann. Bot., 1904, 18, 205. 
 
 9. The Origin of a Land Flora. London, 1908. 
 
 10. Note on Ophioglossum palmatum. Report Brit. Assoc. Adv. Sci., Sheffield, Sept. 1910. 
 Boodle, L. A. 
 
 I. On some Points in the Anatomy of the Ophioglossacea?. Ann. Bot., 1899, 13, 377-394. 
 Brebner G. 
 
 1. On the Mucilage Canals of the Marattiaceae. Journ. Linn. Soc, London, 1895, 33, 444. 
 
 2. On the Prothallus and Embryo o( Danaa simplicifolia Rudge. Ann. Bot., 1896, 10, 109. 
 
 3. On the Anatomy of Danea and other Marattiaceae. Ann. Bot., 1902, 16, 517-552. 
 Britton, E. G. 
 
 I. A Revision of the North American Species of Ophioglnssum. Bull. Torrey Bot. Club, 1897, 
 24. 545-5.^9- 
 Bruchmann, H. 
 
 1. Ueber das Prothallium und die Keimpflanzc von Opiiinglossujn vulgalum L. Bot. Zeit., 1904, 
 
 62, 227-247. 
 
 2. Ueber das Prothallium und die Sporcnpflanze von Botrychium lunaria L. Flora, 1906, 96, 
 
 203. 
 Burlingame, L . L. 
 
 I. The Sporangium of tlie Ophioglossales. Bot. Gaz., I<)07. 44. 34. 
 
 219 
 
220 BIBLIOGRAPHY 
 
 Campbell, D. H. 
 
 1. On the Affinities of the Fihcineae. Bot. Gaz. 1890, 15, i. 
 
 2. On the Relationships of the Archegoniat;c. Bot. Gaz., 1891, 16, 323. 
 
 3. Observations on the Development o{ Maratlia doughsii Baker. Ann. Bot., 1894, 8, I. 
 
 4. The Structure and Development of the Mosses and Ferns. London and New York, 1895. 
 
 Second edition, 1905. 
 
 5. The Origin of Terrestrial Plants. Proc. Amer. Assoc. Adv. Sci., 1903, 52, 463-482. 
 
 6. Antithetic versus Homologous Alternation. Amer. Nat., 1903, 37, 153. 
 
 7. Affinities of the Ophioglossacea; and Marsiliaceae. Amer. Nat., 1904, 38, 761-775. 
 
 8. Studies on the Ophioglossaceae. Ann. du Jardin botanique de Buitenzorg, 1907, 2e ser., 
 
 6, 138-194- 
 
 9. The Prothallium of Kauljussia and GUichenia. Ibid., 1908, 2e ser., 8, 69-102. 
 
 10. The Prothallium and Embryo of Datura. Preliminary Note. Ann. Bot., Oct. 1909, 23. 
 
 11. The Embryo and Young Sporophyte oi Angiopteris and Kaulfussia. Ann. du Jardin botanique 
 
 de Buitenzorg, 1910, 2e suppl., 3, 69-82. 
 Cardiff, I. D. 
 
 I. Development of the Sporangium in Botrychium. Bot. Gaz., 1905, 39, 340. 
 Charles, Grace M. 
 
 I. The Anatomy of the Sporeling of Marattia alata. Bot. Gaz., 191 1, 51, 81-101. 
 Cavers, F. 
 
 I. On Saprophytism and Mycorrhiza in Hepatica. New Phytologist, 1903, 2, 30-35. 
 Celakovsky, L. 
 
 I. Untersuchungen iiber die Homologien der generativen Produkte der Fruchtblatter bei den 
 Phanerogamen und Gefasskryptogamen. Pringsheim's Jahrb. fur wiss. Bot., 1884, 
 14. 291- 
 Christ, H. 
 
 1. Die Farnkrauter der Erde. Jena, 1897. 
 
 2. (and K. Giesenhagen). Pteridographische Notizen. Flora, 1899, 86, 72-85. 
 Christensen, C. 
 
 I. Index Filicum. Copenhagen, 1906. 
 Chrysler, M. .\. 
 
 I. The Nature of the Fertile Spike in the Ophioglossace:?. Ann. Bot., 1910, 24, 1-18. 
 Copeland, E. B. 
 
 I. Ferns of the Malay Asiatic Region. Philippine Journ. Sci., 1909, Sec. C. Bot. IV. 
 Davenport, G. E. 
 
 I. Vernation in Boirychium. Bull. Torrey Bot. Club, 1881, 8, 100. 
 Farmer, J. Bretland. 
 
 1. The Embryogeny oi Angwpteris evecta Hofm. .Ann. Bot., 1892, 6, 265. 
 
 2. (and W. G. Freeman). On the Structure and Affinities of Helminthostachys zeylanica. Ann. 
 
 Bot., Sept. 1899, 13. 
 
 3. (and T. G. Hill). On the Arrangement and Structure of the Vascular Strands in Anowpteris 
 
 evecta and some other Marattiaceas. Ann. Bot., 1902, 16, 371-402. 
 GoEBEL, K. 
 
 1. Beitrage zur vergleichenden Entwickelungsgeschichte der Sporangien. Bot. Zeit., 1880, 545: 
 
 1881,681. 
 
 2. Vergleichende Entwickelungsgeschichte der Pflanzenorgane. Schenk. Handbuch der Botanik, 
 
 1884, 3. 
 
 3. Outlines of Classification and Special Morphology. (Translation of the German edition.) 
 
 Oxford, Clarendon Press, 1887. 
 
 4. Organographie der Pflanzen. 4 pts. Jena, 1898-1901. 
 CJrisebach. a. H. R. 
 
 I. Flora of the British West Indian Islands. London, 1864. 
 Gwynne-Vaughn, D. T. 
 
 1. On an Unexplained Point in the Anatomy of Ht'lmintluistachys zeylanica. Ann. Bot., 1902, 
 
 16, 170. 
 
 2. On the Anatomy of Archangiopteris henryi and other Marattiaceae. Ann. Bot., 1905, 19, 
 
 411. 
 Hannig, E. 
 
 I. Ueber die Staubgriibchen an den Stammcn und Blattstielen der Cyatheaceen und Marrat- 
 tiaceen. Bot. Zeit., 1898, 56, 9-33. 
 
lilULIOGRAl'irV 221 
 
 Hill, T. G. 
 
 1. See Farmer (3). 
 
 2. On Secondary Tliickinini; in ./ni^iopWris c'vala. Ann. Hot., U)02, 16, 17!. 
 HlLLI-liRAND, W. F. 
 
 I. Flora of the Hawaiian Islands. 1H88. 
 
 HOFMEISTER, W. 
 
 I. The Higher Cryptoganiia. Ray Soeiet)-, 1.S6;. This loni.iins a iranslatinn of the "Ver- 
 gleichende Untersuchungen," as well as the later p.i|His ii|hmi iIk- .ArirheKoniatii-. 
 
 HOLLE, J. G. 
 
 1. Ueber Bau und Entwickelung der Vegetationsorgane der ( )|il)i(),i;l(isseen. Mot. Zeif., 1875, 241. 
 
 2. Vegctationsorgane der Maratllaeeeii. Mot. Zeit., 1S76, 21 i. 
 HoLTZMAN, C. L. 
 
 I. On Apical (irowtii of the Stem, aiul the Developineiit of the Sporan.minii of Biilrycliium 
 virginiuniim. Bot. Gaz., 1S92, 17, 214. 
 HiioKER, Sir J., and J. G. Baker. 
 
 I. Synopsis Filicuni. London, 1874. 
 Janse, J. M. 
 
 I. Les cndophyte.s radicaiix de certaines planter ia\anaises. Ann. du |.iidm hot jiiii|iie de 
 Buitenzorg, 1895, 14, 54-201. 
 Jeffrey, E. C. 
 
 1. T\\eG-dmcto^hytcoi Bolrychiumvirginianiim. I'roc. Caiiad. Inst., 1898, 5. 
 
 2. The Development, Structure, and Affinities of the (ieniis lu/uiifliun. Mem. Boston Soc. 
 
 Nat. Hist., 1899, S, no. 5. 
 
 3. The Structure and Development of the Stem in l'teridoph\ tes and (Jymnosperms. I'hil. 
 
 Trans. Roy. Soc, 1902, ser. B, 195, 1 19-146. 
 JoNKMANN, H. F. 
 
 1. La generation sexuee des Marattiacees. Arch, nt'eriand., etc., 1880, t. 15, 199. .Also in 
 
 Bot. Zeit., 1878, 129. 
 
 2. Over de Keimung van Kaulfussia tcscu/ijolia. Xederland. Kiuidkdg. Archief, 1879, ser. 2, 
 
 deal 3, stuck 2, p. 262. 
 
 3. L'embryogenie de I'Angioptcns ct du Muruttia. Arch, neerland., t. 50, 213-230. 
 
 (Separate, not dated.) 
 Koch, L. 
 
 Ueber Bau und Wachstum der Wurzelspitze von Jngioph-ris evccta. I'ringsht im"s Jahrb. fur 
 wiss. Bot., 1895, 28, 369. 
 Klihn, R. 
 
 1. Untersuchungen uber die Anatomic der Marattiaceen und anderen Gefasskryptogamen. 
 
 Flora, 1889, 72, 457. 
 
 2. Ueber den anatomischen Bau von Duiiu'u. Flora, 1890, 7!, 147 
 Lanc. W. H. 
 
 1. On the Prothalli of Opliioglojsum pendulum and Ihimintlwstachys zexlanica. .Ann. Bot., 1902, 
 
 16, 23-56. 
 
 2. On a Suspensor in Helminthostachys. Ibid., 1910, 24, 61 1. 
 
 LUERSSEN, C. 
 
 1. Spaltoffnungen von Kaulfussia. Bot. Zeit., 1873, 625. 
 
 2. Ueber die Entwiekelungsgeschichte des Marattiaceenvorkeims. Bot. Zeit., 1875, 535. 
 
 3. Handbuch der systematischen Botanik. Bd. i: Kryptogamen. Leipzig, 1879. 
 
 4. Die Farnpflanzen oder Gefassbundelkryptngamen (Rabenhorst, Krvptogamcnflora). Leip- 
 
 zig, 1 884- 1 889. 
 Lyon, H. L. 
 
 I. A New Genus of Ophioglossace.e. Bot. (iaz., 1905, 40, 455. 
 Magnus, VV. 
 
 I. Studien an der endotrophen .M)corrhiza von AU-oltia nidus-avis L. Fringsln im'.s I.ihrb fur 
 wiss. Bot., 1900, 35. 
 Mettenius, Georg. 
 
 1. Filices Horti Botanici Lipsiensis. Leipzig, 1856. 
 
 2. Ueber den Bau von AngiopUris. Abhandj. der. K. Sachs. Gesill. der Wiss., 1864, 6, 501. 
 MOTTIER, D. M. 
 
 Fecundation in Plants. Publication no. 15, Carnegie Institution of W'.ishington, 1904. 
 
 I'cllRAl'l.T, G. 
 
 1. ^UT I'OphiiigldSSum vulgalum. Jnurii. de hoianiciue, 1S91, 6. 
 
 2. Rccherches anatomiques sur les cryptogames vasculaires. .Ann. des sci. naturelles, 1893, ser. 
 
 7, t. 18, 113. 
 
222 BIBLIOGRAPHY 
 
 POTONIK, H. 
 
 I. Fossil Ptcridopliytes. Engler and Prantl. Die naturliclieii Pflanzenfamilien. I'll, i, Abt. IV. 
 Prantl, K. 
 
 1. Helminthnstachys zeylanica in ilire Bezieliung zu Ophwglussum und Botrythium. Ber. d. 
 
 deutsch. bot. Gesell., 1883, i, 155. 
 
 2. Systematisches Uebersicht der Ophioglosseen. Ibid., 348. 
 Raciborski, M. 
 
 I. Die Pteridophyten der Flora von Buitenzorg. Leiden, 1898. 
 ROSTOWZEW, S. 
 
 1. Beitrage zur Kenntniss der Gefasskryptogamen. I. Umbildung von Wurzein in Sprosse. 
 
 Flora, 1890, 73, 155. 
 
 2. Opliioglossacea? (Russian). Moscow, 1892. 
 Russow, E. 
 
 1. Vergleichende Untersucliungen, betrelFend die Histologie der vegetativen und sporenbildenden 
 
 Organe und die Entwickelung der Sporen der Leitbundelkryptogamen. Mem. de 
 I'Acad. imp. des sciences de St. Pctersbourg, 1S72, ser. 7, 19, no. i. 
 
 2. Betraclitungen uber das Leitbundel und Grundgewebe aus vergleichend morphologischem 
 
 und physiologischem Gesichtspunkte. Dorpat, 1875. 
 
 3. Developpement des tubes cribreu.x. Ann. des sci. naturelles, 1882, ser. 6, 14, 167. 
 Sadebeck, R. 
 
 I. Der Embryo von Equisclum. Bot. Zeit., 1877, 44. Also, Pringslieim's Jalirb. fur wiss. Bot., 
 1878, II, SIS- 
 
 SCHOUTE, J. C. 
 
 Die Stelartheorie. Groningen, 1902. 
 Scott, D. H. 
 
 I. Studies in Fossil Botany. London, 1900. 
 
 Second edition, 1909. (Also many monographs on fossil plants.) 
 Seward, A. C. 
 
 I. Fossil Plants. 2 vols. Cambridge, 1897-1910. 
 Shaw, W. R. 
 
 1. The Fertilization of Onoclea. Ann. Bot., 1898, 12, 251-285. 
 
 2. Ueber die Blepharoplasten bei Onoclea und Marsilia. Ber. d. deutsch. bot. Gesell., i8g8, 
 
 16, 177-184- 
 Shove, R. F. 
 
 I. On the Structure of the Stem o{ Angiopteris evccla. Ann. Bot., 1900, 14, 497. 
 Solms-Laubach, H., Gra^ zu. 
 
 I. Fossil Botany. Oxford, Clarendon Press, 1891. 
 Strasburger, E. 
 
 1. Die Befruchtung bei den Farnkrautern. Pringsheim's Jalirb. fur wiss. Bot., 7, 390. 
 
 2. The Periodic Reduction of Chromosomes in Living Organisms. Ann. Bot., 1894, 8, 281. 
 
 3. Ueber Reductionsteilung. Sitzungsber. der K. Preuss. Akad. der Wiss., 1904, 18. 
 Tansley, A. G. 
 
 I. Lectures on the Evolution of the Filicinean Vascular System. Cambridge, 1909. 
 Ternetz, Charlotte. 
 
 I. Ueber die Assimilation des atmosphaerischen StickstofFcs dutch Pilze. Pringsheim's Jahrb. 
 fiir wiss. Bot., 1907, 44, 353. 
 Treub, M. 
 
 I. Etudes sur les Lycopodiacees. Ann. du Jardin botani<iue de Buitenzorg, 1884-1890, 4, 5, 7, 8. 
 Underwood, L. M. 
 
 1. Our Native Ferns and their Allies. 6th ed. New York, 1900. 
 
 2. A Review of the Genus Danaa. Bull. Torrey Bot. Club, 1902, 29, 669. 
 Van Tieghem, Ph. 
 
 1. Sur quelques pointes de I'anatomie des cryptogames vasculaires. Bull, de la Soc. bot. de France, 
 
 1883, 25, 169. 
 
 2. Sur la limite du cylindre centrale et de I'ecorce dans les cryptogames vasculaires. Journ. 
 
 de botanique, 1888, 369. 
 
 3. (and Duliot) Recherches comparatives sur I'origine des membres endogenes dans les plantes 
 
 vasculaires. Ann. des sci. naturelles, 1888, ser. 7, t. 8, i. 
 
 4. Remarques sur la structure de la tige des Ophioglossees. Journ. de botanique, 1890, 365. 
 J. Traite de botanique. 2. ed. Paris, 1892. 
 
LIST OF PLATES. 
 
 Plate i. 
 
 1. Two spores o{ Ophioglossiim moluccanum Schlecht, showing range in size, x 500. 
 
 2. Two-celled ganietopliytc of 0. molucranum. 
 
 3. A three-celled stage. sf>, ruptured spore membrane. 
 
 4. A four-celled stage seen from above. 
 
 5. Two transverse optical sections of a four-celled gametopbyte. 
 
 6. Three-celled gametopbyte of 0. intermedium Ilk. 
 
 7-9. Young gametophytes of O. pendulum L., showing mycorrbizal infection, m, niycorrhizal 
 filaments. 
 10. Adult gametopbyte of O. mo/ufct(fium. x 10. ^, basal tuber; 5 , antberidia. 
 H-14. Gametophytes of O. pendulum. X j. /;, adventitious buds; sp, young sporopbytc. 
 
 15. Upper part of ripe antberidium of O. pendulum. X 275. 0, opercular cell. 
 16-20. Development of the spermatozoid in 0. moluccanum. x 950. hi, blepharoplasts. 
 
 Plate 2. 
 
 21-29. Spermatogenesis in Uphwglossum pendulum, x 950. 21 shows spermatocyte before final 
 nuclear division, with two blepharoplasts, bl ; 22, 23, nuclear spindle from side and pole; 
 24-29, development of spermatozoid; n, nucleus of spermatozoid; c, cilia; v, protoplasmic 
 vesicle. 
 30-32. Three stages in the development of the archegonium of O. pendulum, x 275. i, basal cell; 
 «', n", neck canal cells. 
 ^2- Venter of the archegonium, showing apparently a ventral canal cell, v; n, one of the nuclei 
 of the neck canal cell, x 650. 
 34-37. Spermatogenesis in B. virginianum; 38-41, in Kauljussia; 42, 43, in Damta. All x 950. 
 
 44. Nearly ripe spermatozoids of Angioptcris. x 950. n, nucleus; bl, blcpharoplast; c, cilia; 
 
 v, vesicle. 
 
 45. Nearly ripe archegonium oi Danica, showing apparent ventral canal cell, v. 
 
 46-48. Recently fertilized egg-cells of Dantea. sp, spermatozoid (?) within the egg-nucleus. 
 
 Plate 3. 
 
 Ophinglnssum mi>lucciinum Schlecht. 
 
 Several types collected at Buitenzorg, Java, and referred to O. moluccanum Schlecht. There are 
 at least three species. Fig. 5 is the typical O. moluccanum. All figures about natural size. 
 
 Plate 4. 
 
 A. Six plants of Ophioglossum (Ophiodcrmn) intermedium Hk. reduced one-half. Collected at 
 
 Buitenzorg, Java. 
 
 B. Ophioglossum {Ophioderma) pendulum L., collected in the Hanuclla Reserve I'orest, Ceylon. 
 1,2. Full-grown sporophytes of 0. /)£^ni/u/«m, much reduced. In 2 the sporangia! spike, sp,\s 
 
 forked. 
 3-6. Details of the sporophyll and sporangiopborc. 
 
 Plate 5. 
 
 Ophioglossum (Cheiroglossn) patmntum L., collected in Jamaica. 
 I. The whole plant. 
 2, 3. Sporophylls. In 3 there is a single median spike. 
 
 Plate 6. 
 
 A. I, Botrychium obliquum Miibl.; 2, B. lunaria L. 
 
 B. B. virginianum (L.) Swz. 
 
 Plate 7. 
 
 Botrychium stlaijolium Pr. California. 
 
224 LIST OF PLATES 
 
 Plate 8. 
 
 A. Young plant o( Hclmititliostachys zeyltinitu (L.) Ilk. Rcdiictd about onc-lialf. 
 
 B. Full-grown plant oi Hehninihostachys, reduced about two-thirds. Collected in the Hanwclla 
 
 Reserve Forest, Ceylon. 
 
 Plate g. 
 
 A. I, small plant of Df/Hrcrt jinmani Underwood; 2, rhizome of the same species; 5, apt.x of sterile 
 
 leaf; 4, fertile leaf; 5, sterile leaf of D. jamaicensis Underwood. 
 
 B. I, young plant of Danwa elliptica Sm. (.''); 2, sporophyll of D. jamaicensis; ^, 4, very young 
 plants of/), jimuiiccnsis {'>.). 
 
 Plate 
 
 All specimens collected in Jamaica. 
 
 10. 
 
 Adult sporophyte of Daiuea illiptun. Ja 
 
 Kauljussia ivstulijolia Bl. 
 
 1. Young plant showing the rhizome. 
 
 2. Sporophyll. Specimens collected near Buitenzorg, Java. 
 
 Plate 12. 
 
 A. Marattia iilata Swz. 
 
 1. Part of fertile leaf. 
 
 2. A single fertile pinna. 
 
 3. Fallen leaf-base with adventitious buds, k. 
 
 B. Marattia alata. 
 
 I. Leaf from young plant developed from an adventitious bud. 
 2, 3. Young sporophytes arising from fertilization. 
 All specimens from Jamaica. 
 
 Plate 13. 
 
 Angiopterts evecta (Forst.) HofFm. 
 1-3. Specimens from Australia; 4, 5, specimens from Peradeniya, Ceylon. 
 
 I, caude.x of a small plant; 2, part of a sporophyll; 3, cross-section of the petiole of a full-grown 
 leaf; 4, young sporophyte arising from the prothallium; 5, single leaf from a similar sporo- 
 phyte. 
 
A-} 
 
 /S 
 
 
 ■c^.^. 
 
 .■\ 
 
 .6 
 
 
 ^^K 
 
 ^^m 
 
 1. Two spores of O. moluccanum, showing range in size. X 500. 
 
 2. Two-celled gametophyle of O. moluccanum. 
 
 3. A three-celled stage. 5p, ruptured spore membrane. 
 
 4. A four-celled stage, seen from above. 
 
 5. Two transverse optical sections of a four-celled gametophyle. 
 
 6. Three-celled gamctophyte of O. intermedium Hk. 
 
 7-9. Young gametophytcs of O. pendulum L., showing myconhizal 
 infection, m. mycorrhizol filaments. 
 
 10. Adult gamctophyte of O. moluccanum. X 10. /, basal tuber; 
 3 , antheridia. 
 11-14. Gametophytcs of O. pendulum. X 3. A, adventitious buds; 
 sp, young sporophytc. 
 15. Upper part of ripe antheridium of O. pendulum. X 275. 
 
 16-20. Development of the spermatozoid in O. moluccanum. X 950. 
 hi, Itlepharoplasts. 
 
''^yj^ft*' 
 
 # „ i 
 
 30-32, 
 33. 
 
 Spermatogenesis in Ophioglossum pendulum. X 950; 21 shows 
 spermatocyte before final nuclear division, with two 
 blepharoplasts, hi; 22, 23, nuclear spindle (rora side and 
 pole; 24-29, development of spermatozoid ; n, nucleus of 
 spermatozoid ; c, cilia; V, protoplasmic vesicle. 
 
 Three stages in development of archegoiuum of O. pendulum. 
 X 275. A, basal cell; n'. n-', neck canal cells. 
 
 Venter of archegonium showing apparently a ventral canal cell, t). 
 n, one of the nuclei of the neck canal cell. X 650. 
 
 34-37. Spermatogenesis in B. cirginianum; 38-41, in Koulfussia; 
 42, 43, in Daniea. All X 950. 
 
 44. Nearly ripe spetmatozoids of Angiopleris. X 950. n, nucleus : 
 
 A/, blepharoplast; c, cilia; V, vesicle. 
 
 45. Nearly ripe archegonium of Danaa, showing apparent ventral 
 
 ell. 
 
 46-48 Recently fertilized egg-cells of Da 
 within the egg-nucleus. 
 
 oid (?) 
 

 \ 
 
 A. Six plants ot Ophiogh 
 
 Builenzorg, Java. 
 
 B. Ophioglossum {Ophiodcrma) pendulum L. C. 
 
 1, 2. Full-grown sporophylcs o( O. pendulum, 
 i-6. Details ol the spotophyll and sporangiophc 
 
 (OpIiioJcrma) inlcrmedium Hk., reduced 
 
 cted in the Hanwella Reserve Forest, Ceylon. 
 
 ich reduced. In 2 the sporangial spike, sp, is forked. 
 
Ophioglossum {Cheiroglossa) palmalum L. Collected in Jamaica. 
 I . The whole plant. 2, 3. Sporophylls; in 3 there is a single median spike 
 
A. I. small plant o( Danaa jenmani Underwood; 2. rhizome of the same species; 3, apex ot sterile leaf; 
 
 4. fertile leaf; 5, sterile leaf of D. Jamaicensis Underwood. 
 
 B. I. young plant of Dancta MpUca Sm.; 2. sporophyll ot D. Jamaican,),: 3. 4. very young 
 
 plants of D. jamaicensis (?). 
 
 All Specimens from Jamaica. 
 
Adult sporophyle of Dansa clUptica. Ja 
 
Kaultussia oesculifolia Bl. 
 
 1, young plant showing the rhizome; 2, sporophyll. 
 
 Specimens collected near Buitenzorg. Java. 
 

 ^^n 
 
 lis 
 
31 
 
 g ^ 
 
 X g 
 
 
 Library 
 N". O. State Gollfl«r« 
 
INDEX 
 
 Adder-tongue fern. See Ophioglossuni. 
 Adventitious buds. See Buds. 
 Alternation of generations, 209 
 Aneura, 120 
 
 •Angiopteris, 4, 117, 118, 120, 121, 126, 139, 140, 
 141, 146, ISO, 152, 153, 154, IS5, 
 158, 164, 196, 198, 199, 201, 202, 
 203, 204, 20s, 207 
 Figs. 88, 89, 95, loi, 109, no, iii, 112, 
 124, 130, 134, 180, 181, 182, 183, 187 
 Plate 13 
 Angiopteris, species of, 196 
 Angiopteris evecta (Forst.) Hotfm., 119 
 Angiopteris pruinosa var. hypoleuca, 122 
 Annulus of sporangium, 208 
 Antheridium: 
 
 Angiopteris, 126 
 Anthoceros, 42, 122, 211 
 Botrychium, 18, 24, 21; 
 Danasa, 126, 128 
 Equisetum, 24 
 Helminthostachys, 20, 21 
 Kaulfussia, 127 
 Lycopodium, 24 
 Marattia, 126, 128 
 Marattiaceae, 24, 129 
 Ophioglossum, 22, 23, 24 
 .■\nthoceros, Anthocerotes, 42, 122, 210, 211 
 Apical growth. See Leaf, Root, Stem. 
 Apogamy (in Botrychium), 52 
 Archangiopteris henryi, 4, 117, 203, 204 
 
 Fig. 184 
 Archegonium: 
 
 Angiopteris, 131 
 Botrychium, 30, 31 
 Danxa, 132, 133, 134 
 Helminthostachys, 20, 21 
 Kaulfussia, 13 
 Marattia, 126 
 (Iphioglossuni, 28, 29 
 .Archesporium: 
 
 Botrychium, 1 15 
 Helminthostachys, 111; 
 Marattiaces, 206 
 Ophioglossum, 115 
 Asplenium nidus L., 10, 13 
 Blepharoplast, 26, 128, 129 
 
 Botrychium, 3, j, 16, 17, 18, 19, 25, 27, 28, 30, 
 31, 32, 39. 46, 47. 48, 49. 50, .S2. 66, 
 82, 83, 99, loi, 102, 103, 109, no, 
 136, 142, 213 
 Classification of, 99 
 Botrychium lanuginosum Wall., 99, loi, 10? 
 Fig. 74 
 
 I Botrychium lunaria (L.) S\v., 6, 16, 17, 28, 34, 46, 
 48, 50, 52, 65, 83, 99, m 
 Fi<;s. 8, 36, 71 
 Plate 6 
 B. matricari;cfolium A. Br., 16 
 B. obliquum Muhl., 16, 34, 39, 53, 54, 103, 136, 
 142 
 Plate 6 
 B. silaifolium Pr., 99 
 
 Plate 7 
 B. simplex Hitchcock, 5, 6, 16, 83, 99, 100, 109, 
 IIS 
 Fig. 71 
 B. ternatum (Thbg.) Sw., 100, 103 
 
 Fig. 71 
 B. virginianum (L.) Sw., 6, 16, 17, 18, 27, 28, 32, 
 34. 46. 47. 50. 52, 53, 54. 59. 60, 
 61, 63, 67, 69, 83, 99, 100, loi, 103, 
 no, 112, n4 
 Figs, 7, 14. IS. 17. 28, 29, 32, 33. 40, 
 41. 42, 43, 44. 71. 72, 74. 76, 77. 85 
 Plate 6 
 Botryopterideiv, 3, 214 
 
 Bryophytes (see also Mosses, Liverworts), 212 
 Buds: 
 
 in Botrychium, loi 
 on leaf of Dan.-ea sintensis, 178 
 on gametophyte of Marattiace*, 121 
 on gametophyte of Ophioglossum, 14 
 on primary root of Ophioglossum, 40 
 root buds in Ophioglossum vulgatum, 58, 59 
 Calcium pectate in Marattiaceae, 204 
 Calcium oxalate crystals in Marattiacea?, 204 
 Cambium, 65, 82 
 Canal, so-called "canal" in Helminthostachys, 
 
 74. 79 
 Cauline fibro-vascular bundles (see also Commis- 
 sural strand), 174, 175, 184, 191, 
 197 
 Cheiroglossa (see also Ophioglossum palmatum), 
 5.88 
 Fig. 70 
 Plate 5 
 Chlorophyll in gametophyte of Ophioglossum, 9, 
 
 Christensenia = Kaulfussia, 122 
 
 Collateral bundles, Primitive nature of, 214 
 
 CoUenchyma in: 
 
 Angiopteris, 202 
 
 Kaulfussia, 186 
 
 Marattia, 194 
 Commissural vascular bundles in — 
 
 Angiopteris, 197 
 
 Dan:ca, 174, 175 
 
226 
 
 Commissural vaseular bundks in — 
 Marattia, 191 
 Kaulfussia, 184 
 Commissure (of stipules), 187 
 Completoria complens, 22 
 Corallorhrza, }} 
 Cotyledon of — 
 
 Angiopteris, 146, 147, 151 
 Botrychium, 50, 51, 52, 64, 67, 8j 
 Danaea, 148, 150 
 Helminthostachys, 54, 69 
 Kaulfussia, 1 47, 149 
 Marattia, 138, 147, 150, 151 
 Ophioglossum, 35, 42 
 Crystals. See Calcium oxalate. 
 Cycads, 30 
 
 Dana?a, 47, 53, 82, 84, 117, iiS, 124, 125, 
 128, 135, 142, 151, 154, 160, 
 164, 205, 212 
 Danaea elliptica Smith, 124, 125, 126, 132, 
 142, 148, 154 
 Figs. 92, 93, 103, los, 117, 125, 131, 
 147, 151, 152, 161. Plates 9, 
 D. jamaicensis Underwood, 124, 126, 136, 
 144, 146, 148, 154, 155, 156 
 Figs. 91, 93, 98, 99, 100, 104, 114, 
 116, 118, 119, 120, 121, 125, 
 138, 140, 143, 144, 145, 146, 
 149, 150, 153, 160, 161, 189 
 Plate 9 
 D. jenmani Underwood, 124 
 
 Figs. 91, 154, 15s, 156, 137, 162 
 Plate 9 
 D. simplicifolia Rudge, 124, 127, 136, 142, 
 
 160, 175, 176, 206 
 D. sintensis, 178 
 D. trichomanoides Spruce, 178 
 Dehiscence of — 
 
 Antheridium, 24, 2;, 129 
 Sporangium, 109 
 Synangium of Marattiaceae, 207 
 Dichotomy of — 
 
 Prothallium in Marattiacea", 121 
 Root in Ophioglossum, 93 
 Dicotyledons, 65 
 Dictyostele in — 
 Angiopteris, 197 
 Dana;a, 175 
 Kaulfussia, 185 
 Marattia, 192 
 Drosera, 32 
 Embryo of — 
 
 Angiopteris, 139 
 Botrychium, 46, 47, 48, 51. 136 
 Danaea, 118, 136, 142 
 Helminthostachys, 54, 67 
 Kaulfussia, 141 
 Marattia, 13;, 136, 137 
 Marattiacea?, 118, 135 
 Ophioglossum, 34, 35, 36, 37, 38, 43 
 
 Endodermis: 
 
 Angiopteris, 164, 201 
 
 Dan^a, 163, 167 
 
 Helminthostachys, 72 
 
 Kaulfussia, 164, 182, 183 
 
 Marattia, 164, 193 
 
 Ophioglossum, 92 
 Endarch bundles in Helminthostachys, 77 
 Endophyte. See Mycorrhiza. 
 Entomophthoreie, 22 
 
 Eqiiist tLim, KiiuisL-tinea", 26, 27, 39, 132, 139 
 Eubotrychiuni (see also Botrychium), 100 
 Euophioglossum (see also Ophioglossum), 83, 
 
 86, 89, 91, 93 
 Eusporangiatae, 3 
 
 Comparison with Bryophytes, 209, 210, 212 
 
 Nature of vascular system, 214 
 
 Relation to Leptosporangiatx, 217 
 Fegatella, 33 
 Ferns, origin of, 209 
 Fertilization: 
 
 Botrychium, 32 
 
 Marattiaceae, 134 
 
 Ophioglossum, 31 
 Fibro-vascular system (see also Leaf-trace, Stele) : 
 
 Angiopteris, 196, 198, 199, 201 
 
 Botrychium, 60, 62, 66 
 
 Danxa, 160, 162, 163, 164, 166, I.72, 174, 
 .76 
 
 Helminthostachys, 73, 76, 77, 78, 106 
 
 Kaulfussia, 181, 182, 184, 185 
 
 Marattia, 189, 191, 193 
 Fossils: 
 
 Bryophytes, 209 
 
 Ferns, 3 
 
 Marattiaceae, 117 
 Gametophyte: 
 
 Angiopteris, 121 
 
 Botrychium, 16, 17, 18 
 
 Danaea, 124, 125 
 
 Helminthostachys, 20, 21, 22 
 
 Kaulfussia, 122, 123 
 
 Marattia, 119, 121, 122, 210 
 
 Ophioglossum, 6, 10, 11, 12, 13, 14, 15, 211 
 Germination of spores: 
 
 Marattiace;e, 119, 1 20 
 
 Ophioglossaceae, 7, 8, 9 
 Gleichenia, mycorrhiza in, 33 
 Gymnogramme, 26 
 Hairs: 
 
 Marattiaceae, 1 50 
 
 Ophioglossum palmatum, 99 
 Helminthostachys, 5, 8, 10, 19, 20, 21, 25, 31, 32, 
 54, 67, 68, 69, 70, 71, 72, 73, 74, 75, 
 76, 77. 78, 79, 80, 81, 82, 84, 104, 
 103, 106, 107, 108, 112, 115, 116, 
 213. 214 
 Figs. 10, 1 1, 45, 46, 47, 48, 49, 50, 51, 52, 
 
 53, 54, 78, 79, 80, 86 
 Plate 8 
 
227 
 
 Horned liverworrs (sec also Antliocerotes) 
 
 209 
 
 Indusium, 204 
 
 
 
 Isoetes, 205, 206 
 
 
 
 Kaulfussia, 4, 82, 84, 117, 118, 119, 122 
 
 123 
 
 132, 
 
 141, 147, 149, 152, 153, 
 
 154. 
 
 iSS, 
 
 157, 158. 164. 178, 179. 
 
 180, 
 
 181, 
 
 182, 183, 185, 186, 187, 
 
 206 
 
 208 
 
 Fifts. 90, 96, 102, 113, 123, 
 
 128, 
 
 I3f', 
 
 141, 163, 164, 166, 167, 
 
 168, 
 
 170, 
 
 171, 1S8 
 I'l ire 1 I 
 
 
 
 Laeiin;c: 
 
 
 
 Intcrnodal in 1 lelniinrliostaelns. 
 
 75 
 
 
 Inrernodal ,n Kaulfussia, .84 
 
 
 
 Leaf; 
 
 
 
 Ansiopf^ii^. -01 
 
 
 
 l^otrychium, 61, loi 
 
 
 
 Danaea, 176, 178 
 
 
 
 Helminthostachys, 67, 70, 74 
 
 
 
 Kaulfussia, 185 
 
 
 
 Marattia, 194 
 
 
 
 Marattiaceae, 118, 157, 159, 214 
 
 
 
 Opliioglossum, 57, 84, 87, 88 
 
 
 
 Leaf, anatomy: 
 
 
 
 Angiopteris, 202 
 
 
 
 Botrvchium, 102 
 
 
 
 Danxa, 160 
 
 
 
 Helmintliostacl)\s, 107 
 
 
 
 Kaulfussia. 187 
 
 
 
 Marattia, 194, 195 
 
 
 
 Opliioglossum, 93, 96 
 
 
 
 Leaf, ferrile (see also Sporopliyll), 5, 85, 
 
 109, 
 
 204, 
 
 -:i,l 
 
 
 
 Leaf, gaps, 78, 173 
 
 
 
 Leaf, trace, 66, 72, 80, 173, 186, 191, 197 
 
 
 Leptosporangiata?, 3, 217 
 
 
 
 Liverworts, mycorrliiza in, 33 
 
 
 
 Lycopodium, 9, 10, 24 
 
 
 
 L. cernuum, 9, 10 
 
 
 
 Mantle cells (of antheridium), 24, 129 
 
 
 
 Marattia, 117, 119, 120, 122, 126, 128, 
 
 132. 
 
 I3S. 
 
 136, 137, 147. 149, 152. 
 
 IS3. 
 
 15 s. 
 
 156, 164, 188, 189, 192, 
 
 193- 
 
 194. 
 
 195, 204, 206 
 
 
 
 M. alata Swartz, 138, 188, 191, 192, 
 
 193. 
 
 194 
 
 Figs. 175, 176, 177. Plate 12 
 
 
 
 M. cicutasfolia Klf., 119, 132, 136 
 
 
 
 M. douglasii (Pr.) Baker, 119, 120, 122, 
 
 126, 
 
 I3S, 
 
 136, 137, 147, 149, 151, 
 
 53, 
 
 56, 
 
 188, 189, 190, 205, 210 
 
 
 
 Figs. 87, 88, 94, 106, 107, 108, 
 
 122, 
 
 27. 1 
 
 13^ 13.3. 14'' "7i. 171. 
 
 74- 
 
 77. 
 
 IS:;, 190, 191 
 
 
 
 M. fraxinea Sniiih, 120, 18S, igi, i,,. 
 
 
 
 Figs, 87. 186 
 
 
 
 M. salicifolia Schrad., 193 
 
 
 
 M. sanibucina Blunie, 120, 188 
 
 
 
 FiK. 87 
 
 
 
 M. weinmannixfolia Liehm., 136 
 
 
 
 Marattiacea?, 3, 35, 39, 48, 53, 61, 83, 84, 117, 
 129, 130, 134, 135, 209, 212, 214, 
 
 215 
 
 Macroglossum, 117, 204 
 M. alidie Copeland, 204 
 Mesarch vascular bundles in I lelininrliosraclns, 
 
 78 
 Monotropa, 33 
 Mosses, 209 
 
 Mucilage cells, 168, 170, 173, 177, 182, 185, 187, 
 I 190 
 
 j Mucilage ducts, 168, 170, 173, 177, 182, 185, 187, 
 190 
 Mycorrliiza, 6, 9, II, 15, 18, 19, 20, 21, 22, 32, 
 
 33,127,179,181,212 
 Operculum (of antheridium), 24, 25, 128, 129 
 j Ophioderma (see also Opliioglossum), 5, 87, 94 
 I Ophioglossacea;, 3, 5, 6, 7, 34, 82, 86, 108, 109, 
 208, 213, 214 
 Ophioglossum, 5, 6, 7, 9, 10, 11, 12, 13, 14, 15, 
 22, 24, 26, 27, 28, 35, 39, 42, 43, 56, 
 60, 82, 83, 85, 86, 87, 89, 91, 207, 
 209,210,213,214,215 
 O. bergianum Schlecht, 5, 87, 92 
 O. capense Sw., 92 
 O. californicum Prantl., 5 
 O. ellipticum Hk., 92 
 O. intermedium Hooker, 88, 97, 
 Fig. 69 
 Plate 4 
 O. lusitanicum L., 86, 87 
 
 O. moluccanum Sclilecht., 5, 6, 7, 11, 12, 26, 28, 
 34- 35- 36, 38, 39. 40, 42. 45. 55. 
 56. 57. 59- 86, 90, 94, 1X2, 210, 212 
 Figs. 2, 4, 12, 18, 22, 23, 24, 2c;, 26, 27, 
 37. 38, 39, 55. 56, 57. 58. 59. 60, 61. 
 62, 81, 83, 192 
 Plates I, 3 
 O. palniatum L., 88, 89, 98, 99 
 Fig. 70 
 Plate 5 
 O. pedunculosuni l)es\-., 10. n. 13, 34, 36. 39. 40, 
 42. 57 
 Fig, 3 
 O. penduhuii L., 5. 6. 7. 9, 10. 14. 18, 26. 28. 29, 
 36, 37, 38. 40, 42, 44, 56, 88, 95, 96, 
 109, 112 
 Figs. I. 3, 4, 13, 16, 20, 21, 63, 64, 65, 66, 
 
 67, 68, 82, 84 
 Plates I, 2, 4 
 O. reticulatum L., 6, 40, 90 
 O. simplex Ridley, 88. 98, 109 
 
 Fig. 71 
 O. vulgatum L.. 3. 5. 6. 10. 13. 15, 27, 28, 2<>, 34, 
 43.44,57.58,59.86,87,92 
 Figs. 3, 5, 19 
 Orchids as humus saprophytes, 33 
 Osmunda, Osmundaceae, 121, 123. 207 
 Osmundopteris. See Botrychium \irginianum. 
 Pali.sade tissue, 107, 178. 195, 202 
 
228 
 
 Pellia, 122 
 
 Periderm, 66, 8o, 193 
 
 Peronosporeae, 22 
 
 Phyllotrichium (see also Botrychium), 100, 102 
 
 Pith, in Helminthostachys, 76, 79 
 
 Polypodium quercifolium, 10 
 
 Prothallium. See Gametophyte. 
 
 Pro-ophioglossum, 211 
 
 Protocorm, 53 
 
 Protophloem in Angiopteris, 201 
 
 Protostele, 175, 188 
 
 Protoxylem: 
 
 Angiopteris, 201 
 
 Helminthostachys, 79, 82 
 
 Kaulfussia, 185 
 Pteris cretica, 53 
 Pythium, 22 
 Rhizoids: 
 
 Botrychium, 18 
 
 Danrea, 127 
 
 Helminthostachj's, 20 
 
 Kaulfussia, 123 
 
 Ophioglossum, 12 
 Rhizome (see also Stem) : 
 
 Botrychium, loi 
 
 Dansa, 175, 176, 177 
 
 Helminthostachys, 106 
 
 Kaulfussia, 186 
 
 Ophioglossum, 91, 94 
 Root: 
 
 Angiopteris, 140, 202, 203 
 
 Botrychium, 47, 51, 64, 103, 104 
 
 Danaea, 157, 164, 177, 179 
 
 Helminthostachys, 70, 80, 81, 105, 107, 108 
 
 Kaulfussia, 187' 
 
 Marattia, 138, 156, 192, 195 
 
 Ophioglossum, 35, 39, 43, 56, 83, 87, 89, 93, 
 97>98 
 Root apex: 
 
 Botrychium, 49 
 
 Danaea, 178 
 
 Helminthostachys, 72, 80 
 
 Marattia, 138, 156, 195 
 
 Ophioglossum, 94 
 Root hairs, multicellular, of Kaulfussia, 183, 187 
 Scales, epidermal: 
 
 Danaea, 151 
 
 Helminthostachys, 74, 75 
 
 Kaulfussia, 180 
 
 Marattiaceae, 150 
 Sceptridium. See Botrychium obliquum. 
 Sclerenchyma, 170, 171, 194 
 Secondary wood: 
 
 Angiopteris, 204 
 
 Botrychium, 65, 82 
 
 Helminthostachys, 78 
 Sex-organs. See Antherldium, Archegonium. 
 Sieve tubes, 151, 163, 201, 202 
 Siliceous deposits in Marattiaceae, 204 
 
 Siphonostele: 
 
 Angiopteris, 197 
 
 Kaulfussia, 184 
 
 Marattia, 188 
 Spermatogenesis: 
 
 Botrychium, 28 
 
 Marattiaceae, 129, 130 
 
 Ophioglossum, 26, 27 
 Spermatozoids: 
 
 Botrychium, 28 
 
 Marattiaceae, 130 
 
 Ophioglossum, 26, 27 
 Spike of Ophioglossaceae. See Sporangiophore. 
 Sporangium (see also Synangium): 
 
 Angiopteris, 205 
 
 Botrychium, 109, III, 1 14, 1 15 
 
 Danaea, 205, 206, 207 
 
 Helminthostach)'s, iii, 116 
 
 Kaulfussia, 207 
 
 Marattia, 205 
 
 Ophioglossum, 109, in, 113 
 Sporangiophore: 
 
 Botrychium, 109, no 
 
 Helminthostachys, 105, 106 
 
 Ophioglossaceae, 5, 85, 108, 109, 208 
 
 Ophioglossum, 109, no 
 Spore: 
 
 Marattiaceae, 208 
 
 Ophioglossaceae, 6, 7 
 Spore division, 114 
 Sporophyll: 
 
 Angiopteris, 204, 206 
 
 Archangiopteris, 204 
 
 Botrychium, 100, 102 
 
 Dan«a, 204 
 
 Helminthostachys, in, 116 
 
 Kaulfussia, 204, 208 
 
 Marattia, 204 
 
 Marattiaceae, 204, 208, 214 
 
 Ophioglossum, 85, 88, 93, 109, no, 112, 214 
 Sporophyte (see also Embryo) : 
 
 Angiopteris, 196, 198, 199, 20l 
 
 Botrychium, ji, 59, 60, 62, 63, 99 
 
 Danaea, 163, 178 
 
 Helminthostachys, 54, 67, 68, 77, 104, 105 
 
 Kaulfussia, 179, 180, 186 
 
 Marattia, 188, 189 
 
 Ophioglossaceae, comparison of young spor- 
 ophyte, 82 
 
 Ophioglossum, 38, 39, 44, 45, 55 
 Staubgrijbchen of Marattiaceae, 204 
 Stele: 
 
 Botrychium, 82 
 
 Helminthostachys, 52, 75, 82 
 Stem, stem apex: 
 
 Angiopteris, 153, 198, 199 
 
 Botrychium, 49, 53, loi, 103 
 
 Danaea, 118, 136, 156, 170 
 
 Helminthostachys, 70, 71, 74 
 
 Kaulfussia, 185, 187 
 
229 
 
 Stem, sttm apex: 
 
 Marattia, 192 
 
 Marattiaceae, 154, 162 
 
 Ophioglossum, 42, 50, 56, 89, 91, 94, 97. 99 
 Sterilization of sporogenoiis tissue, 209 
 Stipule, stipular sheath: 
 
 Botrychium, 6l, loi 
 
 Dan;ea, 170 
 
 Helminthostachys, 71, 73. 105 
 
 Kaulfussia, 185, 187 
 
 Marattia, 195 
 
 Marattiacex, 118 
 
 Ophioglossaceae, 83, 84 
 
 Ophioglossum, 91 
 Stomata: 
 
 Botrychium, loi 
 
 Dan^ca, 151 
 
 Helminthostachys, 106 
 
 Kaulfussia, 152, 187 
 
 Marattia, 195 
 
 Ophioglossum, 96 
 Suspensor: 
 
 Botrychium, 47, 48, 53, 136 
 
 Danaea, 118, 136 
 
 Synangium, 204, 205, 206, 207, 20 
 Tannin cells: 
 
 Angiopteris, 146 
 
 Dannea, 165, 169, 171 
 
 Helminthostachys, 79, 84 
 
 Kaulfussia, 185 
 
 Marattia, 189 
 
 Marattiaceae, 150 
 Tapetum, ill, 116, 206 
 Trabeculae, 207 
 Tracheary tissue: 
 
 Angiopteris, 201 
 
 Botrychium, 66 
 
 Helminthostachys, 68, 79, 82 
 
 Kaulfussia, 185 
 
 Marattia, 189 
 Ventral canal-cell, 30, 31, 133. I34. 
 
 201, 213 
 Vestigial leaves: 
 
 Botrychium lunaria, 51 
 
 Helminthostachys, 69, 81 
 
 Ophioglossum vulgatum, 35 
 
 176, 187, 194,