= potters ea 25) sated) asses Eras beh Tre Sopot sae Eyobescasess seats Sciences ages en ee ShIETOO TOEQ g UMA GYMNOSPERMS STRUCTURE AND EVOLUTION THE UNIVERSITY OF CHICAGO PRESS CHICAGO, ILLINOIS —— THE BAKER & TAYLOR COMPANY NEW YORK THE CAMBRIDGE UNIVERSITY PRESS LONDON THE MARUZEN-KABUSHIKI-KAISHA TOKYO, OSAKA, KYOTO, FUKUOKA, SENDAI THE COMMERCIAL PRESS, LIMITED SHANGHAI GYMNOSPERMS meRUCTURE AND EVOLUTION By CHARLES JOSEPH CHAMBERLAIN, Pu.D., Sc.D. WITH 397 FIGURES THE UNIVERSITY OF CHIGAGO: PRESS CHICAGO *ILLINGIS COPYRIGHT 1935 BY THE UNIVERSITY OF CHICAGO ALL RIGHTS RESERVED. PUBLISHED MARCH 1935 FIFTH IMPRESSION FEBRUARY 1937 COMPOSED AND PRINTED BY THE UNIVERSITY OF CHICAGO PRESS CHICAGO, ILLINOIS, U.S.A. TO MY WIFE AND DAUGHTER cy ' 7) a f v4 fm a 2 . m > CS / s 1 { ere / L | Ms \ we \ oan \O, \ XS 4 AS PREFACE A few centuries ago people believed that the earth was created suddenly, and that plants and animals were created just as they are today; but no one with any scientific training now believes in such an origin of the earth or its inhabitants. The first living things were simple and, in some way or another, originated from non-living mat- ter. Such simple forms gradually developed into more and more com- plex organisms. No one believes that any organism as complex as a fern, or even as complex as a moss or liverwort, ever developed di- rectly from non-living matter; they came from simpler forms. We may take it as a fact that the plants and animals on the earth today are the lineal descendants of plants and animals which lived hun- dreds of millions of years ago. Line after line of these ancient forms became extinct; but, here and there, an individual varying in some way from the main line, with some variation which made it able to withstand changing conditions which were fatal to its neighbors, sur- vived and became the progenitor of a new race. Consequently, if we are to understand the gymnosperms, or any other great group of plants, we should study not only those which exist today, but also those extinct ancestors whose fragmentary rec- ords can be read in the rocks. And beyond the available records, we should try to imagine the missing parts of the life-histories, and try to picture to ourselves the still more ancient progenitors. Since this book is intended to be of service not only to those who would gain some knowledge of the gymnosperms, but also to those who would go farther and do productive research in this great group, a few admonitions may be helpful. Read, and read widely, that you may know what has already been accomplished; and read critically. But no one can read critically whose knowledge comes entirely, or even principally, from reading. You must have a first-hand knowledge of the material, must study it in the field and in the laboratory. The student who has not had sufficient experience to make a first-class preparation for microscopic vii Viil PREFACE study cannot safely interpret slides made by others. He is in the same class with the one who claims he sees it but can’t draw it; while the real trouble is not in his hand, but in his head. Studies in the field bring a kind of knowledge which cannot be gained in any other way; and the technical work of making slides and drawings, if prop- erly done, affords a great stimulus to mental development. A chapter on ‘‘Alternation of Generations” has been added be- cause the gymnosperms stand in such a suggestive place in the evolu- tion of the sporophyte and, particularly, in the reduction of the gametophyte. There is a single alphabetical Bibliography, numbered consecu- tively. The small numerals throughout the book, both in text and illustrations, refer to this Bibliography. It is a pleasure to acknowledge helpful advice and criticism, espe- cially from Dr. A. C. Noé and Dr. Freppa D. REED, who have read the entire manuscript and have definitely improved the paleobotani- cal features. Dr. Joun T. BucuHuotrz read and criticized the chapter on “Embryogeny of Conifers.” Iam indebted to Dr. E. J. Kraus for suggestions in regard to the anatomy of conifers. Corrections and suggestions will always be welcome. CHARLES JOSEPH CHAMBERLAIN UNIVERSITY OF CHICAGO August 1934 CONTENTS CHAPTER I. INTRODUCTION CHAPTER II. CyCADOPHYTES—CYCADOFILICALES Distribution Life-histories General remarks IPRMIOpeny ei oe kl. Speculative reconstruction CHAPTER II]. CycADOPHYTES—BENNETTITALES Distribution Life-history . Phylogeny CHAPTER IV. CyCADOPHYTES—CYCADALES . Geographic distribution CHAPTER V. CYCADALES (continued) . The life-history CHAPTER VI. CyCADALES (continued) The female gametophyte . The male gametophyte Fertilization CHAPTER VII. CycapDALEs (continued) Embryogeny The seedling Hybrids . CHAPTER VIII. CycapALes (continued) . Phylogeny Taxonomy CHAPTER IX. CONIFEROPHYTES—CORDAITALES Distribution Life-history . The sporophyte The gametophytes . Phylogeny 525/78 PAGE x CONTENTS CHAPTER X. CONIFEROPHYTES—GINKGOALES Extinct members Ginkgo Life-history . Phylogeny CHAPTER XI. CONIFEROPHYTES—CONIFERALES Taxonomy oe Geographic distribution The sporophyte—vegetative . CHAPTER XII. CONIFEROPHYTES—CONIFERALES (continued) . The sporophyte—reproductive CHAPTER XIII. CoNIFEROPHYTES—CONIFERALES (continued) The male gametophyte CHAPTER XIV. CONIFEROPHYTES—CONIFERALES (continued) The female gametophyte . CHAPTER XV. CONIFEROPHYTES—CONIFERALES (continued) . Fertilization CHAPTER XVI. CONIFEROPHYTES—CONIFERALES (continued) The embryo CHAPTER XVII. CONIFEROPHYTES—GNETALES—EPHEDRA Geographic distribution The sporophyte—vegetative . The sporophyte—reproductive Fertilization Embryogeny CHAPTER XVIII. CoNIFEROPHYTES—GNETALES—WELWITSCHIA . Geographic distribution The sporophyte—vegetative . The sporophyte—reproductive The male gametophyte The female gametophyte . Fertilization Embryogeny CHAPTER XIX. CONIFEROPHYTES—GNETALES—GNETUM . The sporophyte—vegetative . The sporophyte—reproductive PAGE 184 184 186 186 215 217 226 231 234 275 275 306 306 321 321 331 331 342 342 361 362 364 368 377 381 384 384 385 392 400 401 404 404 408 408 412 CONTENTS The male gametophyte The female gametophyte . Fertilization and embryogeny CHAPTER XX. PHYLOGENY CHAPTER XXI. ALTERNATION OF GENERATIONS BIBLIOGRAPHY INDEX xl PAGE 420 421 423 427 440 446 481 \z . ae \&, Vas i = =. CHAPTER I NN # INTRODUCTION The gymnosperm line is one of extreme age, reaching back at least two or three hundred millions of years—so far back that its origin is lost in that distant past. But when we get our first glimpse of the group there were already two distinct lines, the Cycadophytes and the Coniferophytes, differing from each other in easily distinguish- able characters. Whether these two lines, millions of years back of any records yet discovered, may have had a common origin, we do not know. We simply know that the earliest material which has been found and studied shows the two lines about as sharply sepa- rated as their living representatives are today. That the gymnosperms did not originate as seed plants we believe to be self-evident. The old Greeks believed that Minerva, dressed in full armor, sprang from the head of Jove; but such an explanation of the origin of a group of plants would hardly satisfy a modern biolo- gist. For the more immediate origin of the gymnosperms we look to the Pteridophytes, for we believe that they were the ancestors of the gymnosperms. Of course, all the material for a study of the early gymnosperms is fossil, a record left in stone. By far the greater part of this record consists of impressions, and most of the impressions are those of leaves and stems, with some roots; but there are some impressions of reproductive structures. A leaf would fall into the sand or clay; then the sand or clay would become solid stone; all organic parts would be dissolved, and only the form would remain. Some material behaved differently. The most valuable material is in the form of calcareous nodules, called “coal balls,’ which are found scattered through coal seams like raisins in a cake. Coal balls were probably produced when water, with calcium or silica in solu- tion, invaded a swamp which formed a coal bed. The solution satu- rated plant tissues and preserved them, just as fixing solutions do today. Then the material solidified so that the plants were imbedded I 2 GYMNOSPERMS in solid rock. During the swampy conditions leaves, stems, roots, together with sporangia, spores, and seeds, falling into the swamp, where the water contained calcium or silica, became changed into stone. Starting with a piece of stem or leaf and some sticky clay, the mass would pick up other pieces of stem and leaf, and also spores or seeds, just as one sees the process going on today, until balls as large as one’s fist, or even larger than one’s head, would be formed. The mass, soaked with lime water, would become calcified, preserving the material so perfectly that not only cell walls but often cell contents can be studied about as in living material. To get such material into condition for microscopic study, the time-honored method, and still probably the best method, although very tedious, is to saw through a coal ball, polish a sawed surface, cement this smooth surface to a piece of glass; then saw as close to the glass as possible, thus leaving a thin section firmly cemented to the slide. This thin section is then ground and polished until it is thin enough to be studied with a microscope. It is then cleaned and dried, some balsam and a thin cover glass are added, and the mount is ready for study. A later method, much more rapid, and for some purposes almost as good, has come into favor because it is so easy and saves so much time. After sawing through a ball and polishing a surface, the pol- ished surface is immersed in 4 per cent hydrochloric acid for a minute or two; the acid is gently rinsed out with water; the water is removed with 95 per cent and 1oo per cent alcohol; the surface is flooded gently with ether alcohol, and then with a rather thick solution of celloidin in ether alcohol. When the ether alcohol evaporates, the celloidin can be peeled off in a thin film, which is placed in equal parts of bergamot oil, cedar oil, and carbolic acid, and then mounted in balsam. Such mounts look much like those prepared by the other slow, laborious process; but all calcareous structures are lost. ‘This method proves that the old theory, that all organic material was changed completely into stone, is incorrect. If it were correct there would be no plant structures in a celloidin peel. After taking off a peel, the surface from which the peels are being made is polished for half a minute, then cleaned and dried. It is then ready for another peel. One great advantage of the method is INTRODUCTION 3 that serial peels can be made with com- paratively little loss, for it is possible to get five, or even ten, peels to the millimeter. When sections are cut with a saw, they are generally a millimeter in thickness; and the saw, with the polishing for the next section, takes an- other millimeter. Consequently, serial sections are seldom less than two millimeters apart. However, it must be remembered that in the peel method all material which is soluble in the acid is etched away, leaving the insoluble material standing erect. The celloidin surrounds the standing material and hardens. When the peel is pulled off the stand- ing material is torn loose from the sur- face and, being held in place by the hardened celloidin, can be mounted and studied. While the method is very popular, its limitations must be recog- nized. Instead of celloidin, many paleobot- anists are now using gelatin in making peels. From impressions, and from sec- tions and peels of petrified material, plants which lived hundreds of mil- lions of years ago have been studied and described until paleobotany has become a major subject of critical importance in any study of evolu- tion or phylogeny. Besides, plant remains have been so_ thoroughly identified that they serve to identify geological horizons, and are of great MILLIONS of YEARS QUATERNARY TERTIARY ‘ND 62 CRETACEQUS 70 PERMIAN 40 CARBONIFEROQUS 80 DEVONIAN 50 i ORDOVICIAN 70 CAMBRIAN 100 PROTEROZOIC 500 ARCHEOZOIC 1000 Fic. 1.—Diagram of geological time.—Compiled by Dr. A. C. Noé from various sources. RECENT NP PLEISTOCEN PLEIOCENE MIOCENE 4, : " q y 4 wy OLIGOCENE EOCENE WZ) id % NWAANMAAWA cateterninie er he: oo BNENAAS BAN eeet eee Zz Gene "4 Al y ti UPPER CRETACEOUS NUNAN AAAAAAAAA QARANN ARRY AMT a NeNciewen L5///) | MIDDLE JURASSIC Drs 5 = NaS SR ae S, Ss ve - 30 Fics. 27—31.—Hypothetical development of a heterosporous carboniferous fern an- cestor of the Cycadofilicales. Fig. 27, a megasporangium with spore mother-cells, at this stage indistinguishable from a homosporous sporangium. Below are aborting spo- rangia and the sorus is surrounded by an envelope which might be called an integument. Fig. 28, one mother-cell has divided. Fig. 29, one megaspore has germinated and has produced two eggs, and has developed a heavy spore coat. Fig. 30, the female gameto- phyte is protruding and has an embryo. The sporangium wall has been ruptured. Fig. 31, the megaspore, with its embryo, has been shed from the sporangium and is germinat- ing outside the sporangium. Therefore it has not quite reached the seed stage. 38 GYMNOSPERMS the drop not only brings the pollen into the pollen chamber, but seals the chamber, so that the male gametophyte is protected during its further development. Fics. 32, 33.—Hypothetical sections of megasporangia of a member of the Cycado- filicales after it has passed beyond the heterosporous fern stage and become a seed plant. Fig. 32,a megasporangium at the time of fertilization. Three pollen grains in the pollen chamber. No pollen tubes. The megaspore coat is thinner than in the heteros- porous fern stage. Fig. 33, the seedling is developing while the megaspore is permanent- ly retained within the megasporangium. The seed stage has been reached. Fic. 34.—Hypothetical development of the male gametophyte in a member of the Cycadofilicales, showing, in order, a microspore; a prothallial cell and a fertile cell; both vegetative and fertile portions developing; differentiation into spermatogenous cells and wall cells; mature sperms; sperms escaping. In early stages, in phylogeny, the male gametophyte, by the breaking down of a few cells at the top of the sporangium, came into contact with the female gametophyte, and there was no formation of a pollen tube. When the sperms were shed, they were in as close CYCADOFILICALES 39 contact with the eggs as in a living cycad after the development of its pollen tube. After fertilization, the degree of development of the embryo or the presence or absence of a resting period are not con- cerned in the definition of a seed. In the development of the male gametophyte there was probably not much difference between the heterosporous fern and the early seed plant. In the homosporous fern, the development was probably about like that of the homosporous fern of today, with practically Fics. 35-36.—Development of an embryo with walls forming from the beginning. Fig. 36, Development of an embryo with a free nuclear stage preceding wall for- mation. all of the gametophyte outside the spore.. In the early development of heterospory, the gametophyte probably protruded considerably and became green, as it sometimes does in living heterosporous forms. As heterospory advanced, the gametophyte tissue became more and more included within the spore, protruding only enough to crack the spore coat and shed the sperms (fig. 34). The pollen tube, at first only a haustorial organ, is a comparatively modern develop- ment. The Cycadofilicales had not reached this stage. A guess at the development of the embryo might be hazarded. As the homosporous fern condition was passing into the heterosporous fern condition, while the fertilized egg was still quite small, there may have been no free nuclear period; but as the megaspore and its eggs became larger, there would be a free nuclear period, followed by the formation of walls and body regions (figs. 35 and 36). 40 GYMNOSPERMS Why no embryos are found is still a mystery. They are abundant in the Bennettitales, and structures as delicate as these embryos may have been are well preserved. Although it may seem presumptuous to dispute the claims of the two greatest living paleobotanists, Scorr* and SEWARD, nevertheless, a lifetime study of the comparative morphology of living plants, especially Pteridophytes and Gymnosperms, makes me feel confident that the course of evolution has been from homospory, through het- erospory, to the seed. The genetic line must have been homospo- rous Filicales, heterosporous Filicales, Cycadofilicales. *D.H. Scott died on January 29, 1934. CHAPTER HI CYCADOPHYTES—BENNETTITALES Just as the Carboniferous has been called the ‘“‘Age of Ferns,” the Mesozoic has been called the “Age of Cycads.”’ In the Carboniferous, the Cycadofilicales reached their highest development and began to decline. While they were still plastic, probably in the Upper Carboniferous, two great lines, which later became prominent, began to differentiate from the plexus, while the ancestral lines became weaker and weaker and, finally, in the Trias- sic and early Jurassic, became extinct. One of the new lines is called the Bennettitales; the other, the Cycadales. The Bennettitales de- veloped rapidly, reached their greatest display in the Jurassic, and became extinct in the Upper Cretaceous; while the Cycadales, al- though not so prominent a feature of the vegetation, still flourish in various tropical and subtropical regions. What caused the Bennettitales to become extinct, while the Cyca- dales survived, can only be conjectured. Both were probably eaten by the immense herbivorous dinosaurs of the Jurassic and Triassic. Their leaves and trunks, the parts constantly exposed to the weather, were very similar; the flower buds of the Bennettitales would seem to be even better protected than those of the Cycadales and, edaphi- cally, the two groups faced the same conditions. But the covering of the strobili by bud scales may not have been entirely advantageous for, if the seeds were as short-lived as those of the living Cycads, many of them would have died before they were shed from the plant. The seeds of the Cycadales are larger and their thick stony layer may have made them more resistant than the small seeds of the Ben- nettitales. At any rate, one group died while the other lived. Many have made contributions to our knowledge of this group; some toits taxonomy, some to morphology, and some to phylogeny; but the name of WIELAND will always be most prominently asso- ciated with the Bennettitales, for he has collected more than all 41 42 GYMNOSPERMS others combined, has prepared his own material for microscopic study, developing excellent methods, and has made his own draw- ings, photographs, and photomicrographs. Three large quarto vol- umes, with scores of plates and numerous drawings in the text, to- gether with a clear and interesting literary style, present the results of a lifetime of productive research. And besides there are numerous shorter papers. DISTRIBUTION At this time, any account of the geographic distribution of the Bennettitales must be regarded as merely a beginning. Whenever WIELAND makes a trip into a Mesozoic region, a new locality for fos- sil cycads is added to the list. Wherever members of this group have flourished, they are likely to be preserved, because their armored trunks and tough leaves, with heavy cuticle, make sharp impressions even when no internal structure is preserved. In many localities the material is silicified. American forms were first brought to notice in 1860. Various peo- ple picked them up between Baltimore and Washington and kept them as curiosities, calling them fossil bee hives, fossil wasps’ nests, etc. Ward®s?-°>*:°59 first described them and 60 were taken to the Wom- an’s College in Baltimore but have been removed to the Museum of Natural History in Washington. Specimens have also been found in North Carolina, Pennsylvania, Kansas, Colorado, Texas, and Cali- fornia; but the most extensive collections have been made in the Upper Jurassic of Wyoming and the Black Hills of South Dakota. To protect the richest part of the Black Hills region, WIELAND took up 360 acres as a claim, and then succeeded in having it made a national preserve. More than 700 trunks from this locality are now in the Yale Museum, where, with many specimens from other lo- calities, they constitute the largest collection in the world. WIELAND has recently discovered a rich locality (not yet described) in Arizona. Prince Edward Island, in Canada, has silicified Triassic material. WIELAND,**” in 1909, visited the Mixteca Alta, near Oaxaca, Mexi- co, and discovered a region wonderfully rich in fossil cycads. The result of this investigation, published (in Spanish) by the Mexican government, is illustrated with 50 excellent plates made from photo- graphs. The principal genera in this locality are Ptilophyllum, Ptero- BENNETTITALES 43 phyllum, Otozamites, and Williamsonia. Many of the species are new. While the impressions of leaves and fruits are abundant and wonder- fully clear, there are no silicified specimens to show the internal structure. In England and Scotland well-preserved specimens have been found in various horizons in the Triassic, Jurassic, and Cretaceous. Germany, Belgium, Poland, Russia, Italy, and the Isle of Wight have yielded material, and India has long been a famous fossil cycad locality. LIFE-HISTORY Most of the material from which the internal structures of the Cycadofilicales have been studied is calcified. In the Bennettitales the material is silicified. In making peels, hydrofluoric acid must be used, and cutting sections is more laborious; but the harder surface allows a higher degree of polish, and very satisfactory studies, and even photomicrographs, can be made from cut surfaces without any balsam or cover. The stem.—If any form can be considered typical of the Bennetti- tales, it is the unbranched stem with a crown of leaves at the top; but branching was probably as common as in the living cycads. The axillary strobili gave the stems a very characteristic appearance so that they could not be mistaken even for those Cycadales in which the strobili are axillary (fig. 37). The tuberous habit was common, and this, with frequent branch- ing, gives the specimens a striking resemblance to the branching plants of Zamia. The stems and branches, which are so similar that it is often difficult to determine which is the main stem and which are the branches, are usually as much as 15 cm. in diameter, and speci- mens often reach a diameter of 25-50 cm. The branching tuberous forms are not very tall, usually not exceeding half a meter. Some columnar forms, which are not likely to be branched, have reached a meter in height. The tallest on record is Cycadeoidea gigantea (Sew- ard), 1.18 M. high and about one-third that figure in diameter. The smallest known specimen is Bennettites scotti, 8.5 cm. in height and about two-thirds as much in diameter; this, however, was prob- ably a bud from a larger specimen. 44 GYMNOSPERMS Williamsonia gigas was tall and slender, while Williamsonia (Ano- mozamiles) angustifolia was slender and profusely branching. The Williamsonia (Anomozamites) angustifolia type of stem shows, on the outside, what the Dioon type of stem shows on the inside; for the outwardly unbranched stem of Dioon is really profusely branched Fic. 37.—Cycadeoidea wielandii: upper part of a trunk found near Hermosa, South Dakota: some of the strobili are projecting and some have fallen out, leaving cavities which are dark in the illustration. The height of the specimen is 54 cm.—From a photo- graph by THIESSEN. on the inside (fig. 38). If material of this Williamsonia were avail- able, just at the base of the terminal cone we should expect to find newly formed meristems which would give rise to branches. Each of these branches, with its leaves, would soon be terminated by a cone, at the base of which a new meristem would form, and so a profusely branched plant would be developed. It should be noted that the BENNETTITALES 45 leaves are only once pinnate, and that in Williamsoniella coronata the leaves are simple. In Williamsonia gigas the stem is outwardly unbranched and the internal structure, judging from Williamson’s reconstruction, would also be unbranched. The stems were covered by an armor of persistent leaf bases, as in most of the living Cycadales. The numerous axillary cones, often hundreds, gave the stems a striking appearance. Fic. 38.—Williamsonia (Anomozamites) angustifolia: branching trunks with leaves and strobili at the forks; about 4 natural size—After NATHORST. The vascular system is an endarch siphonostele, the highest type found in gymnosperms, and characteristic of the Archichlamydeae and Sympetalae in the angiosperms. No seedlings have been ex- amined and, consequently, it is not known whether a mesarch condi- tion might be found, as in the living Cycadales. In transverse section the stem shows the large pith, scanty wood, and thick cortex, which are so prevalent in the whole Cycadophyte phylum (fig. 39). The scanty zone of wood, as in most of the living Cycads, shows no growth-rings. Cycadeoidea jenneyana is exceptional in having a 46 GYMNOSPERMS zone of wood, reaching, near the base of the trunk, a radial thickness of 8 cm. Growth-rings in this specimen are very distinct. WIE- LAND thinks the trunk is polyxylic, although he admits that there may have been a persistent cambium. His excellent photograph of the transverse section is so identical with the appearance of Dioon spinulosum and Dioon edule, where the rings are undoubtedly formed by a persistent cambium, that we feel certain that the same condition is present in Cycadeoidea jenneyana. Besides, the polyxylic condi- tion, with its successive cambiums in the living cycads, where it is Fic. 39.—Cycadella sp.: transverse section of top of small trunk from Wyoming, showing large pith, woody cylinder, and cortex; m, pith; x, xylem; c, cambium; cb, leaf traces; r, ramentum; h, leaf (or peduncle) traces; /, base of leaf; natural size —After WIELAND.*% well known, makes the successive zones or rings look very different from the growth-rings formed by a persistent cambium. SCOTT says that in Cycadeoidea yatesii there is evidence for the presence of two or more zones of xylem and phloem, like those in Cycas. Whether the rings, whatever their character, are annual, may be very doubt- ful, for, in the living cycads, rings looking very much like those of Cycadeoidea jenneyana may be formed every other year, or at inter- vals of 20 years or more. The leaf-trace bundles pass directly through the cortex into the leaves, there being none of the girdling which is such a prominent feature of the living cycads. There are no bundles in the pith, nor, with axillary cones, would any be anticipated. However, there are BENNETTITALES 47 numerous mucilage cavities in the pith, some of them several cells wide. The histological details of xylem and phloem are beautifully pre- served, but the delicate cells of the cambium are broken (fig. 40). Transverse sections at about the middle of the xylem region and the middle of the phloem region are shown in figs. 41 and 42. mh) s = Pixms] >. \w) ¥ Y Inte ‘= oats wu Ar cy le voy an ( Oy. li a " J un’ CG aX ea A Sei AOA @0n, | | Bors 20 Re fe ly 8 seeSe, Ye icy xe DA~S- < la \: ( Py se x Od Dosis I h~soS Cory LS CHA 7 pe & one, i Ps NAY ‘a Pe 8! oa Mi aa, t+, Fic. 40.—Cycadeoidea wielandii: c, cambium mostly broken down; above it the phloem with many thick-walled tracheids, and below, the secondary wood. xX 100.— After WIELAND.*8° Most of the tracheids are sclariform, but those of the protoxylem, next to the pith, are spiral. No pitted tracheids were observed in Cycadeoidea wielandi, although the preservation is so excellent that they could hardly escape notice, if present. However, LIGNIER?° found bordered pits on the radial walls of tracheids of Cycadeoidea micromela, and there are some bordered pits in Cycadeoidea painei and dartont. The markings on the tracheids should receive most criti- 48 GYMNOSPERMS cal attention, for the Cycadofilicales had so generally reached the pitted condition. The question may well be raised whether a plant whose tracheids have reached the pitted condition could give rise to Calyite®: fee Wa p) 4; nites! i's Fic. 41.—Cycadeoidea wielandii: transverse section of secondary wood; X150.— After WIELAND.®°” —\ i “Bot Me = le in = line =) san Oboe IC Fic. 42.—Cycadeoidea wielandii: transverse section of phloem showing thin-walled cells and thick-walled tracheids; * 150.—After WreLanp.™ a line with sclariform tracheids. Atavism is a familiar phenomenon; but, as we understand it, its influence is potent only for a short time, not persisting from one geological age to another. If the sequence, spiral, sclariform, pitted, is fundamental, we should not expect a form ional le. In the lower part only occas , but because the natural coloring does not innu transverse section of p Fic. 43.—Cycadeoidea ingens (type) thick-walled cells are shown, not because of poor preservation make them d 680 /IELAND. = X140.—After W ? istinct 50 GYMNOSPERMS like Cycadeoidea wielandi to be derived from any of the Cycadofili- cales which have already reached the pitted condition. But if the sequence is due to physiological conditions, especially rate of growth, the kind of markings would have little or no significance as far as relationships are concerned. wv Fic. 44.—Cycadeoidea jenneyana: tangential section of a trunk showing numerous transverse sections of strobili among the foliage leaves; 3.—After WreLANnpb.* We should anticipate much more of the pitted condition than has yet been found; but it must be remembered that Stangeria still re- tains the sclariform tracheid of its very remote fern ancestors, and that Dioon, except in young plants, has sclariform tracheids in the protoxylem. BENNETTITALES 51 The leaf —The leaves of the Bennettitales and the Cycadales are so similar that we doubt whether the two groups could be separated on the basis of the leaf alone. In both, the dominating leaf is the pinnate leaf of their filicinean progenitors. Leaves more than once pinnate are rare, found only in Bowenia in the Cycadales, and we have not found any record of more than once pinnate leaves in the Bennettitales, if we except the pos- sibility that the synangia are modi- fied pinnules. If they are modified pinnules, the microsporophylls are bipinnate. In Wailliamsonia coro- nata the leaves were simple and entire. In size, the leaves are about like those of living Cycads, reaching a length of more than 3 meters in Cycadeoidea ingens and only about JUL 6 cm. in Williamsonia coronata; paved | most of them, however, ranged in FOUL size between these two. In the abundant material from the Lias of the Mixteca Alta of Mexico, many Fic. 45.—Cycadella ramentosa: of the leaves of Ptilophyllum and ae eng ae Soma nee Otozamites were less than 25 cm. in length, and in Ptilophyllum acutifolium var. minor the leaflets are less than 10 cm. long. The venation is “parallel,” which, in this phylum, means that it is dichotomous, with very little forking beyond the base of the leaf- let. There are no midribs in any of the leaves, but it is interesting to note that, associated with leaves of the Bennettitales in the Mix- teca Alta region, there are leaflets with midribs (Stangerites oaxa- censis and Sagenopteris rhoifolis var. mexicana) resembling leaflets of the living African Stangeria. In this connection, it might be men- tioned that a South African specimen, named Zamites, in the muse- um at Grahamstown, South Africa, is hardly distinguishable from the living Dioon edule of Mexico. The histological structure of the leaf is typically xerophyllous, x — = KA iD Amur vy PA TL YY) G GYMNOSPERMS eal N with strongly cutinized epidermis. The leaves are thick, and their internal structure shows that they were leathery and well fitted to withstand adverse conditions (fig. 43). On the upper side there are some thick-walled hypodermal cells and a well-marked palisade. Near the middle is a single row of mesarch bundles with a strong bundle sheath connected with the upper part of the leaf by scleren- chyma cells. Below the bundles is a wavy line of thick-walled cells, Fic. 46.—Cycadeoidea ingens: photograph of a model in the Field Museum of Nat- ural History in Chicago. The model is almost entirely of glass and was made by Mr. SELLA, with the aid of criticism and advice by WIELAND. beneath which is a thick layer, extending to the lower epidermis, of thick-walled cells. In the figure, only a few of these are clearly shown, and although the preservation is good, the lack of natural coloring makes most of the cells very indistinct. In the upper part, the photomicrograph was retouched. The root.—Scarcely anything is known about the root of this group. A paper by Dr. Stopes describes rootlets in a section identi- fied as Bennettites saxbyanus, but these rootlets were probably ad- BENNETTITALES 53 o ventitious, from an adventitious bud. They were about 1 mm. in diameter, and had abundant root hairs. There were many rootlets in the section, cut in various directions. In longitudinal and in oblique views there were ‘three or four rather large sclariform tra- cheids. There was no satisfactory view in transverse section. N Fic. 47.—Cycadeoidea: diagrammatic view of unexpanded strobilus. Outside the sporophylls are numerous bracts densely covered with ramentum; about natural size.— After WIELAND.*° The strobilus—The most characteristic feature of the Bennetti- tales is the strobilus. The strobili were borne on the upper part of the plant, usually in great numbers, hundreds having been counted in several cases. They are all axillary, in some cases a strobilus being 54 GYMNOSPERMS borne in the axis of every leaf. A tangential section of a trunk of Cycadeoidea shows transverse sections of the bases of foliage leaves and numerous strobili surrounded by scale leaves (fig. 44). At the bases of foliage leaves and entirely covering the scale leaves is a dense ramentum consisting of scales several cells wide and often more than one cell in thickness (fig. 45). This is a dominant fern character and not at all present in the Cycadales, in which the ramentum consists aoe it O Nae ys ee 2 ima Anette. = (eee 2m * Pa aoa Fone a ~ apie Pte AW fi emg aie » Res Fis zs g ‘a G Fic. 48.—Cycadeoidea dacotensis: section of an unexpanded microsporophyll show- ing, in most cases, transverse views of the two rows of loculi. At the upper right, parts of two microsporangia are cut nearly longitudinally. Some of the loculi contain micro- spores; X 25.—After WreLANp.® of long unicellular hairs. Occasionally there is a transverse partition, but even where the ramentum is equally abundant in the Cycadales the two groups could be distinguished by this character alone. The strobili of the Bennettitales are typically bisporangiate, with leaflike microsporophylls as loosely arranged as the megasporophylls of the living Cycas, and in the center a cone consisting of innumerable small megasporophylls, each bearing a single terminal ovule. The megasporophylls are slender peduncles without a trace of lateral pinnae or ovules. Intermixed with the megasporophylls are sterile leaves with a thickened top and with no lateral leaflets. A splendid reconstruction of the bisporangiate strobilus of Cyca- deoidea ingens has been made for the Field Museum by Mr. SELLA, of BENNETTITALES 55 the Museum staff, with suggestions and corrections by WIELAND (fig. 46). The leaflike microsporophylls surround the comparatively small ovulate strobilus. The time of flowering has been the subject of considerable specu- lation. It has been noted that in some of the tallest specimens there are strobili only in the upper part, and that these strobili are of ap- proximately the same age. Furthermore, in some there are no vegetative leaves at the extreme tip, and no indication that yy there would be any. If such leaves had Gi jy been present, they would have been dis- Ci! lg N covered, as WIELAND found them in Cy- AY iD cadeoidea ingens. From a very thorough ES AS study, WIELAND concludes that many of Si A | =; the Bennettitales flower but once, and Telia Q US, then die. Such a behavior is well known np. ’ ; Is in angiosperms. The century plant, A gave ey 0 iS americana, flowers but once and dies; but =| iS it does not wait a hundred years. Plants ai ia from ‘suckers,’ about 3 years old, flower ip ®}, is in 5-7 years. The palm, Corypha umbra- ra a) le culifera, attains an age of about 40 years a | = and a height of 60 feet, when the axis is Bh i MF suddenly prolonged another 40 feet as aS S an immense flower stalk bearing thou- TTS sands of flowers. The plant then dies. Fic. 49.—Cycadeoidea daco- ; : tensis: longitudinal section of Yucca whipplet flowers but once, and ’ ; ; sporangium showing stalk and dies; while other species of the genus two loculi containing micro- flower year after year. None oftheliving spores; X40.—After WIE- : : LAND. 8 cycads die on account of flowering. The micros porophyll—The microsporophylls retain more of the ancestral fern-leaf characters than any of the rest of the Cycado- phyte line above the Cycadofilicales. They have about the same topography as the vegetative leaves, but are much reduced in size. The pinnules bear synangia on both sides, so that the pinnule looks again pinnate. If the synangia are really reduced pinnules, as WIE- LAND believes, the sporophylls are twice pinnate. Some strobili, 56 GYMNOSPERMS younger than the one shown in fig. 46, have been sectioned, yielding younger stages in both micro- and megasporangia (fig. 47). The microsporangia are multilocular. We should hesitate to call them “‘synangia,”’ since the term implies a fusion of separate sporan- gia to form a single large synangium. The structure, however, is practically identical with that in the Marattiaceous ferns (fig. 48). There is an outer layer of thick-walled cells, followed by thin-walled CD DOE €C DAdQEO @ @ 222 Gr [9 3 Gwe & €o CoCo ae Fic. 50.—Cycadeoidea dacotensis: microspores indicating multicellular structure.— After WIELAND. cells and then, doubtless, a tapetum. There is a cleft between the two rows of loculi, and each loculus opens into the cleft, as in Ma- rattia. A complete longitudinal section of the sporangium is shown in fig. 49. Nothing very definite is known of the internal structure of the microspore, but it seems practically certain that there were several cells. The fact that so many show a circular marking would indicate that there is a prothallial cell (fig. 50). No pollen tubes have been observed. There is some nucellar tis- sue, but whether it breaks down so as to allow the pollen grains to come into direct contact with the gametophyte, as in the Cycado- filicales, is not known. BENNETTITALES 57 The megas porangium.—The solitary megasporangium is terminal on a slender stalk, which represents the rachis of the sporophyll. In the Cycadales, all the ovules are lateral, while, here, all are terminal (fig. 51). In the two great lines, evolution has taken two distinct NWN S ~ TS AWN TM A A / i--~\. NW SSR NAAMUANTAUN Fic. 51 Fic. 51.—Bennettites gibsonianus: diagram of female strobilus, showing terminal seed with dicotyl embryos. The sporophylls are long and slender, and they alternate with sterile bracts (sporophylls). The whole strobilus is surrounded by bracts.—Modi- fied by Scorts4 after Sorms-LAUBACHS7 and PoToNnré.455 Fic. 52.—Bennettites morieri: longitudinal section of a seed; a, micropylar tube; 0, prismatic layer; c, pulpy tissue; d, ‘‘corpuscular” mass; e, interseminal scale; f, embryo space; g, remains of nucellus; /, chalaza; 7, tubular envelope; k, micropylar canal; /, nu- cellar beak; m, pollen chamber; n, fibrous stratum; 0, basal expansion of n; p, pedicel bundle.—After WIELAND’s®* reproduction of LIGNIER’S39 figure. courses. Starting with sporophyll, bearing ovules both laterally and at the apex, the lateral ovules, in the Bennettitales, became sterile and disappeared; while in the Cycadales the top of the sporophyll became sterile and the lateral ones remained. 58 GYMNOSPERMS The ovules are prevailingly small, many of them not more than 5 mm. in length, and few of them exceeding a centimeter. Just what the structure is and how it compares with that of groups above and below has not yet been determined very definitely. The best illustrations available are drawn from later stages in Cycadeoidea darloni and Bennettites moreri (figs. 52 and 53). The outer layer is thick and palisaded at the top; the middle layer, shown in black in Fic. 53.—Cycadeoidea dartoni: longitudinal section of ripe seed with a dicotyl em- bryo; X 20.—After WIELAND.*87 the figure, is stony, and the thinner inner layer is probably mem- branaceous. The thin line, at a little distance from the embryo, may be the megaspore membrane. The complicated micropylar end of the ovule is a striking feature. The embryo.—It will be remembered that in the Cycadofilicales no embryo has yet been found. The almost constant presence of dicotyl embryos in the Bennettitales is in striking contrast. As shown in figs. 51 and 53, the hypocotyl is more extensive than in the living cycads, and the suspensor, if present, could not have been very prom- inent. eo ——— — —————<—— Oe BENNETTITALES 59 PHYLOGENY Paleobotanists agree that the Bennettitales have come from the Cycadofilicales. The leaves of some of the lower forms, like Pizlo- phyllum and Williamsonia, are so identical with the leaves of Dioon, that the similarity may not be accidental. It may mean that both Bennettitales and Cycadales inherited the leaf from some very re- mote ancestor; or it may be possible that some of the unattached leaves actually belong to the Cycadales. The habit of the trunk in both groups may have been similarly inherited. But the strobilus is so different in the two lines that the Bennettitales could not have given rise to the Cycadales. The lateral leaflets of the megasporo- phyll, entirely lost in the earliest known Bennettitales, could not have been transmitted to the Cycadales. The vertebrate paleontolo- gist illustrates this law of evolution by saying that a tooth, lost in phylogeny, is lost for good. The Bennettitales came from the Cyca- dofilicales, probably in the later Carboniferous. Although no defi- nitely recognizable material dates that far back, the first recogniza- ble specimens are so advanced that they must have been separated from the main line for a long time. That the Bennettitales may have given rise to any of the angio- sperms we regard as not only improbable but impossible. One at- tempt to connect them makes the connection with the Sympetalae, a group obviously derived from the Archichlamydeae, and not only cyclic, but tetracyclic. A more recent attempt to connect them with the angiosperms, and a connection which seems more reasonable, has been with the magnolias; but, again, the superficial resemblance of the bisporangiate fructification of the Bennettitales to the magnolia type of flower does not seem adequate for establishing such a rela- tionship, and so we must conclude that the Bennettitales were the last of their line; they left no progeny. CHAPTER IV CYCADOPHYTES—CYCADALES All the forms described in previous chapters are extinct. The Cy- cadofilicales had already begun to decline before the end of the Car- boniferous, so that it was a dwindling group which struggled through the Triassic and touched the Jurassic. The Bennettitales were domi- nant in the Jurassic, even forming forests, if a dense vegetation of such small plants could be called a forest; and well into the Creta- ceous they were still abundant. But none of them persisted until the end of that period. The Cycadales, like the Bennettitales, were derived from the Cycadofilicales, probably being differentiated from that group in the later Carboniferous. Consequently, throughout the Mesozoic, until the Bennettitales became extinct, the two groups must have been growing side by side. Why are the Cycadales so scantily represented as fossils, while the Bennettitales are so abundant? The strobili of the Cycadales usually begin to decay as soon as they reach maturity. Sometimes, under dry conditions, the cones shrivel and become very hard; but under such conditions, fossilization is unlikely to take place. Of course, the leaves are practically identical in the two groups, so that it is quite possible that some of the unattached leaves which have been assigned to the Bennettitales really belong to the Cycadales. The trunks also are often very similar, and the resem- blance would be most striking when the Cycadales trunk bore nu- merous axillary strobili. Neither the Bennettitales nor the extinct members of the Cyca- dales have left any fossils of such great size as some of the living Cycadales; for many reach a height of 2 meters, while 4 or 5 meters is not rare, and occasional individuals have been measured up to a height of 18 meters. While the Cycadales have not left as complete a record in the rocks as we might wish, still, there is enough to prove their presence 60 CYCADALES 61 practically throughout the Mesozoic, and there were cycad-like leaves in the Permo-Carboniferous. Megasporophylls of Cycadospa- dix hennoquet, from the lower Liassic, differ less from those of the living Cycas revoluta than those of Cycas revoluta differ from some of the other living species of the genus. Records are more abundant in the Jurassic; and the group reached its widest distribution and greatest display in the early Cretaceous. Then it began to decline, so that the living cycads are less abundant than their prede- cessors. The lack of a satisfactory geological record is partly compensated for by the fact that the cycads have come down from the remote past with so little change that, if one could be transported back a hundred million years, he would doubtless recognize some of the genera. The cycads of today may well be called “‘living fossils.” There is only one family, the Cycadaceae, with only nine genera, four of which belong to the Western Hemisphere, and five to the Eastern. In a recent monograph, SCHUSTER*”’ recognizes 65 species; but with his subspecies, varieties, and forms—categories of no in- terest to the morphologist—the number is much larger. We prefer to regard a species as a norm which may vary considerably in many directions. Giving names to variants, which may never occur again, merely makes confusion and burdens the nomenclature. If all genera and species had been described by competent observers who had studied them in the field, taxonomy would not have so many species to deal with. GEOGRAPHIC DISTRIBUTION In the Mesozoic, the cycads were world-wide in their distribution, as cosmopolitan as Pieris and Typha are today; but now they are confined to tropical and subtropical regions, and even there they occur in scanty patches, in out-of-the-way places, so that collecting involves hard tramping, often over rocky, inhospitable ground, where natives as well as nature seem to conspire against success. The great cycad regions of the world are Mexico and the West Indies, in the Western Hemisphere; and Australia and South Africa in the Southern Hemisphere. Only two genera occur outside of these regions: Zamia, in the West, and Cycas in the East. 62 GYMNOSPERMS THE WESTERN CYCADS Two of the western genera, Dioon and Ceratozamia, are entirely confined to Mexico, and Zamia is abundant there. Microcycas is found only in western Cuba; but here, again, Zamia is abundant. The cycads grow so luxuriantly in most tropical and subtropical regions that they are popular decorative plants on lawns and in patios. Consequently, people who write from hearsay, without actu- ally visiting the localities, may greatly increase the range of a plant. Dioon edule.—The finest station for Dioon edule is at Chavarrillo, 15 miles east of Jalapa (fig. 54). It grows in the blazing sun, asso- ciated with small cacti, small acacias, bromeliads, and small oaks 30-60 cm. in height. Although plants are abundant and cone freely, the patch does not seem to be spreading. Farther down the line, at Palmar and Colorado, there is another large stand; but there does not seem to be any extension beyond Rinconada. There are plants at Huatusco, south of Jalapa; and at Rascon, between San Luis Potosi and Tampico, Dioon edule is so abundant that many cattle die from eating the poisonous leaves. It has also been reported, probably correctly, as growing in rocky places in the states of San Luis Potosi, Nuevo Leén, and Tamaulipas. There are doubtless many other stations, but any reports not coming from experienced botanists should be checked, because the plant is so often used to decorate the patio. Dioon spinulosum.—Dioon spinulosum grows much farther south than D. edule, appearing first about 55 miles south of Vera Cruz, and becoming more abundant southward (fig. 55). On the Hacienda de Joliet, near Tierra Blanca, there are hundreds of large plants, some of them 10-16 meters in height. South of Tuxtepec, about 100 miles south of Vera Cruz, there are forests of this species. Since it is very popular as a decorative plant, any reports of its geographic distribu- tion should be confirmed. In the first two localities from which I obtained material, the plants were nursery specimens. The plant does not occur, except as a nursery specimen, in either of the two localities mentioned in the original descriptions. Dioon purpusii.—This comparatively new species has a much more restricted range. It is found near the railroad at Santa Cata- rina; also in the Tomellin Canyon, north of Oaxaca, and in the Sierra CYCADALES 63 Mixteca, Puebla. It looks like Dioon edule and might be mistaken for it. Years before the new species was described, WIELAND figured Fic. 54.—Dioon edule: a female plant at Chavarrillo, near Jalapa, Mexico (Septem- ber, 1906). The trunk is about 1.5 meters in height, and shows the armor of leaf bases. It is about 1,000 years old.—After CHAMBERLAIN." it as Dioon edule, in the first volume of his American Fossil Cycads*° (his fig. 101). The drawing is so accurate, even in minor details, 64 GYMNOSPERMS that, from his drawing alone, there is no doubt that the plant is Dioon pur pusii. Ceratozamia.—This is a variable genus and several species have been described; but most of them could probably be raised from the Fic. 55.—Dioon spinulosum: on the Hacienda de Joliet, near Tierra Blanca. The plant in the center is about 10 meters in height.—After CHAMBERLAIN." seeds of a single cone of Ceratozamia mexicana, the dominant species. C. mexicana is abundant a few miles out from Jalapa in the direction of the extinct volcano, Naolinco (fig. 56). On the precipitous moun- tain side opposite the volcano, growing in dense shade, are hundreds CYCADALES 65 of mature plants. How far they extend was not determined. This species also grows at Huatusco. Colipa and Mirador are also cited as stations for species of Ceratozamia. Fic. 56.—Ceratozamia mexicana: on the steep mountain side opposite the extinct volcano, Naolinco, near Jalapa, Mexico (September, 1906).—After CHAMBERLAIN.™ Microcycas.—This genus is monotypic, with Microcycas calocoma as the only species, and western Cuba as the only locality. It is a tall, arborescent form, often 2 or 3 meters in height, and occasionally 66 GYMNOSPERMS reaching a height of 8 or 10 meters (fig. 57). The name, Microcycas, is unfortunate, for it is one of the largest of the cycads and does not look at all like Cycas. In sharp contrast with the generic name, the specific name is very fitting, for the crown of leaves is exceptionally beautiful. In striking contrast with the forms before mentioned, iy 57.—Microcycas calocoma with male cone: near Herradura, Cuba (October, 1914). Microcycas does not appear in cultivation in the open, and conserva- tories have not been able to make it flourish. Zamia.—The remaining western genus, Zamia, with more than a third of all the species in the family—ScuusTER says 28 species—has a wide distribution, its various species ranging from Florida through the West Indies, through Mexico, Central America, the northern CYCADALES 67 part of South America, and down the Andes into Chili (fig. 58). Some of the species are local, well marked, and easily recognizable, while others range widely and are so variable that identifications are un- certain. The smallest cycad known, Zamia pygmaea, grows in the Microcycas region. The leaves of adult coning specimens are some- EA & s # % tae Fic. 58.—Zamia floridana: at Miami, Florida. The male cones on the left are nearly ready to shed pollen; the female cone at the right is a year older. It has reached its full size, but the seeds are not quite ripe. times only 4 or 5 cm. long. Associated with Z. pygmaea, but in soil not quite so bad, is Z. kickxit. It is quite possible that Z. pygmaea might become Z. kickxii under equally favorable conditions. The largest species of Zamia are arborescent, and have leaves over a meter in length. THE EASTERN CYCADS The remaining five genera are oriental. Two of them, Macroza- mia and Bowenia, are confined to Australia, while a third genus, Cycas, is abundant in Australia, but extends beyond through various 68 GYMNOSPERMS islands, touching India and China, and reaching its northernmost limit in the southern part of Japan. The other two genera, Encepha- lartos and Stangeria, belong exclusively to South Africa. Fic. 59.—Macrozamia denisoni: this large plant 18 feet (5.5 meters) in height, is on Tambourine Mountain, near Brisbane, Australia. The residents call it the “Great Grandfather Peter Tree.”—From a photograph by Miss H1tpA GEISSMANN. Macrozamia.—M acrozamia is abundant from New South Wales to the northernmost part of Queensland (fig. 59). ScHUSTERS’ allows CYCADALES 69 it g species; but there are certainly more, some of them, like M. fra- seri, with its unique cones, and M. macdonnellii, with its immense seeds, being very sharply marked. Perhaps the most variable species is Macrozamia spiralis. At Avo- ca, near Sydney, it forms almost impenetrable thickets, extending almost to the high-tide line of the ocean. New South Wales has gummy nn ™ Fic. 60.—Macrozamia moorei: a poisoned specimen at Springsure, Queensland, Aus- tralia, showing the notch and the hole into which arsenic was placed to kill the plant. — After CHAMBERLAIN.™® several other species, bearing more or less resemblance to M. spi- ralis and probably related to it. Some of them extend as far west as the Blue Mountains. All of these have tuberous, more or less sub- terranean, stems. The most abundant species with a tuberous stem, north of this region, is M. miquelii, in Queensland. It is at its best around Rockhampton and Byfield. The arborescent species flourish in Queensland. Macrozamia deni- soni, on Tambourine Mountain, has been called the most beautiful of all cycads. It is often more than 2 meters high. M. hopei is the 70 GYMNOSPERMS tallest of all cycads, reaching a height of nearly 20 meters. It is found in northeastern Queensland. M acrozamia moorei, at Springsure, in Queensland, about 200 miles west of Rockhampton, is one of the most remarkable of all cycads. The range is limited, but plants are so abundant within that range that they cause great havoc among cattle, which eat the poisonous green leaves (fig. 60). Fic. 61.—Bowenia serrulata: a fully grown plant at Byfield, near Rockhampton, Australia. In the western part of Australia, Macrozamia fraseri may be the only species. It is tuberous or short stemmed and both micro- and megasporophylls are prolonged into spines, sometimes as much as 15 cm. long, giving the cones a very characteristic appearance. Bowenia.—This is another genus strictly confined to Australia and, as far as we know, not occurring outside of Queensland. There are two species, Bowenia serrulata, at its best around Byfield and Maryvale, near Rockhampton, and B. spectabilis, in the Cairns re- gion. The bipinnate leaves mark it off sharply from all other cycads (fig. 61). CYCADALES 71 Cycas.—This genus, like the American genus Zamia, has a wide distribution, its various species ranging beyond Australia into the ‘S ae <- 5s el ee WA nef Fic. 62.—Cycas media: a female plant with ripe seeds, at Frenchman’s Creek, on the farm of Mr. SypNEy W. SNELL, near Rockhampton, Australia. The trees are Eucalyp- tus (November, 1911). islands north, even to the southern part of Japan; while a couple of species are found on the mainland of India and China, and one even in Madagascar. GYMNOSPERMS i) a | The dominant species in Australia is Cycas media, most abundant around Rockhampton, but found along the Burnett and Dawson rivers, at Cape Upstart, Rockingham Bay, and Mount Elliott. Itisa tall species, 2-3 meters in height, with occasional specimens reaching 6 meters (fig. 62). Cycas kennedyana is found in the Normanby Ranges near Port Denison, and C. normanbyana in the mountains about the mouth of the Burdekin River. C. cairnsiana is in the Newcastle Range. Fic. 63.—Encephalartos friderici-guilielmi: on the mountain overlooking Queens- town, South Africa. The low plant at the left is Aloe ferox. The big rock is dolerite.— From CHAMBERLAIN, The Living Cycads’ (University of Chicago Press). The four species just mentioned are endemic in Queensland. Three others have been described from western Australia. They are not very well known, and ScuusTeEr‘” lists them as varieties of Cycas media. North of Australia, in the various islands and on the mainland of China and India, are several species, some of them none too well de- fined. Among these, Cycas circinalis is the most widely distributed, and may be the most variable. It is so popular in cultivation that reported habitats need to be checked. CYCADALES 73 Widely separated from the other species is Cycas madagascariensis, in Madagascar. It looks like the variable C. circinalis. The best known of all cycads, and the most widely cultivated, is the Japanese Cycas revoluta. It is endemic in the southernmost part of Japan, and is at its best around Kagoshima. It is so well marked that it is easily recognizable. In cultivation it is the most popular of all cycads, and is so hardy that it might easily become established in such places as Southern California and around the Gulf of Mexico. Fic. 64.—Encephalartos latifrons: at Trapp’s Valley, near Grahamstown, South Afri- ca (February, 1912). The two cycads at the left were said not to have grown “any” in the past 40 years; the two at the right—Encephalartos altensteinii—were said to have grown, in that time, “about six inches.” While ScHuUSTER*® recognizes only eight species in this genus, there are doubtless more and perhaps even twice this number. The remaining genera, Encephalartos and Stangeria, are endemic in South Africa. Encephalartos.—There are at least 14 species of this dominant African genus, most of them in Cape Colony. Encephalartos alten- steinii, the most familiar species in cultivation, is abundant and is found around East London, at Kentani in the Transkei, at Trapps Valley, at Kranz Kloof and other places near Durban, and as far 74 GYMNOSPERMS north as Mozambique. £. villosus is rather abundant in the East London region, growing in the bush veldt, while £. altensteinii grows in rocky places in the open. It is also abundant at Kentani and oc- curs in Pondoland and Uganda. E. hildebrandtii, which resembles E. villosus, is farther north, extending as far as Mombasa. E. fride- rict guilelmi makes a great display on the mountains near Queens- ; ] : os Vane 4" meh Re macy @ Fic. 65.—Stangeria paradoxa: near Mtunzini, Zululand, South Africa (January, 1912).—From CHAMBERLAIN, The Living Cycads™ (University of Chicago Press). M7 poae ae. town and, farther south, on the Windvogelberg at Cathcart (fig. 63). At Junction Farm at the junction of the Zwartkei and Greatkei rivers, it is associated with E. lehmannii, and, in the brush, there are occasional specimens of E. villosus. E. caffer makes its greatest dis- play at van Staadens, near Port Elizabeth. E. horridus, a very char- acteristic species, is abundant at Uitenhage, not far from Port Eliza- beth. E. latifrons, a remarkable, slow-growing species, is not abun- dant anywhere, but scattered specimens occur at Trapps Valley, a short distance southeast from Grahamstown (fig. 64). E. barteri be- CYCADALES 75 longs to tropical Southwest Africa, where it is found along the lower part of the Niger River. E. septentrionalis, the northernmost species, is in central Africa in the Niam-Niam region. Information is rapidly accumulating for a much more thorough and accurate account of the geographical distribution of this genus than has been written. Stangeria paradoxa.—The final genus of the family, Stangeria, was long classed with the ferns, being so near like Lomaria, a common tropical genus of the Polypodiaceae, that it was not even described as a separate genus (fig. 65). When seeds were discovered, it was named Stangeria and was given the specific name because it looked like a fern but was not a member of that assemblage. It is abundant in Zululand, and may occur a little farther north, and extends to the neighborhood of Port Elizabeth. It is probably monotypic, although forms growing in the bush veldt and those in the grass veldt look different. It does not get very far from the coast. In Zululand it is associated with Encephalartos brachyphyllus, and, in the East Lon- don region, with E. villosus and E. altensteinit. CHAPTER V CYCADALES—Continued THE LIFE-HISTORY Studies of the life-histories of extinct plants are necessarily incom- plete because the preservation is never perfect, and delicate parts, like the gametophytes, the young embryos, and the meristems, have usually decayed before fossilization began. Consequently, the parts best known are the mature vascular structures and the harder parts of sporangia and seeds. In the living cycads, life-history studies are handicapped only by the difficulty in getting material. There are stages in the life-history when collections should be made almost every day; for other stages, once a week is often enough; and for more than half of the year, once a month is sufficient. Since the material is tropical or subtropical, and trained histologists in those localities are scarce, the difficulty is not in the existence of material, but in securing a well-fixed series of stages. Of course, one can fix his own material, but financial con- siderations prevent botanists from staying a year in a cycad locality. However, cycads are easy to transport, and large cones may be in fine condition three weeks after being taken from the plant. THE SPOROPHYTE—VEGETATIVE The vegetative features of the life-history will be considered under the topics, “‘stem,” “leaf,” and ‘‘root.” The stem.—The cycads have sometimes been described as plants with branched leaves and unbranched stems. All have branched leaves, but there is considerable branching in the stem. The typical habit is an unbranched stem with a crown of leaves at the top, so that the plant looks like a small palm or tree-fern (fig. 66; see also fig. 54). All of the arborescent forms are called palms by the natives. Dioon is the Palma de Dolores; Microcycas is Palma Corcha; Encephalartos is the Bread Palm; etc. Even in localities where tree- ferns are familiar objects, the natives do not call them ferns. 76 CYCADALES 77 Adult plants of Dioon, Ceratozamia, Microcycas, and Cycas are always arborescent; Bowenia and Stangeria are tuberous, with all or most of the stem subterranean; while in Zamia, Macrozamia, and Encephalartos, some species are arborescent and others tuberous. The tallest of all cycads is Macrozamia hopei. The Queensland botanist, F. M. BaILry, gave the height as ‘20 to 60 feet,” a little more than 18 meters. A similar height assigned to Cycas media, in Fic. 66.—Macrozamia moorei: at Springsure; about 200 miles west of Rockhampton, Queensland, Australia (November, 1grr). ENGLER’S Pflanzenfamilien, was a mistake, since Cycas media does not reach more than one-third of that height. Someone probably saw the tall Macrozamia, thought it was Cycas, and so started the mis- take. Dioon spinulosum grows in the dense rainy forest and is 6-10 meters high, with occasional specimens reaching a height of 15 me- ters. Microcycas calocoma sometimes reaches a height of nearly 1o meters. None of the other genera are nearly so tall. Any arborescent cycad more than 2 meters in height may be regarded as a tall speci- men. 78 GYMNOSPERMS The trunk in all of the arborescent forms is covered by an armor of leaf bases (fig. 67). The leaf does not fall off as in most deciduous plants, but loses its leaflets, bends down, and decays to a point a few centimeters from the cortex, when an abscission layer appears and cuts the rachis off cleanly, leaving a few centimeters of it to form the armor. Beneath the original abscission layer, embryonic layers ap- .— WV. Fic. 67.—Dioon edule: portion of trunk of an old plant, showing armor of leaf bases. The trunk is smaller below than above. It also shows three zones, marking prolonged dormant periods.—From CHAMBERLAIN, The Living Cycads’® (University of Chicago Press). pear in succession and cut off thin membranous sheets so that finally the trunk may have a smaller diameter near the base than it has at a short distance below the crown. In most arborescent forms the scaling-off of these thin laminae does not progress far enough to obscure the original leaf bases and, consequently, the number of leaves which a plant has borne can be counted even on plants more than a thousand years old. The age in all such forms can be determined with considerable accuracy, if the CYCADALES 79 duration of the crown is known. Unfortunately, this period is seldom known. People in cycad localities know that new crowns appear every year, but have not noticed how often any individual plant forms acrown. Records of conservatory plants are worthless because leaves last much longer than in the field. The duration of the crown and the average number of leaves in a crown, under natural condi- tions, furnish a basis for an approximate estimate of the age. In Dioon edule a new crown is formed every other year, and the average number of leaves in the crown of an adult plant is about 20, so that the average is 10 leaves a year. If the number of leaf bases is 10,000, the plant is about 1,000 years old. The number of leaves produced while the plant is young is much smaller and as it reaches the coning age it may produce a cone and a crown of leaves at the same time, thus reducing its vitality so that it may go into a dormant condition, producing neither cones nor leaves for several years. It is evident that estimates made in this way are likely to be conserva- tive. Plants of Dioon edule less than 2 meters in height, like the speci- men shown in fig. 54, may be 1,000 years old. Plants in protected ravines may be much older. Dioon spinulosum, more than 10 meters in height, may not be more than 200 years old. Three plants of Encephalartos altensteinit and two of E. latifrons, in front of a residence at Trapps Valley, South Africa, under prac- tically field conditions, had been under observation for 46 years in 1912. The owner said that the £. altensteinit might have grown 6 inches, but that the E. latifrons did not seem to have grown at all, although all of the plants had had green leaves all the time and had occasionally produced cones. There seems to be no way of estimating the age of those tuberous forms which have no persistent armor of leaf bases. Naturally, the tuberous forms are comparatively small, but it is possible that they may reach a great age. The stem of Bowenza serrulata is more or less spherical, and is often 20 cm. in diameter, with occasional specimens twice that size. Bowenia spectabilis does not reach so great a diam- eter, but is often 30 cm. in length (figs. 68, 69). Stangeria and the tuberous species of Zamia have the same general shape of stem as Bowenia spectabilis. 80 GYMNOSPERMS The smallest of all cycads is Zamia pygmaea, with adult stems 1 or 2 cm. in diameter, and rarely reaching 3 cm. In transverse section the stem shows a large pith and large cortex with a scanty zone of wood between (fig. 70). The vascular cylinder is an endarch siphonostele in the adult plant, but the seedling shows ( 69 cone. a, apogeotropic root. The elongated, fusiform \ stem contrasts sharply with the short, broad stem of af! A \ B. serrulata.—After CHAMBERLAIN."™S Fic. 69.—Bowenia serrulata: short, broad stem with 68 numerous branches at the top, some bearing male cones.—After CHAMBERLAIN."3 (| } Fic. 68.—Bowenia spectabilis: plant with a female a distinct mesarch condition, and there is some centripetal wood in cone axes and in the stalks of sporophylls. The amount of xylem in most stems is surprisingly small. A ma- ture plant of Zamia floridana, with a stem 15 cm. in height and 6 cm. in diameter, had a zone of xylem 2 mm. wide. The zone of phloem had the same diameter. A plant of Ceratozamia mexicana, 30 cm. high and 15 cm. in diameter, had a zone of xylem 3 mm. wide and phloem 2 mm. wide. Dioon edule, 60 cm. tall and 21 cm. in diameter, had zones of xylem and phloem each 5 mm. in width. CYCADALES 81 Dioon spinulosum is exceptional in the amount of wood. A stem 6 meters in height and 33 cm. in diameter had a zone of wood 10 cm. in width, with phloem 1.4 cm. in width (fig. 71). What the amount of wood might be in plants with twice that height and diameter has not been determined. The strong medullary rays are a conspicuous feature of the transverse section. Fic. 70.—Zamia floridana: transverse section of stem, showing large pith and cortex and scanty zone of wood. Parts of leaf traces are also shown. A strange feature in the anatomy of most cycad stems is the ab- sence of growth-rings. Stems of Zamia, Stangeria, and Ceratozamia, which may be more than 50 years old, show no trace of growth-rings. On the other hand, in Dioon spinulosum and in D. edule, growth- rings are so well marked that they can be seen and counted without a lens (fig. 72). The rings are formed by a persistent cambium and they look like the familiar annual rings of dicotyls, but they are not annual or even seasonal. In D. spinulosum they are formed every other year, the stimulus which produces a new crown of leaves mak- ing a growth-ring in the stem. In D. edule even the stimulus which 82 GYMNOSPERMS results in the formation of a new crown of leaves or a cone is not sufficient to make a new ring, but when the plant has gone into a prolonged resting period of several years, the stimulus which brings Fic. 71.—Dioon spinulosum: transverse section of trunk, showing large pith and cor- tex and prominent medullary rays. The growth-rings show faintly. The zone of wood is the broadest ever described in a cycad—From CHAMBERLAIN, The Living Cycads”° (University of Chicago Press). Fic. 72.—Dioon spmmulosum: the growth-rings are quite clear at the right.—After CHAMBERLAIN. it out of its resting period produces a new ring, so that the rings may appear at intervals of 10, 20, or even more years. The rings are due to the alternation of larger and smaller cells, as in dicotyls, but the CYCADALES 83 difference in size, under the microscope, is so slight, that it might almost escape notice (fig. 73). Stems with a vascular cylinder developed from a persistent cam- bium of the familiar dicotyl type are called ‘“‘monoxylic.”” When there is more than one zone, each with xylem and phloem, the stem is polyxylic (fig. 74). Cycas, some species of Macro- zamia, and some of Encephalartos are of this type. The zones are formed at ir- regular intervals of, probably, many years, and may mark the number of times the plant has had prolonged dor- mant periods. Where the zones succeed- ing the first stelar zone originate is not settled definitely. JEFFREY? claims they arise in the pericycle, while Mit- LER*™ claims that pericycle is indistin- guishable from the endodermis and the rest of the cortex. SistER HELEN ANGELA’ found cambiums in the cor- tex of Ceratozamia, thus proving that embryonic tissue can arise in the cortex. In some angiosperms, like the beet and some other members of the Chenopo- diaceae, similar zones arise from a well- marked pericycle. In Boerhavia (Nycta- ginaceae) the zones are particularly well marked. The problem is hard to settle because it would kill a valuable Hos gs Aas : histology of the growth-rings; plant to get material. the slow growth, corresponding A stem of Cycas media, 3 meters in to the summer wood in an or- height, showed at the base three zones ‘inary dicotyl, is at g.—After - CHAMBERLAIN.!9 of xylem and phloem. A piece of stem of Cycas pectinata Griff., 20 cm. in diameter, had 14 zones, doubtless a very high number. The protoxylem, in seedlings, has spiral markings, but in older plants, where elongation is extremely slow, even the protoxylem 84 GYMNOSPERMS consists of scalariform tracheids. The secondary xylem consists of tracheids with bordered pits, except in Zamia and Stangeria, which still retain the scalariform tracheid of their remote fern ancestry. In cone axes and in sporophylls the scalariform tracheid is also retained, but probably not without exception. The pitting is usually multiseriate, sometimes with as many as four or five rows of pits. SIFTON,‘’ in 1920, investigated pitting in the cycads and discussed the literature. He believes the bordered pit is derived from the scalariform and cites transitions in Dioon spinu- losum as evidence. The pitting at the ends of tracheids is more primi- fyi red gare seater “as ee ay Bi! # Pini Ke wg st eas Tic. 74.—Cycas media: transverse section of stem, showing three zones of wood tive than that throughout the rest of its length. He also found ter- tiary thickenings on the walls of tracheids, resembling those which are so characteristic of taxads. This subject will be considered more fully when dealing with the Coniferophytes. SIFTON claims that bars of Sanio are present in both primary and secondary wood, while HALr”° claims that they do not occur in cycads. Differences in interpretation, rather than in observation, seem to be responsible for the discrepancy. The whole subject of markings on cell walls needs investigation, and a thorough cytological study may yield more reliable results than the usual examination of mature structures. Dr. GRACE BARK- LEY’s® work on Trichosanthes anguina showed how the spiral thick- ening arises from the alternation of dense protoplasm with small vacuoles and less dense protoplasm with large vacuoles. That the simple pit was in some way connected with vacuoles in the proto- _— = CYCADALES 85 plasm was already known. Intermingled with the tracheids are some- times rows of thin-walled cells, as in Dioon spinulosum (fig. 75). The rays in Dioon spinulosum are usually 1 cell wide, but often 2 cells, and occasionally 3 cells, wide. Longitudinally, they vary from t cell to 20 cells or more (fig. 76). In Cycas media some of the rays are even 6 or 7 cells in width, and are correspondingly long. The cells of the rays contain a large amount of starch but, instead of the YOUYUYU UY OP O® Ove Fic. 75.—Dioon spinulosum: longitudinal section of mature wood, showing multi- seriate bordered pits, and also one of the thin-walled cells (¢); X 390.—After CHAMBER- LAIN.1°9 starch, a cell may include a large crystal of calcium oxalate. The thin-walled cells of the xylem, which also contain starch, are usually in contact with the rays (fig. 76). The cambium, with a little of the xylem and phloem, is shown in fig. 77. The bast tracheids have the same width as the other cells of the phloem. While the phloem has not been studied critically in enough forms to warrant a generalization, it may be that such a com- parative study of Cycadofilicales, Bennettitales, and Cycadales would be worth while. 86 GYMNOSPERMS The large ray, or leaf gap, is immensely larger, in Dioon spinulo- sum 20 or more cells in width and 50 or more in height. The gap con- tains the leaf trace and, above it, a mucilage duct (fig. 78). The re- D i er rea OO OY & 25° oe a==8 ie is i °° (<>) RE DO © ° ° qo © Cos = Om 0 O Fic. 76.—Dioon spinulosum: longitudinal Fic. 77.—Dioon spinulosum: trans- tangential section of mature wood, showing verse section of stem, showing cam- thin-walled cells of the xylem (¢) containing _bium (c); thin-walled cells (¢), contain- starch (s), the xylem tracheids, and the small ing starch (s); calcium oxalate crystals medullary rays containing starch and calcium (x); phloem with some thick-walled oxalate crystals; X125.—After CHamBer- _ tracheids (d); and medullary rays (r); LAIN." 125.—After CHAMBERLAIN." ticulate arrangement of the tracheids, characteristic of the cycads, is well shown in the figure. The leaf trace, as it appears near the base of the leaf gap, is small, but in the cortex the trace becomes very conspicuous. The girdling of leaf traces in the cycads has long attracted attention, especially ane, = {<-> SSS \{@ 77 TKO LD ) ER LA Si EO tas aa a Y 2 a BVO ow CO OE art ea TS peal Nab eae se 00 S022 D=_—! cl o> = OD Be — gg SS Fic. 78.—Dioon spinulosum: longitudinal section of mature wood, showing a leaf gap with a leaf trace connected with the base of the gap by tracheids; above the leaf trace is a mucilage canal; there are many small medullary rays, mostly one cell wide. The dark spots are calcium oxalate crystals. m, mucilage duct ; p and x, phloem and xylem of leaf trace; s, scalariform tracheid.—After Dr. La Dema M. LANGDoN.343 88 GYMNOSPERMS since the Bennettitales have a leaf trace passing straight from the stele to the leaf. The most critical studies of the leaf trace have been made by Turrssen,*4 Dorety,’” and LANcpon,* (figs. 78, 79, 80, and 81). The leaf traces pass through the cortex, not horizontally, but rising a little, so that in a thin transverse section the girdling feature might be overlooked. The leaf trace often passes half-way around the stem, so that it enters the leaf almost opposite the start- ing place. The traces from the leaf gaps keep joining the girdling trace as it passes through the cortex, making the trace larger and \ —> —= == -\ Tix r\\} mut ~ff-| 7A ear a 1 —-| ame | aT =A ae 2 S ZL. fin Fic. 79.—Dioon spinulosum: reconstruction of vascular system of stem of a seedling, showing the girdling of the leaf traces. The two groups of five bundles each (parallel at the top), are passing out to two leaves.—After SIsteR HELEN ANGELA Dorety.'? larger as it nears the leaf. The trace, as it appears in a thick trans- verse section, is shown in fig. 79. The traces consist almost entirely of scalariform tracheids. This feature and their union with the xylem of the stem is shown in fig. 81. Such a girdling of the leaf trace is not confined to cycads, being found in many angiosperms which have “radial”’ leaves. A curious feature of the stem, when one has an opportunity to see a longitudinal section of the entire trunk of a genus which has termi- nal cones, is the cone dome. The first cone borne by such a plant is actually terminal, but all succeeding cones, although apparently ter- minal, are really lateral. The meristem is entirely used in the forma- tion of the cone, and a new meristem appears at the base of the pe- duncle, and from the new meristem crowns of leaves are formed CYCADALES 89 until another cone is produced, when a new meristem appears at the base of the peduncle of the second cone, and the process is repeated every time a cone is produced (fig. 82). A median longitudinal section of the upper part of a large trunk of Dioon spinulosum shows very clearly these domes, each one of which was, in its turn, the apex of the plant (fig. 83). This structure, as presented dia- grammatically by Dr. F. Grace SmituH,5% bears a striking resemblance to Williamsonia (Anomozamites) angustifolia (fig. 84). Naturally, there are no cone domes in the female plant of Cycas, and in those species of Macrozamia and Encephalartos which have axillary cones. There should be cone domes in the male plant of Cycas. What the con- dition may be in Encephalartos and Macro- zamia, when they bear a single apparently terminal cone, remains to be seen. The leaf —The beautiful crown of graceful leaves makes the cycad look like a palm of the Phoenix type, for the leaves are pinnate in all Fic. 80.—Dioon spinu- losum: semidiagrammatic view of vascular system of top of stem of a seed- ling, showing the girdling except Bowenza, in which they are bipinnate. In most cycads the leaves come in crowns (fig. 85). The leaves are formed in spiral suc- cession, but it is only when the leaves are very young that there is any noticeable differ- ence in size. In a young crown of a dozen leaves, while the oldest leaf is 30 cm. long, of leaf traces of the first leaf, and also how the leaf trace is built up by traces from the leaf gaps: 2, and 73, ventral strands apparently united with the two dorsal strands.— After Dr. LA Dema M. LANGDON.343 the youngest may measure only 4 or 5 cm. The length of the leaf varies from about 3 meters in Cycas circinalis down to 5 or 6 cm. in Zamia pygmaea. Dioon spinulosum has beau- tiful leaves, often 2 meters in length, and in many cycads the leaves are a meter long. The number of leaflets on each side of the rachis varies from more than a hundred in species with large, long leaves down to 3 or 4 in the smallest leaves. In seedlings, even of those species which have a hundred leaflets when mature, there may be only one or two pairs of leaflets. The number then increases gradual- ly with the age and size of the plant. —= —- ee eee —S> a Pen . a a ee Tt, Ee fae = ads 08 IM Pee atten PTT o* 9°58! in Ge a SEER eet ect B Antec ee inl tS 3 L \_L J eel EERE ars cS } = Gn G Assis, aiad LaDema M. Langdon, del Fic. 81.—Dioon spinulosum: radial longitudinal section, showing union of leaf traces with the vascular cylinder. Most of the tracheids of the leaf trace are scalariform; those of the cylinder are pitted; * 55.—After Dr. La Dema M. Lancpon.58 ey a + ot EP a Ted ie ie frre Pee. N . yan 6b be ee Fic. 82.—Dioon edule: photograph of surface view of apex of stem, showing a cone dome with its bundles going to a cone, and, to the left of it, a similar series of bundles going to the new apex, which is producing leaves; below is a cone dome with some of the bundles cut across.—After CHAMBERLAIN.™ Med _CYCADALES gr Leaflets vary greatly in size. In Ceratozamia longifolia they reach 4o cm. in length and 2.5 cm. in width; in Dioon spinulosum, 8.5—20 cm. in length and 10-17 mm. in width; in Dioon edule, 11-15 cm. in Fic. 83.—Dioon spinulosum: photograph of a surface view of the top of a large plant cut longitudinally through the middle. The second cone dome from the top shows the peduncle of the cone, and part of the peduncle can be seen in the cone dome just below it, and in the lowest cone dome; one-half natural size—After CHAMBERLAIN."™! length and 4-8.5 mm. in width. The widest leaflets are those of Zamia skinneri, 19-29 cm. long and 3-10 cm. wide. The narrowest is in the Cuban Zamia angustissima, with leaflets about 7 cm. long and 1 mm. wide. 92 GYMNOSPERMS The vernation, in Cycas, is circinate both in the rachis and the leaflets (fig. 86). In Zamia, Ceratozamia, Bowenia, and Stangeria the rachis is somewhat circinate (subcircinate) but sometimes looks al- most reflexed. In Dioon, Macrozamia, and Encephelartos both rachis and leaflets are perfectly straight (fig. 87). In the family, as a whole, there is more of the erect vernation than of the circinate, the pre- Yj Vic. 84.—Zamia floridana: diagram of apex of stem, showing relation of cones, crown, and axis, with the axes much lengthened.—After Dr. F. GRAcE Smiru.s‘s vailing type in ferns. It is interesting to note that both Cycadeoidea and Cycadella have the erect type of vernation. The leaves of the various genera are so characteristic that they can be identified by this feature. StstER Mary Atice LAMB3** con- structed a key based entirely upon the leaf. Only two genera, Cycas and Stangeria, have a midrib in the leaflet. In Cycas, there is only a midrib without any side veins; while in Stangeria there are side veins from the midrib. Bowenia has a bipinnate leaf, marking it off easily from all the rest. In Macrozamia there is an obvious gland at the base of the leaflet in most species, and a histological examination might show it in the rest. In Dioon the bases of the leaflets are as wide or even wider than the middle of the leaflet. In Microcycas, the Fic. 85.—Dioon spinulosuwm: new crown of leaves nearly erect; the previous crown nearly horizontal. In the greenhouse of the University of Chicago. . SE t sj 2G all hsihae4 5 -“ Yes i roe ~S “oy te ~ = ; - Pe. =e MAA at AO, ~<* - ‘ - young crown, showing circinate vernation of both rachis greenhouse of the University of Chicago. —Cycas revoluta: 86 and leaflets. In the Fic. CYCADALES 95 leaflets are reflexed on the rachis. The other three are not so easy. In Ceratozamia the leaflets are always entire (integerrima), a char- acter very rare in Zamia and found only in a part of the species of Encephalartos. By adding the histological characters of the leaf, any indefiniteness in the identification could be removed. The leaflets of young and adult plants are usually quite different, so that some botanists regard the juvenile form as evidence in favor Fic. 87.—Dioon edule: young crown, showing perfectly erect vernation. In the greenhouse of the University of Chicago. of the theory of recapitulation, which means that ontogeny recapitu- lates phylogeny, or the history of the individual recapitulates the history of the race. Dioon spinulosum, in the seedling and in young plants, has a leaf with a long naked petiole and lowest pair of leaflets nearly as large as the rest; while in older and adult plants the leaflets are more and more reduced until, at the base of the leaf, they are scarcely more than spines. In Dioon edule leaflets of seedlings and young plants have a spiny 96 GYMNOSPERMS margin, especially near the tip, while in the adult plant the margins are entire (fig. 88). If the theory of recapitulation holds, it would mean that Dioon spinulosum, with its spinulose leaflets, is the ancestral form and that D. edule shows the spinulose leaflets, in its younger stages, on ac- count of its spinulose ancestry. C Fic, 88.—Margins of leaflets: A, part of leaf of seedling of Dioon edule, showing spiny leaflets; B, adult leaf with entire margins; C, adult leaf of Dioon spinulosum with spiny leaflets—From CHAMBERLAIN, The Living Cycads° (University of Chicago Press). In most, and perhaps all, of the cycads, the leaves of young plants differ from those of the adult, in some the change taking place after the plant is 50 or more years old. Taxonomists have been trapped into identifying two species from leaves taken from a single plant of Encephalartos altensteinii, just as it passes from the spiny to the en- tire leaflet condition. Some of the leaves, at that time, will have spiny leaflets, while in others the leaflets will be perfectly entire, as CYCADALES 97 in E. caffer. A bud coming from a wound in an old E. altensteinii plant shows typically spiny leaves. The leaves of all the cycads have a strongly xerophytic structure (fig. 89). The cells of the epidermis have thick walls and are heavily cutinized. The stomata are sunken and are mostly confined to the under surface, except in Bowenia and Macrozamia. In Microcycas there is not much thickening of the hypodermal cells, except above and below the bundles and near the margins of the leaflets; in Macro- 5 7 ore NSS E , \ 2 . irs ~ WAS \ EP) Sis if = Se / Was NEB Fic. 89.—Dioon edule: transverse section of part of leaflet. c, cuticle; e, epidermis; h, hypodermis; , palisade; s, suberized cell—From CHAMBERLAIN, The Living Cy- cads!° (University of Chicago Press). zamia moorei the hypodermal cells are thick walled throughout; and in Encephalartos altensteinii the thick-walled hypodermal region is several cells thick. There is usually a well-marked palisade layer and a more or less spongy parenchyma beneath. The bundles are usually surrounded by thick-walled cells, and such cells often extend from the bundle to the hypodermal cells above and below. Many of the thick-walled tra- cheids around the bundles, and nearly all thick-walled tracheids scat- tered through the thin-walled cells between bundles, are bast fibers. 98 GYMNOSPERMS Altogether, the structure is such that the leaves are strong and leathery. Many of them keep their fresh, green color a long time Fic. 90.—Dioon edule: seedling; all the part bearing secondary roots is the primary root. The stem, bearing the leaf and scale leaves, is so small at this stage that it is hidden by the emergent part of the cotyledons. — From CHAMBERLAIN, Elements of Plant Sciences (Mc- Graw-Hill Book Co.). after being cut off from the plant, so that they are popular for decorative purposes in cycad regions, and the beautiful leaves of Cycas revoluta have become a standard nursery stock for funeral wreaths and for Palm Sunday. The root.—The primary root may be very large, even as large as the stem; and in the seedling the root is much larger than the stem which, at this stage, is rather inconspicuous (fig. go). After the seedling has become thor- oughly established, the stem begins to grow more rapidly, and in the adult plant is usually much larger than the root. Roots may attain a great length. A root of Dioon spinulosum, at a distance of 12 meters from the stem, hanging exposed over a rock, was still 3:cm. in diameter when it entered a crevice and could not be followed any farther. The root is tetrarch. Secondary growth be- gins early and more or less irregularly, so that the topography, as seen in transverse sections, differs markedly from that of familiar dicotyl roots. All cycads have remarkable apogeotropic roots (fig. 91). These grow up, instead of down, branch dichotomously and profusely, forming coralloid masses above ground. The vascular structure is about the same as in normal roots, but bacteria, or “bacterioids,” get in very near the tip and cause some distor- tion, which seems to prepare the way for the entrance of a blue-green alga, Anabaena. The alga multiplies rapidly, so that there is a blue-green zone midway between the vascular cylinder and the epidermis (fig. 92). While the CYCADALES 99 zone is usually only one cell wide, the cells are so enlarged radially that the zone is easily visible to the naked eye. These roots are al- most universal in seedlings and are much more prevalent in the green- house than in the field. THE SPOROPHYTE—REPRODUCTIVE No plants are more absolutely dioecious than the cycads (fig. 93). SCHUSTER reports one case in Cycas revoluta where a plant was cut into two longitudinal pieces, which were taken to different places. It is claimed that one piece produced a female strobilus and the other, a male. Ona lawn in Australia there were several plants of Cycas revoluta. It was re- ported to me that one of these pro- duced a female strobilus and, a few years later, a male strobilus. It is also claimed that a bud from a fe- male plant of Cycas circinalis, in the Garfield Park Conservatory at Chi- cago, reached the coning stage and produced a male cone. In 30 years of study in the field and in green- houses I have never seen anything to indicate that the cycads are not absolutely dioecious. It will be remembered that in the Bennettitales the strobili are pre- vailingly bisporangiate. The reduc- tion from the bisporangiate condi- tion to the dioecious is a general tendency in plants. Fic. 91.—Cycas revoluta: coralloid masses of root tubercles on erect (apogeotropic) roots.—From CHAM- BERLAIN, The Living Cycads'° (Uni- versity of Chicago Press). The female strobilus.*—The largest cones that have ever existed * The term “strobilus” is used to include both the crowns of loose sporophylls, like the female sporophylls of Cycas and the male sporophylls of Cycadeoidea, and compact 100 GYMNOSPERMS are found in the living cycads (fig. 94). Cones of Macrozamia deni- sonii are often 70 cm. in length, with a weight of 30 kilos; and they sometimes reach a length of nearly a meter with a weight of 38 kilos. A cone of Encephalartos caffer in the park at Port Elizabeth, South Africa, weighed 42 kilos, and cones of this species, even when two or three are borne at the same time, reach a weight of 20 kilos. In Dioon spinulosum the cone reaches a length of 50 cm. and a weight of 15 kilos. The cones of Dioon edule are smaller, about 30 cm. in length and weighing five or six kilos. Mucrocycas has a large slender cone occasionally reach- ing 94 cm. in length and a weight of 9.5 kilos, but most of its cones are not nearly so long or heavy. Cones of other cycads are small- er. In Ceratozamia the average cone is about 26 cm. in length, and in the remaining genera con- siderably shorter. Zamia pyg- maea has the smallest cones, Fic. 92.—Cycas revoluta: transverse about 2 cm. long and I.5 cm. in section of root tubercle, showing promi- djameter. nent algal zone; X 20.—After Life.3ss The evolution of the compact cone from a loose crown of sporophylls is shown very clearly in the living cycads. In Cycas revoluta the female strobilus consists of a crown of sporo- phylls arising spirally in acropetal succession and as loosely arranged as the male sporophylls of Cycadeoidea (fig. 95). The upper part of the sporophyll has numerous leaflets, one or two of which are oc- casionally replaced by small ovules. The ovules are not transformed leaflets, but the leaflet is probably very much shortened, and bears a terminal ovule. Below the leafy portion there are usually three pairs of ovules, sometimes two pairs, and occasionally four pairs. Both sporophylls and ovules are covered with yellowish hairs, but as the cones. As long as the terms ‘“‘male”’ and “female” are applied to the 2x generation in animals, there should be no objection to applying the same terms to the corresponding generation in plants. CYCADALES IOI ovules get large, some of the hairs are lost, and the ripe seeds have a soft orange-red color. In the various species of Cycas there can be traced a gradual re- duction in the size of the sporophyll, a reduction of the leaflets until the sporophyll has merely a serrate margin, and a reduction in the number of ovules to a single pair, the number characteristic of all the other genera (figs. 96 and 97). Fic. 93.—Dioon edule: two male plants at the left and a female plant at the right. Chavarrillo, Mexico (September, 1906). The taller plants, to the top of the leaves, are about 7 feet high——From CHAMBERLAIN, Elements of Plant Sciences’ (McGraw-Hill Book Co.). In Dioon edule the sporophylls have lost even the serration, but are broad and loosely compacted into a cone (fig. 98). The final stage, with much reduced sporophylls and very compact cone, is well illustrated by Zamia (fig. 99). The various genera show various reductions from the leafy char- acter to a peltate sporophyll bearing scarcely any resemblance to a leaf (fig. 100). In Macrozamia the rachis of the sporophyll remains as a more or less prolonged, tapering spine, especially in the upper part of the cone. In Ceratozamia and Encephalartos the terminal 102 GYMNOSPERMS part of the rachis is suppressed, but there are often serrations repre- senting the pinnae. In the rest, reduction has gone still farther, so that the sporophyll is merely a thick peltate structure bearing scarce- ly any resemblance to a leaf. -% Fic. 94.—Macrozamia denisoni: female cone nearly a meter (37 inches) long, and weighing 38.5 kilos (85 pounds).—From a photo taken on Tambourine Mountain, near Brisbane, Australia, by Miss Hrtpa GrissMANN. It will be remembered that in the Bennettitales the strobili are prevailingly bisporangiate. The reduction from the bisporangiate condition to the dioecious is a general tendency in plants. The re- CYCADALES 103 duction of the sporophyll from the leafy condition to the peltate structure, which reaches its extreme in Zamia, can be traced in great detail through the various genera and species of the living cycads. There is no doubt that Cycas revoluta shows the most primitive sporophyll condition in the family, producing a crown of sporophylls just as it, and the other cycads, produce a crown of foliage leaves, still leaving in the center a meristem to produce more leaves or Fic. 95.—Cycas revoluta: strobilus consisting of a loose crown of sporophylls still retaining many pinnae in the upper portion. sporophylls. Thus, the original meristem continues from the em- bryo to the death of the plant. Proliferation is seen, occasionally, at the top of a male cone, and sometimes in other genera, where vegeta- tive leaves, greatly reduced, but sometimes bearing leaflets, appear instead of sporophylls. The megasporangium.—The megasporangia, or ovules* as they are usually called, are all erect and have a single massive integument. In Cycas circinalis and Macrozamia denisonii they reach a length of * The term “ovule” (little egg) was mistakenly devised to apply to the entire mega- sporangium. It is short, convenient, and in general use but, like the term “cell,”’ has nothing else to recommend it. 104 GYMNOSPERMS 6 cm., about the size of the largest seeds of Trigonocarpus. Most cy- cads have seeds from 3 to 5 cm. in length. Zamia kickxii has very small seeds, only a centimeter long, while Zamia pygmaea, with seeds from 5 to 7 mm. in length, has the smallest which have been meas- ured. The principal features of the ovule are well illustrated by a longi- tudinal section (fig. ror). Only the upper part of the nucellus is free "4 7 " Fic. 96.—Cycas circinalis: crown of sporophylls not yet expanded. The apex of the sporophyll has become merely serrate, there are no separate pinnae, as in Cycas revoluta. from the integument. After the ovule reaches its full size, as in the illustration, the stony layer, with a fleshy layer on each side of it, is very conspicuous. Later, as the female gametophyte grows, it ab- sorbs most of the inner fleshy layer, so that it disappears, except as a dry papery membrane closely applied to the megaspore membrane. Usually two strong vascular strands enter the ovule. The outer Fic. 97.—Cycas circinalis: a later stage, after sporophylls have expanded.—Garfield Park Conservatory, Chicago. 106 GYMNOSPERMS strands branch immediately, before they reach the level of the stony layer, and, from that point, extend almost to the micropyle without any further branching. The number of these outer bundles is quite Fic. 98.—Dioon edule: sporophylls loosely compacted into a cone, 30 cm. in length.— After CHAMBERLAIN," constant for any species and usually does not exceed a dozen. The inner strands, after reaching the inner fleshy layer, fork repeatedly so that they are much more numerous than the outer bundles (fig. CYCADALES 107 102). Usually they end before they reach the nucellus, but sometimes extend beyond the free portion of the nucellus into the inner fleshy layer of the integument. The outer fleshy layer remains fleshy and, in the ripe seed, be- comes variously colored, bright red in Encephalartos altensteinii, Fic. 99.—Zamia floridana: a very compact cone, with megasporophylls so regularly arranged that they appear to be in vertical rows. pale yellowish in Encephalartos horridus, orange-red in Zamia flori- dana, blood red in Zamia latifoliolata, salmon pink in Microcycas, and nearly white in Dioon and Ceratozamia. The color seems to be constant for any given species. Fic. 100.—Diagram showing the reduction of the megasporophyll and the evolution of the cone. A, Cycas revoluta, with leaflike megasporophyll; /’, megasporophylls in a loose crown, not compacted into a cone; B, Cycas media, leaflets of the megasporophyll reduced to serrations; G, the grouping of megasporophylls in early stages makes the crown of sporophylls look like a cone; C, Dioon edule, the leaflet character is lost, and G, the sporophylls, are compacted into a loose cone; D, Macrozamia miquelii, the midrib of the leaf represented by a spine, and J, the sporophyll compacted into a tight cone; E, Zamia floridana, sporophylls with hardly any resemblance to a leaf, and J, compacted into a very tight cone.—From CHAMBERLAIN, Elements of Plant Science*s (McGraw- Hill Book Co.). CYCADALES 109 The upper part of the ovule, at the time of fertilization, is shown in detail in a later figure (fig. 147). This figure shows the nucellus with pollen tubes which have digested their way completely through. The sharp beak is characteristic. At this stage the inner fleshy layer has become a thin, dry membrane, sticking to the inner border of the stony layer. There are many mucilage ducts in the outer fleshy layer and many cells (shaded in the drawing) filled with tannin. The younger stages in the development of the ovule have not been studied very thoroughly on account of the difficulty in getting material. In most cycads these stages occur while the young cone is still covered by scale leaves and there is uncertainty whether a cone or a crown of leaves is developing. Naturally, in conservatories, it would be difficult or impossible to get permission to cut out the top of a rare plant. There are few, if any, trained histologists in cycad regions, and when a histologist reaches such a place, on a hasty trip, the young stages might not be available. The most thor- : ough study was made by Dr. F. Grace Bee ye ep adiealer SMITH,° who, after sending repeatedly gitudinal section of ovule, and getting little except to learn the ap- shortly after pollination; ™, proximate time for various stages, went BUCTOpY’ mae mucedae) erene oe i sperm; s, stony layer; p, basal and spent a month in the Zamia floridana papilla; i, inner vascular bun- region, and fixed material. Previously, dles; 0, outer vascular bundles; Lano#" had figured a row of three cellsin % @bscission layer; x 2.—After : : CHAMBERLAIN. Stangeria, the lowest of which was cer- tainly a megaspore, and TrREuB™ had found a similar stage in Cera- tozamia. Dr. SmitTH found, quite regularly, a row of four megaspores in Zamia, three of which aborted, while the fourth germinated and formed the functioning female gametophyte (figs. 103, 104, 105). The formation of the megaspore brings to a close the 2x genera- tion, which in all plants, from the Bryophytes up, and also in many Thallophytes, is also the sporophyte generation. The reduction of 18 fe) GYMNOSPERMS chromosomes, which, of course, takes place during the formation of megaspores from the megaspore mother cell, has not yet been de- scribed, but counts in microsporogenesis and in other phases of the life-history have constantly shown that the x and 2” numbers are 12 and 24. As the megaspore germinates, the cells next to the developing fe- male gametophyte become differentiated into a layer of “‘spongy tissue,” looking like sporogenous tissue on account of the dense cell Fic. 102.—Dioon edule: A, transverse section of ovule near the middle; B, inner vascular system, treated with eosin and photographed after the female gametophyte and part of the inner fleshy layer had been removed; C, ovule photographed from above; o, outer bundles (the eosin has diffused some in A, and considerably in C); m, micro- pyle; s, stony layer; 7, inner vascular system; p, basal papilla; e, female gametophyte; n, inner fleshy layer; X 2.—After CHAMBERLAIN," contents. This layer nourishes the gametophyte in early stages, then weakens and finally becomes almost indistinguishable. Such a layer, very highly developed in cycads, is prevalent in gymnosperms. The male strobilus (fig. 106).—The male, or microsporangiate, strobilus is not so large as the female and there are no leafy sporo- phylls like the megasporophylls of Cycas revoluta. All of the strobili are compact cones, even in Cycas. Occasionally there is a slight pro- liferation of the axis, producing a few much reduced leaves, but the cone ripens, dies, and any further development comes from a new meristem. This applies to stems bearing a single terminal cone, as in Dioon and many others. When cones are axillary, as in Macrozamia i re CYCADALES III mooret, the original apical meristem persists from the embryo to the death of the plant. While the cones do not reach the great length and weight of some of the female cones, they are nevertheless the largest living male Fics. 103-105.—Megaspores of cycads. Fig. 103, Stangeria paradoxa; the lower cell of the row of three is the functioning megaspore; X 250.—After LANG." Fig. 104, Cera- tozamia mexicana, similar stage; X 206—After TREuB.* Fig. 105, Zamia floridana, row of four megaspores; X930.—After Dr. F. Grace Smiru.5% cones and are probably larger than those of their geological an- cestors. 112 GYMNOSPERMS The largest cone described is that of Macrozamia denisonii, which in extreme cases reaches a length of 80 cm. and a diameter 20 cm. In Encephalartos altensteinii the longest reported was 60 cm., with a diameter of 12 cm. Other measurements are, Cy- cas circinalis, 45 cm.; Cycas revoluta, 40; Dioon s pinulosum, 40; Dioon edule, 30; Dioon pur pusti, 20; Zamia floridana, 10; Bowenta serrulata, 5; and Zamia pygmaea, 2 cm. These are the maximum measurements. The average cones are not much more than half as long. Measurements of the same cone, taken 48 hours apart, might be very different; for just before shedding the pollen, the cone elongates im- mensely and rapidly, so that the sporangia are freely ex- posed. The microsporophylls are spirally arranged in acropetal succession, but the arrange- ment is so absolutely regular that, in surface view, they often look as if they were in vertical rows, like the grains of corn on a cob (fig. 107, and Fic. 106.—Dioon edule: male cone, photo- see also fig. 106). graphed at Chavarrillo, a short distance east While the arrangement is of Jalapa, Mexico. September, 1906. One- third natural size. After CHAMBERLAIN.™® spiral and acropetal, so that the sporophylls at the top are the last to be formed, they are the first to ripen their pollen, probably because the ripening depends largely upon drying, and CYCADALES 113 these sporophylls are farthest from the water supply. A large cone may shed its pollen from the upper sporophylls several 3 es =e oe nes * na ig Fic. 107.—Zamia portoricensis: male cones of various ages on one plant. The micro- sporophylls, although in strictly spiral arrangement, look as if they were in vertical rows. The University of Chicago greenhouse (January, 1933). days before the lower ones dehisce (fig. 108). Like the cones, the microsporophylls vary greatly in size; in Cycas circinalis, from 3 to 5 cm. in length and from 12 to 23 mm. in width; in Dioon edule, from 114 GYMNOSPERMS 1o to 28 mm. in length and 7 to 19 mm. in width; in Ceratozamia mexicana, from 10 to 15 mm. in length and 7 to 8 mm. in width; in Zamia floridana, about 6 mm. long and 5 mm. wide; and in Zamia pygmaea, about 4 mm. long and 3 mm. wide. lic. 108.—Encephalartos villosus: top of male cone, with sporophylls at the top spreading apart and exposing the sporangia. The University of Chicago greenhouse. From a photograph by SepGwick. The micros porangia.—The sporangia, always on the abaxial sur- face of the sporophyll, are arranged radially in sori, the number vary- ing from 5 in the Cycas region of the family down to 2 or 3, with oc- casional single sporangia in the Zamia end of the family (figs. 109, TIO). CYCADALES II5 The sorus arrangement is the typical fern arrangement. There are no “‘synangia,”’ like those of all the Marattiaceae except Angiopteris. The number of sporangia on a sporophyll is largest in the Cycas region, and decreases to the Zamia end of the family. The numbers of sporangia on a sporophyll, averaged from several counts are as follows: Cycas media, 1,160; Dioon spinulosum, 770; uae eres caffer, 567; Macrozamia miquelii, 503; Dioon edule, 295; Microcycas calocoma, 245; Cera- tozamia mexicana, 191; Stangeria paradoxa, 153; Bowenza serrulata, 67; and Zamia flori- dana, 25. Sporophylls near the top and near the bottom of the cone have fewer sporan- gia, in Zamia, often only 2 or 3 on each side, with a sterile area between. By hunting, one can sometimes find a sporophyll with only one microsporangium on each side, so that it looks like a small megasporophyll with its two ovules. The more reduced spo- rophylls at the extreme top and bottom of Rees iid ude: the cone are entirely sterile. ae wiew audeciew of ab- In histological structure the microspo- axial surface. The sori are rangia bear a striking resemblance to those ™stly in fours and threes; ¥ 5 x $—After CHAMBERLAIN." of Angiopteris (figs. 111, 112). In both, the spores are very numerous, the stalk is massive, there are several layers of wall cells between the epidermis and tapetum, dehiscence is similar, and there is some ramentum. In the cycad, the ramentum is unicel- lular, with rarely a cross wall; while in the fern, two or three cross walls are common. The tapetum consists of very small cells in the cycad and of rather large ones in the fern. The structure of the sporangia, in both cases, is much like that of the eusporangiate ferns and of the Cycadofilicales of the Carboniferous. In all the cycads the microsporangium is strictly unilocular, in striking contrast with the multilocular (synangium) type of the Bennettitales. The development of the sporangium is of the eusporangiate type. There is a hypodermal archesporial cell or, in large sporangia, there may be a row or plate; the archesporial cell divides, forming a pri- mary wall cell and a primary sporogenous cell; from these, the thick 116 GYMNOSPERMS wall and numerous sporogenous cells are produced. The tapetum becomes clearly distinguishable rather late, so that it cannot always Fic. 110.—Ceratozamia mexicana: four microsporophylls; a, with sporangia not yet dehisced; b, sporangia at the upper part de- hisced and the pollen still holding together in balls; c, later stage, with nearly all of the pollen shed. The sorus arrangement (mostly threes) is more easily seen in the later stages; X 2.—After CHAMBERLAIN." be determined with certainty whether it is coming from the progeny of the primary wall cell or from that of the primary sporogenous cell. But, what- ever its origin may be, it is in contact with spores which it is to nourish. As the spore mother-cells enlarge, the tapetum breaks down and appears as a mass of nucleated protoplasm sur- rounding the growing mother- cells, which absorb not only the tapetum but also the wall cells between the tapetum and the epidermis. The epidermal cells become very much thick- ened, especially at the bottom and along the sides, while re- maining thin at the top, where the cell contents break through and escape (fig. 113). The spore mother-cells round off and each divides twice to form four micro- spores, thus bringing the 2x, or sporophyte, generation to a close and initiating the x, or gametophyte generation. In the first mitosis 12 pairs of chromo- somes are easily distinguished, and in the anaphase of the second division the number 12 is easily counted. WS O) PRC} 4 Fic. 111.—Dioon edule: longitudinal section of microsporangium, showing numerous microspores and typical structure of a eusporangiate sporangium. Two ramental hairs, each with one transverse wall, are shown on the lower left part of the sporangium.— From CHAMBERLAIN, The Living Cycads'° (University of Chicago Press). Fic. 112.—Angiopteris evecta: a typical eusporangiate sporangium, much like that of the carboniferous Cycadofilicales —From CHAMBERLAIN, The Living Cycads?° (Uni- versity of Chicago Press). Fic. 113.—Stangeria paradoxa: section of part of microsporangium with micro- spores. The cell walls of the outer layer are becoming very thick; the tapetal cells have broken down into a tapetal plasmodium. Finally, all the cells between the spores and the outer layer will be absorbed by the growing microspores. CHAPTER VI CYCADALES—Continued The reduction of chromosomes, in both megaspore mother-cells and in microspore mother-cells, brings back the original « number, which is the only number in plants below the level of sexuality. Since the terms gametophyte and sporophyte are coincident with x and 2x generations, in plants from the Hepaticae through the angiosperms, dents are likely to think the terms are synonymous. It is customary to call the x generation the gameto- phyte because, in the higher plants, it produces only gametes; and the 2x generation, the sporophyte, be- cause, in the higher plants, it is the only generation producing spores. THE FEMALE GAMETOPHYTE The megaspore is the first cell Fic. 114.—Zamia floridana: En- of the female gametophyte. It en- larging pPeespOre; 93 een Ties larges considerably, absorbing some F. Grace Smitu.s 4 of the neighboring cells before it divides (fig. 114). By the time the first division of the nucleus has been completed, there is a great change in the surrounding cells, which form a layer, quite common in gymnosperms, called “spongy tissue’ by STRASBURGER (fig. 115). The earlier free nuclear divisions are simultaneous, and the nuclei are crowded outward by a large cen- tral vacuole, so that most of the protoplasm, containing the free nuclei, is in a thin peripheral layer (fig. 116). Free nuclear division, in Dioon edule, continues until there are about 1,000 free nuclei, be- fore walls begin to form. Cell walls appear first at the periphery, and wall formation pro- ceeds toward the center until the entire gametophyte becomes cellu- 118 and in many Thallophytes, stu- ~ ge CYCADALES 119 lar (fig. 117). Long before the cellular stage is reached, a nutritive layer, one or two cells thick, is developed in contact with the gameto- phyte. It has been called the endosperm jacket, and is so conspicu- ous that it can be seen with the naked eye. It is shown in both the preceding figures. It functions like a tapetum, passing nutritive ma- terial from the cells next to it into the growing gametophyte. The arrangement of cells in the young gametophyte is extremely regular, radiating from the center to the periphery, as shown in hig. 2. Fic. 115.—Zamia floridana: First division of megaspore; a, actively nutritive cells; b, tissue of closely packed cells; c, flattened cells; *930.—After Dr. F. GRAcE Smitu.s® Soon after this stage, when the gametophyte has reached a diam- eter of about 3 mm., some of the cells at the micropylar end become larger, and their nuclei move from their central position to the pe- ripheral end of the cell (figs. 118-122). These are archegonium ini- tials. They are very numerous in comparison with the number which develop up to the fertilization stage. In most of them the nucleus does not divide at all, and, in Dioon edule, not more than ten ever reach the fertilization stage, and the usual numbers are only three, four or five. In other cycads there may be a couple more or a couple less. Microcycas is exceptional. CALDWELL® found scores, and some- 120 GYMNOSPERMS times hundreds of archegonia, some of them on the sides of the game- tophyte, and some even at the base. They are often as crowded as in the Cupressaceae. Dr. LILLIAN REYNOLDs,‘” studying the de- velopment of the archegonia in Microcycas, found the large numbers reported by CALDWELL, but found that only the group at the micro- pylar end of the gametophyte progressed up to the ventral canal mitosis, and that only in connection with this group was there any archegonial chamber. oe) ngs, v. rs Fic. 116.—Dioon edule: ovule soon after pollination, free nuclear stage of the female gametophyte—From CHAMBERLAIN, The Living Cycads° (University of Chicago Press). Fic. 117.—Dioon edule: the female gametophyte has become cellular throughout.— From CHAMBERLAIN, The Living Cycads° (University of Chicago Press). In Dioon edule the archegonium initials can be seen early in No- vember (figs. 118-122). They soon divide, forming the central cell and the primary neck cell, which divides almost immediately, so that December material shows the two neck cells characteristic of all the cycads. The central cell enlarges rapidly without a corresponding increase in the amount of protoplasm. Consequently, there is a large vacuole /2/ Fics. 118-122.—Dioon edule: development of the archegonium: fig. 118, archegoni- um initial; fig. 119, primary neck cell and central cell; fig. 120, the two neck cells; fig. 121, nucleus of central cell has divided, forming the ventral canal nucleus and the egg nucleus; fig. 122, mature archegonia showing archegonial chamber; figs. 118-120, X98; fig. 121, X50, fig. 122, X8.—From CHAMBERLAIN, The Living Cycads° (University of Chicago Press). 122 GYMNOSPERMS pressing the protoplasm against the wall of the cell. This single large vacuole is filled with a colorless sap. Material from the outside passes into the central cell, the amount of protoplasm increases rapidly, and soon the entire cell is filled with a beautifully vacuolate proto- plasm, with very large vacuoles in the central portion and smaller and smaller ones toward the periphery. There is a gradual grada- tion in the size of vacuoles from the large ones, more than 500 microns in diameter, down to the smallest ones, which can be seen with a 1.5 mm. oil immersion ob- jective and a 20X eyepiece; and probably the gradation continues beyond the resolving power of present-day microscopes. The central cell and its nucleus grow for about 6 months before the nucleus divides, the division taking place about the middle of April. The achromatic figure in this division is rather scanty, and there is not the slightest trace of the formation of a cell wall Fic. 123.—Dioon edule: mitosis giv- B ing rise to the ventral canal nucleus between the two nuclei (fig. and egg nucleus, showing that there is 123). no cell plate between the two nuclei; <350.—After CHAMBERLAIN.!° The ventral canal nucleus soon disorganizes, and, at the time of fertilization, is usually unrecognizable. Occasionally, however, it en- larges and goes through the same kind of development as the egg nucleus. SEDGWICK*3? observed several cases of this sort in Encepha- lartos villosus, and suggested that the egg nucleus might be fertilized by the ventral canal nucleus, as has been observed several times in Ginkgo, Pinus, and Picea. The fact that embryos have been found in conservatory material where no male cones were present, indicates that such a fertilization may occur. HemsLey’s*" reported hybrids between Ceratozamia longifolia and C. mexicana, in which 3-year-old pollen was used, also suggests such a fertilization; for cycad pollen probably does not retain its vitality more than a month—probably not so long. CYCADALES 123 The failure of wall formation at the ventral canal mitosis indicates a more advanced stage in the reduction of the archegonium than that found in Ginkgo, Pinus, and other forms which develop a defi- nite wall between the two nuclei. It is interesting, in this connection, to recall the reduction of the archegonium from forms with long necks, numerous neck canal cells, and a definite ventral canal cell, as in Marchantia; through forms like Riccia, with four neck canal cells, a ventral canal cell and an egg cell; through forms with only one neck canal cell and a ventral canal cell and an egg cell, like Marsilia, where the next step brings the condition in Ginkgo, with no neck canal cell, but with a ventral canal cell and an egg cell. In this series one finds occasional binucleate neck canal cells and, in related forms, a smaller number of neck canal cells, so that there is first the failure of a wall to be formed be- tween two nuclei, and then the failure of the division itself. In this reduction, the cycads have gone a step farther than Ginkgo, which still retains the wall between the ventral canal nucleus and that of the egg. It is only when viewed in this way that it becomes a matter of any evolutionary importance whether there is a ventral canal cell and an egg cell, or merely two nuclei not separated by a wall. Comparative morphology leaves no doubt that the neck canal cells and the ventral canal cell are homologous with the egg, so that the archegonium, phylogenetically, contained several eggs. The neck itself keeps pace with the reduction of the neck canal cells until, in the cycads, there are almost always just two neck cells; but in Encephalartos villosus SepGwick*37 found that the neck cells rather frequently divide. One of his figures shows six neck cells. The neck cells grow rapidly and, in later stages, become very turgid. To return to Dioon edule, which seems to be a typical cycad: the nutrition of the egg is practically the nutrition of the central cell, for the division of the nucleus of the central cell to form the ventral canal nucleus and the egg nucleus takes place only a few days before fer- tilization. So the process may be called the nutrition of the egg. For the first two or three months after the appearance of the archegonium initials, food materials are received from the surround- ing cell by the usual method of transferring substances from one cell 124 GYMNOSPERMS to another; but, as the central cell becomes large, a definite layer of cells, called the “archegonial jacket,’ appears. As the jacket be- comes more and more differentiated, the egg membrane thickens and finally becomes so tough that it retains its form and connection with the suspensor, even when embryos half the length of the seed are dis- sected out. In the ripe seed this resistent egg membrane can still be recognized. In an early stage it is easy to see that it has numerous large pits into which the turgid protoplasm of the egg presses. For some time the pits are covered by a thin membrane, the middle lamella between the wall of the jacket cell and that of the egg. The cells of the female gametophyte are full of starch, proteid, and prob- ably other materials which find their way through the archegonium jacket and into the egg in the usual manner. But the growing egg becomes so turgid that the papillae, or haus- toria, as they may be called, break the thin membrane, which closes the pit, thus leaving the haustoria of the egg in direct contact with the protoplasm of the jacket cell, so that materials can pass from the jacket cell into the egg as readily as from one part of a cell to another (figs. 124-20). The turgidity of the female gametophyte, and, later, the turgidity of the central cell, and, still later, the egg cell, is extreme. In trim- ming material for fixing, if one cuts too near the jacket, the young gametophyte breaks through. At the fertilization period, and for a month before that time, if one cuts too near the jacket, there may be a rupture several millimeters in length. If one cuts into the endo- sperm too near the archegonial jacket, there will be a small rupture of the egg membrane and cells nearest to it, and the liquid contents of the egg will spurt out, sometimes to a distance of 20 cm. Shortly before the nucleus of the central cell divides, the tissues around the archegonial region grow rapidly, leaving the archegonia in a depression called the “‘archegonial chamber.” Immediately after the mitosis which gives rise to the ventral canal nucleus and the egg nucleus, the ventral canal nucleus begins to dis- organize and soon disappears, while the egg nucleus moves toward the center of the egg, increasing immensely in size, until it sometimes reaches a diameter of 500 microns, easily visible to the naked eye. During this time the chromatin, easily distinguishable in the late CYCADALES 125 telophase, becomes obscured by other nuclear contents until it be- comes unrecognizable. The contents of the egg become very dense. Most of the vacuolate structure of the protoplasm disappears, apparently by the breaking down of the thin sides of the vacuoles, resulting in a more or less 127 Fics. 124-129.—Pits in the egg membrane and haustoria: fig. 124, Cycas revoluta, 150; fig. 125, the same, X 375; fig. 126, Dioon edule before the haustorium has broken the pit closing membrane; in figs. 127 and 128, the membrane has been broken and ma- terial is passing directly from the jacket cell into the haustorium, X 800; fig. 129, En- cephalartos lehmanii, shallow pit with protoplasmic continuity between the haustorium and the jacket cell; X 1100. Figs. 124 and 125, after IkENO;7® figs. 126-128, after CHAM- BERLAIN;?® fig. 129, after STOPES and Fujm.5 fibrillar appearance. Starch, proteids, and oil can be identified. In the living condition, at this stage, the outer border of the egg, which may be called the ‘‘Hautschicht,” is as colorless as water, while the interior is slightly turbid. GYMNOSPERMS 126 As in all seed plants, the megaspore, with its contained female gametophyte, is never shed from the sporangium. The spore coats of the heterosporous ferns, the ancient ancestors of the cycads, were very thick. As the megaspore became retained, its spore coats be- came thinner, but are easily recognized even in rock sections of the extinct Cycadofilicales (figs. 16, 19, 20). In the Cycadales, the mega- spore membrane stains brilliantly with safranin. THomson*®s found that the membrane consists of two layers, which may be called the en- dospore and exospore, the outer of which is suberized, while the inner consists of a substance closely related to pectin. The inner layer is dense and homo- geneous. The outer layer consists of club-shaped bodies with enlarged ends, so that under moderate magni- fication there seem to be three layers, a middle layer being the stalks of the club-shaped bodies (figs. 130, 131). The membrane, in Dioon edule, is strongly developed as early as Fic. 130.—Cycas revoluta: endo- spore and exospore about equal in thickness. Thickness of entire mem- brane is 4.5 microns.—After R. B. THOMSON.®33 Fic. 131.—Dioon edule: The megaspore membrane; the inner part is dense and homogeneous, while the outer part consists of ovoid, stalked bodies; q00,—After CHAMBERLAIN.!°° November, and continues to increase in thickness up to the germination of the seed, when it reaches a thickness of 10 microns. In fresh material, one might mistake the jacket of the female gametophyte for the mem- brane, but the jacket is coarser and can be stripped off entire with for- ceps. The membrane is more deli- cate, but pieces several millimeters in length can be stripped off. The membrane is thicker at the middle of the gametophyte than at the top or bottom. As the archegonial chamber develops, the membrane in that region is ruptured, so that the pollen chamber and archegonial chamber form a continuous cavity (figs. 132, 133; see also fig. 147). CYCADALES 127 THE MALE GAMETOPHYTE The microspore is the first cell of the male gametophyte. It has a very definite polarity and two very definite spore coats. The exine is thick at the bottom, moderate on the sides, and thin at the top; while the intine is thin at the bottom and top and very thick on the sides (figs. 134-37). The microspore, in all the cycads, begins to germinate while still contained in the microsporangium. At the first mitosis, a prothallial cell is cut off. This cell does not degenerate, as in most gymnosperms, 132 ‘ud S33 Fic. 132.—Dioon edule: megasporophyll with two ovules; one-half natural size. Fic. 133.—Dzioon edule: section of ovule at the stage shown in fig. 132; X 2. Both figs. 132 and 133 from CHAMBERLAIN, The Living Cycads” (University of Chicago Press). but becomes very active, although it never divides. The other cell divides, forming a generative cell and a tube cell. At this three- celled stage the microspore is shed from the sporangium. There have been many reports of insect pollination, but in a rather extensive field study in which all of the genera have been examined, most of them at the time of pollination, nothing has been observed which would indicate anything but the wind pollination so charac- teristic of the whole group of gymnosperms. The pollen is light and dry, and easily blown by the wind. The entire mass of pollen in a microsporangium, for a while, hangs together in one mass, being held by a kind of membrane formed by the disorganized wall and tapetal cells. When this breaks, the dry pollen is shed. As soon as the microsporophylls begin to crack apart, insects arrive and can be 128 GYMNOSPERMS seen crawling over the sporangia and doubtless feeding on the pollen. While the insects are so abundant in the male cones, they are rare in the female cones, except that a beetle is common about species of Encephalartos. This beetle, however, bores into the female gameto- phyte, destroying many of the seeds rather than furthering seed Tm “ill TOY n a % ae BO = 2 AANA 2, Wi) 5a fe CP LETARD RO Pr HT 136 Fics. 134-137.—Dioon edule: fig. 134, microspore; fig. 135, germinating microspore; p, prothallial cell; g, generative cell: fig. 136, exine ruptured by the young pollen tube: fig. 137, later stage, with more starch: the generative cell has divided, forming a stalk cell (s) and body cell (6). The prothallial cell (p) is protruding into the stalk cell; figs. 134-136, X1260; fig. 137, X 1000.—After CHAMBERLAIN.'% production. Many of the insects observed are flying species, but fertilization more than a hundred meters from a male plant is rare, and, in collecting material, it is well to select female cones within 10 or 20 meters of a male plant. Nucelli from a female cone within 4 or 5 meters of a male cone may show t5 or 20 pollen tubes, while those at 100 meters may show only 2 or 3, and female cones at a distance of 200 meters may show only 2 or 3 good seeds or none at all. Any CYCADALES 129 claim that there is any insect pollination should be supported by the most critical field study. At the time of pollination a large pollination drop appears at the micropyle. Cells at the top of the nucellus break down and some of their contents ooze out as a mucilaginous drop. The pollen grains fall on the drop and, as it dries, are drawn into the young pollen chamber below. Further drying seals the chamber, and the top be- comes very hard, forming the nucellar beak. Pollen germinates readily in sugar solutions, in thick juice of pear preserves, and in many syrups. The pollen tube soon appears, and grows to several times the length of the pollen grain, but the genera- tive cell does not divide. Cultures kept for a month show no division; but in material taken in the field, probably not more than a week after pollination, the division has taken place, giving rise to a stalk cell and a body cell, the name ‘“‘stalk cell” being given from a sup- posed homology with the stalk cell of the antheridium of Pterido- phytes. The pollen tube, coming from the upper end of the pollen grain, grows into the nucellus, and acts as a haustorium, conveying food material to the basal end of the tube. In Ceratozamia, in addition to the usual haustorial pollen tube, there are numerous haustoria ex- tending downward from the base of the pollen tube. The genus could be identified by this feature as positively as by the two horns on the sporophyll, which give it its name. The pollen tube is a haustorium, not a sperm-carrier as in angiosperms. There is no further cell division for a long time, the division of the body cell taking place almost immediately before fertilization. The interval between pollination and fertilization is about four months in Cycas revoluta, five months in Zamia floridana, and about six months in Dioon edule. During this long period the pollen tube digests its way downward, enlarging the pollen chamber until it finally extends completely through the nucellus and the pollen tubes hang free in a cavity which is partly pollen chamber and partly archegonial chamber (fig. 147). When the body cell, or spermatogenous cell, is first formed, no blepharoplast or centrosome has been demonstrated; but, as the body cell enlarges and elongates, first one and then two very small 130 GYMNOSPERMS blepharoplasts appear. They are at the side of the nucleus opposite the prothallial cell, and one of them soon moves around to the oppo- site side of the nucleus. They remain in this position and grow rapid- ly. As the elongated body cell increases in size, it gradually becomes spherical, and the two blepharoplasts move go degrees, so that a line drawn through them would be perpendicular to the long axis of the pollen tube (figs. 138, 139). The blepharoplasts are at first very Fics. 138 and 139.—Dioon edule: fig. 138, December material, body cell elongated and blepharoplast parallel with the long axis of the pollen tube; fig. 139, May material with blepharoplasts rotated until they have become transverse to the long axis of the pollen tube; X 237.—From CHAMBERLAIN, The Living Cycads*° (University of Chicago Press). dense and homogeneous, but later become very vacuolate and reach an immense size, from 16 to 18 microns in diameter in Dioon edule, and 20-27 microns in Ceralozamia mexicana, the largest yet known. The radiations surrounding the blepharoplasts are very conspicu- ous and often extend to the wall of the cell. Shortly before fertilization the body cell divides, and in each of the two resulting cells a single sperm is developed. As the sperm grows, the radiations begin to disappear and the blepharoplast breaks up into a great number of small granules (fig. 140). Some of these be- come attached to a beaklike protuberance of the nucleus, which in- creases immensely in size and rotates so that the mass of granules, CYCADALES 131 fusing into a band, is drawn out into a spiral with several turns (fig. 141). From this spiral band, lying just beneath the Hautschicht, thousands of cilia are developed. They pierce the Hautschicht and extend into the cell cavity. The topography of the nucellus with its pollen tubes during these stages is shown in figs. 142 and 143. In all of the cycads, except Microcycas, there are two sperms in each pollen tube. In Microcycas, according to CALDWELL,* there Fic. 140.—Stangeria paradoxa: the body cell has divided to form two sperms, and each blepharoplast has broken into numerous granules which will form the spiral band. The nuclei of the sperms are still comparatively small (January 11, 1913). 400. are usually 16 sperms. Dr. DorotHy Downie found 8 to 11 body cells, so that the number of sperms would range from 16 to 22. Spermatogenesis in Microcycas was studied in great detail by Dr. Downie’ and the results not only show how the large number of sperms originates, but suggest what is probably the real nature of the generative cell and stalk cell (figs. 144 and 145). After the prothallial cell has been formed, the next mitosis gives rise to a tube cell and a generative cell. The latter divides in the usual way, forming a stalk cell and a body cell; but from this point, the behavior is peculiar to Microcycas. The stalk cell divides, giving rise to another body cell, 132 GYMNOSPERMS which later gives rise to two sperms, while the cell beneath it divides again, producing another body cell, and the process is repeated until as many as 8 and sometimes even 11 body cells have been formed, each of which produces two sperms. ~ _— - a . be, “#y . Fic. 141.—Dioon edule: two young sperms. The nuclei of the sperms have become larger, and, in the one at the left a part of the spiral band can be seen attached to the beak of the nucleus; * 350. Dr. Downie"? regards the generative cell as the primary sperma- togenous cell, and the stalk cell as a spermatogenous cell still active in Microcycas, but having lost the division potentiality in other gymnosperms. This interpretation seems to be sound from the stand- point of comparative morphology and phylogeny. Sperms of cycads are remarkably large (fig. 146). In Dioon edule, CYCADALES 133 while still in the pollen tube, they measure about 200 microns in diameter and 275 microns in length. After leaving the tube, they Fic. 142.—Stangeria paradoxa: nucellus with pollen tubes after the tubes have bro- ken entirely through the nucellus. In one pollen tube, the body cell has divided to form the two sperms (January 19, 1913); X50.—From CHAMBERLAIN, The Living Cycads”° (University of Chicago Press). Fic. 143.—Stangeria paradoxa: two sperms have been formed from each body cell (February 2, 1913); X50.—From CHAMBERLAIN, The Living Cycads™° (University of Chicago Press). increase in size, reaching a diameter of 230 microns and a length of 300 microns. Consequently, they are easily visible to the naked eye. 134 GYMNOSPERMS The 2 sperms begin to move while still within the sperm mother- cells, but soon the peripheral part of the wall between them breaks down and leaves them free within the wall of what was called the “body cell” before its division. This cell increases in size, and the sperms, which stick together, move around in the cavity. The cilia, at first, move slowly and then more vigorously, so that the pair of sperms roll about in the rather small space. The movement of the Fic. 145 Fics. 144 and 145.—Microcycas calocoma: fig. 144, pollen tube with four body cells, and stalk cell in telophase of division to form another; fig. 145, pollen tube with three body cells and stalk cell in prophase of division to form another; #, prothallial cell; s, stalk cell; b, body cells. No centrosomes have been found in the stalk cell or in the body cell immediately after its formation; 1280.—After Dr. Dorotay G. Downte.*% cilia is accompanied by pulsating and amoeboid movements, and when the apex of a sperm strikes the wall, there is a sudden, convul- sive movement which makes one think of Vorticella. They swim for an hour or more in the cavity of the body cell before they separate, and then for another half-hour before they escape into the general cavity of the pollen tube. When free from each other, the general movement is straight ahead, with a rotation upon the longer axis. They swim up into the tube as far as the diminishing diameter will permit, and then come back. When a nucellus is inverted, with a OO —_— Fic. 146.—Ceratozamia mexicana: photomicrograph of a section of a sperm, show- ing the large nucleus, thin sheath of protoplasm, and numerous cilia; X 360. The photo- micrograph was made by Miss Erne, Tuomas.—From CHAMBERLAIN, The Living Cycads° (University of Chicago Press). 136 GYMNOSPERMS drop of strong sugar solution and still further protected by a bell jar, the movements have continued for 5 hours. How much longer they Fic. 147.—Dioon edule: a reconstruction, from several sections, of an ovule at the time of fertilization. The pollen tube on the left shows the body cell still undivided; the one in the middle shows two sperms and the remains of the prothallial and stalk cells; the one on the right shows the two sperm mother-cells and the spiral ciliated band be- ginning to develop. Two pollen tubes have discharged their sperms. A sperm has en- tered the egg on the left; the one on the right still shows the ventral canal nucleus. Two sperms, in the thick liquid discharged from the pollen tube just above them, are ready to enter the egg. The dark line below the nucellus is the megaspore membrane.—After CHAMBERLAIN.?% would move under natural conditions is conjectural. Efforts to keep the sperms alive after escaping from the pollen tube were not suc- CYCADALES 137 cessful. In weak sugar solutions they almost explode. In a to per cent solution, they quickly die; in a 20 per cent solution, they live a little longer. FERTILIZATION Some have imagined that the archegonial chamber is a cavity in which the sperms swim for a while before entering the egg. This view is entirely incorrect. At the time of fertilization the pollen chamber and the archegonial chamber are merely moist. There is not a trace of any free liquid. In studying the living condition and in fixing material for de- tailed study, hundreds of ovules have been cut, and always there is the moist membrane, but nothing more. A careful study of a drawing will help to understand the process of fertilization in a cycad (fig. 147). In the later stages of the development of the sperms, the basal end of the pollen tube becomes very much swollen and so turgid that it finally bursts, discharging the sperms, with liquid from the pollen tube, into the archegonial chamber. This liquid is the only medium in which the sperms can move while in the archego- 3 Fic. 148.—Stangeria para- nial chamber. doxa: fertilization; the sperm As before noted, the egg becomes ex- nucleus is entering the top of tremely turgid during the final stages the ese nucleus; the ciliated y 8 8 8 band is at the top of the egg; of its development, in fact, so turgid y fo sien Coimarerain = that the contents of the egg would spurt out into the archegonial chamber were it not for the turgid neck cells. The liquid discharged from the pollen tube has such a high pressure that sperms discharged from the pollen tube into a 30 per cent solution of cane sugar move about freely. The liquid from the pollen tube, coming into contact with the neck cells, lowers their turgidity, and some of the contents of the upper part of the egg escape into the archegonial chamber, leaving large 138 GYMNOSPERMS vacuoles at the top of the egg. A sperm is then drawn into the egg so violently that the protoplasm, with its ciliated band, is torn off and left near the top of the egg, while the nucleus moves downward to unite with the egg nucleus. The ciliated band remains in the top of the egg throughout the free nuclear stages of the embryo, and sometimes can be distinguished still later. The cilia become indis- tinguishable from the protoplasm of the egg, as if they actually be- come a part of the egg protoplasm; but the dense band, from which they arise, seems solid as long as it can be recognized. The behavior of the chromatin, from the entrance of the sperm up to its contact with the egg nucleus, has not been satisfactorily de- scribed in any gymnosperm. Both nuclei are filled with a substance which stains deeply with iron haematoxylin, but most of that sub- stance is certainly not chromatin (fig. 146). STRASBURGER called it metaplasm, because he did not regard it as chromatin or as proto- plasm. Even after the sperm nucleus begins to enter the egg nucleus the metaplasm is conspicuous, and the chromatin does not seem to be recognizable (fig. 148). The term “fertilization” is used rather loosely and perhaps it is better not to attempt to define it too closely. Any stage with the sperm inside the egg is likely to be called fertilization, and any stage with the two nuclei in contact passes for fertilization. Details are more thoroughly worked out in Coniferales, and so the subject will be treated again, especially in Pinus, Abies, and Juniperus. CHAPTER VII CYCADALES—Continued EMBRYOGENY Just where fertilization is completed and embryogeny begins, is indefinite; but the stage shown in fig. 148 is certainly regarded as a stage in fertilization, while the stage shown in fig. 149 shows the first division of the nucleus of the fertilized egg and, conse- quently, is a 2x mitosis. So somewhere be- tween the stages shown in these two figures fertilization has been completed, and the sporophyte generation has started. Around the first mitotic figure there is a fibrillar area many times as large as the fig- ure itself. The fibrillae seem to be the same as the spindle fibers and radiations of later figures. The chromosomes, at metaphase of the first mitosis, are not hard to count, but it is surprising to find that the num- ber is 12, the same number counted at the mitosis which gives rise to.the ventral canal nucleus and the egg nucleus, and also at various stages in the development of the gametophyte. The anaphase of this mitosis was not found, but anaphases of later mi- toses showed 24 chromosomes, and root tips showed the same number. The micro- spore has 12 chromosomes, so there is no doubt that in Stangeria, and probably in the rest of the cycads, the « and 2% num- bers are 12 and 24. HUTCHINSON’S?’5 work Fic. 149.—Stangeria para- doxa: first division of the nu- cleus of the fertilized egg. At the top is the spiral band of the sperm which fertilized the egg, and also three sperms which got through the neck but failed to penetrate the egg; X42.—After CHAMBER- LAIN." on the first division in Abies, which will be considered in the proper place, offers an explanation of the condition found in Stangeria. 139 140 GYMNOSPERMS Free nuclear stage.—Following fertilization there is a period of free nuclear division. The divisions are simultaneous, so that the number of nuclei is, theoretically, 2, 4, 8, 16, 32, 64, 128, 256, 512, and, insome cases, 1,024. The earlier divisions are very regular and the numbers of nuclei are about what should be anticipated; but, in Dioon edule, the eighth, ninth, and tenth divisions are irregular, some of the Fics. 150 and 151.—Stangeria paradoxa: free nuclear stage of embryo; fig. 150, 8- nucleate stage, showing simultaneous mitoses in two groups of nuclei; very definite polarity with simultaneous mitoses in both groups of nuclei; fig. 151, polarity with mitoses in upper group but not in the lower; 42.—After CHAMBERLAIN."9 nuclei failing to divide, especially in the upper part of the embryo. The number beyond the 256 nucleate stage is likely to be less than the theoretical estimate, because some of the nuclei fail to divide. In Stangeria there is often a distinct polarity, the nuclei being in two groups (figs. 150, 151). If the nuclei in the lower group divide, those at the top do also; but those at the top may divide without a corresponding division in the lower group. A glance at the illustra- CYCADALES 141 tions of free nuclear division in Stangeria will show that the nuclei in the upper part of the embryo, during the earlier mitoses, may be- come much more numerous than in the lower part. Fic. 152.—Zamia floridana: photomicrograph of a small portion of a section showing an early free nuclear division; X 413. Negative by Miss Ere, Tuomas.—From CuHam- BERLAIN, Methods in Plant Histology (sth ed.) (University of Chicago Press). During the earlier mitoses, the spindle fibers and polar radiations are very striking, but there are no centrosomes at any stage (fig. 152). The number of nuclei reached in the free nuclear period varies. 142 GYMNOSPERMS In Dioon edule the egg often reaches a length of 5 mm., and the num- ber of free nuclei is about 1,000—theoretically 1,024. In Zamia flori- dana, with an egg about 3 mm. in length, the number is quite regu- larly 256. In Bowenta serrulata, LAWSON reports 64, the lowest number ever found in a cycad. In other gymnosperms, with much smaller eggs, the number comes down to 4, and in Sequoia a wall is formed at the first mitosis. In the homosporous ferns there is no free nuclear period, and none in living heterosporous ferns, although it is probable that there was Le et OE eT Oc ya RAS nee € paiaee Pe te WP tae me Le BY KTH 58.8 So sa Fic. 153.—Stangeria paradoxa: simultaneous division of nuclei in lower part of em- bryo; *140.—After CHAMBERLAIN."? a free nuclear period in extinct heterosporous ferns, and, in our opinion, there was an extensive free nuclear period in some of the Cycadofilicales. A free nuclear period arose as a consequence of the enlarging eggs. The mass of protoplasm became so large that the early mitotic figures could not segment it. In Cycas and Stangeria the nuclei at the base of the embryo under- go a vigorous simultaneous division after the nuclei of the rest of the embryo have ceased to divide (fig. 153). The suspensor, and all the rest of the embryo below it, come from this second period of simul- taneous division. Formation of cell walls in embryo.—After the nuclei have increased in number until there is a comparatively small amount of proto- Fics. 154-156.—Dioon edule: evanescent segmentation; fig. 154, lower part of em- bryo; fig. 156, diagram of entire embryo; fig. 155, lower part of embryo, showing perma- nent walls at the bottom and free nuclei above; m, the thick membrane of the egg; fig. 154, 200; fig. 156, X20; fig. 155, X 108.—After CHAMBERLAIN.” 144 GYMNOSPERMS plasm about each one, segmentation begins. In some cycads, like Dioon and Stangeria, an evanescent segmentation throughout the entire embryo takes place before the permanent walls appear (figs. 154-506). In early stages they are stronger at the base of the embryo and weaker and weaker above (figs. 156, 157). Following the free nuclear period, in the type of embryogeny which we should regard as the most primitive, the embryo becomes Fic. 157.—Stangeria paradoxa: permanent walls forming below; above, weaker walls and then free nuclei.—After CHAMBERLAIN.”9 cellular throughout, as in some species of Cycas,“° Encephalartos, and Macrozamia'® (fig. 158). In these, the cells break down in the central portion, leaving two or three layers of cells at the outside. In others, including those with evanescent segmentation, the upper part remains in the free nuclear condition, while cell division con- tinues in the basal region. Differentiation into body regions—Hundreds of cells are formed before there is any differentiation into body regions. The first dif- ferentiation to appear is an elongation of cells which are becoming the suspensor. Mucilage cavities also appear at this time (figs. 159- CYCADALES 145 61). The body regions differentiate very slowly, the cellular condi- tion, except in the suspensor region, being somewhat uniform even after the topography of cotyledons and coleorhiza become easily rec- ognizable (fig. 162). That the dermatogen has not yet become com- pletely differentiated, even after the appearance of cotyledons, is shown by the frequent occurrence of periclines. The suspensor is a remarkable feature of the cycad embryo. In an early cotyle- don stage the suspensor is not a simple structure derived from a single embryo, but is made up of the suspensors of all the embryos which have come from the ferti- lized eggs of that group. In some of the embryos, the suspensor elongates but lit- tle; in others, it elongates more; and in the embryo which is to be the only one to reach full development the later stretches of the suspensor belong only to the ma- ture embryo. In a good preparation the various embryos, each at the end of its own suspensor, can be seen in spite of the coilingand twisting ofthe compound struc- ture made up of all the suspensors (fig. 163). In Ceratozamia, in the living condi- tion, the suspensors can be pulled out toa Beery ae Cranes: ; young embryo, indicating that length of 7 or 8 centimeters. Whatever the entire egg may have seg- the primary function of this suspensor mented and that the central may be, it certainly thrusts the growing me pene Gop 25. embryo down into the gametophyte. Another remarkable feature of the cycad embryo is the coleorhiza. It appears early in the development of the embryo, and in later stages, after the embryo has reached the full length of the seed, it becomes extremely hard. The mature seed is rather uniform throughout the family (fig. 164). There is an outer, fleshy seed coat which is white or creamy, or which may be variously colored, with red or various combinations 1 | ey SOS Nese gee. SIS OXY &. s4 ATT 1) J Utes C) Seratuanetueestani a Saw Ty wD) "> ERA eceate ? Fics. 159-161.—Dioon edule: fig. 159, embryo showing mucilage cavities (m) and be- ginning of suspensor; fig. 161, later stage, but with dermatogen not yet fully differen- tiated; fig. 160, diagram with embryo at the left a little more advanced than that in fig. 159, and the one at the right, in the same stage as that in fig. 161.—After CHAM- BERLAIN."”° CYCADALES 147 of red and orange predominating. Beneath the fleshy layer is a hard, tough, stony layer; and, within this, the inner, fleshy layer, which rr lg eudl N opeee Wen i OETA (VLD) 2a Capes, PAX CITY | Hat TRIO S TS EEE A Ln S Ce Ha here Mic alllieitratsaines LO ST OL TAI ES CATT I ST DUS EY RIT Ba DER TT A Ed) Hira er sg SRB AST je Sesth: PeCsgr hy Fic. 162.—Dioon edule: early cotyledon stage of embryo; the enlarged portion back of the cotyledons is the coleorhiza; periclines show that the dermatogen is not yet fully differentiated.—After CHAMBERLAIN.?” soon gives up most of the cell contents to the female gametophyte and growing embryo. This layer, containing the inner vascular sys- tem, becomes a dry, papery membrane. Just underneath the top of it the remains of the nu- cellus can be seen, forming a dry, papery cap. The embryo, now extending the whole length of the seed, inside the seed coats, has two cot- yledons, which are not always equal in size; one leaf with usually one or more scale leaves; and the long suspensor coiled and packed against the micropylar end. Fic. 163.—Ceratoza- mid mexicana: young embryos with suspen- sors; the rounded bodies at the top are the tough egg membranes; X1.5. —After CHAMBERLAIN.” 148 GYMNOSPERMS Occasionally there are more than two cotyledons, three being re- ported for Encephalartos; and Ceratozamia has regularly only one. The cotyledon situation in Ceratozamia is particularly interesting. SISTER HELEN ANGELA,"” noticing a few tracheids in the part of the embryo opposite the cotyledon, suspected that they might represent Fic. 164.—Dioon edule: mature seed: The darkly shaded part at the top is the coleorhiza (c); the two cotyledons and first leaf are shown; the dotted part is a female gametophyte; the hatched part is the stony layer (s) of the seed, outside of which is the outer fleshy layer (0); m, micropyle; natural size. From CHaAm- BERLAIN, The Living Cycads° (University of Chicago Press). the missing cotyledon. Ceratozamia is unique in that the cone decays and sheds its seeds before the cotyledon stage. She revolved seeds on a clinostat during the entire embryogeny, and all these seeds developed two cotyledons, while more than 100 seeds, grown in the usual way, had only one cotyledon, which developed on the side next to the ground. In all cases, the cotyledons, in the basal region, form a continuous cotyledonary tube. At the tips, cotyledons are sometimes lobed or divided, giving an impression that there are more than two. THE SEEDLING There is no resting stage in a cycad seed: de- velopment is continuous from fertilization to the death of the old plant. Seeds do not long retain their vitality. Those with the hardest stony layer and toughest outer fleshy layer may be germi- nated a year after they fall out from the cone. Seeds of Macrozamia moorei, which have an ex- ceptionally thick and hard stony layer, have ger- minated after 2 years. If aseed rattles when shak- en, the chances are against any germination. Germination of the seed.—Cycad seeds are nearly always elongated. When they fall out from the cone they do not sink into the ground, but lie on the surface. To secure the best germination in the green- house, they should not be covered with soil, but pressed in lightly— about half-way—with the long axis of the seed parallel with the sur- face of the soil. As the embryo within the germinating seed elongates, it fractures the stony coat, the hard coleorhiza protecting the delicate young CYCADALES 149 root-tip (figs. 165-67). After the coleorhiza has fractured the stony coat, the root tip digests its way through the coleorhiza and begins to turn down. Entering the soil, it grows rapidly, while the stem remains inconspicuous. Usually only one leaf appears. Part of the cotyledon protrudes from the seed but the greater part remains in- side, absorbing all of the female gametophyte and passing it on to 165 166 Fics. 165-167.—Dioon edule: seedling; fig. 165, the coleorhiza has fractured the stony coat and the root tip has digested its way through the base of the coleorhiza; fig. 166, the root end is turning down: about two-thirds of the protruding part is cotyle- don and the lower part is coleorhiza, with the root visible at the base; fig. 167, later stage with 3 leaves between the cotyledons; c, coleorhiza; 7, root; s, cotyledons; all natural size.—From CHAMBERLAIN, The Living Cycads (University of Chicago Press). the seedling. The cotyledons then become dry, but the stony coat of the seed, with the withered cotyledons inside, remains attached to the plant for a year or even longer. Anatomy of the seedling.—Many have studied the anatomy of the cycad seedling and have described its principal features. THIEs- SEN,“ Matre,3” and SISTER HELEN ANGELA Dorety’? have pre- sented the most detailed accounts. THIESSEN’s account of Dioon edule and SISTER HELEN ANGELA’S of Dioon spinulosum have already been mentioned in connection with the girdling of the leaf trace. 150 GYMNOSPERMS Tuiessen*™ found that the vascular cylinder of the stem is very short, so that it may be called a vascular plate rather than a cylinder (fig. 168). The plate is squarish and has a protoxylem group at each corner, from each of which a strand extends downward, forming the protoxylem of the tetrarch root. Four strands extend in the op- posite direction, soon forking and entering the cotyledons, so that each cotyledon gets four strands. For each of the first leaves four strands leave the vascular plate at or near its four protoxylem points. An interesting feature is the change from the endarch to the exarch condition. The cotyledonary bundles, as they leave the vascular plate, are endarch; but, farther along in the cotyledon, they become mesarch. The leaf traces also, when leaving the vascular plate, are endarch; but, in the leaf base, centripetal xylem appears, so that the bundle becomes mesarch. From this point the centripetal xylem in- creases and the centrifugal decreases until the bundle becomes en- tirely exarch before it reaches the region of leaflets. Beyond this point the bundles are exarch. The vascular cylinder of the embryo is a protostele; but in older stages it gradually becomes an endarch siphonostele, except in Mi- crocycas.*° SIstER HELEN ANGELA™ also found that in Microcycas the vascu- lar strands of cotyledons and leaves are endarch near the base and exarch in the upper portions. The vascular plate, from its first ap- pearance, is an endarch siphonostele, probably the only siphonostele in early stages of a cycad seedling (fig. 169). Matte,*” and others who have studied the embryogeny of the cycads, agree that the stem consists of leaf bases. This would mean that the adult stem is a mass of leaf bases, in which secondary growth has produced the familiar trunk. This interpretation is suggestive and may have a wider application. The seedling usually has just one leaf. Leaf after leaf appears at irregular intervals, and it is likely to be several years before leaves begin to appear in crowns. It is rare for a cycad to produce a cone before the tenth year, and some are probably much older before the coning stage is reached. The demarcation between seedling and adult plant is as indefinite as that between baby and boy—and between boy and man. - — ~ / oe mee ee we wm ew we wm ew ew ww ew we ww ow we we owe os wwe So? 1 ee ee we we ee oem ewe eee = ce os ee a - man Pod - - =. -- ww eew wee ee Fic. 168.—Dioon edule: semi-diagrammatic reconstruction of vascular system of embryo, showing girdling: col, cotyledon; ¢b, tubular part of cotyledon; cs, cotyledon- ary strands; /} to /{, four strands of first leaf; /} to /!, strands of second leaf; vp, vas- cular plate; a, xylem from protoxylem of the plate to form protoxylem of primary root. —After TurEssEen.™4 152 GYMNOSPERMS HYBRIDS Hybrids between Ceratozamia longifolia and C. mexicana were claimed as early as 1882; and there was also a claim that hybrids had been secured between Ceratozamia robusta and C. brevifrons. In both cases, seeds were secured which germinated. But, as HELMSLEY** remarked, ‘‘There is only one thing certain in all this, and that is the uncertainty of the so-called species.”” Doubtless, HELMSLEY was Fic. 169.—Microcycas calocoma: transverse section of stem just above the cotyle- donary plate, showing the siphonostele condition: A, B, C, D, the four cotyledonary strands; f', f?, f3, f4, the leaf traces still in procambial condition —After SisteR HELEN ANGELA Dorety."9 right, for all of these so-called species could probably be raised from seeds of a single cone of Ceratozamia mexicana. In the first-mentioned cross, pollen from a male cone of Cerato- zamia longifolia was preserved for 3 years and then used to pollinate C. mexicana. It is possible that the mere irritation of dead pollen in the pollen chamber might stimulate development; but it is certain that the 3-year-old pollen was dead. However, the cycads hybridize freely. Of more than a dozen suc- cessful pollinations made at the University of Chicago, two will be described here. Zamia latifoliolata * Zamia pumila.’3—On December 20, 1921, I pollinated female cones of two plants of Zamia latifoliolata with pol- CYCADALES 153 len of Z. floridana. The sporophylls of the female cones were tightly closed, and it is more than doubtful whether any of the pollen sur- vived the usual spraying. A week later sporophylls of both female cones opened and Dr. Paut J. SepGwicxk pollinated them with Zamia pumila. One cone produced three seeds, and the other, four. All were planted, and four of the seven have developed into vigorous F, plants, two of them female and two male. The first of these four Fic. 170.—Ceratozamia mexicana X Zamia monticola, F,; seedlings: s, seed; c, cotyle- dons; b, bud; 7, primary root; ar, apogeotropic root; co, coleorhiza; a, egg membrane: A, B, and C, with two cotyledons: D, with only one; C’, embryo in female gametophyte; C’’, transverse section of embryo with two cotyledons.—After CHAMBERLAIN.!3 to cone produced a small male cone which shed its pollen in January, 1928. So, the male cone appeared while the plant was only 6 years old. In 1929 three of the F, plants produced cones, one female and two males. One of the males shed all of its pollen before the sporophylls of the female opened; but the second male cone was still shedding pollen at the right time, and from this pollination good seeds were secured and several F, plants are thriving. Ceratozamia mexicana X Zamia monticola.’3—A generic cross be- tween Ceratozamia mexicana and Zamia monticola was particularly successful. One cone was pollinated March 23, 1924, and another, on a different plant, on April 25, 1924. From these two cones more 154 GYMNOSPERMS than a hundred seeds were secured, and now (December, 1934) more than fifty thriving F, plants look as if they were large enough to pro- duce cones. It will be remembered that, normally, Ceratozamia has only one cotyledon, while Zamia has the usual two. In 1925, when the seedlings were being repotted, a careful examination was made and of the fifty-six which had survived up to this stage, forty-seven showed two cotyledons; three had one cotyledon, and in the other six the cotyledon situation could not be determined without sacrific- ing the seedlings (fig. 170). It is evident that, in this generic hybrid, Zamia is dominant as far as the cotyledons are concerned. At this time (1934) not a single one of the F, plants has coned. The leaves now look like Ceratozamia mexicana leaves, and some of them would be diagnosed as C. longi- folia. Miss SopHIA PAPADOPOULOS**4 made a comparison of the leaflets of the two parents and the F, generation, and found that in some features the F, resembled one parent and in some resembled the other; but that in the stomata there were features of both parents. When cones are produced it will be interesting to find whether the sporophylls have the two strong horns of Ceratozamia or the truncate apex of Zamia. CHAPTER VIII CYCADALES—Continued PHYLOGENY A study of the life-history of any group should make the investi- gator try to determine what its ancestry may have been and try to find whether it has left any progeny. The “living fossil” character of the Cycadales makes them par- ticularly favorable for a study of phylogeny, because the ancestors, which must have flourished in the Upper Carboniferous, are the best known of any fossil plants. Without going into any detailed discussion, it may be assumed that the Cycadales have come from the Filicales either directly or through the Cycadofilicales. It is true that the lycopod line was very well represented in the Carboniferous, but paleobotanists agree that this line has not given rise to either the Cycadales or the Ben- nettitales. The fern leaves of that period are nearly always pinnate, more often twice or thrice pinnate than once pinnate; and the Cycadofili- cales are similar, with more than once-pinnate leaves prevailing. As far as the leaf is concerned, the cycads might have come from either group. But, as shown in an earlier chapter, we think it has been proved, as far as anything can be proved in phylogeny, that the Cycadofili- cales came from the Filicales. Lines of evolution do not progress at the same rate: one organ may progress rapidly while another re- mains stationary. The cycads retain the swimming sperm, but have lost the wall between the ventral canal nucleus and the egg nucleus; while in the pines the sperms have lost the swimming character but still keep the wall between the ventral canal nucleus and the egg nucleus. In Stangeria, the leaf is so fernlike that the genus was placed in the Polypodiaceae, of the true ferns—even in the genus Lomaria; but it has lost the wall between the ventral canal nucleus and the t59 156 GY MNOSPERMS egg nucleus. Extremely rapid changes may take place in reproduc- tive structures without any noticeable changes in the leaves. Most of the apples on the market today were unknown 30 years ago; but there has been little corresponding change in the leaves. Since this is true, there is no occasion for surprise if the leaf of the true ferns is carried over into the Cycadofilicales, with little or no change, while the reproductive structures become so modified that they form the basis for a great phylum. As we have already remarked, leaves, stems, and roots might orig- inate independently in different groups. Nearly all botanists believe that the land flora had an aquatic ancestry, and no one doubts that algae preceded vascular plants. When algae transmigrated to the land, it became necessary to develop protective, conductive, and supporting structures, which had not been produced while the plant was surrounded by water. By mutation, by smaller variations, by natural selection, or in some other way, an efficient conducting sys- tem was developed. Later, more algae may have transmigrated to the land, and, hav- ing similar conditions to contend with, may have made similar re- sponses and developed a vascular system much like that of the pre- vious transmigrants. Because two systems of this sort resemble each other, it has been assumed that one of them transmitted it to the other. Paleobota- nists are dependent, to a great extent, upon the evidence of vascular anatomy. Resemblances may be due to heredity, but it is possible that some of the similar structures may owe their similarity to en- vironment. And so there is a logical possibility that the leaves of Filicales and Cycadofilicales may owe their similarity to similarity of conditions. But we doubt whether details in venation and margins would be so identical if developed in this way. There are such things as environ- mental anatomy and hereditary anatomy. The former is more quan- titative than qualitative and much more subject to change. If en- vironmental anatomy were as powerful as hereditary, corn seeds and pumpkin seeds planted in the same hill should not produce such dif- ferent plants. We believe that the striking similarity of the leaves of Filicales CYCADALES | and Cycadofilicales is due to heredity, so that the Cycadofilicales represent only the further evolution of some of the Filicales. The strongest evidence for the derivation of the Cycadofilicales from the Filicales, as we have remarked before, is furnished by the seed. The Evolution of the Seed would make a good title for a book. One would have to transport himself back to the days of fairies and giants, of wonderful lamps and carpets, to believe that the seed originated without any ancestry. The heterosporous Pteridophytes of today show unmistakably their homosporous ancestry; and it is only reasonable to suppose that heterosporous forms of ancient times arose from homosporous in the same way, some of the sporogenous cells failing to reach full development and giving up their substance to nourish the one or more spores which thus reached a higher stage in evolution. Mega- spores became larger and larger until a free nuclear period arose within the spore, the gametophyte being retained within the spore, just as the spore itself, later, became retained within the sporan- gium. This, we believe, was the mode of origin of the seed habit. Naturally, it might take place without any great accompanying changes in the leaf. In Selaginella, a lycopod, but nevertheless showing what probably took place in the heterosporous ferns, the megaspore is shed with its contained female gametophyte in various stages of development. Usually, the gametophyte has passed the free nuclear stage and en- tered the stage of cell formation. In Selaginella apus, a semi-aquatic species, archegonia may be formed and fertilization may take place before the megaspore falls out from the sporangium. In extreme cases there is little dehiscence of the sporangium and the megaspore remains inside, so that the shoot with its cotyledons and stem tip, and the root, break through the sporangium. In this extreme case the term “‘seed” is strictly applicable to Selaginella. Very probably, such a situation was common in the Carboniferous, where so many “seedlike”’ structures are found. The line between ferns and advanced members of the Cycadofili- cales, like Trigonocar pon, was as sharp as the line between ferns and seed plants of today. The place where it would be hard or impossible to draw the line would be where the enlarging megaspores were just 158 GYMNOSPERMS beginning to be retained within the sporangium, sometimes falling out—when the plant must be classed as an heterosporous fern—and sometimes remaining inside the sporangium—when the plant must be classed as a seed plant. When some sporangia on a plant shed the /72 Fic. 171.—Reduction of the megasporophyll in Cycadales: A, theoretical ancestor of Cycas; B, Cycas revoluta; C, Cycas circinalis; D, occasionally in Cycas media, and usually in Cycas normanbyana; E, Dioon edule; F, Macrozamia; G, Ceratozamia; H, Zamia. Fic. 172.—Reduction of the megasporophyll in Bennettitales: A, B, C, stages in the reduction in some hypothetical ancestor; D, the usual condition found in fossils; 2, the sporophyll has become entirely sterile; /, fertile and sterile sporophylls with about the arrangement found in Bennettites gibsonianus. megaspore, while others on the same plant retained it, as sometimes happens in Selaginella apus, only a taxonomist, by carefully selecting small portions of the frond, could get satisfactory specimens for his herbarium. The reduction of the megasporophyll in the Cycadofilicales has lt CYCADALES 159 already progressed so far that it suggests the Cycas revoluta type. In Lyginopteris, the tips of leaflets are regularly sterile, while the seeds are borne farther back. In Neuropteris, while many of the seeds are terminal, some are lateral. From such a type, both the Sinn, ” Wy, w%S3 es ) Fics. 173-175.—Theoretical form (fig. 173) which might have given rise to the Bennettitales by the abortion of lateral ovules (fig. 174), and to the Cycadales, by the abortion of terminal ovules (fig. 175). Cycadales and Bennettitales condition could be derived; the Cyca- dales by the abortion of the terminal sporangium, and the Bennet- titales by the loss of the lateral sporangia (figs. 171, 172). We can imagine that the ancestor of both the Bennettitales and the Cycadales looked something like fig. 173. Inside the crown of 160 GYMNOSPERMS vegetative leaves is a crown of much reduced leaves, the male sporo- phylls, bearing sporangia, much as they are borne in a fern. Inside the crown of male sporophylls is a crown of still farther reduced leaves, the female sporophylls, bearing the ovules. The change from the bisporangiate condition to the dioecious is one which can be seen in all stages of transition in living angiosperms; and the evolution of the compact cone from a loose crown of sporophylls is seen in the living cycads. From the condition shown in fig. 173, the Bennettitales may have developed as in fig. 174; and the Cycadales as in fig. 175. From the Cycadofilicales to such a form as Cycadospadix henno- quei, the transition is not hard to imagine; and Cycados padix might well have been included in the living genus, Cycas, for its megasporo- phyll differs less from that of Cycas revoluta than the sporophyll of C. revoluta differs from those of some of the other species of the genus (fig. 176). Cycadospadix milleriana,*® with sporophylls loosely ar- ranged, but stopping the growth of the axis, shows a condition in- termediate between the loose crown of sporophylls of Cycas revoluta, and the compact cone (fig. 177). The theory that the Cycadales may have come from the Bennet- titales has already been referred to. The trunk and leaves are similar in the two lines and some of the Cycadales, like the Bennettitales, have axillary strobili. But the ovules, even in the earliest known Bennettitales, have followed a very different line of evolution. They have, in all cases, retained the terminal ovule and lost the lateral ones; while all of the Cycadales, even those known only as fossils, have lost the terminal ovule and retained the lateral. The Bennet- titales could not have transmitted what the line had already lost. The multilocular microsporangium of the Bennettitales is so dif- ferent from the unilocular sporangium of the Cycadales that they could hardly be related. Lines characterized by unilocular and mul- tilocular sporangia may have been as distinct in the Paleozoic as they are in Angiopteris and Marattia today. Both types, in the Pa- leozoic, may have developed heterospory; and the heterospory of the fern may have progressed into the seed habit of the Cycadofilicales, so that one section of the Cycadofilicales may have given rise to the Bennettitales, and the other, to the Cycadales. CYCADALES 161 Have the Cycadales left any progeny? Something has left some progeny; for an abundant progeny, both gymnosperm and angio- sperm, is very visible and very much alive. What groups could have been responsible? If we consider only the nine living genera of cycads, the answer is easy: They are not responsible; they are the last of their race, re- stricted in geographical dis- tribution, restricted in num- bers, and struggling for their eS very existence. What progeny exists for which an ancestry should be found? The only possibilities are the Cordaitales, Gink- goales, Coniferales, Gnetales, and the angiosperms. We do not believe that any of these 177 owe their origin to the cy- Fic. 176.—Cycados padix hennoquei: a meg- cads. but shall venture some asporophyll rather closely resembling that of é P ‘ 7 : Cycas revoluta.—After ZEILLER.77 speculations in dealing with eee f Fic. 177.—Cycados padix milleriana: mega- the Coniferophytes. sporophylls collected into a loose cone, which We should feel consider- must have stopped the growth of the axis. The able confidence in conclud- axis could not have persisted as in the domale 2 Pree: lind: Filicales plant of Cycas revoluta—After RENAULT.4°S ms ) ) Cycadofilicales, and Cycadales, is a genetic line, as thoroughly established as any line extending over so great a period. TAXONOMY The ideal taxonomy should be based upon life-histories and phy- logeny. Unfortunately, the taxonomy of the Cycadales has suffered more than that of most groups, because the people who have made most of the descriptions have never studied cycads in the field. Microcycas, so large that one can climb around among its branches, and not looking at all like Cycas, would never have received such a name if the taxonomist had seen it in the field. But even such mis- nomers are not so bad as names given to commemorate those whose names would otherwise be lost; or those who would be remembered 162 GYMNOSPERMS without spoiling the name of a plant. It is to be hoped that the growing practice of decapitalization will discourage commemorative names in botany, as it has in zodlogy and geology. How much better Encephalartos horridus and Ceratozamia mexicana than Smithia jones- iana and Wangia yenii. Cycads, in the field, vary with age and with other factors. One who has studied cycads in the field would hesitate to determine most of the species from herbarium specimens. The latest monograph is that of ScHuSTER.S” There is a splendid bibliography and some original work, but what seems to be a total lack of study in the field. The keys are in Latin, as taxonomists claim they should be. The rising generation knows little or no Latin, but must know English, German, and French. Keys in any of these three languages would be more useful to nearly all of the people who might want to do some identification. Cycads have several characters which could be made the basis for taxonomic keys. Most of them are composite, using both vegetative and reproductive features. In cultivation, where individuals are few and coning is rare, a key based entirely upon vegetative features would be desirable. SistER MARY ALiceE LAMB3%° devised a key to the genera based upon the leaves. Only Cycas and Stangeria have a midrib in the leaflet; and Cycas has only a midrib with no lateral veins, while Stangeria has a midrib with lateral veins. Bowenza is the only cycad with twice-pinnate leaves. Dioon is the only cycad with the insertion of the leaflet as broad as any other part of it. Macrozamia is the only one with a gland at the base of the leaflet; but in some species of this large genus the gland is obscure or may be absent. However, its presence identifies most of the species as belonging to this genus. In Microcycas the leaflets are reflexed on the rachis; in the rest they are either flat or turn up a little. Of the other three, Zamia has the rachis subcircinate in vernation; while in En- cephalartos and Ceratozamia it is erect. In Ceratozamia, the leaflets are long, narrow, and taper gradually to a point, and are always en- tire (inlegerrima). In Encephalartos the margin of the leaflet is very jagged in some species and the lower leaflets are more and more re- duced until, at the base, they become mere spines. In Ceratozamia there is no such reduction. CYCADALES 163 With the addition of histological characters the key devised by SISTER Mary Atice LAms could be made sufficiently complete for identification of the genera. In the large genera, she studied several species. A key based upon the male gametophyte could probably be de- vised. Some of the genera can be determined, at a glance, by the cones. The two strong horns of Ceratozamia, the long spine of Ma- crozamia, the elongated peltate top of Mzcrocycas, the loose cone of Dioon, and the crown of megasporophylls in Cycas provide an easy identification. With all characters available, it is easy to make keys. For the convenience of students the following key, somewhat modified from PILGER’s key in ENGLER and PRANTL, Pfanzenfami- lien, may be as good as any: KEY TO THE GENERA OF CYCADACEAE A. Leaflets with a midrib, but no side veins. Mega- sporophylls in a crown, through which the axis continues to grow. Sporophylls with several avules along, the sides—Oriental . . 2:.:..2.(: ae ane asei eee 1. Cycas B. Leaflets with a midrib and pinnate side veins. SME ELCA a ods ais a’ «6 exe.ni sw 0 io ee ola eee 2. Stangeria C. Leaflets parallel veined, no midrib. Megasporo- phylls in cones, each with 2 ovules @)eeanets bipinnate—Australia: > .iscn.seaeereoe sce 3. Bowenia b) Leaflets once pinnate 1. Ovules on stalklike protrusions of the meg- asporophyll and arranged in a loose cone— UES 6160 en nei reine eg Saloons ooo rr Aire 4. Dioon 2. Ovules sessile. Ovules with (a) shield-shaped top with two strong horns Mexico: . 5. 4. sea eeee cen = a 5. Ceratozamia (b) Sporophylls shield shaped, without horns (1) Cones small; sporophylls in longi- tudinal rows; leaves developed AMECPICE..6 Ssaacass Ole eS cane 6. Zamia (2) Cones large; leaves in crowns; mostly large plants—Africa............. 7. Encephalartos 164 GYMNOSPERMS KEY TO THE GENERA OF CYCADACEAE—Continued (c) Sporophylls with a long median spine Australias 5 32 Nig . ve woe “ Ms oor Fic. 195.—Cordaianthus williamsonii: longitudinal y section of ovule, showing the VG, nucellus free from the integ- oe aan ument. The lower part of the nucellus was much de- cayed before silicification AN YY} ] LZ y F f/f Y Fic. 194.—Cordaianthus williamsonii: occurred, but it still shows longitudinal section of female strobilus, that there was an elongated showing sterile bracts and two ovules; X 10. female gametophyte; X35. —After RENAULT. 49 —After RENAULT.‘ its living members. In the Coniferophytes, the earlier members also have a large pith, scanty zone of wood, and a large cortex; but there is a tendency, even in the Cordaitales, to reduce the pith and cortex 182 GYMNOSPERMS and increase the proportion of wood. The later fossil members and » all the living members have a small pith, large zone of wood, and a comparatively small cortex. Whether these resemblances in stems mean genetic relationship is conjectural. Helianthus, Sambucus, and Cucurbita have a large pith, scanty zone of wood, and small cortex; but that does not mean that any one of them inherited the condition from another or that any of them got it from any gymnosperm. Phylogenet- ic anatomy is one of the best indicators of genetic relation- ships; but it is not always easy to distinguish between phylo- genetic anatomy and ecologi- cal anatomy. Environment is a powerful factor. If there had been no changes in this factor, the Cordaitales should still be with us, and their life-stories would be as well known as that of Pinus. While the leaf is peculiarly Fic. 196.—Cordaianthus Grand d’Euryi: Ye to -changesas Spee section of beak of nucellus; wedged in the tions, and may. change from passageway are two large pollen grains, the simple to compound in the lower one showing part of the surface of the Jifetime of an individual, we exine and also the multicellular interior; dee hee Remauyie ie should regard the compound leaf as an outstanding char- acter of the Cycadophytes; and the simple leaf as just as strong a characteristic of the Coniferophytes. The two phyla existed side by side in the Paleozoic, and the Cordaitales have been recognized much farther back than the Cy- cadofilicales; but it is more than doubtful whether the Cordaitales were the ancestors of the Cycadofilicales. Both must have come from heterosporous Pteridophytes. There was so much variety among the Paleozoic Pteridophytes that it is not necessary to make CORDAITALES 183 both come from the same section of that group. Some Pteridophyte with compound leaves and a stem with large pith, scanty wood, and large cortex, may have given rise to the Cycadophytes; while an- other, with similar stem structure, but with simple leaves, may have been the ancestor of the Coniferophytes. When the life-histories of the Paleozoic forms become better known, there will be a firmer foundation for theories, and theorizing will not be so easy. With only a few facts to reconcile, any theory will do; but when facts become numerous, a theory reconciling all the facts is more difficult to formulate. CHAPTER X CONIFEROPHY TES—GINKGOALES This order is still represented by one living genus, Ginkgo. Its ancestry can be traced back to the Paleozoic, and, like the cycads, it may be called a living fossil, for it still retains the swimming sperm which characterizes the ferns and which probably characterized all of the Paleozoic seed plants. Leaves are abundant in the Permian, and leaves which may be- long to the order, but which may belong to the Cordaitales, were abundant in the Carboniferous. From the Triassic the order de- veloped rapidly and can be traced with confidence, reaching its greatest abundance and widest distribution in the Jurassic; but be- fore the end of the Jurassic, it began to wane, and is now represented by the single genus, Ginkgo. EXTINCT MEMBERS As extinct members of this order, aside from Baiera, which all recognize as nearly related to the living Ginkgo, P1ILGER, in Die natiirlichen P flanzenfamilien, lists seven genera as more or less prob- ably related, and six more as more or less doubtfully belonging here. The living genus, Ginkgo, is not rare in the Liassic, and in the middle Jurassic was abundant and world-wide in its distribution. Several extinct species have been described, but, since they are based principally upon leaves, it seems just as well to put all of them under Ginkgo biloba. Its leaves vary so much with age, position on the tree, and environment, that most, if not all, of the leaf impressions which have tempted taxonomists to make more species could be duplicated by leaves of the living species. Baiera*®’ was abundant in the lower Permian of Europe and North America. In the Jurassic it was not so abundant as Ginkgo. Im- pressions have been found in the Rhaetic of South Africa. The genus became extinct in the Lower Cretaceous. 184 GINKGOALES 185 Usually only the leaves are found, and they are very seldom attached. They differ from the living Ginkgo in being more deeply incised and in having narrower lobes. The venation is as distinctly Fic. 197.—Baiera gracilis: the leaf is bilobed, but each of the two lobes is repeatedly dichotomous; natural size.—After RENAULT.4%5 dichotomous as in the ferns (fig. 197). The seeds are borne on a branching axis, usually with several seeds on an axis. Some of the figures, which seem to give the shape very accurately, do not show any collar (fig. 198A). The microsporangia were in strobili consisting of sev- eral sporophylls, each with several microsporangia. In younger stages, the sporo- A Fic. 198.—Baiera miinsteriana: A, five un- ripe seeds; B, portion of male strobilus with three sporophylls bearing several microspo- rangia, shortly before shedding pollen; C, similar specimen at a later stage. Rhaetic from Bayreuth.—After ScuimprrR-SCHENK.5! phylls, with their microsporangia, closely resemble those of the liv- ing Ginkgo, except that there are generally six or more sporangia on a sporophyll (fig. 198B and C). 186 GYMNOSPERMS GINKGO Probably no other living tree can trace its ancestry so far back as Ginkgo, for, as we have noted, it can be recognized in the Liassic. It is doubtful whether it exists today in the wild state. Travelers claim that they have seen it growing wild in the forests of western China, where, they say, it reaches a height of more than 30 meters, with a diameter of 1.3 meters. If extinct in the wild state, it must have be- come so recently, perhaps in the last two or three thousand years, or the date may have been much later, even in the last hundred years; for the passenger pigeon is extinct, although people now living can remember when it was so abundant that immense flocks cast shadows like big clouds. Ginkgo was kept alive by priests in China and Japan, who cultivated it in temple grounds. Now, in cultiva- tion, it is world-wide and hardy even where the winter temperature reaches 20°F. below zero. It is a beautiful tree of various aspects, for it may be tall and slender; or it may have a trunk a meter in diameter, soon breaking into widely spreading branches, so that the breadth may exceed the height. Whether it ever reaches the reported 30 meters in height is doubtful. A botanist should identify the tree which is being measured. THE LIFE-HISTORY The life-history, especially its spermatogenesis and embryogeny, has been so thoroughly investigated that the principal features are well known. The stem.—In most cultivated specimens the trunk is strongly excurrent and the outline steeply pyramidal: but as the tree gets old—so or 60 years—it is likely to broaden, so that the outline is rounded at the top. If the top of a young tree is cut off, the branches spread widely; so that the trunk of a tree a meter in diameter may not be more than twice that height before it breaks into large spread- ing branches (figs. 199, 200). As in many Coniferophytes, there are two kinds of branches, the long branch and the short spur. The first growth is always of the first type; and as the branch, or the main axis, increases in length, it grows for a year as a long branch before any spurs appear. The Young trees with strongly excurrent trunks Fic. 199.—Ginkgo biloba FIG. 200.—Ginkgo biloba: ovulate tree at Imperial University of Tokyo.—From a photograph by Miyake. GINKGOALES 189 long branch grows rapidly, even half a meter in a year, while a spur 2 or 3 centimeters long may be several years old. A spur with its leaf scars and scale leaf scars, and with’half a dozen leaves coming out from the top, at nearly the same level, recalls the cycad trunk, with its armor and crown of leaves. A spur, even after reaching an age of 5 or 10 years, instead of pro- ducing a crown of leaves, may grow out into a long slender shoot with widely scattered leaves. Fic. 201 Fic. 202 Fic. 201.—Ginkgo biloba: transverse section of long shoot; X8 Fic. 202.—Ginkgo biloba: transverse section of spur shoot; X8 In transverse section, the topography of the long shoot and the spur is very different (figs. 201 and 202). The long shoot has a com- paratively small pith and cortex, its wood is harder, and there are not so many mucilage cavities. Even in old spurs, where the quanti- ty of wood is at its maximum, it is not difficult to cut sections. In both shoots there is an abundant development of secondary cortex. Annual rings, while not so prominent as in Pinus and other gym- nosperms of the Chicago region, are well marked. Chicago is near the northern limit for Ginkgo. Seeds planted in the open scarcely ever survive the first winter. The seedlings are kept inside the first winter and then, for several years, are put out earlier and earlier in the spring and taken in later and later in the autumn, until they are rugged enough to be set out permanently. Even then the tree is not as hardy as most native trees, for a few weeks of mild weather, too early in the spring, will cause the buds to swell. When cold weather 190 GYMNOSPERMS follows, as it usually does, many leaves fall, and those remaining on the tree do not reach their full size and are likely to fall early in the autumn. Such an untimely swelling of the buds is usually accom- panied by the formation of a growth-ring; not a strong ring which might cause uncertainty in estimating the age, but nevertheless a ring which can easily be seen. Mucilage cavities are abundant. They are found in the pith and in the primary cortex of the stele, in the root, in the petiole and blade of the leaf, in the seed and its peduncle, and in the sporangium. They seem to be absent from the secondary cortex. Tannin cells and cal- cium oxalate crystals are also abundantly and widely distributed. The vascular cylinder of the stem is an endarch siphonostele. The protoxylem consists exclusively of spiral elements, which are much more abundant in the spur shoots than in the long shoots. In the long shoots, in transverse section, in the radial direction, there are from one to five spiral elements; while in the spur shoots there may be as many as ten, with five or more quite common. In both shoots, the spiral elements border directly upon the pitted tracheids. The tracheids of the secondary wood have one or two rows of bordered pits on the radial walls, not crowded as in the Cordaites group, but more or less scattered. Their arrangement is either opposite or alternate. The last tracheids of a year’s growth have tangential pitting, where they border upon the spring wood of the next year, a feature of considerable interest, because paleozoic stems of the Cordaites type, when they have rings, do not show any radial pitting. Bars of Sanio are easily demonstrated in the secondary wood, but do not occur in the primary woods” (fig. 203). The pits are often so scattered that no bars are formed. The trabeculae of Sanio also occur, but are easily distinguished from the bars, because they cross the lumen of the tracheid, while the bars are covered by a . secondary thickening of the cell wall. The medullary rays are characteristically short in vertical extent, many of them being only one cell in height, and they are nearly al- ways only one cell in width. Rays one cell wide and two cells high predominate in the trunk and in the long shoots, but some rays are three, four, or even five cells in height. In the spur shoots the rays, GINKGOALES IQI although only one cell wide, range from one to sixteen cells in height, with rays three to six cells high very common (figs. 204 and 205). The rays remain alive for a long time, retaining their nuclei, proto- plasm, and starch for 30 years or more. With the exception of the medul- lary rays, there are no parenchyma cells, like the thin-walled cells of the cycads, interspersed with the woody tracheids. The leaf —The beautiful leaf, with its symmetrical dichotomous vena- tion, has given Ginkgo its colloquial name, the Maidenhair Tree, because the leaves on the spur shoots resem- ble those of Adiantum, the Maiden- hair Fern. The leaves on the long shoots are mostly bilobed, the fea- ture which suggested the specific name; but the leaves on the spurs usually have only a wavy margin, with none of the deep lobing. Leaves at the top of the tree, on first-year long shoots, and especially the leaves of seedlings, are very deeply lobed; and besides the two deep primary ®) (*) b) KOO lO ot es Ete Fic. 203.—Ginkgo biloba: longi- tudinal section of mature wood show- ing pitting and “Bars of Sanio.”— After JEFFREY.3% lobes may have two or three secondary lobes on each side, so that they approach the deeply and narrowly lobed leaves of the extinct Baiera (figs. 206, 207). With so much variation in the leaves, species based upon leaf characters alone are open to more or less suspicion. The dichotomous venation of the leaf is very regular. In the petiole there are two strands, each, by repeated forking, forming the venation of its side of the leaf (Fig. 208). It may be that the more vigorous growth of seedlings and long shoots may be responsible for the bilobed character of their leaves, while the leaves of the slow growing spurs are seldom bilobed. 192 GYMNOSPERMS The veins occasionally have a few centripetal tracheids, and thus have an indistinct mesarch structure. A single strand enters the cotyledons, distinctly mesarch below, but becoming exarch higher up. FIG. 204 FIG. 205 Fic. 204.—Ginkgo biloba: longitudinal tangential section of long shoot, showing med- ullary rays, one to three cells in longitudinal extent; X 180. Fic. 205.—Ginkgo biloba: longitudinal tangential section spur shoot, showing medul- lary rays, one to fifteen cells in longitudinal extent; X 180. The mucilage cavities in the leaves are elongated, 1-5mm. in length, and are nearly as conspicuous as the veins themselves. They take the same direction as the veins, and are usually spaced about half-way between adjacent veins. Many cells contain a large crystal of calcium oxalate, and tannin cells are abundant. The two strands in the petiole are endarch, with spiral protoxylem elements. At a distance of several cells from the protophloem there is a sheath of thick-walled cells, and, just within it on the protoxylem side, there are often some thick-walled suberized tracheids (fig. 209). GINKGOALES 193 The tissue is rather uniform throughout the leaf, there being no well marked palisade in the leaves of the spur shoots. The leaves on Fic. 206.—Ginkgo biloba: tracings of leaves from a single tree at the University of Chicago. The leaves in the upper part of the figure are from long shoots; and the ten lower leaves are from spur shoots; one-third natural size. the long shoots, especially the larger leaves, have a well-marked palisade. The stomata are on the abaxial surface and are slightly sunken (fig. 210). The root.—The root has a diarch cylinder, except when there are three cotyledons, in which case it is triarch. size. GINKGOALES 195 A short distance back from the root tip, the outer layers of cells are suberized, and just within the suberized layers, tannin is very abundant. Starch, as in most roots, is very abundant. SS Oe: Fic. 209.—Ginkgo biloba: part of one of the two strands in the petiole of the leaf; co, calcium oxalate crystals; s, sheath; pp, protophloem; ~, phloem; c, cambium; x, xy- lem; 7, rays; px, protoxylem; X 180. Within a millimeter of the tip, the endodermis can be distin- guished, and the next layer of cells outside it have characteristic ring thickenings. The mature root looks much like the mature stem, with annual rings and with similar bordered pits. 196 GYMNOSPERMS THE SPOROPHYTE—REPRODUCTIVE The reproductive structures of Ginkgo are the most primitive in living seed plants, except the Cycadales. The male strobilus.—The male strobilus is strikingly like that of the extinct Baiera, the principal difference being that Baiera has as many as six microsporangia on a sporophyll, while Ginkgo almost always has only two. The number two, however, is not entirely ] u) ©, Q aura ae Te [{ © on as a 22. ae i) © oeae ace ee eal Noss Cees (Qe sZeceeta > @ e enaems=: oe NY rk . @, @, eee & , a Poot dite Soro Fic. 210.—Ginkgo biloba: section of a young leaf cut at a right angle to the vein. The spiral protoxylem has lignified but the walls of the rest of the xylem have scarcely begun to thicken. A large mucilage cavity is shown at the right; X 180. rigid, for material grown in the United States often shows three or four sporangia, and SPRECHER®’” cites a case with seven. The slender sporophyll is surmounted by a hump and bears two pendant raicrosporangia (figs. 211, 212). The stalk has two small collateral endarch bundles. The development of the microsporangium follows the usual eusporangiate type, an archesporial cell giving rise to a primary wall cell and a primary sporogenous cell, the latter giving rise to the spo- rogenous tissue and, later, to the spores. Miss ANNA STARR‘ investigated the hump at the top of the sporo- phyll and found that its large mucilage cavity developed like a GINKGOALES 197 microsporangium (fig. 213). COULTER and LANp* had found that in Torreya resin ducts are formed from three of an original seven sporangia, so that the sporophyll is really peltate, like that of Taxus. In Ginkgo the mucilage cavity in the hump shows the type of de- Fic. 211.—Ginkgo biloba: long shoot Fic. 212.—Ginkgo biloba: single with spur shoots bearing male strobili male strobilus, showing sporophylls, and young leaves; about natural size. each bearing two sporangia; X3.5. —After COULTER.! —After CoULTER.'S3 velopment which they found in Torreya. Since Baiera has regularly more than two sporangia on a sporophyll, and Ginkgo occasionally has more than two, the mucilage cavity might represent lost spo- rangia. Against this theory, it can be urged that Bazera, with all its sporangia, still has a hump, and that Ginkgo, in the rare cases with more than two sporangia, has the hump just as well developed as when there are only two. Further, the mucilage cavities in the leaf, in early stages, look like those in the hump. 198 GYMNOSPERMS Fuju,'” who has made an extensive study of abnormalities in Ginkgo, found microsporangia growing on foliage leaves. They were usually near the base of the leaf blade. Fic. 213.—Ginkgo biloba: A, topography of strobilus, showing sporogenous tissue (dotted) and mucilage cavity in the hump (outlined); B, early stage in development of microsporangium, showing two of the primary wall cells and two of the primary sporog- enous cells; C, the hump, with its mucilage ¢avity; A, September 2, X 20; B, beginning of sporogenous tissue, X 485; C, *485.—After Dr. ANNA M. Starrs? In the northern part of the United States the male strobili can be recognized early in July as small papillae in the axils of bracts. Be- GINKGOALES 199 fore winter stops their growth, the sporogenous tissue is well de- veloped, and may reach the mother-cell stage. Reduction of chro- mosomes takes place as soon as growth is resumed in the spring. Eight and sixteen are the « and 2% numbers. The female strobilus.—The ovules are borne, often in great num- bers, on spur shoots (fig. 214). This branch shows an unusually large number of peduncles bearing two full-sized ovules. Usually one of Fic. 214.—Ginkgo biloba: long shoot with spur shoots bearing ovules. Many of the peduncles bear two fully developed ovules. the two ovules aborts early. As may be seen at the extreme left, the long shoot, with its spurs, has come from an earlier spur. It will be noted that the leaves are merely rounded, with none of the bilobed character. The bilobed condition is probably due to a vigorous early growth of the two earliest veins, the leaf traces, which would cause the lobing by growing faster than the parenchyma of the leaf blade. The leaves on the spurs are very immature during the early development of the strobili. It may be that the diversion of food materials to the rapidly developing strobili may cause the leaves to grow slowly and evenly at this early period, during which the contour of the leaf is being 200 GYMNOSPERMS determined. When spurs do not bear reproductive organs, the leaves are often lobed. The young ovules break through the bud scales and pollination occurs late in April or early in May, while the leaves are still im- mature (fig. 215). In northern Ohio the bud of the spur begins to swell about April 1. By dissecting away the outer brown bud scales and the inner Fic. 215.—Ginkgo biloba: spur shoots with ovules and young leaves, shortly after pollination. In the photo at the right, the leaves have been cut away. The collar is con- spicuous at this stage; 1.5.—After COULTER. greenish scales, one can see the pale, cream-colored ovules and the very young greenish foliage leaves. At this time the single integu- ment has appeared, but has not begun to cover the nucellus. By May 1, sections show the megaspore mother-cell surrounded by the spongy tissue characteristic of gymnosperms (fig. 216). On the same tree, at the same time, some of the mother-cells have divided, giving rise to linear tetrads. Sometimes the mother-cell gives rise to a row of three cells, one of the first two cells produced by the mother-cell having failed to divide. In such a case, only two megaspores are produced; and in any case only one megaspore functions. Occasion- ally, there are two megaspore mother-cells, one above the other, GINKGOALES 201 with one or more parenchyma cells between. Miss CAROTHERS counted 8 as the x number of chromosomes. SPRECHER did not determine the exact number, but estimated that it was not less than 7 or more than tro. By the end of April, a pollen chamber has been formed by the breaking down of cells at the apex of the nucellus. Some of the Fic. 216.—Ginkgo biloba: A, section of young ovule, showing nucellus and integu- ment, April 1; B, section of pair of ovules; C, megaspore mother-cell surrounded by spongy tissue, May 1. A, X15; B, X12; C, X325.—After CAROTHERS.” broken down material is mucilaginous and is exuded, appearing just outside the micropyle as the pollination drop. Pollen, caught in this drop, is drawn down into the pollen chamber and is sealed in by the drying of the drop. The further development of the pollen chamber and other parts of the ovule is doubtless stimulated by the presence and development of the pollen. The ovules increase rapidly in size, maintaining a mottled pea- green color. One of the two ovules on each peduncle usually aborts; but when conditions are favorable, it is not at all uncommon for both ovules to develop into good seeds (fig. 214). 202 GYMNOSPERMS A striking feature of the ovule is the collar. It is conspicuous throughout the development of the ovule, but is proportionately larger during the earlier stages (fig. 215). Fuyrt found many ovules Fic. 217.—Ginkgo biloba: abnormal development of ovulate structures; a, bud; 8, nearly normal strobilus; c, leaf with irregular thickenings; d, leaves bearing ovules; e, ovule; f, thickening at base of ovule; g, longitudinal striation along fleshy part of seed.— After Fuyrr.' borne on more or less modified foliage leaves (fig. 217). These ex- ceptional cases led to the conclusion that the collar is a modified leaf- sporophyll bearing an ovule. The fact that the peduncle has four vascular bundles, while the petiole of a leaf has only two, strongly GINKGOALES 203 supports the interpretation that the peduncle is a stem, bearing two sporophylls, the collars, each bearing an ovule (fig. 218). When a peduncle bears three ovules, there are six bundles; and when there are several ovules, the peduncle has double that number of bundles, so that a transverse section has a zone of wood like that in a stem. In general, the development of the ovule resembles that of a cy- cad (figs. 219 and 220). There is a prominent nucellus with a large pollen chamber, surmounted by a beak which becomes brown and hard. The single integument, with the lower part of the ovule, be- comes differentiated into three layers, the outer fleshy layer, and inner fleshy layer, with a stony layer between them. The develop- Fic. 218.—Ginkgo biloba: A, transverse section of petiole of leaf, showing two vascu- lar bundles; B, similar section of peduncle, showing four bundles; X 17. ment of the inner fleshy layer differs from that of a cycad, for its cells are very thin walled and watery until the ovule reaches a considerable size. It then develops as in a cycad, finally becoming a thin, dry, papery brownish membrane. The outer fleshy layer remains fresh, greenish, and juicy until the frost comes in autumn. The ovules then fall to the ground, the fleshy layer becomes wrinkled, but still re- mains watery, and has a characteristic odor which is unpleasant to most people. The ovules are poisonous to some people, causing sores on the hands or other parts which may be touched. The vascular supply of the ovule is very scanty. Two strands enter the inner fleshy layer, and, without any branching, extend to the free part of the nucellus. The entire outer vascular system, so prominent in the outer fleshy layer of the cycad ovule, is entirely lacking. In the rare cases in which the ovule is triangular in trans- verse section, there are three bundles in the inner fleshy layer. 204 GYMNOSPERMS In the very abnormal cases in which several ovules are borne on a branching axis, only one ovule terminates each branch, and its peduncle has only two bundles, still further supporting the theory that the collar is a modified leaf blade-sporophyll, bearing an ovule (fig. 221). Fic. 219 Fic. 220 Fic. 219.—Ginkgo biloba: longitudinal section of an ovule shortly after pollination; o, outer fleshy layer of integument; s, stony layer of integument; 7, inner fleshy layer of integument; c, collar; 0 (lower), abortive ovule. The inner fleshy layer and nucellus are shaded with lines; X 2.—From CouLTEeR and CHAMBERLAIN, Morphology of Gymno- Sperms'4 (University of Chicago Press). Fic. 220.—Ginkgo biloba: longitudinal section of ovule, after the stony layer has be- come hard and the inner fleshy layer has become dry and papery; X 2.—From CouLTER and CHAMBERLAIN, Morphology of Gymnosperms' (University of Chicago Press). THE GAMETOPHYTES The gametophytes, especially the male, bear some resemblance to those of the cycads, but have such definite characteristics that the Ginkgoales and Cycadales could be identified by their gametophytes. The male gametophyte——The microspore is the first cell of the male gametophyte. As in the cycads, it germinates while still en- closed in the microsporangium. At the first mitosis, two very un- equal cells are formed, the smaller one, a prothallial cell. The inner, GINKGOALES 205 larger cell divides again, producing another prothallial cell, and a larger cell which is the antheridial initial. The first prothallial cell soon aborts, but the second is persistent. The antheridial initial di- vides, producing a generative cell, in contact with the second pro- Fic. 222.—Ginkgo biloba: early development of the male gameto- phyte; A, first prothallial cell and inner cell; B, second prothallial cell and inner cell dividing to form the generative cell and tube cell; C, mi- » : tosis in antheridium initial; D, first : prothallial cell (aborting), second >» prothallial cell, generative cell, and a tube cell—the shedding stage. The : exine is shaded with lines. It does Fic. 221.—Ginkgo biloba: axis with not cover the top of the pollen grain. seven ovules, each borne singly on a The intine is represented only by a peduncle, not in pairs. Each peduncle line. It alone covers the top of the has two bundles instead of the four pollen grain; X770.—From CHAM- which appear when the peduncle bears BERLAIN, Methods in Plant Histol- two ovules.—After SPRECHER.S7® ogy? (University of Chicago Press). thallial cell, and an inner cell, the tube cell, which does not divide again. In this four-celled condition the pollen is shed (fig. 222). If fe- male trees are within 200 meters, some pollen will reach the pollina- tion drops and there will be some seed. If there are female trees within 100 meters, pollination is likely to be so abundant that most of the ovules will be pollinated. The exine does not cover the entire pollen grain, being lacking at the top, which is covered only by the intine. Upon reaching the 206 GYMNOSPERMS pollen chamber, the intine protrudes and becomes anchored in the tissue of the nucellus, where it acts as a haustorium, the original function of a pollen tube. As in the cycads, the pollen grain end of the tube advances toward the female gametophyte, breaking down and absorbing the tissue before it, so that the pollen chamber is enlarged until it reaches entirely through the nucellus, and nothing remains between the pollen tubes and the female gametophyte. Hirase?®8 traced the development of the pollen tube from its earliest stages up to the mature sperms (fig. 223). After reaching the pollen chamber, the generative cell divides, producing a stalk cell and body cell. Hrrase’s figures show the nuclei, but not differentiated cells. The blepharoplasts appear only in the body cell, where they increase in size but apparently take no part in the division, unless they may function in the orientation of the nuclear figure. In each of the two sperms formed by the division of the body cell, the blepharoplast becomes attached to the nucleus, a small portion of which seems to be attracted so that it forms a beak. The blepharo- plast is then drawn out into a spiral band, which develops hundreds of cilia. With the exception of the sperms of cycads, these are the largest swimming sperms which have ever been recorded, about 80 mi- crons in length, as estimated from Hirase’s figures. The sperms are more elongated than in the cycads, and the spiral, ciliated band, with fewer turns, is more confined to the apical region. Ginkgo and the cycads are the only known living seed plants which have retained the swimming sperm of their very remote ancestry. The female gametophyte-—The megaspore is the first cell of the female gametophyte. Although four megaspores are usually formed from the megaspore mother-cell, only the lower one develops. The four spores are formed about the first of May, the time of pollina- tion, so that the development of the pollen tube and the develop- ment of the female gametophyte start together. The megaspore enlarges rapidly, its elongated shape changing to nearly spherical. A period of free nuclear division follows, like that in the cycads, with the protoplasm, containing the nuclei, pressed in a thin sheet against the megaspore membrane by the turgidity of the fluid in the large central vacuole. After hundreds of free nuclei have been formed, wall formation begins and progresses from the periph- 2 LYE teed, e agree Weta nsys, Fic. 223.—Ginkgo biloba: A, pollen grain showing the first prothallial cell, aborting; second prothallial cell; and the mitosis which will form the generative cell and tube cell, April 24; B, later stage, same date; C, still later stage, in the pollen chamber, July 10; D, the large nucleus in C has divided, July 13; E, body cell with blepharoplasts, July 27; F, two young sperms, September 12; G, two sperms just ready to be discharged from the pollen tube, Sept. 12; A, B,C, D, E, X 600; F, X144;G, X 372.—After Hrrase.* 208 GYMNOSPERMS ery toward the center. The gametophyte soon becomes elliptical, in both longitudinal and transverse directions. In later stages it takes on a pale-green color, due to the development of chlorophyll, because the stony layer of the seed is rather thin and the outer fleshy layer is so translucent that there is enough light for the pro- duction of chlorophyll. The megaspore enlarges rapidly, its elongated shape changing to nearly spherical as the period of free nuclear Fic. 225.—Ginkgo biloba: detail of peripheral portion of female gametophyte, July 19. The mega- spore membrane is more than four times as thick as on June 5. The membrane about the female game- Fic. 224.—Ginkgo biloba: free nuclei tophyte, entirely independent of in the thin layer of protoplasm lying the megaspore membrane, has been against the megaspore membrane, June formed; X650.—After Caro- 5; X650.—After CAROTHERS.” THERS.” division begins. The first mitoses of the free nuclear stage are si- multaneous, but at the fifth mitosis there are figures in metaphase and telophase, and at the sixth, there will be some nuclei which do not divide. The free nuclear period continues from the second week in May to the first week in July, the number of free nuclei usually reach- ing more than 256 before any walls begin to be formed (fig. 224). Toward the close of the free nuclear period a delicate membrane forms on the outer surface of the thin layer of protoplasm. It is entirely distinct from the megaspore membrane, just as the walls of microspores are distinct from the wall of the microspore mother-cell. GINKGOALES 209 Cell walls now begin to be formed perpendicular to this membrane, and with their outer edges attached to it (fig. 225). As wall forma- tion progresses, the inner side of the innermost cell has no wall; but when the period of wall formation comes to a close a wall is formed on this inner side, so that each of the inner cells has its own wall. (Fig. 226). Consequently, the female gametophyte can be split apart along this central region. During the early growth of the female gametophyte the spongy tissue surrounding it encroaches upon the tissues outside it; but, Fic. 226.—Ginkgo biloba: vertical section of part of female gametophyte at line of closure, showing independent end walls of opposite cells (August 21); X325.—After CAROTHERS.” later, it is itself absorbed by the growing gametophyte and is re- duced to a mass of collapsed and deeply staining cells. Most of the upper part of the nucellus is also absorbed. Archegonium initials appear long before the wall formation in the gametophyte has reached the center. By the middle of June the two-celled neck and the central cell have been formed, and the cen- tral cell is enlarging rapidly. Usually there are only two archegonia, lying in such a plane that longitudinal sections of both can be se- cured by cutting longitudinally and parallel with the flatter face of the gametophyte. Occasionally there are three archegonia. In such cases, the gametophyte and stony layer are triangular in transverse section, and there are three bundles in the inner fleshy layer of the ovule. 210 GYMNOSPERMS The mitosis for the formation of the ventral canal cell and the egg takes place in the second week in September. A definite cell wall is \ Hii} NHN] i}! lips om | AAI W f HII Sx essa | Ni ees 4 besten PULA Fic. 227.—Ginkgo biloba: upper part of female gametophyte, showing two archegonia and the crevice-like arche- gonial chamber, also the tent-pole pro- longation of the gametophyte supporting the nucellus. The swollen ends of two pollen tubes are shown, just ready to discharge the sperms (September 9); < 24.—After Hrrase.33 formed. Consequently, the de- velopment is quite different from that in the cycads, for in none of them is a wall formed at this mi- tosis, the two nuclei lying free in a common mass of protoplasm. As far as this feature is concerned, Ginkgo is more primitive than the cycads. In the final stages of the devel- opment of the female gameto- phyte Ginkgo differs from the cy- cads. Instead of the cup-shaped archegonial chamber of the cy- ? SS QQ e a. SAS SY a a \\ y\ Li AAS nWw 4, Ns ar. , as ‘4 LAN, p, A » As Sty ait > AY i manne YHA SOLE Ve By iy f ne Pr Fic. 228.—Cycadinocarpus angustodunensis: upper part of ovule, showing striking resemblance to Ginkgo; mi, micropyle; int, integument; ar, archegonia; pc, pollen cham- ber with pollen grains; nv, nucellus; pr, female gametophyte.—After RENAULT.4°S cads, Ginkgo has a circular crevice, surrounding a mass of solid tissue upon which the nucellus rests, like a tent on a pole (fig. 227). In general topography, Ginkgo, at this stage, bears a striking re- semblance to Cycadinocarpus angustodunensis, a paleozoic seed assigned to the Cordaitales (fig. 228). GINKGOALES 211 FERTILIZATION In the cycads, at the time when the pollen tubes discharge their sperms, the archegonial chamber is moist but contains no free liquid. There is no liquid in the archegonial chamber except that discharged from the pollen tubes. This feature has been observed by so many in so many cycads that it can be regarded as a well-established fact. In Ginkgo, Htrase? reported free droplets of juice in the arche- gonial chamber, sometimes filling it completely. Personally, I have never been able to make observations upon living material at this stage; but material at the formation of the ventral canal cell, and, later, at the fusion of the egg and sperm nuclei, does not indicate anything different from the well-known cycad condition. It is very desirable that both living material and sections of well-prepared material at this stage should be investigated. The egg and sperm nuclei unite near the center of the egg. IkE- No** observed that in rare cases the wall between the ventral canal cell and the egg breaks down and the ventral canal nucleus fuses with that of the egg, a behavior which has been observed in several of the Coniferales. EMBRYOGENY After fertilization there is a period of simultaneous free nuclear division, asin the female gametophyte. The development, however, is very different, for here the nuclei are evenly distributed through- out the dense protoplasm of the egg; while, in the gametophyte, they are kept in a thin peripheral layer of protoplasm by the large central vacuole. There are usually eight free nuclear divisions, giving rise to 256 free nuclei. The number may be somewhat smaller through the failure of one or more of the nuclei to divide (fig. 229). Walls are then formed simultaneously throughout the embryo, forming cells of approximately the same size. Very little further division takes place in the micropylar region, but there is a vigorous development at the opposite end, and not much between. Conse- quently, there are three not very well-marked regions. The cells at the micropylar end elongate considerably, those in the middle en- large somewhat, while those at the base, through repeated division, become small and numerous (fig. 230). Although there are three 212 GYMNOSPERMS regions, there is no organized suspensor. The further development of the basal region is rapid, and soon there is a differentiation into stem, root, and cotyledons (fig. 231 A, B). During their earlier develop- J bE 4 Stee Fic. 229.—Ginkgo biloba: archegonium and embryo; A, top of archegonium, show- ing the two neck cells; B, transverse section of the two neck cells; C, archegonium just after the formation of the ventral canal cell (v) ; 7, nucleus of eggs; D, free nuclear stage of the embryo; /, the embryo has become cellular throughout; A and B, X 160; C, D, and £, X66.—After STRASBURGER.5% ment, one of the cotyledons is longer than the other and is notched at the tip, while the shorter one is deeply cleft, recalling the bilobed character of the foliage leaf. At maturity, the cotyledons are of nearly equal length. Mucilage canals are abundant in the stem and GINKGOALES 253 large mucilage cavities are abundant in the cotyledons. The mature embryo usually has five leaves, the first two of which are decussate with the cotyledons, while succeeding leaves in the embryo and seedling are irregular in their arrangement (fig. 231 C, D). Fic. 230.—Ginkgo biloba: embryo at stage in which three regions are recognizable. The stem and root will be organized from the small-celled tissue; * 160.—After Lyon.375 Seeds germinate readily, the terminal part of the cotyledons re- maining inside and enlarging considerably. They gradually absorb all of the female gametophyte and pass it on to the growing seedling. There is a strong tap root, and the leaves are deeply lobed (fig. 232). eocswor 2°°" tmees PP corte cces wobee See ewww es = Fic. 231.—Ginkgo biloba: A, embryo in upper part of female gametophyte, X 40; B, topography of same embryo, X10; C, embryo at about the stage shown in fig. 220, 15; D, nearly mature embryo, X 10.—From CouLTer and CHAMBERLAIN, Mor phology of Gymnosperms' (University of Chicago Press). GINKGOALES 215 Where weather conditions are about the same as in Chicago and northern Ohio, scarcely any seedlings grown in the open survive the first winter. They should be kept in the greenhouse the first winter, put out in the spring after frosts are over but while the weather is still cool, and then taken in after the first light frost in autumn. By putting them out earlier and earlier in the spring, and taking them in later and later in the autumn, for four or five years, they be- come hardened so that many will survive the winters. Even where there is some snow and ice, but not such severe cold as in the Chicago region, a consid- erable proportion of the seed- lings survive without any such precautions. PHYLOGENY How far back into geological times the Ginkgoales extend depends upon the interpreta- tion of leaves which may or may not belong to this group. It is certain that as far back as the Lower Permian it was abundant and widely distrib- uted. Baiera, some of the spe- cies of which can scarcely be Fic. 232.—Ginkgo biloba: seedlings, bear- ing a general resemblance to those of cycads, or of Quercus —From CouLTER and CHAM- BERLAIN, Morphology of Gymnosperms™4 (University of Chicago Press). excluded from the genus, Ginkgo, was abundant in the Lower Per- mian. In the Carboniferous, only leaves have been found, but they seem to belong to Ginkgoales rather than to Cordaitales. All of the order, except one species of Ginkgo, have been extinct since the Upper Jurassic, and many think that even the one species is extinct, except as it is kept alive by cultivation. 216 GYMNOSPERMS It would seem that the Ginkgoales have come from the Cordai- tales, or that both groups have come from some common ancestor. The swimming sperm is a Pteridophyte character, which has been lost by all living seed plants except the cycads and Ginkgo. However, some of the Coniferales seem to have lost the swimming habit rather recently, geologically speaking. If Ginkgoales came from Cordaitales, the Cordaitales must have had swimming sperms; for such a character, once lost, could not be regained. It was once thought that the Cycadales came from the Bennetti- tales, but such a derivation is regarded as impossible; and now both groups are generally believed to owe their origin to the Cycadofili- cales. Similarly, it seems quite possible that both Cordaitales and Ginkgoales have a common ancestry. If reproductive organs could be found associated with the Ginkgoales-like leaves of the Carbonif- erous, the problem would be simplified. The extremely fernlike leaf of the Ginkgoales would favor a direct origin from some heterosporous member of the Filicales. An origin from the Cycadofilicales seems to be supported only by the leaf gap. The profusely branching habit, the extensive develop- ment of wood, with a comparatively small amount of pith and cortex together with the simple leaf, indicate a Coniferophyte rather than a Cycadophyte alliance. With the increasing interest in paleobotany in America and the Orient, as well as in England and Europe, and with new fields for collecting being discovered, it is reasonable to hope that material will be discovered which will help to solve the difficult problems of re- lationships and the evolution of those structures upon which theories of relationship are based. CHAPTER XI CONIFEROPHYTES—CONIFERALES The description of all the previous groups has been a description of phyla which have had their origin, rise, culmination, and decline in far away geological eras, with the cycads and Ginkgo as their only surviving remnants. In the Coniferales we have a group which, although many of its members have become extinct, is still the dominant forest-maker of the world. The previous groups have belonged almost exclusively to warm climates. The Coniferales extend from the arctic to the antarctic circle, with a good representation in the tropics, and reach- ing their greatest display where the winters are so severe that the branches droop with snow and ice. The immense dinosaurs of the Triassic and Jurassic reached their maximum size, and, at its culmination, became extinct. There have been trees, some of them very large, ever since the Devonian; but none of them even approached the size of some of the gigantic coni- fers of today. Probably, without protection, the Sequozas, especially Sequoia gigantea, would be on the way to extinction. Gigantism, in both animals and plants, leads to extinction. In the Northern Hemisphere, members of the Abietaceae form immense forests. Formerly, Pinus strobus, the white pine, was a dominant lumber tree, but the forests have been cut so ruthlessly, without adequate provision for reforesting, that this lumber has be- come too expensive for most of its previous uses. Pseudotsuga taxifolia, the Douglas fir, is the principal lumber tree of western North America. The Long-Bell Lumber Company cut 2,000,000 feet of lumber a day, but such efficient reforesting is under way that there is no danger of any scarcity in the future. In Canada, 400 acres of forest are cut every week to supply paper for one newspaper, the Chicago Tribune. No other place in the world has such an interesting display of conifers as the western part of the United States. 217 218 GYMNOSPERMS Pseudotsuga taxifolia is often 190 feet high and 6 feet in diameter, with some trees 200 feet high and ro feet in diameter (fig. 233). A tree 8 feet in diameter will be about 400 years old. Fic. 233.—Pseudotsuga taxifolia: ‘“Top- ping a fir.” Tops are cut off about 30 meters from the ground, and cables are attached so that logs can be carried high in the air, thus saving young trees. This is in the forest of the Long-Bell Lumber Company, south of Seattle, Washington. Pseudotsuga macrocar pa, al- though of little interest as a lumber tree, has a very large cone and a very characteristic spreading of the branches. In Southern California it is asso- ciated with extremely large specimens of Yucca whipplei (fig. 234). Pinus ponderosa, the west- ern yellow pine, reaches a height of 61 meters and an age of 500 years. Pinus lambertiana, the sugar pine, is still larger, sometimes 60 meters tall and 2 meters in diameter. One immense tree near Calaveras Grove, in Cali- fornia, is 66 meters tall and nearly 4 meters in diameter, probably 600 years old, and the largest pine in the world (fig. 235). Abies magnifica, a magnifi- cent fir, deserves its name. It reaches 66 meters in height and an age of 700 years (fig. 236). The most remarkable of all the conifers is Sequoia, named from an Indian chief, who invented a phonetic alphabet and taught his tribe to read and write (fig. 237). There are two living species, S. sempervirens, mostly north of San Francisco and near the coast, and S. gigantea, Fic. 234.—Pseudotsuga macrocarpa: P. macrocarpa, the tree at the left, with Yucca whipplei at right. San Antonio Canyon, near Los Angeles, California. ~e > —~~ gts Se bes, Nea = cs y Fic. 252.—Taxus brevifolia: longitudinal section showing the tertiary spiral thickening. Fic. 253.—Picea nigra: vascular strand of In the tracheid at the right, a one-year stem, showing rows of phloem con- there is a single spiral; in the tinuous with rows of xylem. The section was others, there are two spiral cut from a Christmas tree. It is evident that bands. The material is from the warmth of the house had started cambial San Juan Island, in Puget activity. The outer cells of the phloem have Sound; X 433. begun to disorganize; X 750. downward, through the phloem. Phloem also serves as a storage — region. The phloem, like the xylem, is formed by the cambium. A cam- bium cell divides and one of the two resulting cells—let us say the inner one—becomes a xylem cell. The other cell is now the cambium 247 CONIFERALES cell and at its next division, the outer cell becomes a phloem cell, and the process is repeated. However, there is no such regular alterna- R iS WIS at Be ty eS WAN out iB goua= N\ \t ale \\ <5 . Fic. 274.—Taxodium imbricarium: showing broad trunk bases; near Tallahassee, Florida.—From a negative by Dr. HERMANN Kurz.338 CONIFERALES 269 are either plerome or give rise to it. However, just how definite the differentiation into periblem and plerome may be within the meriste- matic group is not so easy to determine. There is certainly no such marked differentiation as in the Capsella type among dicotyls. It was once claimed that there is no true epidermis in Pterido- phytes except at the root cap, because it was only at this point that Fic. 275.—Taxodium imbricarium: extreme swelling of the lower part of the trunk. The swollen part of the tree at the right is covered with adventitious roots. After Kurz.333 there was a segment of the apical cell giving rise to the special layer. As a matter of fact, the layer which was called periblem was still an embryonic layer which had not yet become differentiated into peri- blem and dermatogen. No doubt, conifers have an epidermis. Whether it is better to say that the epidermis comes from the peri- blem, or that an embryonic region is late in differentiating into peri- blem and dermatogen, may be a problem which different people will decide differently. At any rate, the group of initials, as one sees 270 GYMNOSPERMS it in embryo after embryo, is about as it is shown in the figure. Practically all authorities say that the root cap comes from the peri- blem, and it is obvious that there is no differentiated dermatogen or calyptrogen at the tip of the root. ie LTA aa Ain a ii aH nH Mite hie WW ind AN Ut uN uN aN a YZ Zp y 41) y SORES SSS nhs te a ot i In the root of the seedling, two protoxylem points are prevalent throughout the whole group; but in the dominant genus, Pinus, more than two protoxylem points are the rule (fig. 272). Pinus sylvestris and P. banksiana have diarch roots; in P. laricio and many CONIFERALES 271 others the roots are triarch; in and Pinus edulis, tetrarch. In ad- ventitious roots of Taxodium imbricarium, the pentarch condition is common; but even in the same cluster of roots, some are likely to be tetrach, some, triarch, and many, diarch. Fic. 277.—Pinus edulis: transverse section of a young root after secondary growth has become established. The outer part is scaling off in plates. There is a resin duct op- posite each of the four protoxylem points. The crushed protophloem is still visible, and outside it are cells rich in starch, and just beneath the scaling off layers are numerous resin cells; 118. The protoxylem points, always exarch, appear early, and there are indications of secondary development before the metaxylem 272 GYMNOSPERMS differentiates. As the regions differentiate, it is evident that there is an extensive, many-layered pericycle and a definite, one-layered endodermis (fig. 272). Often there is a resin duct opposite each pro- Se yea Set Set : ee Jr AX Ys SK . eC: SaANLOS an Oa" \ Ds . ~ ©: ie A\\ ry xe NAR CK Fic. 278.—A bies balsamea: transverse section of root, showing diarch primary xylem with a resin duct in the metaxylem. The rays in the secondary wood show the sym- metrical growth curves; X 100. toxylem point; but sometimes there is a resin duct in the metaxylem, as in Abies (fig. 278). The later stages in the structure of the root are much like those in the stem, except that the tracheids are sometimes longer and broader, and the growth-rings usually not so wide. CONIFERALES 273 The origin of the root will be considered under the heading “Embryogeny.”’ The hypocotyl.—In the young seedling the hypocotyl is the most prominent and most extensive part of the plant. In its lower part the structure is much like that of the root, with phloem groups al- 86 ~ me 38 r< ) 1 Lv, \ >< ~ a ; oS ee) < p—: a y y 7) bh —<4 y . Z L] Soe < So. OS Se, KH on ' Sar, ArH ) aibeat gs AY | OTS sun x YY rs oF ae CUS a y) ek eS aan So) a SY 2S ie CU Ne CY THY SOR Repay mQ-eo = at O Orie Calm, GOUlcotor MeO ec SIO RT, ON aE ee AH ae pe sanestueuesssegneeees ; SG e. oa aS Fic. 279.—Pinus laricio: transverse section of hypocotyl, showing four protoxylem points, with four alternating groups of phloem. There are also two resin ducts, not well differentiated at this stage. Two stomata are shown in the epidermis; X 160. ternating with the exarch bundles in typical radial arrangement (fig. 279). But there is no scaling off at the outside, and stomata are present, although not so abundant as in the leaves and young stem. Higher up in the hypocotyl the bundles begin to rotate, making a half-turn, so that near the cotyledons a transverse section shows an 274 GYMNOSPERMS endarch protoxylem. There is no centripetal xylem at any time, the change from an exarch to anendarch position of the protoxylem being due entirely to rotation. The stomata on the hypocotyl of Pinus edulis seem to be a little different from those on other parts of the plant; and those on the cotyledons and leaves are so different from each other that, taken separately, they might seem to belong to different species, or even different genera. The hypocotyl of conifers would seem to be a good place for a thorough study of comparative anatomy. CHAPTER XII CONIFEROPHYTES—CONIFERALES—Continued THE SPOROPHYTE—REPRODUCTIVE The cone is the most characteristic feature of the order and the one to which it owes its name. The staminate strobili are conelike throughout, but the Taxares (Taxads and Podocarps) have some- times been described as conifers without cones because the female strobili do not look like the familiar cones of the other families. The separation of the sexes shows all conditions from bisporan- giate strobili to monoecism and to a complete separation in typical dioecism. The monoecious condition is dominant. Bisporangiate strobili occur only as occasional abnormalities. The Abietaceae and Taxodiaceae are uniformly monoecious. In the Cupressaceae, Fstzroya, Diselma, and Arceuthos are dioecious, and Juniperus communis, usually dioecious, occasionally bears both male and female strobili on the same plant; but, in such cases, either the male or the female is dominant, recalling the condition in Canna- bis sativa and some other angiosperms, which are evidently in transi- tion from monoecism to complete dioecism. The other genera of Cu- pressaceae are monoecious. The Araucariaceae are dominantly dioe- cious, but Agathis australis and Araucaria bidwilli are monoecious. The Taxaceae are dioecious, except in occasional individual cases. In the Podocarpaceae the dominant genus, Podocar pus, is dioecious, although there have been reports—not very convincing—of occa- sional monoecism. Some species of Phyllocladus and Dacrydium are monoecious. The survey shows that the great majority of the genera are monoe- cious, a much smaller number having attained the dioecious condi- tion. In the plant kingdom, from the algae to the highest dicotyls, the course of evolution shows a progressive separation of the sexes, with such intermediates and mixtures that the dioecious condition is evidently the goal. As far as this single feature is concerned, the 275 Fic. 280.—F ’inus contorta: tip of lateral branch showing staminate cones just ready to shed their pollen; an ovulate cone ready for pollination at the tip of the branch; ovulate cone a year older, a little farther down; and, near the bottom, an ovulate cone two years older. Puget Sound Biological Station, Friday Harbor, Washington; June 20, 1929. CONIFERALES 277 dioecious species are later developments than the monoecious; or, rather, they have progressed farther along this line of evolution; but dioecism may be associated with very primitive features. Conse- quently, one could hardly claim that a monoecious family, as a whole, is more primitive than a dioecious family. It is simply more primitive in this one respect. Monoecism, the prevalent condition in the order, is well illustrated in Pinus, the dominant genus of the Northern Hemisphere (fig. 280). This figure shows staminate cones a day or two before shedding their pollen, and ovulate cones of three successive seasons; at the top, a couple of cones just ready for pollination; lower down, an ovulate cone which was pollinated the previous year and which would show young embryos; and near the bottom, an ovulate cone which was pollinated two years earlier than the one at the top, and has shed its seeds. The ovules in Pinus contorta are pollinated in June, the eggs are fertilized the next June, the embryos complete their develop- ment during the summer, and the ripe seeds are shed the following summer. In some pines, like Pinus radiata, the cones remain tightly closed for years and shed the seeds only when there is a fire or an unusually hot, dry season. Pines with hard cones usually shed their seeds the third season, like Pinus contorta, P. banksiana, and many others. BISPORANGIATE STROBILI Bisporangiate strobili are always described as teratological. In the evolution of sex there is a constant tendency to wider and wider separation, so that the theoretical series would be bisporangiate stro- bili, monosporangiate strobili with both sexes on the same plant (monoecism), and, finally, monosporangiate strobili on different plants (dioecism). In the Cycadophytes bisporangiate strobili are common and nor- mal in the Bennettitales, but are unknown in Cycadales, all of which are not only monosporangiate, but are strictly dioecious. In the Co- niferophytes nothing is known of conditions which preceded the Cordaitales, and even in the Cordaitales themselves it can only be said that there no bisporangiate strobili have been found. Gink- goales are monosporangiate and dioecious. 278 GYMNOSPERMS But in the Coniferales bisporangiate strobili occur in many of the genera, and in some they occur so frequently that most botanists who study plants in the field have seen these interesting cones in one form or another. Bisporangiate strobili are most abundant in the Abietaceae, but have been found in all the families except the Taxaceae. The most frequently reported examples are in Picea excelsa, Abies alba, Pinus laricio, and P. maritima; but they have been found in other species of Pinus and Picea, and also in Pseudotsuga taxifolia, Sequoia sempervirens, Juniperus communis, and others. Such cones occur often in cultivated specimens of Picea excelsa, where the ovulate sporophylls are at the top, with the stami- nate below (fig. 281). In Picea the cone axis is long, fleshy, Fic. 281.—Picea excelsa: bisporangiate weak, and turgid. The stami- strobili. Below are staminate sporophylls, nate sporophylls bear one or two each bearing one or two sporangia; above : more or less globular sporangia are the ovuliferous structures, most of the 8 EB 81a, scales bearing two ovules which seem tobe CONtalning some apparently normal. At the base of the ovulate part are good pollen, some dead and a few bracts which have neither ovules nor hj veled. and some in which de- stamens; natural size. : velopment seems to have been arrested at various stages (fig. 282). The bracts at the base of the ovulate part bear neither microsporangia nor megasporangia. GoEBEL?” found hundreds of bisporangiate strobili on a single tree of Pinus maritima (fig. 283). They bear microsporangia below, megasporangia at the top and, between the two, have sporophylls with microsporangia on the abaxial side, and rudimentary ovulifer- ous scales in the axils. A tree of Pinus thunbergia in the Botanical Garden at Tokyo, Ja- pan, bears cones with microsporophylls and megasporophylls inter- mixed from the bottom to the top. STRASBURGER™ observed, in Pinus laricio, branches with stami- CONIFERALES 279 nate cones at the base and ovulate cones at the top, while the part of the branch between bore both kinds of cones and also bisporan- giate cones. The latter had staminate sporophylls at the base, ovu- late sporophylls at the top, with sporophylls (bracts) between bear- ing neither kind of sporangia. In most bisporangiate cones the ovulate sporophylls are at the top with the staminate below; but cases have been reported in which the arrangement was just the opposite. (i ‘ ' ‘ ‘ ‘ ’ ‘ ‘ ‘ ' i) ‘ ’ ‘ ' ‘ Ces yey Cre on) CoO . Fic. 282.—Picea excelsa: A, sections of microsporangia from lower part of a stro- bilus like those shown in fig. 281; B, section of a pollen grain which seems to be normal. The material was merely wrapped in paper and sent from a considerable distance; hence the shrunken protoplasm. A, X20; B, X 300. An abnormal cone of Pinus sp. was once sent to our laboratory as a curiosity. On one side, from the bottom to the top, it looked like a normal cone; while the other side bore spur shoots with two some- what stunted needles. Pines have been noted in which some of the ovulate structures of the cone were replaced by spurs with two needles. Admixtures of male and female characters occur even in dioecious angiosperms. In Salix petiolaris'® the staminate sporophyll may be tipped with a stigma, and carpels, bearing ovules near their bases, may bear microsporangia higher up; and still higher up, may be tipped with good stigmas. 280 GYMNOSPERMS What causes the changes in sex characters has not been deter- mined for either gymnosperms or angiosperms. These peculiarities have been noted more frequently in exotic individuals; but whether the conditions are more likely to occur in such cases, or such cases are more likely to be noticed, might be questionable. It has been claimed that abnormalities occur more fre- quently in nursery material: it is certain that they would be more likely to be noticed. Comparatively little histological work has been done in this field, nearly all the ac- counts dealing with mature cones. It would not be impossible, or even very difficult, to collect a series in various stages of develop- ment, for a tree which produces teratolog- ical cones produces them year after year. In the most frequent teratological cases the arrangement of staminate and ovulate Fic. 283.—Pinus mari- | Sporophylls on the cone axis resembles a fre- tima: bisporangiate strobi- quent arrangement of staminate and ovu- mek price: po late strobili on a tree; for the ovulate cones microsporangia; the next are higher up, with the staminate lower sporophyll above, on each down, as in Abies. In Pinus it is very rare side, has a microsporan- to see a staminate cone at the top of a tree gium in the lower part and : 5 ia a rudimentary ovuliferous OF tip of a branch, while this is the usual scale in the axil; above place for ovulate cones. In seasons in which these are ordinary ovulif- scarcely any cones are produced, two or erous structures.—After e aeag three ovulate cones can usually be found at the top of the vertical shoot at the extreme top of the tree. In Saxagothea the ovulate cone is terminal on a short branch, which bears staminate cones farther back in the axils of its leaves. In general, the ovulate cones are more numerous at the top of the tree or at the ends of the branches, with staminate cones far- ther down or farther back on the branch, suggesting the arrange- ment in the bisporangiate angiosperm flower. THE STAMINATE STROBILUS Throughout the entire order, the staminate strobilus is simple, the sporophylls being borne directly upon the cone axis. There are no CONIFERALES 281 bracts; consequently, the staminate strobilus is a flower, not an in- florescence. The ancient discussion as to whether the ovulate stro- bilus is a flower or an inflorescence will be considered later. The cones vary greatly in size. The cone of Juniperus communis, 2 mm. in length, with nearly mature pollen, but before the rapid lengthening which allows the pollen to be shed freely, is very small; while the cone of Araucaria bidwilli, 10 cm. in length, is extremely large. It is claimed that this cone reaches a length of 20 cm. Dr. GEORGE GRAVES of Fresno, California, where there are scores of fine large trees of this Australian species, has collected a few cones 12.5 cm. in length, but has never seen any approaching the reputed 20 cm. In Araucaria rulei it is reported that cones reach a length of 24 cm. In Araucaria cun- ninghami the cones havea greater diameter and reach a length of 7 cm. before the rapid lengthening begins. They occurinimmense numbers and produce a prodigious amount of pollen (fig. 286). It would not be exag- gerating to estimate the output of a single cone at 10,000,000 pollen grains. The origin of sporophylls from the meri- stem is dominantly spiral, the Cupressaceae being the only family which shows the cy- clic arrangement throughout. The arrange- ment of sporophylls, however, is so geo- metrically regular that they often seem to be in vertical rows, as they would be if the arrangement were cyclic (figs. 284 and 285). Fic. 284.—Pinus contorta: shoot with a large number of axillary staminate cones. The axis of the shoot is prolonged beyond the cones, and, at the base, shows pairs of needles nearly covered by _ bracts; higher up, the needles are en- tirely covered by the bracts. Although the sporophylls ap- pear to be in vertical rows, their arrangement on the cone is strictly spiral; natural size. The sporophylls, like the cones, vary greatly in size (fig. 287). The figure shows typical sporophylls of ten genera. Some of the pines have smaller sporophylls than Pinus laricio; and it is possible that some species of Avaucaria may have larger sporophylls than A. cun- ninghamz; but the figure, with all the sporophylls drawn to the same 282 GYMNOSPERMS scale and magnified 8 diameters, shows the general range of size and appearance. The leafy character is quite pronounced in Araucaria cunning- hami and Picea excelsa, and in Dacrydium elatum it is scarcely dis- tinguishable from the foliage leaf; while in many it is more reduced, and in some there is no more resemblance to the original leafy blade than in the cycad, Zamia. The sporangia are borne on the abaxial face of the sporophyll. However, HAGERuP”® claims that in Dacrydium elatum, a central Sumatran conifer, the sporangia are on the “upper” sur- face. He says that the ovules also are on the up- per face of the sporophyll, and he regards both male and female sporophylls as homologous with those of Lycopodiales. The dominant number of sporangia is two, but many conifers have more (fig. 287). The fig- ure shows several genera with two sporangia, and several with more. The number is 2 in all of Abietaceae; 2-5 in the Taxodiaceae; 2-6 in the Cupressaceae; and many more in the Araucaria- ceae. Araucaria cunninghamia has 13 and they are very large. Agathis australis, the New Zealand Kauri, has 5-15 sporangia on a sporophyll; and in A. bornensis the number of sporangia often reaches 15. Some of the Araucarian sporangia reach a length of 7 mm., and a length of 4-6 mm. Fic. 285.—Podo- 1S not at all rare. carpus with three In Taxus the sporophylls, each bearing, usual- baile eae ly, 6 sporangia, but sometimes as many as 8, are peltate, with the sporangia hanging down, as in Equisetum (fig. 288). In Torreya the condition is similar, but there are only four sporangia borne on one half of the nearly peltate top. CoULTER and LAND’? found, in the sterile half of the top, resin canals which might rep- resent the missing sporangia. In early stages the resin canals look very much like young sporangia. CONIFERALES 283 In the Podocarpaceae there are two sporangia. The sporophylls are generally small, but are often very numerous, in long, slender cones. Fic. 286.—Araucaria cunninghami: a spray of staminate strobili, at Rockhampton, Queensland, Australia; about one-half natural size. The strobili are from 5 to 8 centi- meters in length. The dehiscence is longitudinal, as in Pinus, in the greater number of cases, but often transverse, asin Abies. In some cases it is oblique, as in Picea. Microsporogenesis——The development of the microsporangium and its microspores is the usual development of a eusporangiate spo- rangium, and, in the earlier stages, does not differ much from that 284 GYMNOSPERMS of Selaginella, or even from that of a eusporangiate homosporous form, like Lycopodium. There is a hypodermal cell, or layer of cells, which may be called the archesporium. A periclinal division in the F 1G. 287.—Male sporophylls of conifers: A, Pinus laricio; B, Abies grandis; C, Picea alba; D, Cedrus libani; E, Juniperus communis; F, Cryptomeria japonica; G, Cupressus macrocarpa; H, Agathis australis; I, J, and K, Araucaria cunninghami; I, view from upper surface; J, view from lower surface; K, transverse section showing 13 sporangia in two rows; all 8. archesporium gives rise to two cells or two series of cells, the outer forming the parietal, or wall, layers, and the inner giving rise to the sporogenous tissue. CONIFERALES 285 It has been shown in so many cases that the tapetum, as far as it is derived from either of these series, consists of modified wall cells, that any claim of a sporogenous origin would have to be strongly supported. Of course, much of the tapetum does not come from either of these sources, but is differentiated from whatever cells hap- pen to be in contact with the sporogenous cells. The development of the sporangium in most conifers is very slow, the young cone becoming recognizable in the spring, the archesporium ap- pearing in early summer, and the development of sporoge- nous tissue continuing until the autumn weather becomes too cold, even for a conifer. During the earlier stages of development the cones are entirely covered by scale leaves (fig. 289). The actual dates could not be given, even for a single species, ex- Fic. 288.—Taxus baccata: A, staminate cept in some special locality; shoot with numerous strobili; B, single stami- for, Paekiate for a special nate strobilus; C, ovulate shoot with two ; ovules; A and C, natural size; B X2. stage, like the appearance of the archesporium, for the appearance of the microspore mother cell, for the reduction of chromosomes, and for the shedding of pollen, will vary with the latitude, the elevation, and, in some cases, with the prox- imity of warm or cold ocean currents. Many conifers are very success- ful exotics. The times of shedding pollen of Cupressus macrocarpa at Monterey, California, and-at Auckland, New Zealand, differ by six months. Since this naturally endemic conifer thrives from the lati- tude of Monterey to that of Auckland, the pollen is doubtless shed at various times, and other stages would vary correspondingly. How- ever, in any particular locality, the same stages will appear at about the same time, year after year. In the Chicago region, some conifers pass the winter in the micro- 286 GYMNOSPERMS spore-mother-cell stage (fig. 290). This was easily determined by comparing late autumn, mid-winter, and early spring conditions. In Juniperus virginiana, in the same locality, the mother cells divide, Fic. 289.—Pinus banksiana: young staminate cone with sporangia in early sporog- enous stages at the bottom; at the top, archesporial cells have not yet appeared; near Chicago (September 10, 1931); X45. the exine and intine of the microspores are developed, and the pollen is nearly ready for shedding when winter arrives; but in Juniperus communis, growing close to the other species, the winter is passed in the mother-cell stage. CONIFERALES 287 The reduction of chromosomes.—The time at which reduction oc- curs varies with the species and the locality, and may differ a few days in different years. In the Chicago region, in Pinus laricio, re- duction occurs the first week in May. Dr. MARGARET FERGUSON,’ 28 Opa eo aoe z Ras a0 Sca 6 S Fic. 290.—A, Pinus laricio, microsporangium on October 1; B, the same on Janu- ary 3; C, the same on April 4; D, Taxus canadensis on October 1, showing microspore— mother-cell stage —After CHAMBERLAIN.' whose work on Pinus is the most complete account of the life-history of any genus of the conifers, gives about the same date for reduction in Pinus laricio (austriaca), P. rigida, and P. strobus. Her collec- tions were made in the vicinity of Wellesley, Massachusetts. Hor- MEISTER,’ in Germany, found that in Picea and Abies the pollen 288 GYMNOSPERMS mother-cell stage is reached by the end of November. So they pass the winter in the pollen mother-cell stage. In all of these localities the autumn and winter conditions are ap- proximately the same. In warmer climates the dates would doubt- less be different. We have seen a species of Pinus near Jalapa, in Mexico, shedding its pollen in September. It would be interesting to know the life-history of such a species, with dates for the various FG. 291.—Pinus laricio: reduction of chromosomes in pollen mother-cells, showing stages from the telophase of the first reduction division to young microspores;'s Chi- cago (May 3); X 500. stages. In Araucaria bidwilli, at Fresno, California, where there are many large, luxuriant specimens, reduction takes place the last week in March. In an angiosperm like Lilium, with elongated anthers, a longi- tudinal section at the reduction period shows simultaneous mitosis; but stages at the top, middle, and bottom may be quite different. In Pinus, a section of the sporangium, in any direction, shows as wide a range of stages as are found from the top to the bottom of an anther of Lilium (fig. 291). In an extremely long microsporangium, like that of Araucaria, stages at the top and bottom would probably be different. The number of chromosomes have been counted in several genera. CONIFERALES 280 If any numbers can be said to be dominant, they are 12 and 24 for the x and 2x phases of the life-history. At least 8 species of Pinus, 5 species of Larix, and 2 species of Podocarpus have these numbers. GoopDsPEED™ found 12 and 24 in Sequoia gigantea; but LAwson3“4 found 16 and 32in S. sempervirens. In Abies balsamea?*5 the numbers are 16 and 32; in Sciadopitys verticillata,3° in Cephalotaxus drupa- cea,*° to and 20; and in Taxus baccata,™ 8 and 16. The lowest num- bers, 6 and 12, are reported by SAxTon‘™ for Callitris cupressoides. THE OVULATE STROBILUS Throughout, we have been using the terms “‘strobilus” and “‘cone”’ almost synonymously. All cones are strobili; but not all strobili are cones. The heading, the ovulate strobilus, is more appropriate than the ovulate cone, because it includes the podocarps and taxads, some of whose ovulate structures cannot be called cones, although we do not hesitate to call them strobili. The spore-producing structure of Lycopodium lucidulum is a strobilus, but not a cone. However, it rep- resents a more primitive condition from which the typical cone of Lycopodium clavatum could have been developed, with forms like L. inundatum as intermediates. The term “cone” is shorter, and throughout this work we have used it often, but only when either term, ‘‘cone” or “‘strobilus,” could be employed; and we have used the more comprehensive term, “‘strobilus,’’ when the term “cone,”’ would have been appropriate. In striking contrast with the simple staminate strobilus, the ovu- late strobilus of the Coniferales is compound. The ovule-bearing structures are not borne directly upon the cone axis, as in the stami- nate cone, but in most cases are borne upon a much-discussed struc- ture associated with a bract. This bract is borne directly upon the main axis, and is homologous with the male sporophyll. As in the staminate cone, the parts are formed in spiral succession, except in the Cupressaceae and a few scattered genera in other families. Except in the Taxaceae and some of the Podocarpaceae, the ovu- late strobili are cones, varying in size and appearance from the typi- cal cones of Pinus to the small berry-like cones of Juniperus and plumlike ovules of Torreya (figs. 292-06). Pinus and Araucaria have the largest cones. The longest ovulate 290 GYMNOSPERMS cone in conifers is that of Pinus lambertiana. Supwortn’s illustra- tion was made from a cone 23% inches (nearly 60 cm.) in length, and FIG. 292.—Pinus radiata: (Monterey Pine), a very endemic California species; ovu- late cone, natural size.—After SupwortH. cones from 40 to 50 cm. in length are not at all rare. These immense cones, hanging down from the extreme tips of the branches, can be seen at a distance of a quarter of a mile, and give the tree a charac- teristic appearance (fig. 235). The cone of Pinus coulteri, reaching a CONIFERALES 291 length of 23-35 cm., has a greater diameter, and its average weight is probably greater than that of P. lambertiana. Another western pine, Pinus sabiniana, has a cone from 16 to 26 cm. in length, nearly globular in form, and very heavy. In Araucaria bidwillt the cone reaches a diameter of 30 cm. Since the cones in this family are nearly spherical, this cone may be heavier than any of the longer cones of Fic. 293.—Pseudotsuga taxifolia: ovulate cone; natural size Pinus. In other species of Araucaria and in Agathis the cones are smaller, the cones of the immense Agathis australis, the Kauri of New Zealand, being only 6 cm. in diameter. In Sequoia gigantea, the largest of all trees, the cones are only from 4—6 cm. in length; and in S. sempervivens, the tallest of conifers, reaching a height of 115 me- ters, the cones are even smaller, seldom more than 1.5 cm. in length. The cones of Tsuga canadensis and Larix laricina are also about 1.5 cm. in length. The berry-like cones of Juniperus are still smaller, that of J. com- munis having a diameter of 6-8 mm. In Taxus the seed, including the aril, is from 9 to 13 mm. in diameter. 292 GYMNOSPERMS Between the extremes there are cones of all sizes; of the more or less elongated type, like Pinus, of the more or less globular type, like Cupressus; and of the berry-like type, as in Juniperus; and, besides, there are the ovulate structures, generally small, of the taxads and some of the podocarps, which are responsible for the appellation, ‘“‘conifers without cones.” Structure of the ovulate strobilus—In the Abietaceae the ovulate structures are more or less elongated cones; in the Taxodiaceae the cones range from the elon- ; gated type, asin Sciadopitys wat te Cue and Sequoia, to the spher- nntoralence: ical type, as in Taxodium; in the Cupressaceae and Araucariaceae the spherical type dominates; in the Podocarpaceae such strobili as are called cones are slightly elongated, as in Microcachrys and Saxagothea; while such structures as those of Podo- Fic. 295.—Junip- carpus, Dacrydium, and all of the ¢™"S communis: Taxaceae, make the designation berry-tke cones; natural size. “conifers without cones” seem ap- propriate. However, even in these cases, we believe the ovuliferous structures can be interpreted as very much reduced cones. If one could interpret the ovulate cone of Pinus, the rest of the order would not make much trouble. It seems safe to interpret the so-called “‘bract’”’ as the homologue of the sporophyll of the staminate cone; Fic. 296.— but the structure which bears the ovule is not so easy to Bali oe interpret. The appearance of the much discussed and Gola! variously interpreted structure is shown in fig. 297. At this stage the cone bears striking resemblance to the cone of Lycopodium, the debatable structure resembling the young sporangium of the lycopod. But, in a conifer, this structure is certainly not an ovule. Generally, it seems to be axillary; but oc- CONIFERALES 293 casionally it seems to be borne on the face of the sporophyll. What is its homology? oo ae Fic. 298 Fic. 297.—Pinus banksiana: young ovulate cone still inclosed within the protecting bud scales: 6, the bract, or sporophyll; s, the structure which bears the ovule; X 20. Fic. 298.—Pinus banksiana: longitudinal section of young vegetative shoot, show- ing bracts (b) and spurs (s). The spur (s) on the right has not yet begun to develop scale leaves; the one opposite, on the left, is developing scale leaves; all of the spurs be- low show not only scale leaves, but the two needles. The drawing is somewhat dia- grammatic, since so many spurs would not be shown in median view in a single thin section; X 20. As one traces the evolution of the sporophyte from forms like Riccia, where nearly all of the sporophyte consists of spores, there is strong support for the theory that tracing the evolution is tracing an increasing amount of sterilization of sporogenous tissue. At the ly- 204 GYMNOSPERMS copod level the amount of sporogenous tissue is not only greatly re- stricted, but it appears much later in the individual life-history than in Riccia. In the conifers the comparative amount of sporogenous tissue is still more restricted, and it appears still later in the indi- vidual life-history. Whatever may have been the origin of sporogenous tissue in the liverworts, or, even earlier, in the algae, the spore-bearing structures, from the lycopods up, are modifications of vegetative structures. In Lycopodium lucidulum the sporo- phyllsare like the vegetative leaves, and practically all the leaves are sporophylls; in L. inundatum the spore-bearing leaves are somewhat modified, are confined to the up- per part of the shoot, and are grouped into a loose strobilus. In lycopods of the L. clavatum type the sporophylls are quite different from the vegetative leaves, and are grouped into a compact cone. The cones of conifers, both ovu- late and staminate, frequently pro- liferate, changing from the repro- Fic. 299.—Proliferating cones: A, ductive to the vegetative phase (fig. Larix heterophylla; B, Cryptomeria ja- - - Aone cate aed 299). The proliferating branch may again bear cones. If an ovu- late cone, like that of Pinus, is a modified shoot, the modification has been extreme (fig. 298). At the top of the figure, the young spur (s) looks much like the debatable structure (s) in fig. 297. Histologi- cally, the young spur shoot and the debatable ovuliferous structure look alike (fig. 300). The figure, although drawn from a young ovu- late cone, might pass for a young vegetative spur in the axil of its bract; but, later, the young spur produces scale leaves, and finally a pair of needle leaves, while, in the young cone, the axillary struc- ture produces no scale leaves, but bears two ovules on the face di- rected toward the axis of the cone. The structure has been called an ovuliferous scale, a flattened CONIFERALES 295 branch in the axil of the bract; it has been called an open carpel, a placenta, a ligule, the blended integuments of two ovules; it has been called a leaf of an axillary shoot, the first two leaves of an axil- lary shoot fused by their margins; and if any possible structures have been omitted from this list, it may be assumed that someone has applied them to the ovule-bearing structure. WORSDELL,”” in 1900, collected and discussed the literature, and practically nothing has been added since. In all of the investigations and philosophizing great stress has : & been laid upon abnormal cones, TABOO ES where various intergrades between alagtetae reproductive and vegetative struc- ae tures have been found. Nearly all of these abnormal cones have been studied only in the mature condi- tion. Since trees which produce abnormal cones produce them year after year, material of young stages could be collected, and it would be interesting to make a comparative study of the development of such cones throughout the order, com- Fic. 300.—Pinus banksiana: part of paring especially the conesofforms Young ovulate cone, showing bract (0), : ; with the “ovuliferous scale” (s), in its with and without spur shoots; , ance and comparing those with bract and ovuliferous scale free, as in Abietaceae and Podocarpaceae, with those which have the bract and ovuliferous scale “fused,” as in Taxodiaceae and Cupressaceae. On account of the peculiar relations of the ovuliferous structures, a study of young proliferating cones in Araucariaceae should be particularly interesting. A comparative study of young stages in the ovuliferous structures of the Taxares— Podocarpaceae and Taxaceae—would also be valuable, since it might reveal the presence of lost structures, the absence of which started the designation of ‘“‘conifers without cones.” For many botanists, any structure in the axil of a leaf must be a 296 GYMNOSPERMS shoot, and so a leaf in the axil of another leaf is an impossibility. Consequently the ovuliferous scale could not be a leaf, unless some shoot could be so reduced as to become present only theoretically. In a transverse section through the bract and ovuliferous scale, the bundles of scale and bract show a reversed orientation, the xylem of the bundles facing each other, with the phloem outside; and this is true, whether the bract and scale are free from each other or “fused” into one structure. But the bundles of the bract connect below with the vascular bundles of the cone axis, while the bundle of the ovulif- erous scale connects above; consequently, with this connection, the orientation is not at all peculiar, but only what should be antici- pated. From a study of the bundles, VAN TrEGHEM*? drew the con- clusion that the ovuliferous scale is a leaf on a suppressed branch, a conclusion which was later strengthened by the bundle situation in the ‘‘double” leaves of Sciadopitys. Some claim that the cones of Araucariaceae are quite different from those of the preceding families. The cone of Araucaria looks as if it had a bract and scale, fused as in the Cupressaceae; and Dr. Hannau AasE,' from a study of the vascular anatomy, is inclined to believe that the structure is compound. THomson,"* also studying the anatomy, concludes that there is only a simple sporophyll, so that the cone is simple, like the staminate cone. In the other genus, Agathis, there is no such appearance of bract and scale. Taxonomists separate the two genera on the presence (Araucaria) or absence (Agathis) of a “ligule.’’ The ligule of Araucaria has the position of an ovuliferous scale, and looks like one (fig. 301). At present, we prefer to use the term “ovuliferous scale.” Until some decisive proof of some theory is produced, we shall continue to believe that the bract of the ovuliferous cone is the homologue of the sporophyll of the staminate cone, and shall guess that the ovulif- erous scale—at least in forms with bract and scale—is a modified shoot which, with or without leaves, bears the ovules. Unless some entirely new theory, different from all of these, can be proved, the ovulate cone is compound, and, therefore, is not a flower, but an inflorescence. The megasporangium.—All of the megasporangia, commonly called “ovules,” of the conifers are borne in strobili, which are defi- CONIFERALES 297 nitely organized as cones, except in the taxads and some of the podo- carps, where it is possible that a cone is present, but so reduced or modified that botanists fail to recognize it. There is a single integument. Only by interpreting the epimatium, the aril, or the ovuliferous scale as an integument, can there be two integuments. We should interpret the epimatium as an ovuliferous scale. The nucellus is free from the ovule, at least in early stages of development. Where the two structures are not free from each other, there is said to be a union or fusion. Probably there is not a single case of fusion in the order, but, rather, the same phenomenon appears Fic. 301.—Araucaria bidwilli: the ovuliferous structures: 6, bract; /, the “ligule”’; 0, young ovule in the spore—mother-cell stage; 7, integument; 7. which gives rise to perigyny and epigyny in angiosperms. A growth below the free portions carries them up, and, throughout this region of common growth, the nucellus and integument are said to be united. In this sense, there is considerable union in the Abietaceae, _ where nucellus, above the level of the megaspore, is free from the in- tegument. In the Taxodiaceae there is some union at the base. In the Cupressaceae the two are free, or are more or less united at the base. In the Araucariaceae the nucellus and integument are very free, in Agathis australis the ovule even being stipitate. In Taxus the nucellus and integument are free, even in later stages; but in Torreya, chalazal growth is so extensive that, in later stages, only a small part of the nucellus is free. Ovules are orthotropous, as in Taxus; or anatropous, as in Podo- carpus; or may have intermediate positions. Approximately half of 298 GYMNOSPERMS the genera have more or less anatropous ovules, while, in the other half, they are more or less orthotropous. The free condition is regarded as primitive, and the more or less “united” condition as more advanced. It must be remembered that both conditions existed in the early gymnosperms of the Carbonifer- ous. In Trigonocar pus the nucellus was entirely free from the integ- ument, while, in the best-known of all paleozoic seeds, Lyginopteris (Lagenostoma), the “union”’ was almost complete. JE eapeces se Negeweee OT AAH ES Bera x \\ J) oO HY) é E = sap he Ze ih A) Mgt ‘en ZO W Hf Wy) Up ) ly | Y, T] ° J Y) YH] yy Sao oes Fic. 302.—Pinus banksiana: bract; ovuliferous scale; and ovule. The bract and scale, in such a condition, are said to be free from each other; X 25. Nowhere in the order is there any extensive vascular system in the ovule, like that of the cycads. In most cases, there are no bundles at all in the integument, or in the aril, where this structure is present. In Microcachrys and Saxagothea there are bundles at the base of the integument, and, in Podocarpus, bundles sometimes extend almost to the top of the integument. While there is not such a differentiation of the integument into strongly marked regions, as in the cycads, three regions are often distinguishable, a middle layer which becomes hard, with a fleshy layer on each side. The outer layer may remain fleshy for a longer or shorter time, while the inner, fleshy layer always becomes dry and membranous. The megasporangium, throughout the order, is eusporangiate and massive, and shows great diversity in its appearance (figs. 302-6). CONIFERALES 299 The relation of the ovule to associated structures can be seen more easily than it can be interpreted. In fig. 303 C, in this case, LAw- ro hte nee oe pEee H Fic. 303.—Ovules of conifers: A, Sequoia sempervirens; B, Pherosphaera; s, sporo- phyll; C, Microcachrys; D, Saxagothea cons picua; b, bract; 0, ovuliferous scale; 7, integu- ment; ”, nucellus; A, X95; B, X33; C, X66; D, X25.—A,344 B,35* and C,35° after Law- son; D, after NoREN.4"9 sons claims that the ovule is borne on the face of a sporophyll. The structure which we have marked 3, he regards as the sporophyll, while the one which we have designated as the ovuliferous scale (0), 300 GYMNOSPERMS he regards as a second integument. In Pherosphaeras (fig. 303 B) LAWSON says the single erect ovule is borne on the sporophyll, close to the axis of the cone. There does not seem to be any structure D Fic. 304.—Ovules of conifers: A and B, Taxus baccata—A, at time of pollination and B, somewhat later: 7, integument; », nucellus; a, aril; X15. C, Cryptomeria japonica: b, bract; 0, ovuliferous scale; 7, integument; n, nucellus; X22. D, Thuja occidentalis: i, in- tegument; m, nucellus; X35.—All after HaGervup.?% which could be called an ovu- liferous scale. In Saxagothea*® (fig. 303 D), the ovuliferous scale is comparatively small, and the nucellus protrudes from the micropyle and is more or less glandular, like a stigma. In Microcachrys,3° also, the ovuliferous scale, in early stages, is rather small (fig. 303C). Dacrydium elatum, a Suma- tran species recently described by HaceErvp, seems to bear the ovule on the adaxial face of a sporophyll, which is the homologue of the male sporo- phyll (fig. 305). HacEerup® claims that both microspo- rangia and megasporangia are borne on the upper (adaxial) face of sporophylls, just as sporangia are borne in lyco- pods. The “‘cone” consists of only two sporophylls, and, be- tween them, the sterile tip of the branch which sometimes proliferates. The fact that Hacerup describes and figures an “epimatium” may mean that there is something representing the ovuliferous scale. A detailed study of the development would be interesting. When ovules are terminal on the axis, as in Taxus, there is no sporophyll or ovuliferous scale to confuse the interpretation (fig. 304 A, B). However, in this case, there is an aril, which many regard CONIFERALES 301 as an integument. It can be distinguished at the time of pollination, and, in late stages, becomes red and fleshy, giving the seed its berry- like appearance. In the Podocarpaceae the epimatium is a striking feature of the ovule (fig. 306). It starts like an integument and looks like one. Below it, is the receptacle, which be- comes very large and fleshy. The aril of Taxus and the outer fleshy covering in Torreya start in the same way. In Guetum a delayed integument also starts in this way. Some regard all of these structures as tegumentary Fic. 305.—Dacrydium elatum: A, tip of twig with ovuliferous structures covered by leaves; B, the same with leaves dissected Fic. 306.—Podocarpus: ovule and away, showing two leaves, each bearing an epimatium. The female gametophyte ovule on its inner (adaxial) face; X6.— in an early free nuclear stage. e, After HAGERuP.”% epimatium; 7, receptacle; X10. in their origin, but the epimatium may be homologous with the ovu- liferous scale, rather than with the integument. The megaspore.—In many cases the megaspore mother-cell be- comes recognizable so late in the development of the ovule, and is so deeply placed, that it is impossible to determine just what the origin may have been (fig. 307). Ina species like Pinus, where the mega- spore mother-cell is so deeply placed, one can imagine a row of cells from a hypodermal cell down to the mother-cell. Wherever the ori- 302 GYMNOSPERMS gin can be traced, it is hypodermal. In Taxus, the archesporial cell, or cells, are hypodermal, and can be recognized very easily by the deep staining; and, later, the line of cells between the mother-cell and the periphery is very easily traced (fig. 307A,B). In Taxus sev- a& Care ’ nsasl) HSS Re TATLEACH OY Beale tese,: Fic. 307.—Archesporium and spore mother-cells of conifers: A and B, Taxus bac- cata: A, single archesporial cell; B, two archesporial cells have divided, giving rise to two tapetal cells and two primary sporogenous cells, which are also the megaspore mother-cells; C, Keteleeria fortunei, megaspore mother-cell; D, Larix europea, longitudi- nal section, young ovule showing megaspore mother-cell; E, Pinus rigida, deeply placed megaspore mother-cell; A and B, after DupLer,'® 238; C, after HuTcHINson,?” D, after STRASBURGER,™ 141; E, after MARGARET FERGUSON,™ X 46. eral mother-cells may divide, and several megaspores may germi- nate, producing gametophytes which may reach an advanced free nuclear stage. In Keteleeria, a peculiar Chinese gymnosperm, which has been variously assigned to Pinus, Abies, and Tsuga, the mother- cell is sharply marked (fig. 307C). Perhaps the earliest form in CONIFERALES 303 which the mother-cell was proved to be hypodermal in origin is Larix (fig. 307 D). The cells surrounding the mother-cell soon become modified, and furnish nutrition to it as it increases rapidly in size (fig. 308). These cells constitute the “spongy” tissue, and furnish nutrition, not only Fic. 308.—Pinus laricio: megaspore mother-cell with nucleus in prophase of reduc- tion division. Surrounding it are modified cells, called the “spongy tissue,” which fur- nish nutrition not only to the mother-cell, but to later stages; X 500.—From CouLTER and CHAMBERLAIN, Morphology of Gymnos perms's4 (University of Chicago Press). to the growing megaspore, but even through early stages of the fe- male gametophyte. They are finally entirely absorbed. Occasionally, the megaspore fails to develop. In such cases, the spongy tissue may become very active and its cells may look like megaspores, or like early cellular stages of a female gametophyte. This behavior is rather frequent in Pinus contorta (fig. 309). When the mother-cell is developing normally, the tissue around it 304 GYMNOSPERMS is densely packed with protoplasm, and stains deeply, appearing as in figs. 308 and 309 A. In fig. 309 B, the megaspore has aborted, and is probably represented by the black streak between the two groups Fic. 309.—Pinus contorta: A, megaspore mother-cell developing normally, with spongy tissue around it; B, in this case, the mother-cell has aborted and the surrounding tissue is developing strongly, looking like a group of mother-cells; X 500. of enlarged cells. Whether one or more of these enlarged cells, which, normally, would have been merely nutritive cells, might develop into a functional gametophyte was not deter- mined. It would be worth while for someone, in contact with this western conifer, to work out its entire life-history. The reduction division in mi- crosporogenesis in angiosperms has been studied, described, and philosophized about, until Fic. 310.—Pinus laricio: mega- spore mother-cell as it appears in the Chicago region, about June 1. The spongy tissue has the normal appear- ance.—From COULTER and CHAM- BERLAIN, Morphology of Gymno- sperms (University of Chicago Press). CONIFERALES 305 an appalling literature has accumulated; less is known about the re- duction in megasporogenesis, because it is tedious to get the desired stages. In the gymnosperms not much is known, even about the re- duction in microsporogenesis, and still less is known about reduction in megasporogenesis. Reduction certainly takes place, usually giving rise to a row of four megaspores, the lower one of which functions, while the other three abort (fig. 310). After the first reduction divi- sion, it often happens that only the lower cell divides, giving rise to two megaspores, only the lower one of which functions. The upper cell of the row of three is not a megaspore, because it still has the spo- rophyte number of chromosomes. So many cases of a row of three are reported that it would seem as if there might be a tendency toward the elimination of the division of the upper of the first two cells. With the reduction division, the sporophyte generation comes to a close. The megaspore is the first cell of the gametophyte generation. CHAPTER XIII CONIFEROPHYTES—CONIFERALES—Continued THE MALE GAMETOPHYTE The microspore is the first cell of the male gametophyte. All of the four microspores produced by a microspore mother-cell seem equally vigorous and capable of completing the entire life-history, and nearly all of them germinate and complete more or less of the life-history while still inclosed in the microsporangium. In some species of Cupressus and Juniperus the microspore is shed in the uninucleate condition; but, in such cases, a division occurs soon after reaching the nucellus, before the pollen tube begins to be formed. In these cases there are no prothallial cells, the first division giving rise to a generative cell and a tube nucleus. Consequently, the micro- spore is the antheridial initial. In this more or less advanced stage of development, the micro- spores, or pollen grains, are shed. The whole order is wind-pollinated. In spite of occasional claims, it is very doubtful whether there is a single case of insect pollination in any gymnosperm. Pollen is shed in prodigious quantities. In pine forests, lumbermen speak of “sul- phur showers.” No doubt there are millions of pollen grains in a cone of Araucaria bidwilli, and BURLINGAME* estimated the output in A. braziliensis as high as a billion. In any case, the output is im- mense, and nearly all of the pollen grains die. Very few reach a nucellus, where they may continue their development and form pollen tubes. The microspore always has two spore coats, exine and intine, which vary in their comparative thickness. Usually, the exine is thicker; but in some cases, like Araucaria bidwilli, the intine may be more than twice as thick as the exine. When the pollen is winged, the wings are formed from the exine. STRASBURGER™ believed that the wings arose from a splitting of the exine. Dr. MARGARET Fercuson™ decided that they originate by a separation of the exine 306 CONIFERALES 307 from the intine. A study of very thin sections of Pinus and Abies, sharply stained with safranin and light green, supports Dr. FERGU- son’s conclusion. Only about one-third of the genera have wings or bladders of any sort, the great majority having no such develop- ments of the exine. When wings are present, there are generally two, as in Pinus; but Microcachrys has three, and sometimes four, or even five or six. Most species of Podocarpus have two large wings, but P. dacrydioides has three. Pherosphaera has very small pollen grains with three wings. The male gametophyte is in various stages of development when the pollen is shed, but for any given species in any given locality the time of shedding will vary little, and the stage of development may not vary at all. Very few species, like Taxus canadensis and Cunninghamia sinen- sis, shed the pollen in the uninucleate stage, and these will not have any prothallial cells. The first mitosis, after reaching the nucellus, will be the one which gives rise to the generative cell and tube cell. In species which have prothallial cells, these and the generative cell are formed before the pollen is shed. Pinus is a familiar form with winged pollen and prothallial cells (fig. 311). The figure begins with the second reduction division (A), the tetrahedral arrangement of microspores is shown in B, and the beginning of the wings in C. The first mitosis of the young gameto- phyte is shown in D-F, in F, with the nucleus of the first prothallial cell already degenerating. G and H show the prophase and telophase of the second mitosis, and J, the two prothallial cells. These are usual- ly overgrown by the vigorous intine before the third mitosis (J), which gives rise to the generative cell and tube cell, is completed. When the two prothallial cells and the generative cell (K) have been formed, the pollen is ready to be shed. In Pinus the succeeding stages are found after the pollen has reached the nucellus of the ovule (L). This is a very prevalent course of development of the male gametophyte in conifers. Some differ in having no prothallial cells; some, in having a greater display of prothallial cells; others differ in the extent of development of the male gametophyte when the pollen Fic. 311.—Pinus laricio: stages in the development of the male gametophyte; A, the second reduction division; B, tetrahedral arrangement of microspores; C, the wings are beginning to appear; D, prophase of first division in microspore; & and F, telophase of first division; in F, the nucleus of the first prothallial cell is already disorganizing; G, first prothallial cell and prophase of second division; H, telophase of second division; 7, two prothallial cells; J, telophase of third division; the nucleus nearest the prothallial cell will belong to the generative cell and the other is the tube nucleus; K, the two pro- thallial cells, overlaid by the intine (i), the generative cell and the tube cell; this is the shedding stage; L, pollen tube, as found in nucellus, showing stalk and body cells: ~, pro- thallial cells; s, stalk cell; b, body cell; ¢, tube nucleus; e, exine; A and B, May 3; C, May 10; D-G, May 20; H—J, May 25; K, June 15; L, May 1, nearly a year after the stage shown in A; X600.—From Coutter and CHAMBERLAIN, Mor phology of Gymnos perms.*54 CONIFERALES 309 is shed; and others differ in the number of gametes or in the organiza- tion of gametes. Prothallial cells are a constant feature of the Abietaceae; but in most of the Taxodiaceae (Sciadopitys, Cunninghamia, Sequoia, Cryptomeria, Taxodium), they are generally lacking. In many of the Cupressaceae (Callitria, Widdringtonia, Libocedrus, Thuja, Cupressus, Chamaeceparus, Juniperus), they are also lacking. There are no prothallial cellsin the Taxaceae. Wherever there are no pro- thallial cells, the microspore is the antheridial initial, as in angio- sperms. In all of the Podocarpaceae, except Pherosphaera, there is a vigorous development of prothallial cells. The greatest display of these cells in the whole order is in the Araucariaceae, where the de- velopment of prothallial tissue is far greater than in any living heterosporous pteridophyte (figs. 312, 313). In Araucaria, and others with two or more prothallial cells, the first two cells, and sometimes a third cell, are formed as in Pinus. The large number of prothallial cells is due to the subsequent divi- sion of these two or three cells. More than two cells, formed as in Pinus, are rare. When the first prothallial cell divides periclinally, it looks as if three cells had been formed by the Pinus method. The first division in the primary prothallial cells is nearly always anti- clinal; and even when there are a dozen prothallial cells, there may be no periclines. Where the number is large, sometimes as many as forty, there will be both anticlines and periclines. There can be no doubt that prothallial cells are vestigial. Original- ly they were green, independent plants, bearing the antheridia. In some heterosporous pteridophyte ancestor, the prothallium (gameto- phyte) became included within the spore, became parasitic, and then became reduced, and finally disappeared entirely. In angiosperms there is no normal occurrence of a prothallial cell. Throughout the order the generative cell divides, forming two cells which are called the “‘stalk cell” and “‘body cell,” the stalk cell getting its name from its position in forms like Pinus, where it looks like a stalk, bearing the body cell (fig. 311 L). In the Araucariaceae, in Dacrydium, Phyllocladus, and Podocar pus, there is no stalk posi- tion, but the generative cell divides, forming two cells, one of which aborts, while the other gives rise to gametes. Fic. 312.—Araucaria cunninghamii: development of the male gametophyte; A, uni- nucleate microspore, with nucleus surrounded by starch grains; the exine and intine of about equal thickness; B, the first two prothallial cells; C, the first prothallial cell has di- vided anticlinally and the starch grains are very large; the intine is much thicker than the exine; D, two primary prothallial cells have divided and the generative cell (g) and tube cell (t) have been formed; £, the generative cell has divided into a stalk cell (s) and body cell (b); F’, the body cell has enlarged, the walls of the prothallial cells have broken down, leaving the nuclei free; s is probably the nucleus of the stalk cell and ¢, the tube nucleus; G, the body cell is easily recognizable, ¢ is the tube nucleus, and s may be the stalk nu- cleus; 950. CONIFERALES 311 Just what the stalk cell really is, no one has determined definitely. Where it does not have the stalk position, as in Araucaria and others, the stalk and body cells, at first, look alike, and probably Fic. 313.—Araucaria cunninghamii: male gametophyte: in A-C, the first primary prothallial cell has not divided; in D, it is dividing periclinally; in E—-H, it has divided anticlinally; 7, shows an unusually large number of prothallial cells; ¢, tube nucleus; g, generative cell; s, stalk cell; b, body cell; Xgoo. neither is predestined to become the ‘‘body cell” and produce the two gametes. In Microcycas, Dr. DorotHy Downie” found that the stalk cell divides repeatedly, each time cutting off a body cell, which later produces two gametes. In this case, the stalk cell is spermatogenous. In conifers, it may be, phylogenetically, a sperma- togenous cell, like the body cell. 312 GYMNOSPERMS The microspore in conifers, as in cycads, has a very definite or- ganization. The differentiation into apex and base is just as definite as the sporophyte differentiation into root and shoot. The be- havior, as far as apex and base are concerned, is exactly the opposite of that in cycads and Ginkgo. In the cycads and Ginkgo the pollen tube functions principally as a haustorium—doubtless, the original function of a pollen tube—while the prothallial end of the tube, with the whole pollen grain, grows down into the nucellus, and this pro- thallial cell end of the tube is the one which ruptures and sheds the Fic. 314.—Abies balsamea: pollen grain at time of shedding; b, body cell; s, stalk cell; p, prothallial cell; p/, starch; ¢, tube nucleus; X 535. gametes. In conifers, the prothallial cell end of the tube, with the entire pollen grain, remains where it alights upon the nucellus, and the tube grows down into the nucellus, serving as a haustorium, but also as a carrier of the gametes. In the Araucariaceae, as shown in figs. 312 and 313, the generative cell divides and forms the stalk and body cell before the pollen is shed. This is also true of some others with extensive prothallial tissue, as in Podocarpus, Dacrydium, and Phyllocladus. This division occurs regularly in Abies and Tsuga; but here the stalk cell has a typical stalk position (fig. 314). In these two cases, one or both of the two prothallial cells may divide, so that there may be three or four prothallial cells. In the Abietaceae there is usually a consider- able interval between the formation of the first and second pro- thallial cells, so that the rapidly thickening intine overgrows the CONIFERALES 313 first prothallial cell. Sometimes the interval between the formation of the second prothallial cell and the generative cell is long enough to have the second prothallial cell slightly overgrown by the intine. Where the generative cell divides before pollination, as in Abies, the body cell becomes very large. Pollen grains, in all the conifers, have a rich supply of starch. The body cell does not divide until the pollen tube grows down into the nucellus. Its division, in nearly all cases, gives rise to only two gametes. The male gamete is a short-lived cell. Consequently one expects to find fertilization soon after this division. Just at the time of pollination, in nearly all cases, a pollination drop appears on the tip of the ovule. It is colorless, and looks like a small drop of glycerine. The pollen, falling on this drop, is drawn down to the nucellus, where the growth of the pollen tube begins. In the Araucariaceae the pollen lodges on the ovuliferous scale, or on the ligule, or in the axil of the scale. In Larix and Pseudotsuga the pollen does not reach the tip of the nucellus, but finds a lateral posi- tion. There is a somewhat similar situation in Arthrotaxis and Se- quoia. In such cases, the pollen tube grows somewhat laterally through the nucellus to reach the egg; while in most cases, as in Pinus, the growth is straight down, from the tip of the nucellus to the egg. Dates of pollination and fertilization.—It is interesting to note the time at which various stages in the life-history may be found. While the time at which a certain stage, like pollination, takes place, may vary with latitude and other factors, the intervals between stages do not vary so much. Consequently, if the date of a certain stage is known, the dates for other stages can be predicted approximately. In general, the interval between pollination and fertilization in gymnosperms is longer than in angiosperms, Ephedra, with an in- terval of 10 hours, being a notable exception. On the other hand, there are a few long intervals in angiosperms, the longest being in the oaks, where the interval is more than a year. Other members of the Fagaceae, and also of the Betulaceae and Juglandaceae, have long intervals. But, in general, in angiosperms the interval is to be reckoned in days or hours, rather than in the weeks or months of the gymnosperms. 314 GYMNOSPERMS In the latitude of Chicago, most pines are pollinated about the middle of June, and fertilization takes place a little more than a year later, during the last few days of the following June, or the first few days in July. In Pinus strobus, near Wellesley, Massachusetts, pollination takes place late in May or early in June, and fertiliza- tion a little more than a year later—about the middle of June. HuTcHINSOoN’’s found that in Abies balsamea the interval between pollination and fertilization is only 4 or 5 weeks, and that for the greater part of this time the pollen remains unchanged on the tip of the nucellus. The pollen tube then develops very rapidly, reaching the egg and discharging the sperms in 2 or 3 days. Fertilization occurred June 25 in Ontario, Canada. In Pseudotsuga taxifolia the interval is also very short, pollination taking place early in April, followed by fertilization early in June. In Tsuga canadensis pollination takes place about the middle of May and fertilization about the first of July; so the interval is about six weeks. In Picea excelsa the interval is shorter, with pollination in the middle of May and fertilization the middle of June. In Cedrus deodara pollination occurs late in September, and fertilization about the end of the following May—an interval of 8 months. Several observations have been made in the Taxodiaceae: Sciadopitys is pollinated late in April (Kew) and fertilization occurs about the end of June of the following year—an interval of about 14 months. Sequoia sempervirens is pollinated early in January and fertiliza- tion takes place late in June—an interval of about 6 months. Taxodium distichum is pollinated the middle of March and ferti- lized the middle of June—an interval of 3 months. In Cryptomeria japonica pollination begins in March, and the pollen grains remain unchanged on the nucellus for 4 or 5 weeks; then the pollen tubes grow rapidly, and in 3 or 4 weeks reach the egg, about the first of June—an interval of about 3 months. In Arthrotaxis selaginoides pollination takes place early in April, and fertilization about the first of July—also an interval of 3 months. CONIFERALES 315 Cunninghamia sinensis is pollinated early in April, and fertilized early in July—another interval of about 3 months. In the Cupressaceae there are also long and comparatively short periods: In Actinostrobus pyramidalis the interval between pollination and fertilization is about 3 months—from the middle of July to the middle of October. In Tetraclinis articulata pollination extends over an unusually long period, from April 20 to June 1. Fertilization follows from 3 to 3 months later. In Widdringtonia cupressoides the interval is long, with pollina- tion in January and fertilization 14 or 15 months later—in February or March of the next year. In Libocedrus decurrens, pollinated in California early in April, fertilization occurs about 2 months later—the first week in June. Juniperus communis, in southern Sweden, is pollinated about the middle of June and fertilization takes place more than a year later— the first week in July. Near Wellesley, Massachusetts, pollination occurs about May 11, and fertilization about June 20—also an interval of about 123 months. Juniperus virginiana, with which J. communis is often associated, has a comparatively short interval, with pollination about May 1, and fertilization about June 20 of the same year. In the Araucariaceae, records are not very complete, with almost no reports from anything except exotic material. In Araucaria braziliensis fertilization occurs about April 1—5 or 6 months after pollen falls on the cone. In Araucaria cunninghamii pollen is shed (Queensland, Australia) in December. In Agathis australis pollination occurs in September and October, with fertilization a year later. In the Podocarpaceae, Dr. JOHANNA KILDAHL?* found the pollen of Phyllocladus alpina already shed on November-1; Dr. Mary Younc™ found pollen just ready for shedding on November 13. Both studied New Zealand material. In the Taxaceae records are scanty: In Taxus baccata pollination, at Ziirich, takes place March 1-15. According to DupPLER,'’ the interval between pollination and fertilization is only 1 month, while 316 GYMNOSPERMS in the Ziirich material the interval was over 2 months. In Torrey taxifolia, pollen is shed late in March or early in April, and fertiliza- tion, in 1904 material, took place August 12. In this rather extensive survey, there are slight differences in the dates given by different observers. Most of the dates are for ma- terial growing in its native habitat, but some are from exotic ma- terial. However, this factor is not likely to influence the times of pollination and fertilization as much as would the differences in latitude, altitude, and immediate surroundings. While these factors certainly influence the time of pollination, the interval between pollination and fertilization does not seem to be affected. In conifers with a long period between pollination and fertiliza- tion, there is usually a rapid growth of the pollen tube immediately after pollination. The tube grows down about to the level of the free part of the nucellus, and then stops and winters in this condition, resuming its growth the next spring. In forms with two growth- periods, and in which the division into stalk and body cell does not take place before the pollen is shed, the division into stalk and body cell occurs during the first period. The division of the body cell to form the two sperms takes place shortly before fertilization, usually less than a week before the fusion of gametes. In several closely observed forms, the interval is about 5 days. In almost everything there are exceptions. In Abies balsamea the body cell divides before the tube begins to form; but here the growth of the pollen tube through the nucellus to the egg does not take more than 2 days, and the time may be reckoned in hours. Con- sequently, the interval between the division of the body cell and fertilization is not so long as in cases where the division takes place late in the development of the pollen tube. The sperms.—The sperms, or male gametes, of conifers show con- siderable differences in organization, but there are only two cate- gories: the sperm is either a highly organized cell, or it has lost its cell wall, and, at the time of fertilization, its cytoplasm may have blend- ed more or less with that of the pollen tube. The most primitive form is, doubtless, the highly organized cell, as it is found in Juniperus and other members of the Cupressaceae, Taxodiaceae, Podocarpaceae, and Taxaceae (figs. 315, 316). These CONIFERALES 317 two figures show a sperm almost as highly organized asin the cycads and Ginkgo. Whether each sperm is formed inside a sperm mother- cell, as in the cycads, or arises from the division of the body cell, so Fic. 315.—Body cells and sperms in the Taxodiaceae: A, Taxodium distichum, the two highly organized sperm cells, with the stalk and tube nuclei in advance; X 540; after Coker; B—D, Cunninghamia sinensis: B, pollen grain with generative cell and tube cell (no prothallial cells); C, body cell and stalk and tube nuclei; X 430; after MrvakeE;4? E, F, Cryptomeria japonica: E, body cell and stalk and tube cells; F, the two sperms; X 666; after Lawson.345 318 GYMNOSPERMS that the outer wall of the body cell becomes part of the wall of the sperm cell, does not appear from these figures. At any rate, the sperm lacks only the cilia to make it motile, like the sperms of cycads and Ginkgo. Fic. 316.—Pollen tube structures in the Cupressaceae: A, B, Thuja occidentalis: A, body cell, with stalk and tube nuclei; B, the two sperms; X 560; after Lanp;337 C, Cu- pressus goweniana; group of numerous male cells, with stalk and tube nuclei below; X 350; after JueL;3°” D, Libocedrus decurrens; the two sperms; after LAwsoNn.347 In nearly all cases where the sperms are so highly organized, the sperms appear to be exactly alike in size and structure; but in Taxus, Cephalotaxus, and Torreya, there is great disparity in size. The mi- CONIFERALES 319 totic figure in the body cell is so near the periphery that one of the sperms is only a small lenticular cell which soon aborts (fig. 317). The formation of sperms in pairs is thoroughly established in the liverworts, and probably goes farther back; but, from the liverworts to the sunflowers and orchids the pairing is constant. In the angio- sperms, each sperm has its own function, one fertilizing the egg and the other taking part in the triple fusion which initiates the second period in the development of the female ga- metophyte. In gymnosperms, in many cases, the two sperms differ in size, but whether they differ in chromatin content or in any other way, has not been determined, either in gymnosperms or in any other group. We suspect that a detailed investigation, with adequate technique, would reveal something interesting. In Cupressus goweniana, and also in C. ari- zonica, there are several small sperms (fig. 316C). In the latter species Doax™ found twelve fully developed sperms. Both of these desi ania enc investigators used exotic material. These ino to form two unequal two species recall the condition in Micro- sperms; B, the two sperms; cycas, which has a dozen or more sperms. the stalk and body nuclei It is worth while to note that these highly oe ae eerie ett: 570; after DUPLER.!7° organized sperms, doubtless primitive, are associated with archegonia in which there is no wall between the ventral canal nucleus and that of the egg, a feature which is more advanced than that in Ginkgo and the Abietaceae, where there is a wall between these two nuclei. It is an instructive illustration of the fact that various lines of evolution do not keep pace with each other. In the Abietaceae the sperms are very different. The wall of the body cell is thin, and when its nucleus divides, the daughter-nuclei are left free, with no wall between them. A wall begins to form at the equator of the spindle, but it soon breaks down. The wall of the body cell, although very thin, separates the cytoplasm surrounding the sperm nuclei from the general cytoplasm of the pollen tube, dur- Fic. 317.—Taxus cana- 320 GYMNOSPERMS ing nearly all of the passage of the body cell down the tube (fig. 318). As the tube is about to discharge the sperms, the wall of the body cell seems to have disappeared entirely, so that the sperm nuclei are in a Fic. 318.—Pollen tubes of Pinus: A, Pinus strobus, two sperms, quite unequal; be- low the two sperms, the nucleus of the stalk cell and a little lower down, the tube nu- cleus; X 236; after Dr. MarGARET Fercuson;™ B, P. laricio; the two-sperm nuclei (m); n, nucellus of stalk cell; the other nucleus is the tube nucleus; s, starch; X 500; after CouLtter;"47 C, P. laricio; the two-sperm nuclei with no wall of the body cell visible; at the right of each nucleus, a dense mass which might represent a blepharoplast; X 500; after CHAMBERLAIN." mass of cytoplasm which does not seem to be marked off at all from the general cytoplasm of the pollen tube (fig. 318 C). In such a condition, either as an organized cell or as a naked nucleus, the sperms, together with the stalk and tube nuclei, are discharged into the egg. CHAPTER XIV CONIFEROPHYTES—CONIFERALES—Continued THE FEMALE GAMETOPHYTE The megaspore is the first cell of the female gametophyte. The megaspore mother-cell, in the two divisions by which the four mega- spores are produced and the number of chromo- somes is reduced from the sporophyte number (2x) to the gametophyte («) number, is usually so deeply placed in the nucellus, when it first be- comes recognizable, that its origin looks indef- inite. In all cases in which the origin can be rec- ognized with certainty, the archesporial cell is hypodermal, and its first periclinal division gives rise to a tapetal cell and the megaspore mother- cell. An axial row of four megaspores has been observed in many cases (fig. 319). It happens, frequently, that there is a row of only three cells. In such cases, the upper cell, re- sulting from the first division of the megaspore mother-cell, fails to divide, and does not reach the megaspore stage. The lower cell divides and produces two megaspores, the lower one of which continues to develop, while the other, together with the upper cell of the row of three, disor- ganizes. The failure of a division in the upper of the first two cells was noted in the cycads, Stangeria’” and Ceratozamia;* but Zamia,®™ the- Fic. 319.—Pinus laricio: row of four megaspores, the lower one growing vigorous- ly and the other three aborting; X810.— After Dr. MARGARET FERGUSON.1% most advanced of all cycads, has a row of four. Since, in nearly all cases, only one megaspore germinates, it is natural that there should be a tendency to eliminate the useless spores. The first indication of germination is an enlargement of the mega- spore and its nucleus. After the first division of the megaspore 321 322 GYMNOSPERMS mother-cell, the lower of the two resulting cells is likely to divide first. Although the division in the two cells is said to be simulta- neous, and the two mitotic figures are seen at the same time, the lower Os Mees — SIS) — (> LE ee —spus See is +) i see ‘eeminn TESLA Fic. 320.—Pinus strobus: free nuclear stage in female gametophyte. The proto- plasm, with its nuclei, is pressed outward by the large central vacuole. The gameto- phyte is surrounded by a jacket of spongy tissue; X 62.—After Dr. MARGARET FER- GUSON.1% figure is practically always a little more advanced, and its daughter- nuclei reach the resting condition, while those of the upper mitosis are still in the late telophase. Thus the lower two, and especially the lowest one, get into the resting, or, rather, the working, condition earlier than the upper pair. Anex- cellent example is seen in LAND’s drawing of Ephedra, in which the lowest megaspore is enlarging, the next above is beginning to disor- ganize, while the mitosis in the upper cell of the first “pair” is still in the late telophase (fig. 359 A). The reason for the more rapid de- velopment of the lowest cell is probably because the nutrition comes principally from beneath. In gymnosperms, almost invaria- bly, and nearly as frequentlyin an- giosperms, the lowest megaspore is the one which germinates and de- velopsinto the adult gametophyte. Free nuclear period.—In all co- nifers there is a period of free nu- clear division before any cell walls are formed. Even as early as the close of the first division, a large vacuole may appear between the two daughter-nuclei. With succeeding divisions the outline of the megaspore increases immensely, and the central vacuole, filled with a transparent fluid, presses the protoplasm outward so that it forms a thin layer containing the free nuclei (fig. 320). Mitoses through- CONIFERALES 323 out the free nuclear period are simultaneous (fig. 359 B). Here again, a figure taken from Lanp’s splendid work on Ephedra, will also illustrate what takes place in conifers. A free nuclear period is not confined to conifers, but is likely to occur in any plant where the cell is large in proportion to the nuclear figure. Where the cell is large and the nuclear divisions follow in rapid succession, one division follows another before a wall can be formed on the spindle. As the number of nuclei increases, and the interval between mitoses becomes more prolonged, walls make their appearance. The extent of the free nuclear period depends upon the size and shape of the megaspore outline. If it is long and narrow, the period will be shorter; if more or less spherical, the period will be longer. So there may be only dozens of nuclei when walls begin to come in; or there may be hundreds. JAGER?*5 estimated the number in Taxus baccata at 256, and DupLer'” found the same number in T. canadense. COULTER and Lanp’? also found 256 in Torreya taxifolia, a number already reported for Zamia and Ginkgo. This must be regarded as a very low number for a gymnosperm. Dr. MARGARET FERGUSON’ finds a much larger number in Pinus strobus. She counted 2,000 when walls were coming in. Since these mitoses are simultaneous, that would mean 1,000 free nuclei. Nor- EN*® estimated the number in Juniperus communis as about the same. Period of wall formation.—There are two types of wall formation in the female gametophyte, one of which may be illustrated by Pinus and the other by Taxus. In Pinus, at the final free nuclear division, walls perpendicular to the megaspore membrane are formed on the achromatic fibers which connect the nuclei. A wall is also formed on the side next the megaspore membrane; but the sixth side, toward the center of the megaspore cavity, remains open.’ 345 As division continues, peri- clinal walls come in, but the innermost sides of the innermost cells remain open. This method continues, with the tissue advancing toward the center, even after archegonial initials have appeared at the top. When the central cavity is entirely closed, walls are formed on the centripetal ends of the cells. Consequently, along the line of 324 GYMNOSPERMS Fic. 321.—Taxus canadensis: fe- male gametophyte cellular through- out, but before the appearance of archegonia. The upper gametophyte is in the free nuclear stage, and there is an abortive gametophyte between them, at the right; X286.—After DuPLER.'7® closure, two cell walls are in contact. This is doubtless the reason why many female gametophytes split so easily in the middle. In Taxus,’” the first walls, perpen- dicular to the megaspore membrane, extend to the middle of the mega- spore cavity, so that the cells are like long tubes, which have been called “alveoli.” In nearly all cases the walls on the centripetal ends of the alveoli form late. Subsequent divi- sions divide each of the alveoli into a large number of cells. In Taxus, the gametophyte is cellular through- out before any archegonial initials appear; this is true also of Torreya and others (fig. 321). There are long alveoli in Micro- cachrys tetragona, but here, judging from Lawson’s figures,3° the cen- tripetal end of the alveolus has a wall from the start. In general, the development of the female gametophyte is much as in the angiosperms, except that there are none which are cellular from the beginning, as in the long, narrow embryo sacs of Ceratophyllum sub- mersum and Monotropa hypopitys. It sometimes happens that more than one megaspore germinates. This is rather common in Taxus canadensis, where more than one megaspore mother-cell occurs fre- quently. All four megaspores of a tetrad may germinate, as well as one —— CONIFERALES 225 or two of a second tetrad. DupLErR"” found several ovules in which three female gametophytes had reached the archegonium stage, with a couple more in the free nuclear stage. In Taxus baccata several in- vestigators have failed to find more than one megaspore germinat- ing. DupLeR’s findings are conclusive for T. canadensis. More than one megaspore has been observed to develop in Sequoia,*“ Cunning- hamia,** Sciadopitys 349 Taxodium,® and Cryptomeria.*4s More than one is rare in Pinus.'” In the early stage of the megaspore, no megaspore membrane, distinct from the ordinary cellulose wall, can be distinguished; but as the free nuclear stage advances, and the cellular stage begins, a genuine spore coat appears, and may reach considerable thickness. The thickness of some megaspore membranes, as given by THOM- Son,°8 are as follows: Biota orientalis, 1.7 microns; Sequoia sem- pervirens, 2.7 microns; Pinus sylvestris, 3 microns; P. resinosa, 4.2 microns; Abies balsamea, 4.6 microns; Larix americana, 4.7 microns. These measurements indicate the range of thickness in the order. The chemical composition of the membrane is very complex. On the outside it is suberized, and, on the inside, chemical tests indicate a substance related to pectin. Between the two, there is cellulose which gradually changes into the suberin of the outside: Tuomson® describes two layers, an exosporium and an endo- sporium, and there is no doubt that stained material, which is easily photographed, seems to support his view. However, three layers can be photographed in Dioon edule, where it is doubtful whether there is more than one. It is more than doubtful whether there are two layers originating like the exine and intine of microspores. In spores which are to be shed, spore coats are highly developed, and the megaspore of gymnosperms still retains a highly developed spore coat. When the megaspore began to be retained in the mega- sporangium (ovule) there was no longer need for a protective coat, _ and it became more and more reduced, still appearing in the coni- fers, but not so thick in living forms as in the carboniferous gymno- sperms. In the Gnetales®3 the coat is very thin—1.3 microns in Welwitschia—and in the angiosperms it has disappeared as a recog- nizable spore coat. During its early development, the female gametophyte is sur- 326 GYMNOSPERMS rounded by the spongy tissue, which is gradually absorbed. In later stages, while there is not such a definite surrounding tissue as the “endosperm jacket” of the cycads, the cells in contact with the megaspore membrane are more or less modified, and some have applied the term ‘‘tapetum” to these surrounding cells. The archegonia.—It would seem that most, or even all, of the su- perficial cells at the top of the female gametophyte are, potentially, archegonium initials. The number which becomes recognizable as initials is small, and the number reaching maturity may be still smaller. There are two general types of archegonial groups in conifers. In the Abietaceae and many others, the archegonia are separated by vegetative cells of the gametophyte. In the Cupressaceae and Taxodiaceae they are in contact with each other, forming an arche- gonium complex. In some, the complex is at the apex of the gameto- phyte, as in Thuja, Libocedrus, Tetraclinis, and Juniperus; while in others it is lateral, as in Sequoia, Actinostrobus, Callitris, and Wid- dringtonia. The number of archegonia is much larger where there is an arche- gonium complex than where the archegonia are separated by vegeta- tive tissue. In Thuja the number is usually 6; in Juniperus com- munis, 4-10, usually 7; in Libocedrus decurrens, 10-15; in Biota orientalis, 9-24; in Callitris verrucosa, 17-20; in Actinostrobus pyr- amidalis, 25-30; in Widdringtonia cupressoides, 30-100, variously placed, but never apical; and in W. juniperoides the number reaches 200. In forms with archegonia separated by vegetative cells, the num- bers are smaller and the position nearly, but not quite, apical; for the archegonia are usually in a circle, surrounding the center. In Pinus strobus, P. rigida, and P. resinosa, the usual number of arche- gonium initials is 3; in P. laricio, there are usually 5, occasionally 6, and sometimes only 2 or 3. In Torreya taxifolia,® there is almost invariably only 1 archegonium; but in other genera a single arche- gonium is very rare. Both types, the separated archegonia and the archegonium complex, occur in Sequoia. The archegonium initial first becomes recognizable by elongating and enlarging somewhat, while the adjoining cells divide. Its nu- CONIFERALES B27 cleus moves to the peripheral end of the cell. Early in June, in the Chicago region, archegonium initials are recognizable in Pinus laricio; and in P. banksiana they may be seen a week or 10 days earlier. Within a week after the initial becomes recognizable, its nucleus divides, giving rise to the central cell and the primary neck cell (fig. 322). Fic. 322.—Pinus laricio: development of the archegonium; A, archegonium initial, May 28; B, neck and central cells, June 2; C, two-celled neck and enlarged central cell, June 18; D, mitosis cutting off ventral canal cell, June 21; X 104.—From CouLTER and CHAMBERLAIN, Morphology of Gymnosperms'4 (University of Chicago Press). After the primary neck cell is cut off, it soon divides, but it is nearly a month before there is a division of the central cell. During this period, the central cell enlarges immensely, and a very definite archegonium jacket is formed, which plays for the central cell such a rdle as the ‘‘endosperm jacket” plays for the female gametophyte. With its enlargement, the central cell becomes very vacuolate, and, besides the large vacuoles, a peculiar type of vacuole appears, the “proteid vacuoles,’ which look so much like nuclei that Hor- MEISTER? believed the eggs of gymnosperms differed from those of angiosperms in being multinucleate. HOFMEISTER resented STRAS- 328 GYMNOSPERMS BURGER’S correction of the mistake, perhaps because STRASBURGER was 20 years younger. The primary neck cell early undergoes an anticlinal division, and a second anticline makes a plate of four cells. Later, there is usually a pericline, making two tiers of cells. In each tier there may be two divisions, so that the neck of the archegonium consists of eight cells. One or more of these mitoses may fail to take place, so that the number is usually less than eight. Occasionally, there are more than two tiers, and occasionally only one. In Podocarpus coriacea,™” the number is variable, ranging from 2 to 25. In Tsuga canadensis ,*° 4°5 there is only one tier, and it may be only two-celled. Austrotaxus*® is exceptional in having as many as 16 neck cells, all in one tier. There is no neck canal cell in any gymnosperm. In the evolution of the archegonium the neck canal cell made its final appearance in the heterosporous pteridophytes, all of which have a single small neck canal cell. The division of the central cell to form a ventral canal cell, or nucleus, and the egg, occurs shortly before fertilization, so that the development of the egg is almost entirely the development of the central cell. In the Abietaceae there is a distinct ventral canal cell (fig. 323). In the rest, there is a nuclear division, but no wall is formed between the two resulting nuclei. It is a curious fact that those in which no wall is formed between the two nuclei are the ones in which the sperms are distinct cells. In Pinus, with a distinct ven- tral canal cell, the sperm has become reduced to a naked nucleus. In Pinus, the ventral canal figure is usually small, cutting off a small ventral canal cell which promptly degenerates (fig. 323 A,B). Occasionally, the figure is comparatively large, and cuts off a large ventral canal cell (fig. 323 C,E). When a large cell is cut off, it sometimes happens that the wall between the two nuclei breaks down. The ventral canal nucleus, then being surrounded by the general mass of the cytoplasm of the egg, grows and equals the egg nucleus in size and, apparently, in development. It might be ferti- lized, or it might fertilize the egg (fig. 323 £). HutTcHrInson’’s observed both these cases in Abies balsamea. Lanb**7 found cases in Thuja occidentalis which led him to believe that both the egg and the ventral canal nucleus had been fertilized, CONIFERALES 320 for there were two groups of nuclei, one at the base and the other at the top of the egg. E D Fic. 323.—Pinus laricio: formation of ventral canal cell; A, typical mitotic figure; B, ventral canal cell disorganizing; C, unusually large mitotic figure cutting off a large ventral canal cell; D, a large ventral canal cell; E, the wall between the ventral canal cell and the egg has broken down leaving both nuclei free in the egg: nv, nucleus of ventral canal cell; mo, nucleus of egg; A-C, X500; E, X100.—After CHAMBERLAIN.'™4 In Torreya,'® it is probable that no ventral canal mitosis occurs. The material was so abundant and the observer so competent that 330 GYMNOSPERMS the mitosis could hardly have escaped observation, had it been present. “There can be no doubt that in the evolution of the archegonium there has been a gradual reduction in the length of the neck and in the number of neck canal cells, which, phylogenetically, are prob- ably eggs. In the lower Filicales there are two neck canal cells; in the higher homosporous forms, only one neck canal cell with two nuclei; and in the heterosporous genera, even the mitosis has failed to take place, so that there is only one neck canal cell with one nucleus. In the gymnosperms the mitosis which, in the pterido- phytes, gives rise to the neck canal and ventral series, is suppressed, so that the ventral canal mitosis takes place in the cell which, in Pteris, gives rise to a primary neck canal cell and a central cell. There can hardly be any doubt that the ventral canal cell is a re- duced egg, just as in Taxus, and several other genera with unequal sperms, the smaller sperm is strictly the homologue of the larger one. In tracing the evolution of the archegonium, the bryophytes show a nearly equal division, so that the egg and ventral canal cell can be distinguished only by their relative positions; and, in a few cases, there are several equal eggs. After the ventral canal mitosis, fertilization follows promptly, the interval between the mitosis and fertilization usually not exceed- ing 5 days. i pesultatel OY P8 AD we On {% [Is ~~ < emg hs SRA RY | oO ee a y \% “ase S Se CHAPTER XV CONIFEROPHY TES—CONIFERALES—Continued FERTILIZATION The pollen tubes press and digest their way through the nucellus, destroying much of its tissue; but there is no such complete destruc- tion as in the cycads, where the pollen tubes hang free for a greater part of their length. In conifers, the pollen tubes reach the megaspore membrane, with both gametophytes in various stages of development. In Torreya taxifoliat? the pollen tube completes its passage through the nucellus while the female gameto- phyte is still in an early free nuclear stage, with only 16 or 32 free nuclei (fig. 325). At this stage, the pollen tube has broadened at the base, and the body cell has enlarged, but will enlarge much more before division. Just as the pollen tube is about to dis- charge its two very unequal sperms, it is broader than the egg (fig. 326). It should be noted that the interval between the stages shown in these figures is about 7 weeks. In Callitris verrucosa*™ the pollen tube arrives when the female gameto- phyte is in the 128 nucleate stage. In Taxus canadensis'® the female gametophyte is more advanced when eee oN roo Fic. 324.—Taxus canadensis: nucellus showing seven pollen tubes, three of which show the body cell just ready to divide. The female gametophyte has not yet reached the archegonium initial stage; X92.——After DUPLER.'” the pollen tube reaches it, and is cellular throughout, but the archegonium initials are not yet recognizable (fig. 324). The body 334 332 GYMNOSPERMS cell has rounded off and has enlarged, but its nucleus is still in the resting condition. The arrival of the pollen tube while the female gametophyte is in a comparatively early stage of development is more frequently found where the male gametes are highly organized cells. Fic. 325.—Torreya taxifolia: A, two-celled microspore on degenerating cells at tip of nucellus; pc, pollen chamber; B, enlarged end of pollen tube in contact with the fe- male gametophyte which is still in the free nuclear stage (June 21, 1904); 460.—After CouLter and Lanp.'5? In forms like Pinus, where the sperms are nearly or quite naked nuclei, the cellular condition is reached by the female gametophyte, Fic. 326.—Torreya taxifolia: pollen tube in contact with the egg: m' and m? the two unequal sperms; sé, stalk nucleus; én, tube nucleus; mc, neck cell; 0, egg nucleus; hr, haustorial cells. Cells at the top of the female gametophyte are growing up around the tube; August 12, 1904; *460.—After CoULTER and Lanp.'s? 334 GYMNOSPERMS and the archegonium is in a rather advanced stage of development before the pollen tube comes into contact with the megaspore membrane. In any case, the division of the body cell to form the two sperms takes place shortly before fertilization, so that it is more or less simul- taneous with the division of the central cell to form the egg and the ventral canal cell, or ventral canal nucleus. The neck cells in many cases disorganize, and the megaspore membrane at the top of the female gametophyte is dissolved or be- comes so weak that it is easily broken when the pollen tube arrives. When the pollen tube reaches the egg, the behavior differs. In the Abietaceae the tip of the tube ruptures, forming a pore through which the contents are discharged into the egg. The pollen tube itself does not enter. In the other families more or less of the tip of the pollen tube enters the top of the egg before the discharge takes place. Many of the observations upon which this statement is based should be con- firmed or corrected, for there is, at the time of the discharge, a large vacuole at the top of the egg, and the vacuole often has such a strong plasma membrane that it might be mistaken for the end of a pollen tube. There have been occasional misinterpretations of pollen tubes ever since they were thought to be embryos. The pollen tube, when it comes into contact with the egg, has been growing more and more turgid, so that, when the rupture comes, the contents of the tube are discharged with considerable violence. Some forms, like Pinus, have been studied so thoroughly by so many investigators that anything exceptional is easily distinguished from the normal behavior. A few other forms have been studied al- most as thoroughly by several investigators. Unfortunately, many of the genera have been described by a single investigator in a single paper, and the investigations have too often been based upon exotic material or material growing naturally, but collected and sent to the investigator from a considerable distance. The variations which are known to occur in Pinus should be borne in mind while weighing accounts based upon comparatively limited material. In Pinus both sperms, together with the stalk and tube nuclei and some of the protoplasm, are discharged into the egg. In Taxus CONIFERALES 335 baccata only the larger sperm enters the egg, the smaller sperm and both the stalk and tube nuclei remaining in the pollen tube. In Taxus canadensis all four nuclei enter the egg. In Torreya taxifolia’? only the larger sperm enters. In the Podocarpaceae all four, with some of the prothallial nuclei, enter, but all except the functional sperm degenerate at the top of the egg. In the Araucariaceae both sperms and some small nuclei enter the egg. It should be remem- bered that the stalk, tube, and prothallial nuclei look alike at this time. The sperm nuclei have become so large that they are easily distinguishable from the rest, reaching a length of 150 microns in Araucaria braziliensis. GHOSE?” claims that the pollen tube pene- trates the egg. In Cephalotaxus drupacea and C. fortunei both sperms get into the egg, but one remains at the top and disorganizes there. In the Taxodiaceae generally only one sperm gets into the egg, but in Sciadopitys both get in. Here it should be noted that the archegonia are scattered in Sciadopitys, while in the rest there is an archegonium complex. In Juniperus usually only one sperm gets into the egg. It is large in proportion to the diameter of the egg, and its wall is torn off and remains at the top of the egg, while the nucleus, surrounded by much of its cytoplasm, moves down to fuse with the egg nucleus. When the second sperm enters it remains at the top of the egg, ap- parently undisturbed, recalling the condition in cycads, where even the ciliated band is not disturbed when a second sperm enters the egg. The first sperm opens the way into the egg so that a second sperm enters easily. With the sperm, more or less cytoplasm enters the egg, where it remains at the top and mingles with the general cytoplasm. When the sperm is a highly organized cell, the cell wall comes off soon after the sperm enters the egg, and the nucleus, with most of its surrounding cytoplasm, moves down and unites with the egg nu- cleus. The cytoplasm of the sperm forms a dense sheath, surround- ing the fusing nuclei. The sheath is particularly conspicuous in Juniperus,“ 4 Thuja,3’ Torreya,? and Tetraclinis.% Whether the sheath carries any hereditary characters is doubtful since, in most conifers, only the nucleus reaches the egg. 336 GYMNOSPERMS Nuclei left at the top of the egg nearly always degenerate prompt- ly. Occasionally a nucleus divides, but no embryo is formed, and there seems to be no influence upon the development of the func- tional embryo, which is at the opposite end of the egg. What causes the sperm to move toward the egg nucleus has not yet been determined. In most cases there is a mechanical impulse when the sperm is discharged from the pollen tube, but this is not sufficient to account for much of the movement. Chemotactic ex- periments, which succeed so well with pteridophytes, fail just as they do in cycads. One thing is certain: something brings the nuclei to- gether, and the nuclear membranes are dissolved at the line of con- tact, so that the two nuclei become surrounded by a common mem- brane contributed by both gametes. It is a curious fact that the first two cases of fertilization described for conifers were abnormal. STRASBURGER,5® as early as 1878, in his Befruchtung und Zelltheilung, figured two gamete nuclei of about equal size in Picea vulgaris. Nearly 20 years later, COULTER"? described two fusing gametes of about equal size in Pinus laricio. There can hardly be any doubt now that the “‘sperm,”’ in both these cases, was an enlarged ventral canal nucleus, since normal fertiliza- tion has been observed so often in both these species, and the sperm nucleus has been proved to be very small in comparison with that of the egg (fig. 327). In the nineties it used to be taken for granted that gamete nuclei fused in the resting condition, the assumption being based upon the appearance of this stage in angiosperms. As technique improved, it became evident that this would not hold for Pinus'™ (fig. 327 B, F). Even Coutrter’s'? earlier account, showing two gamete nuclei of nearly equal size, also showed them in the spireme stage, although this fact was not mentioned in the paper (fig. 328). Later, the in- dependence of the two groups of chromatin was observed in Junip- erus,*** 4° and probably any detailed observation, with modern tech- nique, will show such a behavior to be general, if not universal. In Lilium, upon which the most plausible claim for fusion in the resting condition was based, it is now known that the two gametes form two spiremes and that there is not any fusion of chromatin. The most detailed investigation of the behavior of chromatin at CONIFERALES 337 fertilization is HutTcHINSoN’s’’> account of Abies balsamea. He found the separate spiremes of the two gametes and their segmenta- tion forming two groups of chromosomes, each with the x number. Fic. 327.—Fertilization in Pinus: A, Pinus sylvestris, sperm nucleus much smaller than the egg nucleus; June 19; X 135; after BLACKMAN; B, P. laricio, nuclei of egg and sperm in spireme stage; X 500; after CHAMBERLAIN;'4 C-F, P. strobus, later stages than B, all showing that there has been no fusion of chromatin; X 315; after Dr. MARGARET FERGUSON.!% 338 GYMNOSPERMS Two spindles are formed, which soon come together and appear as one. The most unexpected feature in his account is that the chromo- somes become paired (fig. 329). In this diagram, only two of the 16 chromosomes of each gamete are shown. The chromosomes of the egg nucleus are shown in black, and those of the sperm, in outline. Fic. 328.—Pinus laricio: union of two nuclei originally described as male and fe- male, but which are, doubtless, the egg nucleus and the ventral canal nucleus. f, fe- male; m, male, The chromatin is in the spireme stage; X 500.—After COULTER." The chromosomes of the egg and sperm pair, and then twist about each other, as in the third and fourth figures of the diagram. They then split transversely, giving rise to groups of four, which resemble the tetrads of the reduction division (fifth and sixth figures of the diagram). The distribution of the chromosomes then takes place as shown in the final figure of the diagram. CONIFERALES 330 In a form with short, thick chromosomes, at early metaphase of the first division in the fertilized egg, it would be easy to count only the « number of chromosomes; but each of such chromosomes would really be quadrivalent, and in the anaphase the 2x number would appear. This is the only account of such a behavior of chromosomes at fertilization in any gymnosperm, although HuTcHINsoN believes that my own figures and especially Dr. MARGARET FERGUSON’S'”® = Aw vx Fic. 329.—Abies balsamea: diagrams showing the behavior of the chromosomes at fertilization and at the first mitosis of the fertilized egg. The chromosomes of the egg nucleus are shown in black; those of the sperm, in outline. Only two of the sixteen chromosomes of each gamete are shown.—After HUTCHINSON.?75 more detailed drawings, are better interpreted according to his dia- gram. In Puccinia graminis the two gamete nuclei appear at the begin- ning of the aecium and divide in pairs throughout the aecium stage and through the urediniospore and teliospore stages, coming to- gether under one membrane only at the close of the teliospore stage. In the crustacean, Cyclops, the two gamete nuclei also divide in pairs throughout a lesser part of the life-history before they come together under one nuclear membrane. In Abzes the chromosomes of the gamete nuclei are distinct after the two groups of chromo- somes are included within a single nuclear membrane. Then there is a pairing of male and female chromosomes. That such a pairing takes place at the reduction division, every cytologist has observed. In Abies such a pairing also takes place at the reduction division. Whether such a pairing as HUTCHINSON described for Abies may not be general is at least worth a careful study. It should be noted that the chromosomes in A bies balsamea are large and much more favor- able for study than those of Pinus or Juniperus. 340 GYMNOSPERMS A> ] In a recent paper, BEa.s describes the first mitosis in the fertilized egg of Pinus banksiana (fig. 330). He finds two groups of chromo- somes, as others have found them, and finds that the two spindles quickly merge into one multipolar diarch spindle, on the equator of which the chromosomes lie in one group, with the chromosomes of the two gametes indistinguish- able from each other. Each chromosome splits longitudi- nally, and the halves pass to the poles, 24 chromosomes to each pole. He finds no pair- ing or transverse segmenta- tion. Of course, this does not prove that there is no pairing or transverse segmentation in Abies; and the situation de- scribed for Abies does not prove that there is a pairing and transverse segmentation in Pinus. Both accounts need either confirmation or correc- tion. It seems unlikely that Fic. 330.—Pinus banksiana: first division two conifers, in the same of fertilized egg; 150.—From an unpub- ; : lished drawing by J. M. BEALt.35 family, would differ so de- cidedly. Referring again to the first two cases of fertilization in gymno- sperms, described by STRASBURGER,5® in 1878, for Picea vulgaris, and by Coutter,"” in 1897, for Pinus laricio, we must note that in both there was doubtless a fertilization of the egg by the nucleus of the ventral canal cell. My own work on Pinus made this interpreta- tion practically a certainty.'°* Ixeno**" found fertilization of the egg by the ventral canal nucleus in Ginkgo, and SEDGWICK‘*? found that in Encephalartos, in all probability, the same thing occurs. HutTcHINsoNn’’s finds that this occurs occasionally in Abies. CONIFERALES 341 Double fertilization, as it occurs in angiosperms, may not occur in gymnosperms; but it is worth noting that LANb**’ found fertilization of both the egg and ventral canal cell nucleus in Thuja; and Hutcu- INSON?’?5 found the same thing in Abies. In both, there were two similar groups of cells, one at the top and the other—the young embryo—at the bottom. The term “fertilization” is not easy to define. Practically, when teachers order slides showing fertilization from biological supply companies, they accept anything showing the sperm within the egg. CHAPTER XVI CONIFEROPHY TES—CONIFERALES—Continued THE EMBRYO In dealing with fertilization we have already begun a description of the embryogeny, for we have described the fertilized egg and some features of its first mitosis. The fertilized egg is the first cell of the 2x generation. Although some of the early cells of the 2x generation take no part in the formation of root, shoot, cotyledons or leaves, the entire 2x generation, up to the ripe seed, should be included in any account of the embryo. The proembryo.—In all the living pteridophytes, the first division of the fertilized egg is followed by the formation of a cell wall. There is no free nuclear stage for, throughout the entire group, the egg is small and easily segmented. As the egg increased in size, probably in some extinct heterosporous ancestors of the gymnosperms, a free nuclear stage appeared, and became more prolonged as the egg con- tinued to increase in size. The climax was reached in living cycads, where Dioon shows more than a thousand free nuclei before wall for- mation begins. In the higher cycads, like Zamia, the number of free nuclei becomes as low as 256. In the living Coniferophytes, Ginkgo, with 256 free nuclei, has the highest number. In the conifers the eggs are comparatively small, and the number of free nuclei correspond- ingly low. In only one species in the entire conifer line has a wall been reported at the first division of the fertilized egg. In the im- mense Sequoia sempervirens, which has an egg only roo microns in diameter (almost as small as some of those in living pteridophytes), there is no free nuclear stage. A strong cell plate appears at the telophase of the first division of the fertilized egg, and a distinct wall is formed before the second division occurs (fig. 331). In the rest of the conifers, as far as they have been studied, the embryogeny begins with a free nuclear period. Pinus will serve as a thoroughly investigated example. In the free 342 CONIFERALES 343 nuclear divisions the figures are entirely intranuclear, and there is a great display of achromatic structures, especially at the first mitosis (fig. 332). The next two mitoses follow in such rapid succession that the entire free nuclear series may be secured at one collection (fig. 333). These four nuclei, which constitute all of the free nuclear pe- riod, are formed at about the level of the egg nucleus. After the sec- ond mitosis, they pass to the bottom of the egg, where the third mi- tosis, closing with wall formation, takes place. Fic. 331.— Sequoia sempervirens: first Fic. 332.—Pinus laricio: first division division of the fertilized egg. A strongcell of nucleus of fertilized egg. The figure is plate is formed on the spindle, and the entirely intranuclear; X500.—After wall will be completed before the second CHAMBERLAIN.?°%4 division begins; X 500.—After LAwson.344 The positions of nuclei are worth noting. Throughout the develop- ment of the archegonium, from the first appearance of the arche- gonium initial up to the formation of the ventral canal cell, the nu- cleus is at the top. After the ventral canal division, the nucleus moves down to the middle of the egg; and after two mitoses, the four nuclei move down to the bottom of the egg. Perhaps pollen tubes and the disorganization of the nucellus may have something to do with the earlier position, and nutrition, which, in later stages, comes more from beneath, may cause the later movement. When walls come in at the close of the third mitosis they form two tiers of cells with four cells in each tier. The cells of the upper tier Fic. 333.—Pinus laricio: A, the two spirems of the gamete nuclei within the egg nucleus; a detail is shown in fig. 327 B; B, four-nucleate stage; proteid vacuoles are par- ticularly distinct and, at the top, are two large vacuoles which might be mistaken for pollen tubes; C, third mitosis, showing two of the four dividing nuclei; D and £, eight- celled stage; /’, three tiers of four cells each, with nuclei of the lowest tier dividing; G, four tiers of four cells each; H, cells of the suspensor tier elongating; p, pollen tube; , vacuoles; r, rosette tier; s, suspensor tier; A-C, June 25; D-G, July 2; X104.—From COULTER and CHAMBERLAIN, Morphology of Gymnosperms's4 (University of Chicago Press). CONIFERALES 345 then divide, so that there are three tiers; and a division in the lowest tier completes the intra-oval stage of the embryogeny. Some details of these mitoses are shown in Dr. Nitstve KIHLpAHL’s3” study of Pinus (fig. 334). The walls dividing the lower part of the egg into Fic. 334.—Pinus laricio: A, achromatic structures as beginning of third mitosis of fertilized egg; B, late anaphase of third mitosis which will give rise to two tiers of four cells each; C, late telophase of same mitosis with walls nearly complete. The cells of the uppermost tier never have any wall on the sidé next the general cytoplasm of the egg; X350.—After Dr. NrtstnE KrHLpAHt.3?? four tiers are formed on the central spindle in the usual way, while the walls dividing each tier into four cells are formed on the strong threads radiating from the poles of the figure. Root, shoot, cotyledons, leaves, and secondary suspensor cells (embryonal tubes) all come from the lowest tier, the other three tiers contributing nothing to these structures. 346 GYMNOSPERMS Four free nuclei seem to be quite general in the Abietaceae, Taxo- diaceae, and Cupressaceae, although mistakes could be made in stat- ing the number. One might think that there would be eight free nuclei in Thuja; but LaNnp’s37 complete series showed that, while A Fic. 335.—Thuja occidentalis: A, third mitosis in fertilized egg. Walls will come in at the close of this mitosis, forming eight cells; B, organization of tiers; C, elongation of suspensor; 7, ventral nucleus; * 425.—After LAND.337 there are four mitotic figures lying in a common mass of protoplasm, definite walls are formed at the close of this mitosis (fig. 335). The largest number of free nuclei in the conifers is found in the Araucariaceae. BuRLINGAME'S found 32 in Araucaria braziliensis; and Eames" found usually 32 with occasionally 64 in Agathis aus- tralis. CONIFERALES 347 Except for the Araucariaceae, the number of free nuclei in the Taxares is regularly higher than in the Pinares. Phyllocladus,*3 Aus- trotaxus,4°> and Cephalotaxus'* have 8; Podocarpus coriaceus,™ P. macrophyllus, and P. totora,®3 16; and Taxus baccata,*5 16, with an occasional 32, before walls are formed. Torreya'® is exceptional, however, in having only four free nuclei when walls come in. The significance of the longer and shorter free nuclear periods will be discussed in the chapter on ‘‘Phylogeny.” In Pinus the four tiers of cells, with four cells in each tier, are al- most geometrical in their symmetry and arrangement; and their function is just as definite. The upper tier, with the distal ends of its cells open to the general mass of the cytoplasm of the egg, is active in the nutrition of the parts below, as long as any food material re- mains in the egg. The next tier below is called the “rosette,” from its appearance in vertical view. Its cells are actively meristematic and often develop into embryos. The cells of the next tier elongate immensely, to hundreds of times the length of the original tier, and constitute the suspensor, which may be called the primary suspensor to distinguish it from the structures which appear later and might be mistaken for it. Only a part of the cell progeny of the lowest tier develop into em- bryos, and none of the other tiers have any part in the formation of the root, shoot, cotyledons, or leaves. A differentiation of the proembryo into tiers is characteristic of conifers, and even in the massive proembryo of the cycads and Gink- go there is a differentiation into regions, although not so sharply marked. In the Abietaceae there are, prevailingly, four tiers; in the Taxo- diaceae and Cupressaceae, three tiers; in the Araucariaceae there are three tiers (fig. 336), the upper forming the suspensor, the middle one giving rise to the embryo, and the lower merely forming a tem- porary protective cap, which soon disorganizes. In Cephalotaxus three tiers are described: the upper, a rosette; the next below, the suspensor; and the lowest, the group of embryo cells. In the lowest group the lowest cell, or sometimes two cells, acts as a protective cap, as in the Araucariaceae (fig. 337).° Although the development of the embryo is a continuous process, 348 GYMNOSPERMS it is convenient, for the sake of reference, to have some name for the stages before the growing sporophyte breaks through the base of the egg. The term “proembryo” has long been used for this part of the embryogeny. Later embryogeny.—While stages in the development of the em- bryo have been studied, often rather thoroughly, ever since the time Fic. 336.—Araucaria braziliensis: A, the proembryo fills the entire egg: the shaded cells are the embryo proper; the cells above form the suspensor; and the cells below, the protective cap; B, later stage; the protective cap at the left is being pushed off; many of the cells above are “embryonal tubes.” & 153.—After STRASBURGER.®° of HorMEISTER,’” investigations of later stages had been compara- tively desultory, until BucnHorz®’* made a critical examination of nearly all the genera of the order. With a new and ingenious tech- nique, he was able to remove the embryos entire and uninjured. With such preparations, many of them examined in the living condi- tion, and all of them after critical staining, it was easier to observe and safer to interpret, especially in the numerous cases of polyem- bryony. More than one embryo is so common in the gymnosperms that it might be called their most distinguishing characteristic (fig. CONIFERALES 349 338). While more than one, and often many, begin to develop, the ripe seed usually has only one which has developed sufficiently to pro- duce a seedling. More than one embryo may orig- inate in different ways. More than one egg in a gametophyte may be fertilized and one embryo may de- velop to a considerable extent from each fertilized egg. This is simple polyembryony. When more than one FIG. 337.—Cephalotaxus fortunei: A Were Fic. 338.—Pinus banksiana: cleavage — » polyembryony: s, primary suspensor; er, proembryo, still within the egg, showing €2, €3, first, second, and third secondary the rosette, suspensor (s), embryo, and suspensor cells (embryonal tubes); 7, ro- protective cap: B, later stage, after break- _sette; p, basal plate above rosette formed ing through the base of the egg; X63.— from disorganized remains of egg; X 100.— After STRASBURGER.©° After BucnHorz.® 350 GYMNOSPERMS embryo comes from a single egg by the splitting of the product of a single fertilization, the term cleavage polyembryony is used. Both types are widely distributed in the conifers. In the Abietaceae, cleavage polyembryony is a constant feature of Pinus, Cedrus, Pseudolarix, and Tsuga; while simple polyembry- ony is the rule in Larix, Picea, and Pseudotsuga. In Abies, cleavage occurs in rare cases: simple polyembryony is the rule. In the Taxodiaceae, cleavage polyembryony is excessive in Scia- dopitys, is not so frequent in Taxodium, and does not occur in Se- quota. In the Cupressaceae, cleavage polyembryony occurs, at least oc- casionally, in Juniperus, Thuja, Biota, Actinostrobus, Widdringtonia, and Callitris, but not in Tetraclinis. In the Araucariaceae, neither Araucaria nor Agathis has any cleavage polyembryony. In the Podocarpaceae cleavage polyembryony is not so prevalent. A form of polyembryony described as determinate cleavage poly- embryony has been described for Dacrydium,® and in Podocarpus coriaceus some of Coker’s figures suggest cleavage polyembryony, but most species of Podocarpus (for example, P. spicatus’), and also Phyllocladus, Saxegothea,*”’ and Microcachrys*® seem to have simple polyembryony. In the Taxaceae JAGER**s found that in Taxus baccata some of the upper suspensor cells occasionally break away and form small em- bryos; but even this last trace of polyembryony is usually lacking, and it does not occur in the other genera. Simple polyembryony occurs throughout the order except, of course, in forms like Torreya, with only one archegonium. Bucunorz * believes that cleavage polyembryony is a primitive character, and that simple polyembryony has been derived from it. In some forms, like the Araucariaceae, Cephalotaxus, and some spe- cies of Podocarpus, the cap, very probably not a primitive feature, may prevent cleavage; and in others occasional cleavage is inter- preted as a reversion. With this interpretation the Pinus type is very primitive, and the Araucaria type, very advanced. The rosette cells, where they are present, are embryonic, and often CONIFERALES 351 produce embryos (fig. 339 F). They are highly developed in Pinus, where all four of them often produce embryos; so that, with the other four embryos, a single fertilized egg may produce 8 embryos. When three eggs are fertilized, as often happens, 24 embryos may start in a single seed. Since it is possible that six or seven eggs may be fertilized, there could be 48, or even 56, embryos in a young seed. Rosette embryos are also the rule in Cedrus, but rare in Abzes and Tsuga. In the rest of the Abietaceae, except in Pseudotsuga, there are rosette cells, but they do not develop into embryos. Rosette embryos are found in Sciadopitys, but not in Taxodium; they are found in Arthrotaxis, but not in Sequoia. The initials for rosette cells may occur in Podocar pus spicatus, and they are found in Cephalotaxus, where they may develop into embryos; but they are not found in Taxus or Torreya. Rosette cells sometimes elongate like suspensor cells and function like them. CoKER noted that, in Podocarpus coriaceus,™° a single apical cell, and, later a group of cells, are binucleate. BucHHoLz found the same condition in Dacrydium cupressinum, and, in a forthcoming paper, has greatly extended the range of forms which show this character to 12 species, so that all members of the Podocarpaceae, except Saxagothea and Microcachrys, probably have binucleate embryonic cells. A suspensor, at least in the form of embryonal tubes, characterizes the entire order, and perhaps the entire gymnosperm phylum. The primary suspensor cells, which are well illustrated by the four elon- gating cells of the suspensor tier in Pinus, do not divide. When there are two or more cells in the suspensor, looking as if they had come from transverse divisions in the primary suspensor cells, the addi- tional cells have come from transverse divisions in the lowest tier of cells of the proembryo (fig. 339 A and £). Later, such a secondary suspensor cell may divide periclinally. These secondary suspensor cells have been called “embryonal tubes” and in later embryogeny they may become very numerous. In some of his more recent publications BucHHOLz has made use of the term prosuspensor. This part of the suspensor system seems to 352 GYMNOSPERMS be more than the group of greatly elongating cells which emerges from the proembryo. It is a special part of the suspensor system which elongates very greatly, and in such forms as Sciadopitys, may precede the formation of the primary suspensors. Each embryo of Sciadopitys, which appears later, is borne on its own primary sus- pensor. Thus, Pinus may be considered as having no prosuspensor, but a group of four primary suspensors placed parallel to each other, followed by the embryonal tubes, which together constitute the sec- ondary suspensor. Dacrydium has a prosuspensor, but omits the for- mation of the primary suspensor, and forms, instead, a massive sec- ondary suspensor on one or more of the several embryos. The very long suspensors which appear on the early embryos of podocarps would therefore be described as prosuspensors. Prosuspensors seem to be recognizable in conifers above the level of the Abietaceae. There can be no doubt that the highly developed suspensor pushes the growing embryo down into the gametophyte, which is richly stored with starch and other foodstuffs. As the foodstuffs become liquid, the tissues surrounding the embryos break down, so that a large cavity is formed. The elongating suspensor, as it kinks and coils, fills most of this cavity, thrusting the embryo deeper and deep- er into the gametophyte. After the embryo has reached its maximum length, the suspensor, together with the remains of the egg and nu- cellus, forms a dry cap which may protect the root end of the embryo as it breaks through the testa. In the early embryogeny there is frequently a definite apical cell, segmenting like the apical cell of many pteridophytes (fig. 339). In B and C of this figure, the apical cell and its segments are indicated by stronger lines, while the subdivisions of the segments are indi- cated by lighter lines. At the stage shown in D there is no longer any segmentation of the higher fern type, but a growth by a group of initials. There are apical cells in other conifers, Sciadopitys, Thuja, Junip- erus, and others, but the apical cells do not cut off so many seg- ments. Embryos with caps, and many embryos without cleavage polyembryony, like Picea, Larix, and Abies, have no apical cells. In Thuja occidentalis and Taxus, there seems to be an apical cell. is 2X5 o| a] S Re) sy ia fb = re) ee Q Z ih] ag a5 0G [O Peak 6) ** in Die natiirlichen Pflanzenfamilien, estimates the number of species at 35. Pearson“ lists 32, not including the well- marked Ephedra compacta, found near Tehuacan, Mexico. The species are somewhat equally divided between the Western and Eastern hemispheres. PEARSON assigns 17 species to the Old World and 14 to the New, with 6 to North America (there should be 7 with E. compacta), and 8 to South America. In North America they are most abundant in New Mexico, Arizona, and California, although they are found in the states border- ing these on the north. Only one species, Ephedra antisyphilitica, gets as far east as Texas. There are some in northern Mexico, but none have been reported between northern Mexico and the small Ephedra compacta, near Tehuacan. In South America they are most GNETALES—EPHEDRA 363 abundant in the Andes regions of Bolivia and Paraguay, but in Paraguay they also extend across the desert plains even to the Atlantic Ocean. A few are found farther south in Chili and Pat- agonia. In the Eastern Hemisphere, the western limit is the Canaries and Madeira. There are species in the western Alps, in western France, and in some places along the shores of the Mediterranean and Black Fic. 344.—Ephedra: near Palm Springs, California. Ephedra, in this region, often acts as a sand binder. Sea. The genus crosses Arabia and Persia, through northern India, and then northeast through China, across the upper basins of the Yang-tse-Kiang and Hoang-Ho rivers. Ephedra distachya gets into Siberia, very far north for a genus which belongs typically to a warm climate. No species are common to the Western and Eastern hemispheres, and no species seems to cross the Equator. The geographic distribution, with such scattered patches, would indicate great age; but, as already noted, there is scarcely any geological history, no material having been found below the Tertiary. 304 GYMNOSPERMS THE SPOROPHYTE—VEGETATIVE The general appearance of the genus is sufficiently characteristic to make it easily recognizable. Nearly all are straggling shrubs, — - <.+ Fic. 345.—Ephedra trifurca: female plant near Tucson, Arizona. The larger branches at the surface of the ground, and even below it, are 5 cm. in diameter. usually less than two meters in height, and some less than half that height. Ephedra compacta, a very compact, profusely branching form, GNETALES—EPHEDRA 365 near Tehuacan, Mexico, is usually about 30 cm. in height, and sel- dom reaches 50 cm. On the other hand, the stem of the South American E. triandra sometimes reaches a diameter of 30 cm. and a height of several meters. What the age of these large plants might be has not been determined. LANpD3%* reported 40 rings in an old plant of E. trifurca. Since this is the largest number of rings reported for any species, Ephedra is a short-lived plant. , AAU ANIA LLY ASMA TODA OEY IT A eel Pe ES aINd feet ce Zcitetig He eer ane YO [\ q SOQ Y Ni WW) a og © 2201s } HN Vee ares Mt @rgy Fee RU A Sica poe 187 Jeo nar Tene @sMMesy weenie an yANoor Sa n@c@0 (ee EVrigeers Al ernea. | oni earaC oan) IA fag Roe M0) oma OI wOver2aus SAIC ESO i A WAS 10 OA | HORUHSSeOns) ewes! Bl aeeayee MO nig! @ \ Dea PEP CLA ach re Fic. 346.—Ephedra trifurca: transverse section of part of stem about 1 cm. in diam- eter, showing numerous vessels of various sizes. The largest vessels are formed earlier in the season, as may be seen from the vessels above the growth-ring. A large medul- lary ray is shown in the middle of the figure; 170. Branching begins very early, buds appearing even in the axils of the cotyledons. In older plants the first branches are often covered by soil (fig. 345). When the soil is loose, and conditions are favorable, buds at the nodes may grow out into long rhizomes, and buds on the rhizomes may develop into new plants. LAND? found numerous 366 GYMNOSPERMS plants of Ephedra trifurca developing in this way. Several conifers are reproduced by underground buds. The seedling has a strong tap root, which persists for a long time, but is gradually overtaken by adventitious roots. Root hairs are abundant, but there is no mycorrhiza. The leaves are small and rudimentary, and, even when young, are of scarcely any importance in the vegetative economy of the plant. They are opposite or in whorls of three, very rarely of four. The leaves of Ameri- can species are usually in whorls of three. In all cases the whorls alternate. There is hardly anything which could be called a blade. A thick and often fleshy midrib has a thin border, thicker and sheathing at the base, and with scarcely any stomata. The border, at first whitish, soon becomes brown, and withers. An ab- scission layer develops early, and the leaf falls off, leaving a broad scar. The leaf trace is double, and extending into the leaf Fic. 347.—Ephedratrifurca: 8 its only vascular feature. The double longitudinal radial section of trace does not extend beyond the node stem about 1 cm. in diam- above itsinsertion. In both leaf and stem eter, showing one large vessel A and several tracheids; 170. the trace is endarch. Histology.—At the apex of the stem and root there are no separate initials for plerome, periblem, or dermato- gen, and in the root there is probably no dermatogen at all. A general meristem differentiates later into the body regions. The most striking histological feature of Ephedra, and of the other two genera, is the presence of vessels in the secondary wood (fig. 346). The vessels are modified tracheids of all sizes, from that of the ordi- nary tracheid up to several times that diameter, the largest being formed early in the spring, smaller ones later, and still later in the season none at all. The pits, at first bordered, enlarge and lose both torus and border, so that they become mere perforations. Such per- forations often fuse, and when they fuse on the oblique end walls the structure becomes more and more like the continuous tube of the C0 0 OSGo00. ©) O_o SD OOOK OA OED ADAG ©__ DP D_ AGO ao HOLOKG GNETALES—EPHEDRA 367 angiosperms. There is no other gen- eral breaking down of the oblique walls to form continuous vessels. Most of the wood consists of tra- cheids with bordered pits, usually uniseriate, and more abundant on the radial face, but there is also abundant tangential pitting (fig. 348). Medullary rays are very long and very wide (fig. 348). Only about one-third of the two long rays is shown in the figure, with the top of another at the lower left. The cells of the rays are of various shapes and sizes: isodiametric, elongated, and curved. The walls are thick, lignified, and pitted, but the pits are small and simple. Rays in young plants are uni- seriate, but, later, they become multiseriate by longitudinal divi- sions in the uniseriate ray, by the incorporation of adjoining cells, or by the fusion of uniseriate rays. When there is a fusion of uniseri- ate rays, separated by a few tra- cheids, the tracheids are incorpo- rated in the large resulting ray. Such rays have no relation to leaf traces. With practically all of the cells thick walled and lignified, it is natural that the wood should be very hard. In young stems the rigidity is also increased by lignifi- cation of the pith just above the ame OX [S\O) oe Lolo} Sai BLO @) IS ©: O00 Sa S50) Square SS 9] Loe Us ————_ CNS) SEO) __O999509_ BGOGS SO) aS SA (9) {S] O99 O900 50 Hs SS KS c\ Tere SR TPA ee =—— eae x @ nO) ® DIZ ES rS) (Qe35) OOO 9 SO 12 5 OA z So 1° EO) 2 10) \a\lilala} tall 3 aN e Fic. 348.—Ephedra trifurca: longi- tudinal tangential section, showing the ends of two large medullary rays, and a tip of a ray at the lower left. Many cells of the ray are pitted, and in the long tracheids there is both radial and tangential pitting; 170. 368 GYMNOSPERMS nodes, giving the general impression of partitions in bamboos and other grasses. The small, more or less upright, green stems, which give Ephedra such a characteristic appearance, do the photosynthetic work. The long-fluted internodes and reduced leaves make these small shoots look like Equisetum, and their internal structure also resembles that of the ancient pteridophyte (fig. 349). They grow from a small meristem at the base, and many of them fall off at the end of the growing season, thereby effecting a very advan- tageous reduction in the trans- spiration surface. In some there is a vigorous develop- ment of secondary wood, and these, later, produce whorls of small branches. Branches of all sizes are shown in fig. 345. The epidermal cells are thick walled, with a group of very thick-walled cells under the ridges; stomata are distrib- we uted at the bottom or along Fic. 349.—Ephedra trifurca: transverse the sides of the depression be- section of a small upright green twig, about 1 tween ridges. Between these mm. in diameter, showing two ridges, the % ‘i ¥ S ae thick-walled cells and the vas- epression between them, and a stoma at the base of the depression. There are thick-walled cular region is a zone of very cells under the ridges, and a zone of chloro- thin-walled cells with abun- phyll cells extending down to the vascular dant chloroplasts and numer- region; 270. ‘ ous intercellular spaces. The vascular cylinder is an endarch siphonostele. In the center of the pith is a group of rather thick-walled cells with very dense contents, perhaps tannin: there is no resin in the order. THE SPOROPHYTE—REPRODUCTIVE Ephedra, and also the other two genera, are dioecious, but in all of them there are traces of an ancestral bisporangiate condition. By GNETALES—EPHEDRA 369 hunting through a large patch, a monoecious individual can generally be found, and in £. foliata they are rather common. Staminate and ovulate strobili are found on the same plant, and LAND found bi- sporangiate flowers in EF. érifurca. WETTSTEIN’”? found similar flowers in E. campylopoda, and others have reported them for other species (fig. 350). When both sexes are in the same strobilus, the staminate flowers are below and the ovulate at the top, as in the bisporangi- ate strobili already described for Picea and other conifers. The staminate strobilus.—The stami- nate strobili are in whorls of 2, 3, or 4 at the nodes of the small green branches. The whorled condition of the leaves of the vegetative shoot is also present in the floral axis. There is no doubt that the staminate structure is a compound stro- bilus, a feature which, with vessels in the secondary wood, separates the Gnetales : from the rest of the gymnosperms. Fic. 350.—Ephedra campylo- The male strobilus consists of a short 2°42: bisporangia strobilus; <7. : ; : —After WETTSTEIN."7 axillary shoot, bearing 2-8 opposite pairs of bracts, of which the lower one or two pairs are sterile, while the rest bear solitary flowers (fig. 351). In the axil of a fertile bract is a shoot, bearing two thin, opposite scales, which have been interpreted as a perianth. Between these scales the axis is pro- longed, and bears the microsporangia at its top. Whether the scales should be called a perianth, is doubtful; but it is certain that they are reduced opposite leaves, more or less united at the base, covering the axis before the elongation which pushes the stamens out into a favorable position for shedding their pollen. The structure bearing the two scales at the base, and the sporangia at the top, has been called a “‘sporangiophore,”’ a good name because it is efficient for reference, like placenta and receptacle, and is just as noncommittal. In some species one or more—usually more— sporangia are sessile at the top of the sporangiophore. In Ephedra trifurca there are 5 or 6 sporangia, and each one is borne on a filament, which may be called a sporophyll, just as reasonable 370 GYMNOSPERMS an interpretation as to call the structure a branching sporangio- phore. way M0, Fic. 351.—Ephedra viridis: part of staminate flower; 2.—From an unpublished drawing by Dr. S. FLOWERs. The sporangia usually have more than one loculus and have been called synangia; but while the multilocular condition is evident, any actual fusion still remains to be proved. Fic. 352.—Ephedra campylopoda: longitudinal section of tip of staminate strobilus: m, apical meristem of strobilus; b, bract; p, “perianth”’; s, sporangiophore; a, Anlage which is to develop into b, s, and p; X160.—After STRASBURGER.S%® Fic. 353.—Ephedra trifurca: A, longitudinal section at tip of male strobilus, De- cember 20, 1902; B, longitudinal section of male flower, a month later: br, bract; a, sporangiophore; p, perianth; s, microsporangia; ar, archesporial cell; 150.—After LAnp.338 372 GYMNOSPERMS As long ago as 1872, STRASBURGER*” worked out the early de- velopment of the bract, the two scales, and the sporangiophore of Ephedra campylopoda (fig. 352). From his figure it is obvious that the sporangiophore, bearing the two scales, is in the axil of a bract. Fic. 354.—Ephedra viridis: part of ovulate plant; X 2.—From an unpublished draw- ing by Dr. S. FLOWERS. The most complete account of the life-history is by LANp,** 5° who studied abundant material, both living and fixed, of £. érifurca, through spermatogenesis, oogenesis, fertilization, and some stages in the embryogeny (fig. 353). The bract appears close to the stem GNETALES—EPHEDRA 373 tip, and the sporangiophore (a in fig. 353) soon appears in its axil. The two scales (p, in the figure) appear at the base of the sporangiophore, and the hy- podermal archesporial cells become recog- nizable, not at the center of the top, but in a circle around it. From this point up to the microspore, the development is about as in any eusporangi- ate microsporangium. The wall of the ma- ture microsporangium in Ephedra trifur- ca33® 339 ig very thin, with usually one layer of crushed wall cells between the epidermis and tapetum. The tapetal cells are very large, several times the diameter of the epi- dermis and wall cells together. In the vicin- ity of Mesilla Park, New Mexico, the locality of most of LAND’s material, the reduction division, by which the microspore mother- cell gives rise to four microspores, takes place about March 12. At the first division in the spore mother-cell, a wall begins to form, as if the bilateral type, characteristic of monocots, would result; but the incipi- ent membrane disappears and spores of the tetrahedral type are formed. The ovulate strobilus—The ovulate stro- bilus can be distinguished from the stami- nate, even in the early stages, because it is elongated and pointed, while the staminate is globular (fig. 354). The ovulate strobili, like the staminate, are in whorls of 2, 3, or 4, at the nodes of the small green branches. Fic. 355.—Ephedra trifurca: longitudinal section of ovule, showing nucellus with deep pollen chamber (p), inner (z) and outer (0) integuments, and female gametophyte with reproductive storage (s) and hausto- rial (h) regions; X 48.—After LAnp.338 — B one Fic. 356.—Ephedra viridis: longitudinal section of ovulate flower, showing some of the details diagrammed in fig. 355.—From an unpublished drawing by Dr. S. FLowERs There are more bracts than in the staminate strobilus, usually four or more pairs of sterile ones, and a terminal ovule. The ovule has two integuments, the outer consisting of four bracts, coalescent at the base, and the inner, of two bracts, also coalescent at the base, and, at maturity, splitting at the top. The inner integument, at the time of pollination, elongates immensely, and, at its tip, is a spark- ling pollination drop. The topography of the ovule at this stage is well shown in the much copied figure by LAnpb3** (fig. 355), and in a recent figure by FLowEr (fig. 356). STRASBURGER found a row of three megaspores in Ephedra campy- lopoda, and JAccarp found the same number in E. helvetica. LAND,3* in E. trifurca, found about equal numbers of threes and fours (fig. 357). Of course, in rows of three, only the lower two are megaspores: the upper one is still a 2% cell of the sporophyte. THE GAMETOPHYTES Gametophytes of Eastern Hemisphere species have been studied by STRASBURGER,* JaccaRD,?*3 PEARSON,#° BERRIDGE and San- pay,*® and Dr. STEPHANIE HERZFELD,?4 LAnp’s33® 339 work on the American Ephedra trifurca is the most complete account. 374 The male gametophyte.—E phedra trifurca will serve as a type. As in all seed plants, the microspore is the first cell of the male gameto- phyte. At its first division, a prothallial cell is cut off by a wall. At the second division a prothallial nucleus is formed, but there is no wall separating it from the rest of the spore, which is the antheridium initial (fig. 357). The initial then divides, forming a tube cell and a generative cell, which divides to form the nuclei of stalk and body cells. These two nuclei lie in a common mass of cytoplasm, and are never separated by a cell wall. At this stage the pollen is shed. When it reaches the ovule, the generative cell divides, giving rise to two nuclei which are alike in size and appearance. The exine cracks longitudinally, and the intine slips out, becoming prolonged into a short pollen tube (fig. 358). The female gametophyte—The megaspore is the nrst cell of the female gametophyte (fig. 359). Its germination begins, as in all gym- nosperms, with a period of simultaneous free nuclear divisions (fig. 359 B). The free nuclear period lasts about 20 days and walls begin to appear at the 256 nucleate stage (about April 1, in 1903, which was a very dry season). Wall formation begins at the outside and rapidly proceeds toward 375 376 GYMNOSPERMS the center until the gametophyte is cellular throughout. Even before the gametophyte is completely cellular, there is a differentiation into an upper reproductive region, consisting of large elongated cells, and a lower nutritive region, with small cells (figs. 355 and 356). Later, the lower part is further differentiated into an upper storage region, below which the cells act as a haustorium (fig. 355 4, s). Fic. 357.—Ephedra trifurca: development of male gametophyte: # and #r 1, first pro- thallial cell; pr 2, second prothallial cell; ai, antheridial initial; én, tube nucleus; g, gen- erative cell; bn, nucleus of body cell; s/n, nucleus of stalk cell; 1,500.—After LANp.338 Usually there are two archegonia, occasionally only one, and rarely three. The archegonial initial soon cuts off a neck cell, which divides several times, forming four or five tiers, before any anticlines appear (fig. 360). As many as eight tiers were observed, and since each divides anticlinally into four or more cells, there are seldom less than 32 neck cells, and often more. Some of the divisions are so irregular that the neck is not sharply marked off from the surround- ing cells (fig. 360). GNETALES—EPHEDRA When the nucleus of the cen- tral cell divides, no wall is formed between the ventral canal nu- cleus and that of the egg. The ventral canal nucleus may then move down with the egg nucleus, or may remain at the top of the egg. FERTILIZATION A dense mass of cytoplasm, which is evident below the nu- cleus of the central cell, becomes denser, surrounds the egg nu- cleus, and extends for a long distance below it as fertilization takes place (fig. 361). Similar ST Fic. 358.—Ephedra trifurca: A, pollen grain just before formation of pollen tube; the nucleus of the body cell is dividing to form two sperms; B, pollen tube: m: and M2, the two sperms; s, nucleus of stalk cell; ¢, tube nucleus; X500.—After LANp.338 masses have already been noted in connection with both egg and sperm nuclei of conifers. Fic. 359.—Ephedra trifurca: A, row of four megaspores, in which the lower mega- spore is enlarging, the one above it is breaking down, and the mitosis for the formation of the other two megaspores is in late telophase; B, simultaneous free nuclear division to form the 64-nucleate stage of the female gametophyte; m, megaspores; X 500.—After Lanp.338 SH ee or Fic. 360.—Ephedra trifurca: A, neck cell and central cell of archegonium; m, neck cell; B, neck consisting of three tiers, central cell not yet divided; C, later stage, with neck cells not easily distinguishable from surrounding cells; and central cell (c) still not divided: A and B, X 500; C, X112.—After Lanp.338 GNETALES At the time of pollination, the pollen grains fall on the pollina- tion drop and are drawn down the long micropyle. By this time such a deep pollen chamber has been formed by the breaking down of cells in the upper part of the nucellus that the pollen grains fall directly upon the fe- male gametophyte. There would be no place for a pollen tube, were it not for the very long neck of the archegonium. The short pollen tube forces its way through the neck of the archego- nium and discharges all four of its nuclei into the egg. Whether there is any “double fertilization,” like that in the an- giosperms, is questionable; but several observers have noted di- visions at the top of the fertilized egg, as well as at the bottom. Lanb%9 found jacket cells disin- tegrating at the time of fertiliza- tion, so that their protoplasm and nuclei mingled with those of the egg. The second male nucleus and ventral canal nucleus were also in this region. As the ferti- lized egg nucleus divided, there were also numerous divisions at the top of the egg. These small cells are soon absorbed by the growing embryos, functioning like the endosperm of angio- sperms, and LAND*%? believes EPHEDRA 379 Fic. 361.—Ephedra trifurca: fertiliza- tion: pt, pollen tube; 0, egg nucleus; mz and mz sperms; mz is fusing with the egg nucleus; v, ventral canal nucleus; X215.—After LAnp.339 380 GYMNOSPERMS that there is a suggestion of the origin of endosperm in the angio- sperm sense (fig. 362). LAND*? did not show a definite fusion of the second male nucleus with the ventral canal nucleus. Tale eecpanad ~ ‘ we a? R22 peer >. Fic. 362.—Ephedra trifurca: proembryos: A, group of spindles from second male cell and, below them, small cells which have come from jacket cells, or from the division of the non-functioning proembryonal cells, or both; farther down, two functional pro- embryonal cells; B, three of five proembryonal cells, also jacket nuclei in egg cyto- plasm, and free masses of cytoplasm; X 500.—After LAND.339 Miss BERRIDGE and Miss SANDAy* found a similar situation in Ephedra distachya, but thought the nuclei at the top of the egg fused in pairs and formed functional embryonal cells. Nuclei in pairs are GNETALES—EPHEDRA 381 not rare; but LANp%? found many binucleate cells in the normal jacket. Dr. STEPHANIE HERZFELD** figured and described a fusion of the second sperm nucleus with the ventral canal nucleus, and did not hesitate to call it double fertilization in the angiosperm sense (fig. 363). In conifers there have been reports of divisions in the top of the egg, variously attributed to divisions of the second sperm, or the ventral canal nucleus; and Hutcuinson?’> described and figured the fusion of the second sperm with the ventral nucleus in Abies balsamea. That nuclei in the rich protoplasm at the top of the egg should divide, seems very nat- ural, whether the division might be pre- ceded by any fusions or not; and it also seems probable that any such product would serve to nourish the growing embryo; but that there is any genetic continuity between this phenomenon and the double fertilization in angiosperms—in other words, that the angio- - “ ome ; Fic. 363.—Ephedra sperms inherited double fertilization from _., pyldponan double the gymnosperms—seems more than doubtful. fertilization”: e, egg nucleus; s:, first EMBRYOGENY sperm; v, ventral nu- STRASBURGER,” in 1876, outlined the embry- cleus; 52, second ogeny of Ephedra altissima, and LAND, in 1907,°? Reng ee: as made a thorough investigation of E. trifurca. There is a free nuclear period with, usually, 8 free nuclei, which do not sink to the bottom of the egg, but are somewhat evenly dis- tributed throughout the protoplasm (fig. 362). Around each of these nuclei there is organized a cell, which is a young embryo. So there is polyembryony without any cleavage. Usually from three to five of the young embryos begin to develop (fig. 362). The nucleus of the embryo cell divides, but the formation of a wall between the two nuclei is delayed until a suspensor tube has been formed. The em- a” Rake te . ° te Se, i x eee: ~~ ' ct ‘a. Fic. 364.—Ephedra trifurca: early development of embryo: A, unicellular embryo; B, the nucleus has divided; C, embryo and suspensor; D, mode of formation of wall sepa- rating embryo and suspensor; £, division of embryo initial cell; #, embryo and second- ary suspensor; A, B, C, and E£, & 500; D, X1,900; F, X 250; e, embryo nucleus and em- bryo cell; s, suspensor nucleus and suspensor cell; ¢, suspensor tube.—After LAND.339 GNETALES—EPHEDRA 383 bryo nucleus passes into the tube, followed by the suspensor nu- cleus, and a wall is formed between them, beginning at the outside and progressing to the center (fig. 364 D). The embryo cell then divides and the embryo develops. While three or four may reach the stage shown in fig. 364 C, the ripe seed scarcely ever contains more than one embryo. If conditions are favorable, there is no resting period: the seed germinates at once and the seedling becomes established the same season. CHAPTER XVIII CONIFEROPHYTES—GNETALES—WELWITSCHIA The most bizarre of all gymnosperms, if not of all seed plants, is Welwitschia (fig. 365). It has been found only in the coastal region of Southwest Africa, north and south of Walvis Bay, with northern Fic. 365.—Welwitschia mirabilis: large ovulate plant, about 20 miles east of Walvis Bay, on the Namib plateau.—From a photograph by PEARSON. limit 14° S. latitude and the southern, near Hope Mine, 22° 85’ S. latitude, the only extra-tropical station yet reported. So the range, north and south, is about 700 miles. In some places plants have been observed within a few meters of the sea, and they are seldom found more than 20 miles inland. Dr. WEetwitscu discovered the plant in 384 GNETALES—WELWITSCHIA 385 latitude 15° 50’ S. in August, 1860, and sent material to Sir JosrEPH HOooKER,’” with the suggestion that it be called Tumboa mirabilis; but Dr. Hooker, in his classical account, named it Welwitschia, in honor of the discoverer; and practically all of the important litera- ture has treated it under that name, in spite of international con- gresses. At Dr. Hooker’s request, the Brussels congress, in 1910, placed Welwitschia on the list of nomina conservanda, thus making the official name the same as that in the most important literature. A couple of years ago a newspaper report, in typical newspaper reporter style, claimed that a new field of Welwitschia, some 2,500 acres in extent, had been discovered in the Kaokoveld, a vast tract of jungle along the northern frontier of Southwest Africa. Prar- son*° mentions the Kaokoveld, not as a jungle, but as a desert lit- toral, extending some 300 miles south of Choricas, with Welwitschia in particularly large numbers. He says that the region is little known, and is exceptionally difficult for travelers, and that it is likely to contain still unreported Welwitschia localities. On the coast of Damaraland, a few miles northeast of Cape Cross, another patch is reported, and is said to be reproducing abundantly from seed. Still another locality is reported, about 100 miles east of Cape Cross, exceptionally far inland. While the genus extends over a considerable range, north and south, it occurs in scattered patches hard to reach. It is an extreme xerophyte, growing on rocky plains or on beds of streams, which are nearly always dry. In most of its habitats the rainfall does not ex- ceed 1 inch a year; and near Walvis Bay, the best-known locality, the average rainfall for ten years was only about one-third of an inch. However, there are heavy dews. Any botanist, contemplating a trip to Welwitschia localities, had better find out in advance how long it will take and how much it will cost. THE SPOROPHYTE—VEGETATIVE The plant has the shape of a turnip, or of an inverted cone, rarely more than 45 cm. above ground, and frequently almost covered. Prarson“® found plants with only a piece of the rim, with one leaf, above ground. The stem is elliptical, in top view, and the largest 386 GYMNOSPERMS diameter may reach more than 4 feet, 1.2 meters. Below the surface of the ground the stem tapers abruptly, and an extremely long tap root extends downward to the water table; for, in such a dry country, little or no water is taken in, except by the root. The plants sometimes grow in clusters of two or three, or even four or five, probably coming from seedlings of an older plant; and such plants, coming into contact, become fused together, so that points of union are indistinguishable. It might be called a natural graft. The stem.—The tip of the young axis is convex, but growth at the apex soon stops, while it continues near the periphery, making the apex depressed (fig. 366). Hooker?” called the part of the stem above the leaves the crown, and the part between the crown and root, the stock. The whole plant is covered by a periderm, which is thicker on the crown, and every- where harder and darker than the parts beneath. On the crown of an old plant it may reach a thickness of 2 centimeters. On the root it is sometimes loose enough to peel off like the bark. It is thickest at the top of the periphery of the crown, and diminishes to less than half the maximum thickness at the bottom of the depression, and at that level on the outer surface. Its thickness diminishes rapidly near the leaf, and there is none at all in the leaf groove. The periderm grows from a phellogen, or, more probably, from a series of phellogens. Between the leaf and the stem apex the most striking feature of the crown isa series of ridges, which would be somewhat semicircular in longitudinal section, extending the whole breadth of the leaf, almost half of the entire circumference (fig. 366). The ridges are marked by pits and scars, showing the position of the inflorescences which have fallen off. The ridge is certainly not a leaf. It develops from the adaxial surface of the leaf groove, but only a too rigid mor- phology could make it a branch on that account. Whether there is any bract at the base of the inflorescence still remains to be deter- mined. Ridges develop on the stock, below the leaf groove, and are prob- ably just as numerous as those above, but not so conspicuous. PEAR- SON found that they occasionally bore inflorescences. However, they are small and comparatively inconspicuous. GNETALES—WELWITSCHIA 387 In old plants the crown sometimes becomes cracked radially, from the circumference almost to the center. Longitudinal cracks also Fic. 366.—W elwitschia mirabilis: A, young plants with entire leaves; B, older plant with leaf splitting at tip; C, top view, showing depression between leaves; one-half natural size.—After HOoKER.?? appear in the stock, reaching from the leaf groove almost to the root, but they are not so deep as those in the crown. 388 GYMNOSPERMS The leaf.—There are only two leaves, the original pair of leaves of the seedling persisting throughout the life of the plant, which is esti- mated at more than too years. The leaves are entire, until they reach a length of 15-20 cm., when they begin to split at the tip (figs. 365, 366, and 369). Still older leaves, lying on the ground and flap- ping in the wind, are split into ribbons. The leaves are thick and leathery and the venation is parallel; consequently, the splits are long and straight. The cells exposed by the splitting become suber- ized so quickly that the plant suffers no damage. The leaves grow from the base of a groove, which extends nearly half-way around the top of the crown, and functions as a moist chamber. The base of the leaf remains permanently meristematic, while the tip keeps wearing and dying off. In old plants the leaves are about 2 meters long. WELwitscu did not believe that the plant had any leaves at all, but only two cotyledons, which persisted throughout the life of the plant. Hooxker’s material was too old either to correct or confirm that view; but he adopted it tentatively, remarking that it would be easy to determine the facts as soon as anyone got a seedling. These have now been seen by many observers. There are two cotyledons and, at right angles to them, the two leaves (fig. 380). The root.—As already indicated, the root is very long, and must go down to the water table. People have tried to dig it up, but, after digging for several feet, have concluded, probably correctly, that it kept on going down until it reached the water. Dioon spinulosum, growing on a rock, with scarcely any soil, had roots hanging down the vertical face of the rock for 12 m., when—still about 5 cm. in di- ameter—they disappeared in a crack in the rock. The roots of cy- cads, in regions as dry as the Welwitschia country, go down to the water table and keep the plants green, while the grass is yellow, or even scorched. The roots of Welwitschia must go down as deep. Histology.—The most spectacular feature of the histology is the great display of spicular cells. They are usually branching, very thick walled, lignified outside and cellulose inside, and incrusted with crystals of calcium oxalate (fig. 367). These cells, branching and interlocking, are formed in the stem, root, and leaf, and they make free-hand sectioning impossible. One might as well try to cut GNETALES—WELWITSCHIA 389 sections of a thick Scotch plaid blanket as to try to cut a stem of Welwitschia without imbedding. The dry wood is light and not hard, but a treatment of 2 or 3 weeks in 20 per cent hydrofluoric acid should precede any attempt at imbed- ding. With such treatment Dr. La DemMa M. LANGDON cut very satisfactory thin paraffin sections, 3 cm. square, includ- ing the leaf groove, of a plant 15 cm. in diameter. Except in the leaf, the ar- rangement of the bundles is ir- regular. Occasionally, in younger stems, and especially in roots, there is a zonation, A doubtless representing growth- rings; but it is doubtful wheth- er such rings are annual. It is more likely that they were formed some season when there was a heavy rain. We have already noted that the normal rainfall in the Wel- Fic. 367.—Welwitschia mirabilis: spicular witschia country is about 1 cells from the perianth of the staminate inch: but there have been flower: A, branching cell, the branching A oe represented in only one plane; B, a nearly Oods. ' A straight cell; C, transverse section showing A median longitudinal sec- lumen nearly closed and surface incrusted by tion of a stem, with a crown small crystals of calcium oxalate; X 225.— : : From COULTER and CHAMBERLAIN, Mor- 15 cm. in diameter, shows a monies phology of Gymnosperms'4 (University of cup-shaped plate of vascular Chicago Press). tissue about as thick as the leaf groove, and extending from one leaf groove to the other. This is the principal vascular system from which bundles extend upward to the crown and inflorescences, and downward to the stock and root. Although the bundles are collateral, endarch, with a cambium, the xylem of any individual bundle has no recognizable growth-rings. 390 GYMNOSPERMS In such rings as have been mentioned, each ring has its xylem and phloem, as in polyxylic cycads. In transverse section, the bundles are long and narrow, with the xylem cells in rather straight rows for about twenty cells back from the tip, where they become very ir- regular in size and arrangement. It is in this irregular region that one might possibly find some record of seasonal changes. It is not easy to recognize growth-rings in woody monocots; in fact, it has been assumed that they do not exist; but they are easily seen in Aloe bainesii, not quite so easily in A. ferox and Yucca filamentosa. How- ever, the rings are there, and they are seasonal. In the branching inflorescences the bundles are more regular, the transverse section showing rows of xylem usually one cell wide, with thin-walled parenchyma between (fig. 368). The walls of the xylem elements are extremely thick, and, in a bundle like that shown in the figure, more than half of the cells have spiral or annular markings. The protoxylem tracheids of the lower part of the hypocotyl and root are reticulate, and in the leaf, spiral or annular; but all three kinds may be found together. In the secondary wood, the tracheids have bordered pits, often with reticulations between the pits. The cells of the phloem are long and narrow, with prominent nu- clei and sieve areas on the ends, and, also, to some extent, on the side walls. Long, very thick-walled fibers occur singly or in twos or threes in the pith; they are abundant just outside the phloem, and throughout the cortex they form long, straight bundles, showing ten to twenty or more fibers in transverse section, with a single section showing more than 200 such strands. Stomata are rare on the branches of the inflorescences. The outer walls of the epidermal cells are thick, with three easily recognizable layers; the outer, rather thin and, apparently, not cutinized; a mid- dle layer, somewhat cutinized and containing very small crystals of calcium oxalate; and an inner layer of cellulose. Dr. WELwiItscH, in a letter to Dr. HOOKER, said the plant exuded resin like a conifer. There is some exudation, not resin, but more or less mucilaginous, coming from the disorganization of spicular cells, or from lysigenous cavities. The tough, leathery leaves reach a thickness of 2 mm. Stomata are GNETALES—WELWITSCHIA 391 more numerous on the upper side, but there are many on the lower side; PEARSON*4® counted 108-125 to the square millimeter on the upper side, and 87-96 on the lower. »~e¢ Que {x \/ oa + Ss PSA) PRE erst 5 HEY “ec perk si) Cy ee wae Fic. 368.—Welwitschia mirabilis: transverse section of flower stalk, next to the last branch below the cones. Only one or two cells of each xylem row are pitted: the rest have annular, spiral or reticulate marking; X 370. In transverse section, the leaves show a single row of collateral endarch bundles, each surrounded by transfusion tissue, and with strands of thick-walled fibers between the transfusion tissue and the xylem. There is a strong palisade both above and below, with bun- dles of thick-walled fibers alternating with the palisade. Surrounding the bundles is a thin-walled parenchyma, and, everywhere, except in the bundles themselves, great numbers of spicular cells. A leaf could hardly be better adapted to desert conditions. 392 GYMNOSPERMS THE SPOROPHYTE—REPRODUCTIVE Both staminate and ovulate strobili are compound, and are borne in great numbers on branching shoots arising from the rim, just above the leaf groove, although an occasional inflorescence is found below the groove. Fic. 369.—Welwitschia mirabilis: part of the crown of an old plant, showing one of the much split leaves and about a dozen branched flower stalks with staminate strobili at the tips.—From a negative taken by Pearson, January, 1907, on the Nahib Plateau near Walvis Bay, South Africa. The scalpel in the middle of the picture indicates the size. THE STAMINATE STROBILUS The stalk, bearing the strobili, branches more or less dichoto- mously, two, three, or even four times (fig. 369). In January, 1907, when PEARSON made the exposure from which this figure was made, pollen was beginning to shed from the lower stamens of the most ad- vanced cones. The cones are beautifully geometrical in the arrangement of their parts. CHuRcH’s'* drawing of this cone does not exaggerate the GNETALES—WELWITSCHIA 393 regularity (fig. 370). In the lower part of the middle cone the flowers are pushing out beyond their subtending bracts and, especially in the middle row, the six stamens can be seen on the rim of the stami- nate tube, with the funnel-shaped end of the integument in the center of each group of stamens. ay, Fic. 370.—Welwitschia mirabilis: one complete strobilus and parts of two others: X5.2.—After CHURCH.!3? The arrangement of the flowers is strictly cyclic. Two opposite bracts alternate with two opposite bracts at right angles to them, throughout the entire cone, and in the axil of each bract, except a few at the base, is a single flower. More than any other feature, this flower has been called upon to link the Gnetales with the angio- sperms. Its floral diagram might well pass for that of an angiosperm flower (fig. 371). The decussate arrangement of bracts in the cone is 394 GYMNOSPERMS continued in the four parts of the perianth of the flower. There are six trilocular stamens, one opposite each of the smaller bracts of the perianth, and two opposite each of the larger ones. The stamens are Fic. 371.—Welwitschia mi- rabilis; staminate flower: A, longitudinal section of flower with nearly ripe pollen: p, bract of perianth; s, trilocular an- ther; ¢, protuberance on inner face of stamen tube; i, inner in- tegument; 7, nucellus of sterile ovule; b, bract. B, floral dia- gram: lettering as in A; ¢, axis of cone.—From CouLterR and CHAMBERLAIN, Morphology of Gymnos perms'* (University of Chicago Press). borne on a tube, and below each stamen is a protuberance, which can be a nectary when one wishes to emphasize the resem- blance to an angiosperm flower. The most remarkable thing about this flower is that it is bisporangiate. In the center is an ovule, with a well-developed nucellus, but never with any sporogenous tissue. There is an inner integument, which elongates considerably, although it never reaches the extreme length of the inner integument of the ovulate flower. However, the integument of this sterile ovule has an expanded funnel-shaped tip, which has caused some to write about stigmas. Outside the inner integument is the perianth, consisting of four bracts: two at the ends, small and sharply angled at the midrib, and two at the sides, par- allel with the large bract in the axil of which the flower stands. Thus, each flow- er is a strobilus in the axil of a bract, and the whole cone is a compound strobilus. The parts of the flower are in two’s or four’s. In those Cruciferae which have tetradynamous stamens, the positions of the two missing stamens are clearly marked by nectaries. In Welwitschia the two stamens at the ends, in the floral dia- gram, are generally larger than the other four, so that, if two stamens are really missing, they would have alternated with the two stamens opposite each of the larg- er bracts of the perianth. No traces of GNETALES—WELWITSCHIA 395 rudiments have been found, and since insect pollination has never been proved for any gymnosperm, nectaries would hardly be an- ticipated. The anthers are trilocular, resembling somewhat the sporangia of Psilotum. Floral development.—A longitudinal section of a strobilus, about r cm. in length, shows a very complete series of stages in floral de- velopment, from the beginning of the bract to nearly mature pollen. The first protuberance to appear below the apex of the strobilus is a bract. The conical structure which soon appears in its axil is the beginning of the flower, and its apex finally develops into the sterile ovule. The first floral organs to appear are the bracts of the perianth, soon followed by the staminal rings. PEARSON says the two lateral stamens appear first, followed immediately by the other four, which grow faster, so that they all look alike; CHURCH says they appear simultaneously. If the two outer ones appear before the other four, the appearance in succession would favor the view that there are two cycles of stamens and that two of the outer set have been lost. There are six vascular strands in the staminal tube, one going to each filament. The fact that the two strands going to the two lateral stamens are inserted lower down than the other four fa- vors PEARSON’S view. As soon as the rudiments of the stamens can be recognized, the integument appears as a ring about the base of the nucellus. There- fore, the development is acropetal throughout. Later, there is a great development of both staminal tube and the integument. Some of the stages of floral development are shown in fig. 372. In studying the floral development, one cone was sectioned in which the sterile ovule terminated the cone axis. The flower was large and the pollen mother-cells had rounded off, while the next flowers below were in an early sporogenous stage. During the elongation of the staminal tube, the reduction division takes place in the pollen mother-cells, the second division following the first so closely that the young microspores have a typical tetra- hedral arrangement. 396 GYMNOSPERMS THE OVULATE STROBILUS The ovulate strobili, like the staminate, are borne in considerable numbers on branching flower stalks (fig. 373). The young strobili are green, but, at maturity, the color changes to a brilliant red—as brilliant as that of the Christmas Poinsettia. Fic. 372.—Welwitschia mirabilis: floral development: A, longitudinal section of tip of staminate cone; B, bract with young flower in its axil; C-G, older stages; b, bract; p, perianth; 0, ovule; 7, inner integument; d, staminal disk; a, anther; s, “stigma”; m, nec- tary-like swelling of staminal disk; all 50. While the strobilus is compound, its component strobili are much simpler than those of the male, for there is no trace of a bisporangiate condition. A few of the lower bracts have no flowers in their axils, and the flowers in the axils of a few of the upper ones do not mature. Each GNETALES—WELWITSCHIA 397 strobilus produces from fifty to seventy flowers, many of which de- velop more or less; but the number of seeds which will germinate is A Fic. 373.—Welwitschia mirabilis: A, flower stalk with ovulate strobili of various ages, one-half natural size; B, tip of flower stalk with staminate strobili, natural size. This is not a satisfactory representation of the staminate strobili: fig. 370 is more ac- curate.—After HOOKER.?? comparatively small. We have planted about thirty seeds which seemed to have reached the maximum size and which appeared to be viable. These had not been taken from the plant more than 2 398 GYMNOSPERMS years earlier. Only three germinated, and only one lived more than a year. The one shown in fig. 381, was nearly 4 years old when the photograph was taken. PEARSON, in 1910, germinated seeds collect- ed in 1907, and estimated to be at least one year old at that time. In comparing the floral diagram with that of the staminate flower it is evident that it not only lacks any stamens, but there are only HK. Mena e Mase eae Fic. 374.—Welwitschia mirabilis: A, floral diagram of staminate flower: 2, a, axis of strobilus; g, nucellus of sterile ovule; f, integument; e, cycle of stamens; d, inner bract of perianth; c, outer bract of perianth; b, bract in the axil of which the flower stands. B, floral diagram of ovulate flower: 4, embryo sac; other lettering as in A.—After PEAR- SON. 449 two bracts in the perianth. The two broad bracts, which are so con- spicuous in the male, are lacking; but the two bracts at the ends, which are comparatively inconspicuous in the male, are developed into broad, thin wings (fig. 374). The floral development throughout is acropetal. The perianth first appears as a ring at the base of the nucellus, but soon grows GNETALES—WELWITSCHIA 399 more rapidly at the sides, producing the broad, thin wings. While the wings are extremely thin, an abundance of wavy fibers makes them rather tough. They doubtless function in seed dispersal. Fic. 375.—Welwitschia mirabilis: development of ovulate flower: A, longitudinal section of tip of ovulate cone: the lowest flower at the right shows the nucellus, with inner integument (7) and (p) outer integument (perianth); the flower next above shows the same structures in earlier condition; B, older ovulate flower, with the female game- tophyte in an earlier free nuclear stage; 6, bract; C, still later stage, with female game- tophyte (g) in late free nuclear stage; A, X20; B and C, X50. At first the perianth, which PEARSON prefers to regard as an outer integument, much exceeds the inner integument, but, as the time of pollination approaches, the inner integument grows rapidly until it protrudes beyond both the subtending bract and perianth (figs. 375 and 373). 400 GYMNOSPERMS There is a single archesporial cell which divides, forming a pri- mary wall cell and megaspore mother-cell. There is no spongy tis- sue, either at this stage or later. The reduction divisions were not observed, but PEARSON concluded, from early stages in the female gametophyte, that the lower cell of the row develops in the usual way. THE MALE GAMETOPHYTE The microspore is the first cell of the male gametophyte. PEARSON“ followed the devel- opment from the first appear- ance of the archesporial cells up to the shedding of the pollen. The two spore coats, exine and intine, are of about equal thick- ness and usually become more or less split apart (fig. 376). The divisions which take place Fic. 376.—Welwitschia mirabilis: A, inthe microspore before the stage three-nucleate stage of pollen grain, X 940; B, shedding stage, X 600: ex, exine; in, in- shown in fig. 376, have not been tine; n', n?, and 3 interpreted respectively described; but from PEARSON’S ca pihegeninl dR a and tube nuclei. figures and description it is evi- ae dent that the first division forms a prothallial nucleus (PEARSON says there is certainly no prothallial cell) and an antheridium initial, and the antheridium initial, at its first division, produces a tube cell and a generative cell. In this three-nucleate stage the pollen grain reaches the ovule. The pollination drop, characteristic of gymnosperms, appears at the tip of the projecting micropylar tube, which is the greatly prolonged inner integument. Hooker’” suspected that Welwitschia was insect pollinated; and Baines found a parasitic insect, Odontopus ex punctulatus, on plants at Haikamchab. Pearson not only found the insect on both male and female plants, but found pollen on its legs and abdomen. He concluded that it would be impossible for it to crawl over ovulate GNETALES—WELWITSCHIA 401 cones without effecting pollination. But whether the insect is the principal agent in pollination may still be doubtful. It is very prob- able that in most wind-pollinated plants, even in gymnosperms, pollen may occasionally reach the pollination drop on the stigma through the agency of insects. Reaching the pollination drop in the three-nucleate stage, the pollen is drawn down through the long micropyle to the nucellus of the ovule. The pollen tube may begin to develop, even in the micropyle. There is no pore, but the exine splits longitudinally throughout the entire length. In the young pollen tube the generative cell and the tube nucleus enlarge, and the prothallial nucleus can no longer be distinguished. There is no stalk cell. The nucleus of the generative cell soon di- vides, forming the two sperm nuclei, which lie in a common mass of protoplasm, with no wall between them. Except for the prothallial nucleus, this is a typical angiosperm history, and even the prothallial cell occurs sporadically in angio- sperms. THE FEMALE GAMETOPHYTE The megaspore is the first cell of the female gametophyte. The divisions of the megaspore mother-cell have not been worked out, and no chromosome counts have been made anywhere; but PEAR- son’s figures show that the archesporium was doubtless hypodermal and that the lowest megaspore of a row germinates. He figures the mature megaspore, the four-nucleate stage of the female gameto- phyte, and later free nucleate stages. The nuclear divisions are doubtless simultaneous, for simultaneous divisions were observed at two stages. A striking feature of the development of this gameto- phyte is that no large central vacuole is formed, the nuclei being equally distributed throughout, up to what would be the 1,024 nu- clear stage, if all divisions were simultaneous and all nuclei divided (fig. 377). Actual calculation placed the number between 1,015 and 1,360. At the close of the free nuclear period, walls come in irregularly, often inclosing a dozen nuclei in one cell. In this condition, there are occasional nuclear divisions, but nuclear fusions are the rule 402 GYMNOSPERMS and they continue, sometimes giving the impression of amitotic divisions, until nearly all of the cells become uninucleate (fig. 377). VIG. 377.—Welwitschia mirabilis: A, late free-nuclear stage of female gametophyte: n, cell of nucellus; m, megaspore membrane; B, early cellular stage, with most of the cells still multinuclear; at f, nuclear fusions have taken place until the cells are uninu- clear; f, at bottom, shows that cells may divide after the fusions; a, in the middle cell at the left, shows that nuclei may divide before fusion; C, four of the cells, f, have become uninuclear and nuclear fusions are advancing in the rest, m; A, X 305%; B and C, X 350.44—After PEARSON. The cells of the lower part of the gametophyte contain more nuclei than those in the micropylar end, the former often containing a dozen, while the latter contain two or three, or sometimes half a dozen. Fusions are completed earlier in the micropylar region, but GNETALES—WELWITSCHIA 403 finally nearly all of the cells become uninucleate. In their further development, the two regions continue to be different. A remarkable feature in the development of this gametophyte is that there are no archegonia. Cells which might be regarded as archegonial initials, elongate, break through the megaspore membrane, and grow up into the nucellus, looking like pollen tubes, and growing up into the nucellus just as pollen tubes grow down. They are called “‘pro- thallial tubes.”’ They grow about half-way through the nucellus, where they meet the pollen tubes coming down. Nuclear details have not been worked out, but if the nucleus in a prothallial tube has been formed by the fusion of several nuclei, fertilization and chromosome numbers would pre- sent an interesting problem. After the cells in the lower part of the gametophyte have become uninucleate through the exten- sive nuclear fusions, mitosis, with the formation of cell walls, is re- sumed. The chromosome num- Fic. 378.—Welwitschia mirabilis: wpper part of nucellus with prothallial tubes going up to meet the pollen tubes which will be coming down: #, prothallial tubes; s, sterile cells of upper region of female gametophyte; , nucellus; m, megaspore bers, at first Very large, become membrane; * 140.—After PEARSON.439 smaller as divisions continue, and the cells also become smaller. This behavior stops before ferti- lization, but is resumed as soon as fertilization has taken place, a feature quite characteristic of angiosperms. While the gametophyte is characterized by the « number of chromosomes, and the sporophyte is characterized by 2x chromo- somes, we should still regard the prothallium of Welwitschia as gametophyte in which the normal x number of chromosomes has 404 GYMNOSPERMS been modified by extensive nuclear fusions. We see no reason for adopting the term “trophophyte,” which has been proposed. Fertilization.—The pollen tubes growing down into the nucellus and the prothallial tubes growing up into it meet, and fertilization takes place (fig. 378). A pollen tube and a prothallial tube come into contact, the tube walls at the point of contact become dissolved, and FG. 379.—Welwitschia mirabilis: early embryogeny: A, first division of zygote; ps, primary suspensor cell and, below it, the cell from which the embryo is to develop; B, later stage, with inner cortical ring (‘cr) and terminal initial cells (ic); C, still later stage; ps, primary suspensor; icr, cells of inner cortical ring; ocr, cells of outer cortical ring; the embryonic plate (e), ring (x), and cap (c), are each represented in section by two cells; A, X700; B and C, X305.—After PEARSON. #° the nucleus of the prothallial tube passes into the pollen tube, where it fuses with one of the sperm nuclei. PEARSON found no independ- ence of the chromatin of the gametes, as it had already been de- scribed for Pinus and Juniperus, but found a complete fusion, with a resting nucleus, before any division of the zygote. Nearly all of the cytoplasm is contributed by the pollen tube. The zygote nucleus passes from the pollen tube into the prothallial tube on its way down to the endosperm. Embryo.—The zygote elongates and its nucleus divides. As in Ephedra, the wall begins to form at the periphery and closes in to- GNETALES—WELWITSCHIA 405 ward the center. The wall is complete before another mitosis takes place, so there is no free nuclear stage (fig. 379). Fic. 380,—Welwitschia mirabilis: seedling in greenhouse at the University of Chi- cago The upper cell elongates and becomes the primary suspensor: the embryo is derived from the lower cell. The primary suspensor cell elongates but does not divide. The inner cortical ring cells elongate immensely (compare icr in B and C of fig. 379), sur- rounding the primary suspensor cell. The outer cortical ring cells 4006 GYMNOSPERMS then behave like the inner cortical ring cells, elongating and sur- rounding the inner cortical ring cells, so that a transverse section shows the primary suspensor in the middle, surrounded by two cortical rings, the inner consisting of eight cells and the outer of sixteen. Later, a third ring of cortical cells (xin fig. 379 C) is cut off and is added to the suspensor. And finally, even a fourth ring is cut off and is added to the suspensor. Fic. 381.—Welwitschia mirabilis: same seedling as that in fig. 380, two years later The cells (C) undergo division, forming a cap, which is cast off later, as in Cephalotaxus. The stem, leaves, cotyledons and root are formed from the four cells of the plate (e, in C of fig. 379). As the embryo grows down into the endosperm, the suspensor becomes coiled and twisted in typical gymnosperm fashion and, after the cap has been cast off, the outer cells become dermatogen, and the meristems of the root and shoot are organized. The seedling.—The seedling has two cotyledons and two leaves (fig. 380). The cotyledons persist for two or three years and then fall off: the two leaves, at right angles to the cotyledons, persist GNETALES—WELWITSCHIA 407 throughout the life of the plant. The cotyledons and leaves are, at first, erect; but, before the end of the first year, the cotyledons begin to droop and the leaves become horizontal (fig. 381). In the axil of each cotyledon a bud appears and becomes flatter and broader until the two buds meet, forming a continuous plate, under which the stem apex, now arrested, is buried. Meanwhile, the tissues above and below the base of the leaves grow so that the leaf bases become buried in a deepening groove. At the lower part of the hypocotyl, a lateral outgrowth, called the feeder, acts as a haustorium until the food supply of the seed is ex- hausted. The times for various stages are given by PEARSON as follows: most of the stages in the development of both male and female gametophytes can be secured in January. In 1904, at Haikamchab, there was little fertilization before the end of the second week in January. At Welwitsch, in 1907, the last week in January, fertiliza- tion and proembryos were abundant; and in March, 1909, embryos were more advanced. At Haikamchab on May to, 1861, BAINES collected ripe seeds. So the intraseminal growth of the embryo is less than 4 months. CHAPTER XIX CONIFEROPHYTES—GNETALES—GNETUM The final genus in the gymnosperm phylum is Gnetwm—in general appearance almost as much out of place in this assemblage as Wel- witschia, for most of its more than thirty species are lianas. Only a few are trees or shrubs. In striking contrast with Ephedra and Welwitschia, both extreme xerophytes confined to desert habitats, Gnetwm belongs to the luxu- riant tropics, and most of its species, when not in fruit, might be mis- taken for dicotyls. Tropical Asia and the islands between Asia and Australia have nineteen of the thirty-four species given by PEarson.“° The two African species are north of the Welwitschia country, in Cameroons and Angola. Of the twelve American species, three are in the Guin- eas, along the northern coast of South America, seven belong to the Amazon country of Brazil, one to Ecuador, and one to the West Indies. There are none in North America, Australia, or Europe. No species have yet been found common to both Western and Eastern hemispheres, but some in India and the islands between India and Australia are found on both sides of the Equator. THE SPOROPHYTE—VEGETATIVE The lianas, which are characteristic of the genus, twine, climb, or trail over other vegetation, reaching the tops of the tallest trees, with rarely any leaves for the first 40 or 50 feet. Gnetum gnemon, in the Moluccas, and widely cultivated, is a tree, and since it is the best-known species, one is likely to picture the tree habit when one thinks of Gnetum. A species in New Guinea, G. costatum, also has the tree habit; and there are also a few shrubs. The leaf.—The large, oval, entire leaves, with netted veins, make the plants look like dicotyls. There are branches of limited and un- limited growth, but the difference between the two kinds is not so 408 GNETALES—GNETUM 409 extreme as in Pinus or Ginkgo, and in Gnetum gnemon the two kinds look much alike. In climbing species, leaves are borne only on shoots of limited growth. On the long shoots the leaves are usually reduced to scales. All leaves are opposite and decus- sate. There is a short petiole, with no stipules, and the leaves vary so much in size and shape that such charac- ters are of little value in taxon- omy; but they are so typical- ly dicotyledonous in appear- ance that a botanist, not fa- miliar with Gnetum, would guess the plant to be a dicotyl (fig. 382). The plant shown in the figure was raised from a seed at the University of Chicago. When it was about 20 years old and 4 inches in diameter, it was moved from the old greenhouse to the new, where it survived only a few months. A cutting is still alive and is about 2 inches in diam- eter. Many cuttings were made, but only two were suc- cessful. However, as in many conifers, there is occasional vegetative propagation. One season, several leaves from the plant already mentioned produced buds at the margins, Fic. 382.—Gnetum gnemon: young plant raised from seed at the University of Chi- cago. The plant is about two years old.— From Coulter and CHAMBERLAIN, Mor phol- ogy of Gymnosperms'4 (University of Chicago Press). as in Bryophyllum. Some of the buds had a couple of leaves, but none developed beyond this stage. A similar case was noted at the botan- ical garden at Utrecht. 410 GYMNOSPERMS The histological structures are so advanced that they are respon- sible for an undue amount of theorizing. The stem.—A transverse section of a young stem of Gnetum gnemon shows an endarch siphonostele, with strong bundles, large rays, and a cortex, partly of thin-walled parenchyma and partly of suberized fibers, and with a zone of spicular cells just outside the phloem (fig. 383). The protoxylem is very scanty, and the markings are spiral. Vessels in the secondary wood are of all sizes, from that of ordinary tracheids up to those with four or five times that diameter. The pit- ting is multiseriate and extends all the way around and on the end walls, which soon break down, so that there are continuous tubes, as in angiosperms. The end walls of the cells, which are to form vessels, may be quite oblique or perfectly transverse. There is a single perforation, formed by the enlargement or fusion of bordered pits and the disappearance of the middle lamella—a more advanced type than that of Ephedra, where the end walls have several perforations, and more like that of angiosperms, which have a single perforation. It should be noted that the mode of formation of the more or less freely open tube of Gnetum is not like that in the angiosperms, where the single large perforation of the lower dicotyls has been developed from a scalari- form type. THompson®" regards this difference in the development of the vessels as so fundamental that vessels in the secondary wood should not be regarded as any evidence in favor of the theory that angiosperms have come from Gnetales, or that the two have had a common ancestor. In many angiosperms the end walls break down while still very thin, and without any connection with pitting, and even in some (in which secondary thickening has begun), the breaking-down does not seem to depend upon pits. The phloem consists of very uniform cells. There are no compan- ion cells in Gnetum gnemon, although such cells occur in G. latifolium and G. scandens. However, these companion cells are not cut off from a mother-cell, as in angiosperms, but arise independently from the cambium. Here, again, the angiosperm feature, the companion cell, should not be urged as evidence in favor of relationship, be- y Wi WN 7 . : 0) . y) — 2. n iH 2) 4 @ oe a ex e © (5) ea "J TE aseeen Onn @=2 oe a) as | L@ ss = oe U wy Se eee cs a dq I ‘e' Ve Co UW: Gy \ (JO) ea S. DoW Ae OR Tera gS A 20" ie yee Caper ZG i | Th v Nl Se se aw®e oe ae o S25 Coys} yA Sees OOS J OS i, TEC i Y @ L 3} aes <~ OO “Oy UN of L e' ty O\ S20 AR JOINT J = @ Gz en Clos @: : [S) otigees COE wwo% CA 9 &S & Go299I @Y WO IG) (a) ANG, Ab! = q OB: BU A 4 ow, & la ie, ee, i.) < Fic. 383.—Gnetum gnemon: transverse section of bundle of young stem; X17 5 412 GYMNOSPERMS cause the mode of origin is different. Physiologically, the com- panion cells are the same in both groups, but there is no phylogenetic connection. The root.—The root of Gnetum gnemon is diarch, with an extremely scanty amount of primary xylem, only a single row of cells with occasionally a couple of cells in the protoxylem lying side by side (fig. 384). The vessels are of various sizes and are much larger than those of the stem, with several times as great a diameter, as may be seen by comparing figs. 383 and 385, which are both from Gnetum gnemon, and drawn to the same scale. Vessels of lianas are likely to be larger. The phloem, as in the stem, consists of very uniform cells and, just beyond it, are numerous fibers (fig. 385). The rays are broad and thin walled, and are packed with starch, in striking con- trast with those of the stem, which are thick walled and pitted. The difference in structure accounts for the fact that the wood of the stem is very hard, while that of the root is very soft. The tracheids have uniseriate bordered pits, with conspicuous bars of SANTO (fig. 386). The pits of the vessels are smaller and multi- seriate, with few bars of SANTO, or even none at all. The breaking of the end walls of the large cells to form the con- tinuous lumen of the vessels is easily seen in the root (fig. 387). At first, the border of the large perforation may be ragged, but it soon becomes smooth, and sometimes the passage may be even more complete than that shown in the figure. On the whole, Gnetum, in its outward appearance and in its inter- nal structure, has almost reached the angiosperm level; and, since it has no record as a fossil, the temptation to connect it in some way with the angiosperms has often been irresistible. THE SPOROPHYTE—REPRODUCTIVE Like the other two genera, Gnetum is dioecious; but also, like them, it shows very clearly that a condition of complete dioecism has not yet been attained. The staminate strobilus.—The staminate strobilus is a long slender axis, bearing numerous decussate pairs of bracts, which are con- crescent throughout practically their whole length, making them ett SINS, TEA 6) ~ He KW/Z WS é (4 aC ON SelKs ecus SS © SSo2® c O77 Fic. 384.—Gnetum gnemon: transverse section of central part of a root 6 mm. in diameter, showing a single row of primary xylem cells, with protoxylem two cells wide at the extreme tips; 175. Fic. 385.—Gnetum gnemon: transverse section of peripheral part of bundle of root; X175. 414 GYMNOSPERMS look like cups. In their axils, all the way around the axis, the stami- nate flowers are borne. As in Welwitschia, ovulate flowers are associated with the stami- nate, and the ovules develop much farther, producing megaspores, which sometimes germinate. Such gametophytes disorganize in early free nuclear stages; occasionally they develop farther and even pro- duce good seeds. In Welwitsch- ia, the sterile ovules, although reaching the size at which the ar- chesporium appears in functional ovules, show no trace of the begin- ning of sporogenous tissue. The Fic. 387.—Gnetum gnemon: longitudi- Fic. 386.—Gnetum gnemon: longitudi- nal section of two cells of root, showing nal section of tracheids and parenchyma perforation which makes the lumen con- of root, showing bars of SANIO; X 375. tinuous; X 375. staminate flower has a sterile ovule, and thus is bisporangiate. In Gnetum, the association is not so close, for there are no bisporangiate flowers—only ovules and stamens in the axil of the same bract (figs. 388). If the sterile ovule is only slightly developed, it does not appear to inhibit the development of the microsporangia near it; but when the ovule reaches such an advanced stage as that shown in fig. 388 C, the microsporangia near it abort, and only those lower down produce good pollen. In very rare cases, stamens have been found in ovulate strobili. GNETALES—GNETUM 415 In the two African species, Gnetum africanum and G. buchhol- zianum, the staminate strobilus has no trace of ovulate flowers. The staminate flower consists of a stalk with two anthers at the top and a sheathing perianth at the base. The anthers are uni- Fic. 388.—Gnetum gnemon: longitudinal sections with flowers in axils of bracts: A, ovule in ovulate strobilus, showing two integuments and perianth; B, staminate stro- bilus, showing sterile ovule with inner integument and perianth; oldest sporangia at the top; C, staminate strobilus with functional sporangia at the bottom; sporangia near the large ovule, in which the gametophyte has reached the free nuclear stage, are abortive; X23.—From Coulter and CHAMBERLAIN, Morphology of Gymnosperms'4 (University of Chicago Press). locular. In the development of the microsporangium, the tapetum is formed from the sporogenous tissue and not from the wall cells. There is no endothecium. As the microsporangium matures, all cells between the spores and the epidermis break down, and the 416 GYMNOSPERMS epidermis, becoming hard, is the outer protecting layer. This con- dition is universal in Pteridophytes and in the Carboniferous gym- nosperms. The ovulate strobilus.—The ovulate strobilus has received more attention than the staminate, doubtless because it can be collected during several months and is quite conspicuous for some time before the seeds ripen and fall off (fig. 389). The figure shows some ovulate strobili of the plant of Gnetum gnemon (fig. 382) grown at the Uni- versity of Chicago. At this time the plant was about 20 years old and this was the first and only time it ever produced strobili. The characteristic leaves, bearing such a close resemblance to dicotyls, are also well shown in this figure. The axis of the ovulate strobilus in Gnetum gnemon has a pair of opposite sheathing bracts at the base, followed by five or six whorls of ovules, with frve to seven ovules in a whorl (fig. 390). Sometimes there is a terminal ovule. Although there are so many ovules, most of which reach the pollination stage, and even the fertilization stage, only a few, perhaps from two to five, complete the entire develop- ment and become capable of germination. In the development of the strobilus the pairs of concrescent bracts, usually about seven, forming a series of cups, are first to appear, followed by the nucelli of the ovules in a ring in the axils of the bracts (fig. 391). Ovules in ovulate strobili have two integuments and a perianth, while those in the staminate strobili lack the outer integument. In the latter case, a rudiment of the outer integument appears, but soon aborts. In normal ovules the order of appearance is perianth, outer integument, inner integument. As in the other two genera, the inner integument becomes prolonged into the characteristic micro- pylar tube. In the mature seed it becomes reduced to a thin, papery layer. The outer integument finally becomes differentiated into an outer fleshy layer and an inner stony layer, so that the seed has three seed coats, as in the cycads, a stony layer in the middle, with a fleshy layer outside, and a thin, dry, membranous layer inside. These three layers are very common in gymnosperms, and have led some into curious interpretations. The three layers appear in typical form in the cycads, where it is said that the integument is differenti- Fic. 389.—Gnetum gnemon: ovulate strobili on the plant shown in fig. 382. The plant, at this time, was about 15 feet high and had been topped twice to keep its height within the capacity of the old greenhouse——From a negative by Dr. Cuinc YUEH CHANG. 418 GYMNOSPERMS ated into a stony layer in the middle, with a fleshy layer on each side. Such a statement involves the assumption that the nucellus and in- tegument are united, except at the “‘free’’ part of the nucellus, an assumption originating from the fact that some very ancient seeds have the integument and nucellus free from each other throughout. In Gnetum, the outer integument becomes differentiated into an Fic. 390.—Gnetum gnemon: Typical ovulate strobili collected in the Philippines. Three young strobili, like the one from which fig. 388 was made, are shown at the left. The others show ovules in various stages of development; about natural size—From a negative by Dr. W. J. G. LANp. inner stony layer and an outer fleshy layer. In an attempt to unify, some have regarded the outer integument as having been differenti- ated from the inner during phylogeny. They were probably never more united during phylogeny than they are now in ontogeny. In Lygino pleris, the outer layer of the integument becomes stony, while the cupule serves as an outer fleshy layer. Physiologically, the three layers are efficient during the development of the seed and, at ma- turity, they are protective. Their function is as uniform as that of tendrils, and their homology may be as various. The archesporium is hypodermal, dividing into a primary sporog- Paar crt — Fic. 391.—Gnetum gnemon: A, young ovulate strobilus; B, older strobilus with one mature seed; C and D, sections of young strobili; #, hairs; E, young ovule; 1, nucellus; 7, inner integument; 7, rudimentary outer integument; #, perianth; F, nucellus with three megaspore mother-cells; G, advanced ovulate flower; 7, inner integument; 0, outer in- tegument; ~, perianth; H, free nuclear stage of female gametophyte; J, female gameto- phyte with free nuclei above and cellular tissue below; a second gametophyte is shown at the upper right; J, fertilization; t, pollen tube; f, two eggs; and there are several free nuclei of the gametophyte; K, late embryo sac; g, zygote; C, X24; D, X45; G, X33; I, X50; J, X340; K, X50.—After Lorsy,37 except F, which is after STRASBURGER.™3 420 GYMNOSPERMS enous cell, which is also the megaspore mother-cell, and a primary wall cell, which builds up an extensive tissue above the mother-cell. Often, there may be more than a single sporogenous cell.37 The megaspore mother-cell produces four megaspores, of which more than one may germinate.’7" CouLTER found that twelve and twenty-four are the x and 2% num- bers of chromosomes in Gnetum gnemon.*4 THE MALE GAMETOPHYTE The microspore is the first cell of the male gametophyte. The microspore of Gnetum gnemon, up to the shedding stage, is shown in A, B, and C of fig. 392. In this series, by THompson,®° there does not seem to be any prothallial cell, the first division giving rise to a tube nucleus and a generative nucleus, the latter soon dividing to form a stalk nucleus and a body nucleus. No cell walls are either Fic. 392.—Gnelum: male gameto- phyte: A, young microspore: B, a little é : older; ¢, tube nucleus; g, generative cell: figured or described; but in Gnetum C,t, tube nucleus; b, body cell; st, stalk sp.,in Dand E of the same figure, nucleus; D, end of pollen tube; st#, stalk . . nucleus; b, body cell; ¢, tube nucleus: £, ' body cell is organized. end of pollen tube, later stage; ¢, tube In the complete elimination nucleus; m, sperms. A-C, Gnetum gne- of even a single prothallial cell, mon, X1,200; D, E, G, sp.; D, X550; Gnetum has reached the angio- E, X800.—After THompson.®3° mite sperm condition, where there are no prothallial cells, except in extremely rare cases, where they are regarded as reversions. At the first division of the microspore of G. africanum, PEARSON found a very delicate wall, which soon disappeared, and no wall was formed at the second division. It is easy to follow the growth of the pollen tube, because the pollen grains frequently germinate while still in the micropylar tube. The exine is cast off, and the intine grows into a tube, begin- ne ee GNETALES—GNETUM 421 ning near the tube nucleus. By this time, a definite body cell has been organized about the body nucleus, and, with the tube nucleus, it passes into the pollen tube, while the stalk nucleus remains be- hind. PEARSON does not regard this nucleus as a stalk nucleus, but as prothallial. Some of the pollen grains reach the nucellus directly and germinate there. Germination at a distance from the nucellus adds another angiosperm feature to the growing list. The nucleus of the body cell divides as the pollen tube grows down into the nucellus, but no wall is formed between the two nu- clei. They may differ somewhat in size, and THompson® thinks it probable that only the larger one functions, and that the other dis- organizes. THE FEMALE GAMETOPHYTE The megaspore is the first cell of the female gametophyte. As in all other gymnosperms, its germination begins with a series of free nuclear divisions, but, unlike all other gymnosperms, the gameto- phyte—at least the micropylar end of it—remains in the free nu- clear stage, up to the time of fertilization (fig. 393). There has been some difference of opinion regarding the lower part of the gameto- phyte. Lotsy3” described a tissue which he interpreted as homolo- gous with the antipodal cells of angiosperms, only more extensive. CouLTER™ thought that Lorsy37" had mistaken the extensive nutritive tissue at the base of the gametophyte, undoubtedly a sporophytic tissue, for a part of the gametophyte itself. A study of the figures and descriptions of both authors, and also the later work of THompson,®° together with the re-examination of CoULTER’s material, make it certain that there is a considerable amount of cellular tissue at the base of the gametophyte and definitely within the megaspore membrane. THOMPSON shows very clearly the origin of this tissue, which develops from the free nuclear condition, just as in Welwitschia (fig. 394). The first walls are formed irregularly, without any connection with the mitotic figures, and inclose several nuclei in each cell. The nuclei, which are rather small, then begin to fuse, and the cells become uninucleate, with one large nucleus in each cell. Consequently, the tissue corresponds to the antipodal cells of angiosperms, which, in many cases, are quite numerous before 422 GYMNOSPERMS fertilization. This tissue is shown in Lotsy’s figure (391) and in THOMPSON’s figure (395C). CouLTER’s figure (393) shows too young a stage to have the antipodal tissue organized. A later stage which, unfortunately, his material did not contain, would have shown antipodal tissue and the pavement tissue. \y Slo UPS 7 ara Dery Saesese Sail) . Fic. 393.—Gnetum gnemon: A, ovule showing position of female gametophyte and pavement tissue; B, detail of female gametophyte in free nuclear stage, with the nu- tritive pavement tissue beneath; A, 54; B, X 300.—After CouLTer.™® Qualitatively, this gametophyte has nearly reached the angio- sperm level. Although the free nuclear stage is much more extensive than in any angiosperm, it is still in this stage at the time of fertiliza- tion. There is no archegonium, or even the archegonium initial, which is still retained in Welwitschia. Gnetum has reached the final stage in the reduction of the archegonium. Phylogenetically, neck cells and ventral canal cells are gametes and, in the bryophytes, may possibly function as such. In the heterosporous pteridophytes the GNETALES—GNETUM 423 reduction has proceeded until there is only one neck canal cell. There is no neck canal cell in known gymnosperms, and many gym- nosperms have lost the wall between the ventral canal nucleus and that of the egg. In Torreya it is very doubtful whether there is even a mitosis: the central cell becomes the egg. Welwitschia, going a step farther, has lost the mitosis which gives rise to the neck cell and central cell, so that the archegonium initially functions as an egg. In Gnetum there is not even an archegonium initial: a free nucleus, or- ganizing some cytoplasm about itself, functions directly as an egg. Without such a series in mind, some have tried to homologize the angiosperm synergids with various lost structures of the archegonium. There is no relation. The synergids are new ad- vanced structures, having no connection with gymnosperm archegonia. After fertilization the rest of the gametophyte becomes cellular. As yet, there is no evidence that any of the tissue has come from a fusion with a male nucleus, or that a fertilized egg has developed ‘‘endosperm”’ instead of an embryo. Fic. 394.—Gnetum ‘ gnemon: development It used to be said that the endosperm of gym- of tissue at the base of nosperms is formed before fertilization, while the female gameto- phyte; X300.—After that of angiosperms was formed after fertiliza- THompson.®3° tion. In Gnetum most of this tissue is formed after fertilization, and, in this respect, follows the angiosperm meth- od, although no “‘double fertilization,” like that described by HERz- FELD for Ephedra, has been reported. FERTILIZATION AND EMBRYOGENY The most interesting feature about fertilization in Gnetum is that it occurs while the female gametophyte is still in the free nuclear stage. One or more eggs are organized from the free nuclei, looking like the eggs of angiosperms. The pollen tubes, some of which begin their development in the micropylar tube, at some distance from the nucellus, grow through the nucellus and reach the egg. The two 424 GYMNOSPERMS sperm nuclei are formed from the body cell, while the tube is still only about half-way through the nucellus. According to THomp- son,®° the two sperm nuclei differ somewhat in size, and the smaller one may disorganize before the sperms are discharged. Lotsy3” Fic. 395.—Gnetum: A-E, G. gnemon; F, G, G. moluccense: A, upper part of female gametophyte in free nuclear condition, with two eggs (e), and pollen tube with tube nu- cleus (¢) and two sperm nuclei (m); B, fertilized egg (fe) surrounded by free nuclei; C, topographic sketch showing position of B; fe, fertilized egg; p, mass of protoplasm sur- rounding the egg; end, cellular part of the female gametophyte; D, two cells (proembryo, pe) resulting from first division of fertilized egg; Z, suspensor developing from proem- bryo cell; F, end of suspensor with four nuclei which will give rise to the embryo, emb; G, young embryo. A, B, X400; C, X100; D, E, X300; F, X200; G, X50.— After THompson.®° GNETALES—GNETUM 425 described sperm nuclei of equal size, and believed that both were functional, since fertilized eggs were usually found in pairs. Lotsy believed that the number of pollen tubes which had entered an egg could be determined by counting the number of pairs of zygotes. Tuompson’s figure (395 D) is described as a two-celled proembryo. It looks much like a pair of zygotes, as described by Lotsy. If it is a two-celled embryo, as seems likely, there is no free nuclear stage in the embryogeny, and Gnetum, like Sequoia, has reached the angio- sperm level in another feature of its development. Whether the figure represents a two-celled proembryo or a pair of zygotes, the next stage is the development of a tube, the suspen- sor, from each of the two cells (fig. 395 £). The nucleus divides, and, of the resulting nuclei, the one in advance undergoes free nuclear division, producing four nuclei, from which the embryo is organized (fig. 395 F,G). The other nucleus does not divide. It is very desirable that fertilization and embryogeny be worked out in cytological detail, not only in Gnetum, but in the other two genera. Investigators familiar with other gymnosperms may have wondered why Lotsy, PEARSON, and THompson, with material at hand, left some features undecided; but the reason is evident when one attempts to make preparations, especially of the male gameto- phytes of Welwitschia and Gnetum. However, the problems can be solved if fresh material is fixed in very small pieces, for these two genera are no more difficult than Ephedra, in which LAND gave a very complete account of spermatogenesis. The intermediate and later stages in the embryogeny are still to be described. CAYTONIALES At this time no account of the gymnosperms would be quite com- plete without some mention of the Caytoniales inasmuch as they possess some gymnosperm features and are so regarded by some botanists. THomas’®3 account, which refers to the order as angio- sperms, was based upon megasporophylls, fruits, and seeds of Gris- thor pia nathorsti and Caytonia sewardi; microsporophylls or stamens, Antholithus arberi; and fragments of associated leaves, Sagenopteris phillipsi. These investigations revealed a closed carpel with stigma 426 GYMNOSPERMS and anthers with four rows, both characters certainly angiospermic. Later, Harris,” describing material collected from Greenland, came to the conclusion that the carpels were open at the time of pollina- tion and that the pollen grains were drawn in through the micropyle to the nucellus; these are just as certainly gymnospermic characters. For the present, we are not inclined to ascribe the order to either group, but await the finding of additional material, with both re- productive and vegetative structures preserved, which will make it possible to establish the position of the Caytoniales. CHAPTER: XX PHYLOGENY Throughout the preceding chapters, remarks have been made and opinions have been expressed in regard to the relationships of various groups; but it seems best to assemble some of the scattered views, add some arguments, and to consider the evidence upon which botanists founded their theories of phylogeny.* Similarity of structures has been relied upon in all attempts to arrange plants in a natural, or genetic, sequence. Morphologists of the older school believe that most of this evidence, especially that afforded by embryology, is reliable; modern morphologists believe that some of this evidence, but not so much of it, is due to genetic relationship. There are ecologists and physiologists who would go still farther and claim that external form and internal structure are due, almost entirely, to environment and function. Heredity and environment determine the form, structure, and life-history of a plant. Morphologists have overemphasized the influence of heredity, while ecologists and physiologists have over- emphasized the influence of environment. It is becoming more and more evident that some of the things which have been attributed to heredity may be due to environment; but, as an acute observer remarked some 2,000 years ago, you do not gather figs from thistles; and very probably, if there had been con- stant experimentation from that time up to the present, not a single member of the fig family could have been transferred to the sun- flower family. However, it must be recognized that there is such a thing as ecological anatomy, and also that phylogenetic anatomy is equally important. When pumpkin seeds and corn are planted in the same hill, the phylogenetic factor determines that one shall pro- duce pumpkins, and the other, corn. Physiological and ecological * Some of the views, expressed in this chapter, even with their phraseology, were pre- sented in a symposium before the International Congress of Plant Sciences, at Ithaca, New York, August 19, 1926.4 427 428 GYMNOSPERMS factors may have a great influence upon the size and quality of the products, but they are still pumpkins and corn. The apple, pear, and quince resemble each other rather closely, and we believe that their structural similarities are due to genetic relationship; in the same way, the cherry, peach, and plum have a genetic relationship which is indicated by similar structures. In each group the three members have come from a common ancestor in the remote past; and, farther back, these two common ancestors may have come from some still more remote ancestor. The morphologist stresses mutation, smaller variations, hybridiz- ing, and natural selection in accounting for changes, and he is too likely to overlook the influence of soil, climate, and other conditions. The aquatic form of Polygonum amphibium differs so much from the land form that they were described as two species. How far is the capacity for such immediate responses inheritable? Would Proser pinaca palustris, grown in a mesophytic habitat for a thousand generations, lose the capacity for producing dissected leaves when grown in water? Most botanists believe that the land flora came from an aquatic ancestry. When algae transmigrated to the land it became necessary to develop protective, conducting, and supporting structures. The theories of CHuRcH,™3 as expressed in his Thallassiophyta, the views of Scott, as set forth in his Extinct Plantss34 and in the third edition of his Studies in Fossil Botany,>3 and also the views of SEWARD, as presented in his Fossil Plants‘ and in his Plant Life through the Ages,” are suggestive. Let us suppose that some algae became stranded. Those which did not die developed conducting cells, at first merely elongated, then with thickenings and with groupings into bundles. Later, more algae transmigrated, and, having the same conditions to cope with, developed elongated cells with thickenings and groupings like the previous transmigrants. The modern botanist, finding the two descendents resembling each other, assumes that they are related. Paleobotanists are often entirely dependent upon such evidence, because the wood may be the only part preserved; but morpholo- gists, especially anatomists, lay no less stress upon the evidence of anatomy. PHYLOGENY 429 No botanist will deny that heterospory has arisen independently in various groups of plants, and that it has arisen in the same way in the most diverse cases—by the disorganization of young megaspores which have been absorbed by growing megaspores. We have already remarked that the Cycadofilicales were mis- taken for ferns until their seeds were discovered, and it is possible that the discovery of seeds may transfer more of the oldest recog- nized members of the Filicales to the Cycadofilicales. In the angiosperm flower there seem to be some rather definite lines of evolution—from spiral to cyclic arrangement of parts, and, in the cyclic forms, from pentacyclic to tetracyclic, from hypogyny to epigyny, etc. On the basis of these tendencies, the Archichlamy- deae have been arranged in a line from the Amentiferae to the Um- belliferae; the Sympetalae, in a line from Ericaceae to Compositae; and the monocotyls, in a line from the Pandanales to the Orchids. This does not mean that each order must be derived from the one below it; but it is probable that some tetracyclic flowers have come, by direct descent, from pentacyclic flowers. However, the pos- sibility that similar structures may be due to parallel development must always be recognized. It may be that most of the forms in the three great lines of angiosperms have come, by parallel develop- ment, from a few great centers like the Ranunculus—Rose—Legume plexus in the Archichlamydeae. Similarly, there are lines of evolution in the gymnosperms, which does not mean that each family is derived from the one ranked be- fore it, or even that the genera within a family are mentioned in genetic sequence. The sequence of families in an order, and the sequence of genera in a family, is generally based upon the evolution of some one im- portant structure. For example, in the Fucaceae, Fucus, with eight eggs in the oogonium, is placed first, while forms with four, two, and only one egg, come later. There can be no doubt that reduction in the number of eggs in this family is an evolutionary tendency, and, in this respect, Fucus is the most primitive. But, in the most primi- tive form of plant body, conceptacles cover the entire plant. On this basis, Hormoseira should stand at or near the beginning, although only four of its eight nuclei become nuclei of functional eggs. Fucus, 430 GYMNOSPERMS on this basis, would stand not at the beginning but near the end of the line. In the gymnosperms, also, the arrangement will depend upon the structure chosen as a basis. In the cycad, Macrozamia, there is a group of species closely re- sembling M. spiralis. It would be dubious to arrange these species in linear order, each species derived from the one before it. Probably some of the species have come, independently, from the highly variable M. spiralis. However, if two of the species differed some- what from M. spiralis, but resembled each other in the structure of the gland and venation of the leaf, it would seem more probable that one of them had come from the other than that both had developed, independently, identical minor features. Even the open carpel, from which the gymnosperms are named, is not more universally present than the free nuclear period at the germination of the megaspore. This free nuclear period is more characteristic of living gymnosperms than the single cotyledon is of monocots, or the two cotyledons are of dicots; but it would be hard to defend a claim that it is due to inheritance. It is a natural conse- quence of heterospory. Homosporous forms do not have any free nuclear period, and, as heterospory was beginning to emerge from homospory, there may have been no free nuclear period. Some living heterosporous pteridophytes have no free nuclear period, but with heterospory in such an advanced stage as in Selaginella and Isoetes, the early nuclear figures are too small to segment the comparatively large mass of protoplasm, and a free nuclear stage results. After a free nuclear stage has been established, it might persist, even in small megaspores, as in Taxus. A free nuclear stage in the embryo of gymnosperms, immediately following fertilization, occurs in all gymnosperms except Sequoia and, perhaps, Gnetum. It resembles that which occurs at the germination of the megaspore, and occurs for the same reason—the early nuclear figures are too small to segment the comparatively large mass of cytoplasm. The free nuclear periods in the young gametophyte and embryo do not indicate relationships. They are merely similar re- sponses to similar conditions. The seed is the final stage in the evolution of heterospory. There PHYLOGENY 431 were, doubtless, cases in which it would have been difficult, or im- possible, to draw a line between a heterosporous pteridophyte and a seed plant. If the retention of the megaspore makes the sporangium, with its contained megaspore, a seed, while a sporangium which sheds its megaspore, even at an advanced stage of development, has not yet reached the seed condition, a single individual might be partly fern and partly seed plant. Such a condition sometimes occurs in Se- laginella, and may have occurred rather frequently as the early gymnosperms were evolving from the heterosporous pteridophytes. During this period of transition, it would not be surprising to find the leaf remaining at the fern level. In very recent times, numerous varieties of apples, of various aspect, have arisen, while the leaves remained about the same. In the same way, the fern leaf was re- tained by the early gymnosperms. In vascular anatomy the Cycadofilicales are more advanced than the ferns with which they were associated. Circular pits are charac- teristic of the xylem of higher seed plants, while a scalariform marking is equally characteristic of ferns. The known Cycadofilicales have quite generally progressed beyond the scalariform stage, but we should expect to find it in their seedlings, for even the living cycads pass through a scalariform stage, and Stangeria does not get beyond it, except in a few tissues. The genus also retains a very fernlike leaf. This is instructive, for it shows that the fern leaf and vascular anat- omy may be retained after the seed habit has become established. Stangeria was described as a fern, and even assigned to the genus Lomaria, until its seeds were discovered. And so the lower gymno- sperms and plants, which had almost, but not quite, reached the seed condition, would naturally resemble each other so closely that they could not be distinguished by vegetative characters: and the resemblance would be due to genetic relationship. Where complete life-histories are known, we believe that recapitu- lation is a definite help in solving relationships, especially in nearly related forms. Here, we should lay the greatest stress upon the origin and development of heterospory and the seed. Next, we should rank the evidence from vascular anatomy, like the occurrence of spiral, scalariform, and pitted structures, and exarch, mesarch and endarch bundles. The behavior of leaves, like the evidence from 432 GYMNOSPERMS juvenile leaves, we should place last. On the whole, we believe in the recapitulation theory, but we feel certain that it has been called upon to explain things for which it is not responsible. From time to time we have stated or assumed that, in some way or other, the gymnosperms have come from the pteridophytes, and most botanists will agree to this theory. WETTSTEIN,*” a profound student of phylogeny, says that the gymnosperms represent a plant type which is traceable to the pteridophytes, by way of the Cycado- filicales. He thinks that the unity of the entire group is not such as to indicate the remnant of a former race, but rather, a terminal series going back to a common ancestry. WETTSTEIN’S view is accepted by most botanists. Both cycado- phytes and coniferophytes show an unmistakable pteridophyte ancestry. Have they arisen independently, or has one of them given rise to the other? If both have come from a common ancestor, the coniferophytes have progressed much farther and retain fewer of the ancestral characters. As far as any evidence from fossils is concerned the groups were never any more related than they are now. In some way or other the Cordaitales, the lowest of the conifero- phyte stock, came from the pteridophytes. Then the Cordaitales gave rise to the Coniferales and probably also to the Ginkgoales. Long after the Cordaitales became differentiated from the pteri- dophyte stock, the Bennettitales and Cycadales arose, independent- ly, from the Cycadofilicales. The cycadophyte line is much more homogenous than the coniferophyte, and, in our opinion, there is no doubt that it arose from the Filicales. The coniferophyte line is not so homogeneous. Those who lay great stress upon the leaf gap and use the terms “‘lycopsida” and ‘‘pteropsida,” insist that the coniferophyte line must have come from the Cycadofilicales, or at least from the Filicales, because all belong to the pteropsida. Others would minimize the value of the leaf gap in determining relationships, and say that lycopods do not have leaf gaps because the bundles are exarch and the leaf traces, being con- nected with the protoxylem, are already at the periphery of the stele, and naturally would not produce gaps. The lycopods, as a group, have comparatively small leaves, with entire margins, and, in this respect, bear a much closer resemblance to the coniferophytes, PHYLOGENY 433 in which the small leaf, with entire margin, is dominant. Of course, some gymnosperms, notably the Cordaitales, have large leaves; and in some, like Ginkgo, the margin is not entire. But there are no big compound leaves, like those of the Cycadofilicales. Cones were already highly developed in the lycopods, long before they appeared in the cycadophyte line, and they appeared in the Cor- daitales long before they appeared in any of the cycadophyte phylum. As far back as they can be traced, the lycopod and fern phyla seem to be just as distinct as they are now. Throughout our long association in teaching and in conducting research, the late Dr. Joun M. Courter and myself, relying chiefly upon the importance of the leaf gap, assumed that the coniferophyte line has come from the Cycadofilicales. As the survivor of the pair, if I were still in active teaching, I should still teach that theory, but the teaching would lack the dogmatic confidence which characterized the earlier days. So let us assume, but with minds still open, that the Cordai- tales have come, directly or indirectly, from ferns. The major problems which remain concern the origin of the Coni- ferales, Ginkgoales, and Gnetales. The Gnetales, like Minerva, seem to have sprung, full armed, from the head of Jove; but it is possible that the discovery of some new group, like the Caytoniales, may prepare the way for a less fanciful theory of origin. On the other hand, it is possible that thor- ough investigations of complete life-histories in the lower dicots, especially trees, will help to place the Gnetales where they really belong. For a long time the Ginkgoales were classed with the Taxaceae. When motile sperms were discovered, Ginkgo was removed from the taxads, and made the sole living representative of an order. The leaves and long stalks of the female flowers are also characteristic of the new order. The mere fact that the sperms are ciliated does not seem sufficient to warrant a new order. In its early development, the sperm is almost identical with that of Juniperus, and Juniperus usually has a dense mass of kinoplasm, which may represent a bleph- aroplast. There seems little doubt that the sperms of Juniperus are descended rather recently, phylogenetically, from motile sperms; 434 GYMNOSPERMS and the sperms of Taxus do not seem to be very far removed from a swimming ancestry. Ginkgo, like the cycads, has retained the swim- ming sperms, after other structures have advanced. Nevertheless, the swimming sperm is a universal pteridophyte character, which Ginkgo has inherited directly or indirectly from the pteridophytes. On the whole, it seems best to derive Ginkgoales from the Cor- daitales. Some of the seeds assigned to Cordaitales bear a close re- semblance to those of the living Ginkgo, especially in the archegonia and “tent pole” structure of the top of the female gametophyte. Material recently secured, and now under investigation, may throw some light on resemblances between Cordaites and Ginkgo. The seeds of the Cordaitales are already so advanced that they are far removed from the primitive seeds which must have existed while the lower coniferophytes were emerging from their heterosporous pteri- dophyte ancestry. The Coniferales have so many cordaitean characters that they must have come from the Cordaitales, or the two groups must have developed independently from a common ancestor. In this case, it hardly seems necessary to call upon parallel development, as is generally done when there is difficulty in making a direct connection. The Coniferales fall, rather naturally, into two great groups, the Pinares and Taxares, the former containing four families, Abietaceae, Taxodiaceae, Cupressaceae, and Araucariaceae; and the latter two families, the Podocarpaceae and Taxaceae. Most of the literature dealing with the phylogeny of gymno- sperms is concerned with the origin and relationship of these six families, and it is here that the most intensive investigations have been made. Life-histories in these six families are so well known that theories are based upon interpretation of a wide range of established facts. More investigations and more theorizing have been devoted to the Abietaceae and Araucariaceae than to all the other four together. Which of these two families is more ancient? Did both arise inde- pendently from the Cordaitales or some other ancestor, or did one of them give rise to the other? These are questions which are not yet settled. Most of the fossil material consists of stems, roots, and leaves, PHYLOGENY 435 with a comparatively scanty amount of strobili, and still less of gametophytic structures. Consequently, theories have been based almost entirely upon comparative anatomy of the sporophyte. According to JEFFREY, the most profound student of anatomy America has produced, and who, with his students, has developed the subject of comparative anatomy from both the phylogenetic and ecological standpoints, making abundant use of Mesozoic and Paleozoic material, the Abietaceae are the ancestral stock from \\ Sh \\\ g | \ SY pineze \ 2 MI | WZ cost Ni H ” frroveoriine, Arey Zo i he 2g nid ts WY in YY, Ei LZ FS g ee Cordatjoles Fic. 396.—Genealogical tree of the Coniferales, showing their proximity to the Gink- goales.—After JEFFREY.3" which the other five families have been derived. His views in regard to relationships in the coniferophyte phylum are summarized in a diagram taken from his book, The Anatomy of Woody Plants (fig. 396). Some of the evidence for and against a derivation of the Abieta- ceae from the Cordaitales is as follows: In spite of the antiquity of spur shoots in Ginkgoales and early members of the Abietaceae, the spur shoot must be regarded as a later development. The long shoot was the original form of the plant body. There are no spur shoots in the Araucariaceae, or in seedlings of the Abietaceae. As far as the theory of recapitulation is con- cerned, the long shoot is more primitive than the spur. 436 GYMNOSPERMS The crowded pitting and the absence of bars of SANio are claimed to indicate a close relationship between the Araucariaceae and Cordaitales, but, in araucarian seedlings, the pitting is not crowded, and in xylem, near the pith of the ovulate cone, the pits are neither crowded nor alternating. On the contrary, there is a strong tendency to opposite arrangement, and there are definite bars of SANro, which are lacking in adult wood of living araucarians. The Abietaceae and Ginkgoales have opposite pitting and bars of Sanio, both of which features are absent from the Cordaitales. These features were developed later in Pinus, because they are ab- sent from conservative regions, like the inner wood of the cone axis and the inner wood of the root, although they appear farther back from the primary xylem. In these conservative regions ray tracheids are also lacking, although they appear elsewhere. Pinus also lacks a torus in the membrane of the bordered pits near the primary wood, agreeing, in this respect, with Cordaitales. In the Araucariaceae the opposite is true, for the torus is present in the membrane of pits near the primary xylem, and absent elsewhere. If only the structure of the cone axis is considered, Pinus, representing the Abietaceae, has come from an ancestor with wood structure like that of the Cordai- tales. In general, JEFFREY believes that the structure of the wood in- dicates relationship between the Abietaceae and Cordaitales, rather than between Araucariaceae and Cordaitales. The ovulate scale and bract are separate in the Abietaceae. This seems to be a primitive character. Whether they are fused or not in the rest of the Pinares may be questionable. There may be no more fusion than in a gamopetalous corolla. The “fused” condition may be due to zonal growth, which gives rise to perigyny and epigyny. However, in these cases, the condition is regarded as more advanced than in hypogyny. Similarly, the “fused” bract and ovuliferous scale may represent a more advanced condition than the separate bract and scale. Then, in this respect, the Araucariaceae would be more advanced than the Abietaceae. Evidence from the gametophytes has received little attention; but, in our opinion, it is much more definite than the evidence from the anatomy of woody structures. The development of prothallial PHYLOGENY 437 cells in the microspore is more extensive in the Araucariaceae, es- pecially in Araucaria, than in any other known seed plant. There are often twenty or thirty prothallial cells, and as many as forty have been counted. Even the highest numbers recorded for any of the Abietaceae are insignificant in comparison. There is no doubt that the original gametophytes were green and independent. The retention of the gametophytes and their progressive reduction is well known. Pinus has only two evanescent prothallial cells, and the Cupressaceae, none at all. In this feature the Araucariaceae are the most primitive of living conifers. Not much is known about the male gametophyte of Cordaitales, but the best-preserved material indicates that there were many prothallial cells. As far as the evidence from the female gametophyte is known, it is similar, except that in the Abietaceae there is a distinct ventral canal cell, while in the Araucariaceae no wall is formed between the ventral canal nucleus and that of the egg. In the conifers a reduction in the number of free nuclei in the early embryogeny seems to be an evolutionary tendency. The Araucaria- ceae have thirty-two or sixty-four free nuclei, while in the Abietaceae the prevalent number is only four. In this feature the Araucariaceae are more primitive than the Abietaceae. On the whole, it is certain that in gametophytic structures and in early embryogeny the Araucariaceae are more primitive than the Abietaceae, while, on the other hand, the anatomy of the sporophyte would indicate that the Abietaceae are more primitive. If the geological record were complete, the comparative antiquity of Pinus and Araucaria and the rest would be settled definitely, al- though there would still remain the problem of determining whether their similarity was due to parallel development from a common ancestor, or to genetic continuity. ZEILLER and FiicHe7 '7 found pine cones of both the Strobus and Pinaster types already differentiated in the Jurassic of France; and Natuorst‘" concluded, on the evidence of winged pollen grains, that pines existed in the Triassic of Sweden. This would bring the Abietaceae into connection with the Cordaitales. The Araucariaceae are also very ancient, and were believed to have been abundant in the Carboniferous, on account of the Arau- 438 GYMNOSPERMS carioxylon type of wood; but this wood probably belonged to the Cordaitales. The “parallel” veined leaves of Cordaitales once placed the monocotyls in the Carboniferous, and settled “by the sure testimony of history” that the monocots must be older than the dicots. However, the mere fact that the wood structure is so similar in the two cases should have some weight. Scorr believed that the Araucariaceae have the longest fossil history of any of the Conifer- ales, overlapping that of the Cordaitales. Consequently, they might have come from the Cordaitales. Material is being found farther and farther back, and claims are being made that both Abietaceae and Araucariaceae existed in the Permian. The Taxodiaceae and Cupressaceae have not occasioned so much discussion, probably because it is rather generally agreed that they are a branch from the general abietineous stock. Their fossil record does not go as far back. They were abundant in the Cretaceous, and such names as Cupressites, Widdringtonites, and Thujites in- dicate their resemblance to living genera. Cupressinoxylon is Juras- sic, and the family doubtless was more widely represented in that time, but material has not been identified with as much confidence. When F Lorin completes his work on cuticular structures, some of the material which has been identified tentatively may be assigned definitely to its proper place. Both families have the ligneous resin canals and the marginal ray tracheids which characterize the Mesozoic Abietaceae. In tracing the origin of the two families, evidence from the Mesozoic Sequoias should be left out, for they are more like the Araucariaceae, and may belong with them. In gametophytic characters the two families illustrate how one feature may advance while another retains its primitive character. In Juniperus, and in the Cupressaceae generally, there is no pro- thallial cell in the male gametophyte. In the pollen tube there are two male cells, which have progressed but little beyond the swimming sperm level. But in the female gametophyte, there is no wall be- tween tlie ventral canal nucleus and that of the egg. Such features should be considered along with the evidence from comparative anatomy. PHYLOGENY 439 In general, the evidence indicates that the Taxodiaceae and Cu- pressaceae came from the Abietaceae, perhaps originating as a com- mon line which later differentiated into the two families. JEFFREY, who strongly supports the theory that the Abietaceae are the an- cestral stock, believes that the two families were differentiated from that stock earlier than the Araucariaceae. The other two families, Podocarpaceae and Taxaceae, constituting the Taxares, certainly belong together. In the Podocarpaceae the scales of the ovulate cone, and the spo- rophylls and spores of the staminate cone, indicate relationship with the Abietaceae. The wood is similar to that of the Cupressaceae and Taxodiaceae. It seems possible that a thorough study of very young ovulate cones might limit somewhat the application of the old phrase “‘conifers without cones.” In the Taxaceae, Cephalotaxus seems to be the most primitive genus. The ovulate strobilus looks like a cone, and has several ovulif- erous scales with young ovules, but usually only one ovule, the terminal one, develops. In Taxus usually one ovule develops. Oc- casionally two ovules develop and three have been known to start. A study of the development of the young ovulate strobilus in ma- terial which is fruiting superabundantly might be suggestive. Such material was noted on the’estate of REGINALD Cory Esq., at Duf- fryn, near Cardiff, Wales. This Taxus baccata was pruned to form an out-of-door room—without a roof—and the pruning may have caused the extreme production of seed, which gave a reddish tinge to the walls. Possibly such material might show ovules in the axils of the numerous bracts, and indicate the ancestral condition of Taxus. The Taxaceae are not known with certainty below the Cretaceous, although their morphological structure would indicate a greater age. It seems safe to say that from the Carboniferous onward the two great lines, Cycadophytes and Coniferophytes, have been distinct. They have some common characters, but the general outline of the life-history is the same in all seed plants, and the pteridophyte structures from which the seed evolved are similar, even in lycopods and ferns. If the two great lines of gymnosperms had a common origin, it is still to be demonstrated. CHAPTER XXI ALTERNATION .OF GENERATIONS Throughout the book we have been repeating that the micro- spore is the first cell of the male gametophyte, the megaspore is the first cell of the female gametophyte, and that the fertilized egg is the first cell of the sporophyte. This may seem trite, but it is funda- mental and the evolution of these structures in the gymnosperms is so suggestive that it seems worth while to consider them in relation to those which are lower and higher in the scale. That the gametophyte and sporophyte generations alternate in seed plants is accepted by every botanist. Similar alternation is just as readily recognized in the pteridophytes and bryophytes; but, in the thallophytes, there is not the same unanimity of opinion, partly because the average botanist is not so familiar with life-histories in thallophytes and partly from failure to recognize the origin of alter- nation. Below the level of sexuality there is no alternation of generations: merely celJl division. A cell may grow larger than its neighbors, become surrounded by a thicker wall, and tide the plant over un- favorable conditions; but there is no alternation. If such a plant has differentiated chromosomes, we should regard their number as the « number. When a plant reaches the level of sexuality, two cells (or their nuclei) fuse. When they fuse, each contributes the » number of chromosomes and the number is doubled, so that the resulting cell, the zygote, has 2x chromosomes. At first, the fusion was probably only a stimulus to development, and a reduction of chromosomes occurred at the first two mitoses in the zygote. No plant body, except the zygote, was built up. Spiro- gyra and many of the Chlorophyceae show this condition, and many botanists do not recognize any alternation; but alternation has orig- inated, and, from this beginning, the evolution of the 2x generation 440 ALTERNATION OF GENERATIONS 441 and the reduction of the x generation afford a splendid study in the comparative morphology of plants. If zodlogists knew this whole line of evolution and reductions they might find a reasonable in- terpretation for the three polar bodies. In many algae reduction does not take place at the germination of the zygote, but there is a more or less extended period of cell division preceding it. In such cases the 2x body becomes large enough to be seen with the naked eye. It may or may not look like the x body. In Zanardinia the x and 2x bodies look alike; but in Cuitleria, in the same family, they look different. In the Laminaria- ceae the 2x body has become immensely larger than the x body. In Coleochaete scutata the germinating zygote builds up a body which does not look like the structure that produced the gametes; but this new body, usually consisting of eight cells, has the « number of chromosomes, the reduction taking place during the first two mi- toses in the zygote, as in Spirogyra. It would be interesting to know where reduction takes place in those species of Coleochaete, which have more than eight cells in the body produced by the zygote. As many as sixteen and thirty-two have been counted. All these zy- gotes produce spores. Consequently, the eight-celled body, in Coleochaete, although an «x structure, is etymologically a sporophyte. Many «x bodies produce spores, and, in this sense, are sporophytes; but the same plant body may produce both spores and gametes, as in Ulothrix. Fucus, a 2x body, produces gametes. When one’s theories of alternation are based only upon plants from the bryo- phytes up, the fundamentals of the phenomenon may be over- looked. For this reason, we have reiterated: the spore is the first cell of the gametophyte, and the zygote is the first cell of the sporo- phyte. We might have said x and 2x generations; but from the bryophytes up, and in many thallophytes, x and 2x generations are synonymous with gametophyte and sporophyte generations. When plants which he regards as gametophytes produce spores, and those which he regards as sporophytes produce gametes, some botanists become confused. The trouble is that, in these cases, gametophyte and x generation and sporophyte and 2x generation are not synony- mous. The terms “gametophyte” and “‘sporophyte”’ are not as broad as x and 2x generations. 442 GYMNOSPERMS The zygote is the first cell of the 2x generation, whether it divides many times, building up a more or less extensive 2x body, or does not divide at all. The zygote, even in Spirogyra, might be called a sporophyte; for it produces four nuclei, three of which degenerate, while the other remains as the only nucleus of the cell which grows into a new filament. In Mesocarpus the four nuclei organize pro- toplasm about themselves and become recognizable spores, so that the older taxonomists called the whole structure a sporocarp. In Riccia, and in most of the liverworts, the sporophyte is para- sitic upon the gametophyte. In the moss, while still parasitic, it is green and partly independent. In the ferns, and in all plants above that level, the sporophyte, although at first parasitic, finally be- comes entirely independent of the gametophyte. In the gymnosperms, and in all plants which have reached the seed level, the sporophyte is parasitic until the seed germinates. In some of the extinct seed plants it is possible that the male gametophyte may have protruded somewhat from the spore, and may have developed some chlorophyll, but in all known gymno- sperms, both living and extinct, the male gametophyte is contained within the spore until a pollen tube is formed by those gymno- sperms which have pollen tubes. The vegetative part of the male gametophyte of gymnosperms is most extensive in the Araucaria- ceae, where it may consist of as many as forty cells. In most of the Podocarpaceae there is a vigorous development of prothallial cells, but the number is far less than the Araucariaceae. In the Abietaceae there are constantly prothallial cells, but not so many, and in Pinus, the most primitive genus, there are only two small evanescent cells. In the other three families, Taxodiaceae, Cupressaceae, and Taxa- ceae, prothallial cells are practically entirely lacking, so that the microspore, as in the angiosperms, is reduced to an antheridial initial. In Ginkgo there is one evanescent prothallial cell and one more or less permanent prothallial cell. In the Cycads there is one rather permanent prothallial cell. Nothing is known definitely about the prothallial cell situation in any extinct forms; but it is very probable that there were numerous prothallial cells in the Cordaitales, and probably some in the Cycadofilicales and Bennettitales. ALTERNATION OF GENERATIONS 443 The gymnosperms show the end of the series in the reduction of the vegetative part of the male gametophyte. The number of sperms is, dominantly, only two. Microcycas and Cupressus are notable exceptions in having a dozen or more sperms. Probably in extinct forms, near their pteridophyte ancestors, sperms were more numerous. The female gametophyte is, necessarily, parasitic in all seed plants. In all known gymnosperms its development begins with a period of free nuclear divisions, which is followed by a period of wall formation. This sequence was already established in the heteros- porous pteridophytes. Aside from a general reduction in the size of the gametophyte, the main features are the reduction of the archegonium and a delay in wall formation. In the most primitive archegonia, like those of Pinus and Ginkgo, there is a neck, consisting of two or more cells, and a ventral canal] cell separated from the egg by a definite wall. In most, if not all of the Abietaceae, there is a definite ventral canal cell. In the rest of the Coniferales the wall between the ventral canal nucleus and the egg nucleus is lacking, and is also lacking in the Cycadales. Nothing is known about the ventral canal situation in any extinct gymno- sperm; but we should expect to find a well developed ventral canal cell. In Torreya there is not even a ventral canal nucleus, and in Wel- witschia, not even a neck cell, the archegonium initial functioning as an egg. In Gnetum there is not even an archegonium initial. One or more of the free nuclei organize an egg, as in the angiosperms, and fertilization takes place with the upper part of the female gameto- phyte in the free nuclear condition, as in angiosperms. The lower part of this gametophyte, as often in the antipodal region in angio- sperms, is already cellular at the time of fertilization, and, asin angi- osperms, the rest of the gametophyte becomes cellular after fertiliza- tion. Consequently, the reduction of the female gametophyte has al- most reached the angiosperm level, the principal difference being that Gnetum has more antipodal cells and more free nuclei. Not much is known of the female gametophyte of extinct gymno- 444 GYMNOSPERMS sperms; but the female gametophyte of Cardiocarpus, assigned to Cordaites, looks like that of Ginkgo. On the basis of similar struc- tures, Ginkgo and Cordaitales are closely related. The megaspore membrane is highly developed in Cycadofilicales, and the gameto- phytes are cellular throughout. Riccia and all /p Cyanop qygee ow the ae of sexuality most of the Hepaticeae mos] oO Spii vs ist of the Ane Chl/otophyceae /Tosses Dictyota Zanardinia and (many others and many others Angio sperms and some ofhers some others Fic. 397.—Diagrams indicating the comparative extent of the x and 2x generations in various plants. In the second diagram, the extent of the 2x generation is exaggerated; in the sixth and seventh, the x generation is exaggerated. The female gametophyte of gymnosperms thus stands between those of the pteridophytes and the angiosperms. In the alternation of generations the sporophyte of the gymno- sperms has become as dominant as in the angiosperms. The male ALTERNATION OF GENERATIONS 445 gametophyte has reached the angiosperm condition in the final elimination of all prothallial cells. The female gametophyte, while not so much reduced as the male, has nearly reached the angiosperm level. The comparative display of the « and 2x generations may be illustrated by a series of diagrams (fig. 397). The broader line rep- resents the x generation, and the thinner line the 2x generation. In studying the alternation of generations it is interesting to note that the gametophytes, in the early stages of phylogeny, were green and independent, and that they later became parasitic and finally lost their chlorophyll, while the sporophytes, originally green and independent, became parasitic and lost their chlorophyll; then later, in the bryophytes, pteridophytes and seed plants—after pass- ing through a parasitic stage in their ontogeny—again became green and independent. Alternation of generations, viewed as an alternation of x and 2x phases in the life-history, is strictly antithetic. Since we believe that alternation of generations arose from the fusion of gametes, the re- sulting zygote being the first cell from which the evolution of the sporophyte began, we see no place for the old theory of homologous alternation, especially since it is based largely upon the phenomena of apogamy and apospory, which are merely abnormal digressions from the normal life-history. In considering alternation of generations in the gymnosperms, it is to be regretted that our knowledge of the fossil record is so in- complete, especially knowledge of the gametophytic phase in the life-history. It is to be hoped that some new discoveries of material will show what the gametophytes were in the later heterosporous pteridophytes and early gymnosperms which came fromthem. Until facts are available, we can only imagine that the gametophyte generation will become more and more extensive the farther back the record goes, and that the sporophyte—in spite of Devonian trees —will not be quite so far removed from a pteridophyte ancestry. No Io. II. I2. 13. 14. IS. 16. 17. BIBLIOGRAPHY . AASE, HANNAH, Vascular anatomy of the megasporophylls of conifers. Bot. Gazette 60: 277-313. figs. 196. 1915. . Arrourtit, Miss M. F. A., and LA Riviere, Miss H. C. C., On the rib- bing of the seeds of Ginkgo. Ann. Botany 29:591-05. fig. I. 1915. . ALLEN, C. E., The early stages of spindle formation in the pollen mother- cells of Larix. Ann. Botany 17:281-312. pls. 14, 15. 1903. . ARBER, AGNES, On the structure of the paleozoic seed Mitrospermum compressum. Ann. Botany 24:491-509. pls. 37-39. figs. 2. 1910. , Anote on Trigonocarpus. Ann. Botany 28:195, 196. fig. I. 1914. . ARBER, E. A. NEWELL, On some new species of Lagenostoma, a type of pteridospermous seed from the Coal-measures. Ann. Botany 19:326-28. 1905 (abstract). , “The origin of gymnosperms”; a discussion at the Linnean So- ciety. New Phytol. 5:68-76, 141-48. 1906. , On the past history of ferns. Ann. Botany 20:215-32. 1906. , On a new pteridosperm possessing the Sphenopteris type of foli- age. Ann. Botany 22:57-62. pl. 6. 1908. , A revision of the seed impressions of the British Coal Measures. Ann. Botany 28:81-108. pls. 6-8. figs. 8. 1914. , On the fossil floras of the Wyre Forest, with special reference to the neighboring Coal Measure areas. Phil. Trans. Roy. Soc. London B 204:363-445. pls. 26-29. 1914. ARBER, E. A. NEWELL, AND PARKIN, JOHN, On the origin of angiosperms. Jour. Linn. Soc. London Bot. 38:29-80. 1907. , Studies in the evolution of angiosperms: the relationship of the angiosperms to the Gnetales. Ann. Botany 22:48 9-515. 1908. , Beitriige zur Morphologie einiger Gymnospermen. I. Die Ent- wickelung des Endosperms bei Sequoia sempervirens. Bull. Soc. Imp. Nat. Moscou. Pp. 13 pls. 7, 8. 1899. ARNOLDI, W., Beitrige zur Morphologie und Entwickelungsgeschichte einiger Gymnospermen. II. Ueber die Corpuscula und Pollenschliuche bei Sequoia sempervirens. Bull. Soc. Imp. Nat. Moscou. Pp. 8. pls. 10, 11. 1899. , Beitriige zur Morphologie der Gymnospermen. III. Embryo- genie von Cephalotaxus fortunei. Flora 87:46-63. pls. 1-3. 1900. , Beitraige zur Morphologie der Gymnospermen. IV. Was sind die “Keimblischen” oder ‘‘Hofmeister’s-Kérperchen” in der Eizelle der Abietineen? Flora 87:194-204. pls. 6, 7. 1900. 446 18. IQ. 20. at. 22. 23. 24. 25. 26. a 28. 20. 30. ai. 32. 33- 34- 35- 36. 37: BIBLIOGRAPHY 447 , Beitriige zur Morphologie der Gymnospermen. VI. Ueber den Bau der Zellkerne im Embryo von Ginkgo biloba. VII. Die Embryobildung bei Ginkgo biloba. Ann. Inst. Agronomique et Forestiére a Nowoe, Alexan- dria 16:1-22. 1903. , Beitrige zur Morphologie einiger Gymnospermen. V. Weitere Untersuchungen der Embryogenie in der Familie der Seguoiaceen. Bull. Soc. Imp. Nat. Moscou. pp. 28. pls. 2. 1910. Arnotp, C. A., Cordaitean wood from the Pennsylvanian of Michigan and Ohio. Bot. Gazette 91:77-87. 1931. BaiLey, F. M., Comprehensive catalogue of Queensland plants. Brisbane. 1900. BaiLey, Irvinc W., The structure of the wood in the Pineae. Bot. Ga- zette 48:47-55. pl. 5. 1900. , Anatomical characters in the evolution of Pinus. Amer. Nat. 44: 284-93. IQIO. , A Cretaceous Pityoxylon with marginal tracheids. Ann. Botany 25:315-25. pl. 20. 191. —-—,, The structure of the bordered pits of conifers and its bearing upon the tension hypothesis of the ascent of sap in plants. Bot. Gazette 62: Wag —12..pl. Is TO10: BarLey, I. W., and SHEPHERD, H. B., Sanio’s laws for the variation in size of coniferous tracheids. Bot. Gazette 60:66-71. I915. Bartton, H., Recherches organogéniques sur la fleur femelle des Coni- feres. Ann. Sci. Nat. Bot. IV. 14:186-99. pls. 12, 13. 1860. BaANcROFT, Epitu S. W., Pteridosperm anatomy and its relation to that of cycads. New Phytol. 13:41. roro. Bancroft, NELLIE, Rhexoxylon africanum, a new Medullosean stem. Trans. Linn. Soc. London 8:87-103. pls. 10, IT. 1913. , Pteridosperm anatomy and its relation to that of cycads. New Phytol. 13:41-68. figs. 20. 1914. , On some Indian Jurassic gymnosperms. Trans. Roy. Soc. Lon- don B 8:69-86. pls. 7-9. 1913. BARKLEY, GRACE, Differentiation of vascular bundle of Trichosanthes an- guina. Bot. Gazette 83:173-85. 1927. BARTHOLOMEW, Mary, Abnormal prothallia of Pinus sylvestris. Notes Roy. Bot. Gard. Edinburgh 4:253-55. pl. 49. 1900. BartLett, A. W., Note on the occurrence of an abnormal strobilus of Larix europaea DC. Ann. Botany 27:575, 576. figs. 2. 1913. BEAL, J. M., Chromosome behavior in Pinus banksiana following fertili- zation. Bot. Gazette 95:660-66. pis. 2. 1934. Beccari, O., Dei organi dei fiori feminei del Gnetum gnemon L. Nuovo Giorn. Bot. Ital. 7:91-99. 1877. BetajerrF, W., Zur Lehre von dem Pollenschlauche der Gymnospermen. Ber. Deutsch. Bot. Gesell. 9: 280-86. pl. 18. 1801. 45. 46. 47- 48. 49. 50. mis 52. 53: 54. 55. 56. 57: 58. GYMNOSPERMS . Betayerr, W., Zur Lehre von dem Pollenschlauche der Gymnospermen (Zweite Mittheilung). Ber. Deutsch. Bot. Gesell. 11: 196-201. pl. 12. 1893. . Benson, MARGARET, On a new lycopodiaceous seedlike organ. New Phy- tol. 1:58-59. fig. 3. 1902. . —— ., The fructification of Lyginodendron oldhamium. Ann. Botany 16: 575-76. figs. 31. 1902. -, Telangium scottii, a new species of Telangium (Calymmatotheca) showing structure. Ann. Botany 18:161-77. pl. 11. 1904. . ————., On the contents of the pollen chamber of a specimen of Lageno- stoma ovoides. Bot. Gazette 45:409-12. figs. 2. 1908. —, Report of Proc. Roy. Soc. London in Nature 80: 239. 1909. , Cordaites Felicis, sp. nov., a Cordaitean leaf from the Lower Coal Measures of England. Ann. Botany 26:201-7. pl. 22. fig. I. 1912. , Sphaerostoma ovale (Conostoma ovale et intermedium Williamson), a Lower Carboniferous ovule from Pettycur, Fifeshire, Scotland. Trans. Roy. Soc. Edinburgh 50:1-15. pls. 1, 2. figs. 3. 1914. BERRIDGE, Emity M., Fertilization in Ephedra altissima. Ann. Botany 23:509-14. pl. I. 1909. , The structure of the female strobilus in Gnetum gnemon. Ann Botany 26:987-092. figs. 4. 1912. BERRIDGE, Emity M., and SANDAY, ELIzABETH, Oogenesis and embryog- eny in Ephedra distachya. New Phytol. 6:128-34, 167-74. pls. 3, 4. 1907. Berry, E. W., Vascular flora of the Lower Cretaceous. Bot. Gazette 53: 256. 1912. , Some araucarian remains from the Atlantic coastal plain. Bull. Torr. Bot. Club 35:249-60. pls. 11-16. 1908. , Amid-cretaceous species of Torreya. Amer. Jour. Sci. 25:382-86. 1908. BERTRAND, C. E., Remarques sur la structure des graines de pollen de Cordaites. Compt. Rend. Assoc. francaise pour l’avancement des sci- ences. 436-41. 18098. BERTRAND, C. E., et RENAULT, B., Recherches sur les Poroxylons. Archiv. Bot. Nord. France 3:147. figs. 79. 1886. Bessey, E. A., Notes on the spermatozoids of Ginkgo. Science 13:255. IQOl. Bessey, CHARLES E., The morphology of the pine cone. Bot. Gazette 33: 157-59. pl. 8. 1902. Biackman, V. H., On the cytological features of fertilization and related phenomena in Pinus sylvestris L. Phil. Trans. Roy. Soc. London B 190: 395-426. pls. 12-14. 1898. Boopte, L. A., Concrescent and solitary foliage leaves in Pinus. New Phytol. 14:19-22. figs. 4. 1915. Brooks, F. T., and Stites, W., The structure of Podocarpus spinulosa (Smith) R. Br. Ann. Botany 24:305-18. pl. 21. 1910. 59. 60. 61. 62. 63. 64. 65. 66. (7c 68. 60. 70. Wi: We 73- 74- 75- 76. 77: 78. BIBLIOGRAPHY 449 Bower, F. O., On the germination and histology of the seedlings of Wel- witschia mirabilis. Quart. Jour. Mic. Sci. 21:15-30. pls. 3, 4. 1881. , On the further development of Welwitschia mirabilis. Quart. Jour. Mic. Sci. 21:571-94. pls. 22, 23. 1881. , The germination and embryology of Gnetum Gnemon. Quart. Jour. Mic. Sci. 22:278-098. pl. 25. 1882. Braun, A., Vortrag auf dem Congrés scientifique de France, ro iéme Ses- sion (Strassburg), pp. 171, 172. 1842. (Communicated in Braun’s Le- ben von C. Mettenius, p. 335. Berlin. 1882). , BRONGNIART, A., Recherches sur les graines fossiles silicifiés. Paris. 1881. , Recherches sur l’organisation des tiges des Cycadées. Ann. Sci. Nat. I. 16:38 9-402. pls. 20-22. 1829. Brown, Rost., Character and description of Kingia, a new genus of plants found on the S.W. coast of New Holland, with observations on the structure of its unimpregnated ovulum and the female flower in Cycadeae and Coniferae. 1827. Included also in BRowN’s Miscellaneous botanical works 1:1866. Browne, Lapy ISABEL, The phylogeny and interrelationships of the Pteridophyta. V. Filicales. New Phytol. 7: 230-53. 1908. BucuuHo1z, JoHNn T., Suspensor and early embryo of Pinus. Bot. Gazette 66:185-228. pis. 5. figs. 3. 1918. ——-—.,, The suspensor and early embryo of Pinus. Bot. Gazette 66:185- 228. pls. 6-10. 1918. , Studies concerning the evolutionary status of polycotyledony. Amer. Jour. Bot. 6: 106-19. figs. 25. 1910. , Polyembryony among Abietineae. Bot. Gazette 69:153-67. figs. 75. 1920: , Embryo development and polyembryony in relation to the phy- logeny of conifers. Amer. Jour. Bot. 7:125-45. figs. 89. 1920. , Embryogeny of Cephalotaxus Fortunei. Bull. Torr. Bot. Club. 52: BUl—24. pla LO. es, Pr. 12. TO26. , Origin of cleavage polyembryony in conifers. Bot. Gazette 81: 55-71. pls. 3. figs. 2. 1926. , The embryogeny of conifers. Proc. Fourth Internat. Bot. Con- gress. I1:259-92. 19209. , The Ozark White Cedar (Description of Juniperus ashei n. sp.). Bot. Gazette 90:3 26-32. 1930. , The pine embryo and embryos of related genera. Trans. Ill. Acad. Sci. 23%117—25: 1931. , The suspensor of Sciadopitys. Bot. Gazette 92:243-62. 1931. , The suspensor of Cryptomeria japonica. Bot. Gazette 93: 221-26. 1932. GYMNOSPERMS . BucnuHorz, JoHN T., The embryogeny of Chamaecyparis obtusa. Amer. Jour. Bot. 19:230-38. 1932. , Determinate cleavage polyembryony, with special reference to Dacrydium, Bot. Gazette 94:579-88. 10933. , The classification of Coniferales. Trans. Ill. Acad. Sci. 21: (in press). 1934. . Bucnuorz, J. T., and Op, Eprrn M., The anatomy of the embryo of Cedrus, in the dormant stage. Amer. Jour. Bot. 20:35-44. 1933. . BuRLINGAME, L. L., The staminate cone and male gametophyte of Podo- carpus. Bot. Gazette 46:161-78. pls. 8, 9. figs. 9. 1908. , The morphology of Araucaria brasiliensis. Bot. Gazette 55:97- 114. pls. 4, 5. 1913. , The morphology of Araucaria brasiliensis. I1. The ovulate cone and the female gametophyte. Bot. Gazette 57:490-508. pls. 25-27. 1914. , The origin and relationships of the Araucarians. I. Bot. Ga- zette 60:89-114. IQI5. . BuTTERWORTH, J., Some further investigation of fossil seeds of the genus Lagenostoma (Williamson) from the Lower Coal-measures. Mem. and Proc. Manchester Lit. and Phil. Soc. 41:1-4. 1897. . CALDWELL, Otis W., Microcycas calocoma. Bot. Gazette 44:118-41. pls. 10-13. figs. 14. 1907. . CAMPBELL, D. H., Mosses and Ferns. New York. 1895. . CAROTHERS, IDA ELEANOR, The development of the ovule and the fe- male gametophyte of Ginkgo biloba. Bot. Gazette 43:116-30. pls. 5, 6. 1907. . CARRUTHERS, W., On gymnospermatous fruits from the secondary rocks of Britain. Jour. Botany 5:1-21. 1867. , On fossil cycadean stems from the secondary rocks of Great Brit- ain. Trans. Linn. Soc. Bot. 26:675-708. pls. 54-63. 1870. . Caspary, R., Abietinearum floris femini structura morphologica. Ann. Sci. Nat. Bot. IV. 14: 200-209. 1860. . Cavara, F., and Roeast, G., Ricerche sulla fecondazione ed embryogenia dell’Ephedra campylopoda Meyer (review in Bot. Centralbl. 92:5-6. 1902). , Sulla germinazione del polline nelle Ephedra. Boll. accad. Gioenia Sci. Nat. Catania 81:3-9. 1904. . Cetakovsky¥, L., Zur Gymnospermie der Coniferen. Flora. 62:257-64) 272, 283. 1870. ~, Zur Kritik der Ansichten von der Fruchtschuppe der Abietineen. Abhandl. Kgl. B6hm. Gesell. Wiss. VI. 11: 1882. , Die Gymnospermen: eine morphologisch-phylogenetische Studie. Abhandl. Kgl. Béhm. Gesell. Wiss. VII. 4:1-148. 1890. . ———, Nachtrag zu meiner Schrift iiber die Gymnospermen. Engler’s Bot. Jahrb. 24: 202-31. 18098. Ioo. Tol. 102. 103. 104. 105. 106. 107. 108. 109. IIo. DET. Ift2. E13; II4. EUG: TTO. Ery. 118. 119. 120. 12%. 122. E23; 124. 125. 1206. 127. BIBLIOGRAPHY 451 , Die Vermehrung der sporangia von Ginkgo biloba L. Oesterreich. Bot. Zeitschr. 50:229-37, 276-83, 337-41. 1900. CHAMBERLAIN, C. J., Contribution to the life-history of Salix. Bot. Ga- zette 23:147-79. pls. 12-18. 1897. ' , Winter characters of certain sporangia. Bot. Gazette 25:125-28. pl. 11. 1808. , The homology of the blepharoplast. Bot. Gazette 26:431-35. 1808. , Oogenesis in Pinus Laricio. Bot. Gazette 27:268-80. pls. 4-6. 1899. , Nuclear phenomena of sexual reproduction in gymnosperms. Amer. Naturalist 44:595-603. 1910. , The ovule and female gametophyte of Dioon. Bot. Gazette 42: 321-58. pls. 13-15. figs. 9. 1906. , Preliminary note on Ceratozamia. Bot. Gazette 43:137. 1907. , Spermatogenesis in Dioon edule. Bot. Gazette 47:215-36. pls. 15-18. 1909. , Dioon spinulosum. Bot. Gazette 48: 401-13. figs. 7. 1909. , Fertilization and embryogeny of Dioon edule. Bot. Gazette 50: AI5—29. pls. [4-17. 1910. , The adult cycad trunk. Bot. Gazette 52:81-104. figs. 20. 1911. , Morphology of Ceratozamia. Bot. Gazette 53:1-19. pl.r. figs. 7. TOT2. , Two species of Bowenia. Bot. Gazette 54:119-23. 1912. , Morphology of Ceratozamia. Bot. Gazette 53:1-109. pl. 1. figs. 7. IQI2. , Around the world botanical expedition. Popular Science Month- ly 81:417-33. I912. , Macrozamia Moorei, a connecting link between the living and fossil cycads. Bot. Gazette 55:141-54. 1913. , The oriental cycads in the field. Science 38:164-67. 1913. , A phylogenetic study of cycads. Proc. Nat. Acad. Sci. 1:86-go. IQI5. , Stangeria paradoxa. Bot. Gazette 61:353-72. pls. 2, 3. 1916. , The living cycads. University of Chicago Press. 19109. , The living cycads and the phylogeny of seed plants. Amer. Jour. Bot. 7:145-53. 1920. , The origin of Cycads. Science 61:73-77. 1925. —, Hybrids in cycads. Bot. Gazette 81:401-18. 1926. , An evaluation of the structural evidence for genetical relation- ship in plants. Proc. Internat. Congr. Plant Sci. Ithaca, N.Y. 1:473-80. 1920. , Elements of Plant Science. McGraw-Hill Book Co. 1930. , The age and size of plants. Scientific Monthly 35:481-o1. figs. I-9. 1032. , Methods in Plant Histology. University of Chicago Press. 1932. 135. 136. 137- 138. 139. GYMNOSPERMS . Cuavuveaun, G., Le liber précurseur dan le sapin pinsapo. Am. Sci. Nat. Bot. VIII. 19:32I-33. 1904. —, Originé secondaire du double faisceau foliaire chez les sapins (Abies) et les pins (Pinus). Ann. Sci. Nat. Bot. VIII. 19:335-48. 1904. —, Passage de la disposition primitive a la disposition secondaire dans les cotylédons du Pin maritime. Extr. Bull. Mus. Hist. Nat. Paris. I9O2. . Cuick, Eprru, The seedling of Torreya myristica. New Phytol. 2:83-o1. pls. 7, 8. 1903. . Cuurcna, A. H., On the floral mechanism of Welwitschia mirabilis. Phil. Trans. Roy. Soc. London B 205:115-51. pls. 9-13. 1914. , Thalassiophyta and the sub-aerial transmigration. Botanical memoirs, No. 3. Oxford. ro1o. . Curys ter, M. A., Tyloses in tracheids of conifers. New Phytol. 7:198- 204. pl. 5. 1908. , The origin of the erect cells in the phloem of the Abietineae. Bot. Gazette 56:36-50. 1913. , The medullary rays of Cedrus. Bot. Gazette 59:387-06. figs. 7. IQI5. , Vascular tissues of Microcycas calocoma. Bot. Gazette 82: 233-52. 1926. CLARKE, T. SH., and JOHNSTONE, G. R., Polyembryony and germination of polyembryonic coniferous seeds. Amer. Jour. Bot. 18:673-83. pl. 46. 1931. Coker, W. C., Observations on the gametophyte and embryo of Taxo- dium distichum. Johns Hopkins Univ. Circ. 19:45, 46. 1900. , Notes on the gametophytes and embryo of Podocarpus. Bot. Ga- zette 33:89-107. pls. 5-7. 1902. , On the gametophytes and embryo of Taxodium. Bot. Gazette 36:1-27, 114-40. pls. I-II. 1903. , On the spores of certain Coniferae. Bot. Gazette 38: 206-13. figs. 24. 1904. , Fertilization and embryogeny in Cephalotaxus Fortunei. Bot. Ga- zette 43:1-10. pl. 1. figs. 5. 1907. . Constatin, J., et Morot, L., Sur l’origine des faisceaux libéro-ligneux supernuméraires dans la tige des Cycadées. Bull. Soc. Bot. France 32: 173. 1885. . Cook, Met T., Polyembryony in Ginkgo. Bot. Gazette 34:64, 65. fig. I. 1902. . ———.,, Polyembryony in Ginkgo. Bot. Gazette 36:142. 1903. . CouLtTeR, Joun M., Notes on the fertilization and embryogeny of coni- fers. Bot. Gazette 23:40-43. pl. 6. 1807. , The embryosac and embryo of Gnelum Gnemon. Bot. Gazette 46: 43-49. pl. 7. 1908. 169. 170. BIBLIOGRAPHY 453 , Evolutionary tendencies among gymnosperms. Bot. Gazette 48: 81-97. 1909. . CouLTER, JOHN M., and CHAMBERLAIN, C. J., The embryogeny of Zamia. Bot. Gazette 35:184-94. pls. 6-8. 1903. . COULTER, JOHN M., and Curyster, M. A., Regeneration in Zamia. Bot. Gazette 38:452-58. figs. 8. 1904. . COULTER, JOHN M., and Lanp, W. J. G., Gametophytes and embryo of Torreya taxifolia. Bot. Gazette 39:161—78. pls. A, I-3. 1905. . CouLTeErR, J. M., Barnes, C. R., and Cow es, H. C., A textbook of bota- ny. New York. rg1o. . CouLter, J. M., and CHAMBERLAIN, C. J., Morphology of Gymnosperms. University of Chicago Press. 1917. . Cuttinc, E. M., On the meaning of the various forms of the male gametes in the pines and allied conifers. New Phytol. 7:171-76. 1908. . DALLIMORE, W., and Jackson, A. B., A handbook of Coniferae. 2d. ed. 1931. . DANGEARD, P. A., La mode d’union de la tige et la racine chez les Gymno- spermes. Compt. Rend. Acad. Sci. Paris. 110:253, 254. 1890. , Recherches sur les plantules des Coniféres. Le Botaniste 3:126- 204. pls. 13-17. 1892. . DANNEBL, Ueber die Bildung schizolysigener Schleimbehiilter bei Cerato- zamia. Bot. Arch. 29:92-122. 1930. . DE Bary, A., Comparative anatomy of the vegetative organs of the phanerogams and ferns. English translation. 1884. . DE FRAINE, E., On the structure and affinities of Suécliffia, in the light of a newly discovered specimen. Ann. Botany 26:1031-66. pls. 91, 92. 1912. . Dévis®, P., La figure achromatique et la plaque cellulaire dans les micro- sporocytes du Larix europea. La Cellule 32: 249-307. 1922. . Dickson, A., Observations on some bisexual cones occurring in the spruce fir (Abies excelsa). Trans. Edinburgh Bot. Soc. 6:418-22. 1860. . Drets, L., Pflanzengeographie. 3d. ed. 1929. . Dixon, H. N., Fertilization in Pinus sylvestris. Ann. Botany 8:21-34. pls. 3-5. 1804. . Doak, C. C., Multiple male cells in Cupressus arizonica. Bot. Gazette 94:168-82. figs. 15. 1932. . Dorety, HELEN A. (SISTER HELEN ANGELA), The embryo of Cerato- zamia,; a physiological study. Bot. Gazette 45:412-16. figs. 7. 1908. , The seedling of Ceratozamia. Bot. Gazette 46: 203-20. pls. 12-10. 1908. , Vascular anatomy of the seedling of Microcycas calocoma. Bot. Gazette 47:139-47. pls. 5, 6. 1909. The extrafascicular cambium of Ceratozamia. Bot. Gazette 47: 150-52. pl. 7. 1909. 454 173. 190. 19l. GYMNOSPERMS . Dorety, HELEN A. (SISTER HELEN ANGELA), Diagrams in CoULTER and CHAMBERLAIN’S Morphology of Gymnosperms. 1917. , Embryo and seedling of Dioon spinulosum. Bot. Gazette 67: 251- 58. 1919. Downie, Dorotuy G., The male gametophyte of Microcycas calocoma. Bot. Gazette 85:437-50. pls. 19-27. 1928. . Doyte, J., Note on the structure of the ovule of Larix leptolepis. Ann. Botany 30:193-95. 1916. . DucHartreE, P. E., Mémoire sur les embryons qui ont été decrits comme polycotyles. Ann. Sci. Nat. Bot. III. 10:207-37. pls. 7-10. 1848. . Dupier, A. W., The gametophytes of Taxus canadensis Marsh. Bot. Gazette 64:115-36. pls. II-14. 1917. , The staminate strobilus of Taxus canadensis. Bot. Gazette 68: 345-06. pls. 24-26. 19109. . —— , Ovuliferous structures of Taxus canadensis. Bot. Gazette 69: 492- 520. pl. 23. 1920. . Duture, Aucusta V., Anatomy of Gnetum africanum. Ann. Botany 26: 593-602. pis. 3. 1912. . Eames, A. J., The gametophytes of the Kauri. Science N. S. 35: 160. 1912. , The morphology of A gathis australis. Ann. Botany 27:1-38. pls. I-4. 1913. . Eames, A. J., and MAacDANnriets, L. H., An introduction to Plant Anato- my. McGraw-Hill Book Co. 1925. . Ercuier, A. W., Ueber Welwitschia mirabilis, etc. Flora 47:459-604, 473- 79, 489-96, 508-10, 513-20. 1863. , Sind die Coniferen Gymnospermen oder nicht? Flora 56: 241-47, 260-72. 1873. , Coniferae in Engler und Prantl’s Die Natiirlichen Pflanzenfami- lien 2:108. 1889. . Evxins, M. G., and Wi1eLanp, G. R., Cordaitean wood from the Indiana black shale. Amer. Jour. Sci. 38:65-78. pls. 1, 2. 1914. . ENGLEMANN, G., Morphology of the carpellary scales of Coniferae. Amer. Jour. Sci. ITI. 12:469, 470. 1876. . ENGLER, A., Gymnospermae (Geographic distribution) in Engler and Prantl’s Die natiirlichen Pflanzenfamilien. 2d. ed. 1926. . ENGLER and PRANTL, Die natiirlichen Pflanzenfamilien. 2d ed. 13: 1-447. figs. 240. 1926. Enrico, C. V., Contribuzione alla conoscenza della morphologia e dello sviluppo del fascio vascolare della foglie delle Cicadacee. Annales di Bot. 1:109-21. pls. 8, 9. 1903. FANKHAUSER, J., Entwickelung des Stengels und Blattes von Ginkgo. Bern. 1882. 192. 193. 194. 195. 196. 197. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 2009. 210. BIBLIOGRAPHY ASS FarMER, J. B., On the occurrence of two prothallia in the embryo sac of Pinus. Ann. Botany 6:213, 214. fig. 5. 1892. FAuLL, J. H., The anatomy of the Osmundaceae. Bot. Gazette 32:381- 420. pls. 14-17. 1901. FERGUSON, MARGARET C., The development of the pollen tube and the division of the generative nucleus in certain species of pines. Ann. Botany 15:103—223. pls. 12-14. 1OOT. , The development of the egg and fertilization in Pinus strobus. Ann. Botany 15:435-79. pls. 23-25. Igol. , Contributions to the life-history of Pinus, with special reference to sporogenesis, the development of the gametophytes, and fertilization. Proc. Wash. Acad. Sci. 6:1-202. pls. I-24. 1904. FLICHE, P., et ZEILLER, B., Note sur une florule portlandienne des environs de Boulogne-sur-Mer. Bull. Soc. Geol. France IV. 4:787-812. 1904. Fiorin, Rupotr, Untersuchungen zur Stammesgeschichte der Conife- rales und Cordaitales. Erster Teil. pp. 1-588. pls. 58. figs. 111. Stock- holm. Almquist & Wicksells. 1931. Fuju, K., On the different views hitherto proposed regarding the mor- phology of the flowers of Ginkgo biloba. Bot. Mag. Tokyo 10:15-25, 104- to. pl. 1. 1806. , (Has the spermatozoid a tail or not?). Bot. Mag. Tokyo 12:287- 90. 1898. (Japanese.) , (On the morphology of the spermatozoid of Ginkgo). Bot. Mag. Tokyo 13:260-66. pl. 7. 1899. (Japanese.) , (Account of a sperm with two spiral bands). Bot. Mag. Tokyo 14: 16, 17. 1900. (Japanese.) Fuy1oKA, M., Studien iiber den anatomischen Bau des Holzes japanischer Nadelbaume. Jour. Agric. 4:201-36. pls. 18-24. 1913. Futter, Geo. D., Reproduction by layering in the black spruce. Bot. Gazette 55:452-57. figs. 6. 1913. GAGER, C.S., Rare cycads from Australia. Records Brooklyn Bot. Gard. 4:83-92. IQIS5. GEITLER, L., Zur Zytologie von Ephedra. Oest. bot. Zeitsch. 78: 242-50. 19290. : Gerry, ELotsE, The distribution of the “bars of Sanio”’ in the Coniferales. Ann. Botany 24:119-24. pl. 13. IgIo. , Tyloses: their occurrence and practical significance in some American woods. Jour. Agric. Research. 1:445-70. pls. 52-59. 1914. GuosgE, S. L., A contribution to the morphology of A gathis ovata (Moore) Warb. Jour. Indian. Bot. Soc. 4:79-86. pls. 2. 1924. Grpss, L. S., On the development of the female strobilus in Podocarpus. Ann. Botany 26:515-71. pls. 5. 1912. 450 2ift 212. 213. 214. 215. 2106. 217. 218. 219. 220. 221. 222. 22%, 224. 225. 226. 227. 228. 229. 230. 231 GYMNOSPERMS . GorBEL, K., Outlines of classification and special morphology. English translation, p. 316. 1887. ————., Organography of plants. English translation. pp. 153-55. 1900. ———, Morphologische und biologische Bemerkungen. 13. Ueber die Pollenentleerung bei einigen Gymnospermen. Flora 91:237-63. figs. 10. 1902. GoopsPEED, T. H., Chromosome number in the Seguoias. Bot. Gazette 69:348-49. 1920. ; Gorpon, MArjori&, Ray tracheids in Sequoia sempervirens. Bot. Gazette 54:256. 1912. GoROSCHANKIN, J., Zur Kenntniss der Corpuscula bei den Gymnosper- men. Bot. Zeit. 41:825-31. pl. 7A. 1883; Ueber den Befruchtungs-Process bei Pinus Pumilio. 1883. Gotuan, W., Zur Anatomie lebender und fossiler Gymnospermenhdlzer. Abhandl. Preuss. Geol. Landesanstalt. Neue Folge. Heft 44. 1905. , Die fossilen Hélzer von Konig Karls Land. Handl. Kgl. Svensk. Vetensk-Akad. 42: No. ro. GRAND D’Evry, C., La flore carbonifére du Département de la Loire. Mém. Acad. Sci. 24: pp. 624. pls. 38. 1877. , Sur les rhizomes et les racines des fougéres fossiles et des Cycado- filices. Compt. Rend. Acad. Sci. Paris. 138:607-10. 1904. , Sur les graines des Neuropteridées. Compt. Rend. Acad. Sci. Paris. 139: 23-27, 782-86. 1904. , Sur les Rhabdocar pus, les graines et l’évolution des Cordaitées. Compt. Rend. Acad. Sci. Paris 140:995-98. 1905. , Sur les graines de Sphenopteris, sur l’attribution des Codonosper- mum, et sur l’extréme variété des grains des fougéres. Compt. Rend. Acad. Sci. Paris 141:812-15. 1905. , Sur les graines trouvées attachées au Pecopteris pluckenetii Schlot. Compt. Rend. Acad. Sci. Paris. 140:920-23. figs. 2. 1905. ———, Sur les inflorescences des fougéres avec graines, etc. Compt. Rend. Acad. Sci. Paris 143:764. 1906. Groom, Percy, A preliminary inquiry into the significance of tracheid caliber in Coniferae. Bot. Gazette 57: 287-307. 1014. Groom, Percy, and Rusuton, W., Structure of the wood of East Indian species of Pinus. Jour. Linn. Soc. Bot. 41:457-00. pls. 24, 25. 1913. Hacerup, O., Zur Organogenie und Phylogenie der Koniferen-Zapfen. Det. kgl. Danske. Videnskabernes Selokab. Biologiske Meddelelser ro: 1-82. 1933. Harninc, Hupa J., Development of embryo of Gnetum. Bot. Gazette 70:436-45. pls. 39-41. 1920. HALE, J. D., The bars of Sanio. Bot. Gazette 76:241-56. 1923. . Hanpa, M. R., The life-history of Thuja occidentalis. Jour. Burma Re- search Soc, 16:214-19. pl. 1. 1926. 232. 233; 251. 252: BIBLIOGRAPHY 457 Harris, T. M., A new member of the Caytoniales. New Phytol. 33:97- fig NGO. "2) Q7=Lid4s Loss. Harvey, L. R. H., Intra-microsporangial development of the tube in the microspore of Pinus sylvestris. Rep. Michigan Acad. Sci. 19:333-35. IQI7. . Hawker, L. E., Microsporogenesis in Taxus. Ann. Botany 44:535-39. 1930. . Hayata, B., Taiwania, a new genus of conifers. Jour. Linn. Soc. Bot. 32: 330-32. pl. 16. 1906. , On Taiwania and its affinity to other genera. Bot. Mag. Tokyo 21:21-27. pl. I. 1907. . Haypon, WALTER T., The seed production of Pinus sylvestris. Proc. and Trans. Liverpool Biol. Soc. 22:1-32. figs. 16. 1907. Flora fossilis arctica. 1868-83. . HEER, O., Zur Geschichte der Ginkgo-artigen Biume. Engler’s Bot. Jahrb. 1:1-13. 1880. . Hemenway, A. F., Cryptomeria japonica. Bot. Gazette 52:153-54. figs. Ae TOUT: . HemsLey, Hybrid cycads. Gardeners Chronicle. 19: 466-67. 1883. . HERZFELD, DR. STEPHANIE, Die Bliithen der Bennettitalen. Oesterr. Bot. Zeitsch. 62: 289-303. figs. 14. 1912. , Ephedra campylopoda. Morphologie der weiblichen Bliithe und Befruchtungsvorgang. Denkschriften der Akademie der Wissenschaften in Wien. Mathematisch-Naturwissenschaftliche Klasse. 98: 243-68. pls. T2022: , Beitriige zur Kenntniss von Ginkgo. Jahrb. Wiss. Bot. 66:814—-62. pls. 14-17. 1927. , Ueber die Kernteilung im Proembryo von Ginkgo biloba. Jahrb. wiss. Bot. 69: 264-94. 1928. . Hit, T. G., On secondary thickening in recent Pteridophyta. New Phytol. 5: 208-14. 1906. , The germination of Gnetum Gnemon L. Jour. Roy. Hort. Soc. 34: £, 2. 1908: . Hit, T. G., and DE Frame, E., On the seedling structure of gymno- sperms. Ann. Botany 20:471-73. 19006. , On the seedling structure of gymnosperms. I. Ann. Botany 22: 689-712. pl. 8. figs. 35. 1908. , On the seedling structure of gymnosperms. II. Ann. Botany 23: 189-227. pl. 15. figs. II. 1909. , On the seedling structure of gymnosperms. III. Ann. Botany 23: 433-58. pl. 30. 1909. , On the seedling structure of gymnosperms. IV. Gnetales. Ann. Botany 24:319-33. pls. 22, 23. 1910. 265. 266. 267. 268. 260. 270. a7I. 272. 273. GYMNOSPERMS . Hrrasf®, S., Etudes sur la fécondation et l’embryogénie du Ginkgo biloba. Jour. Coll. Sci. Imp. Univ. Tokyo 8:307-22. pls. 31, 32. 1895. ————, (Announcement of the spermatozoid of Ginkgo). Bot. Mag. To- kyo 10:171, 172. May 1896. (Japanese.) . ——— ., (Preliminary report on the spermatozoid of Ginkgo). Bot. Mag. Tokyo 10:325-28. Oct. 1896. (Japanese.) , Etudes sur la fécondation et l’embryogénie du Ginkgo biloba. Jour. Coll. Sci. Imp. Univ. Tokyo 12:103-49. pls. 7-9. 1808. , Nouvelles recherches sur la fécondation et l’embryogénie du Ginkgo biloba. Bot. Mag. Tokyo 32:139-43. 1918. . Hormetster, W., Embryobildung der Phanerogamen. Jahrb. W. Bot. 1: 82-188. pls: 7-10. 1858. , On the germination, development, and fructification of the higher Cryptogamia. English Translation. London. 1863. , Lehre der Pflanzenzelle. 1867. . HoLpEN, RutH, Ray tracheids in the Coniferales. Bot. Gazette 55: 56-66. pls. I, 2. 1913. . ———, Some fossil plants from eastern Canada. Ann. Botany 27: 243-55. pls. 22, 23. 1913. , Cretaceous Pityoxyla from Clifford, N.J. Proc. Amer. Acad. 48: 609-23. pls. 4. 1913. , Contributions to the anatomy of Mesozoic conifers. I. Jurassic coniferous woods from Yorkshire. Ann. Botany 27:533-45. pls. 39, 40. 1913. , On the relation between Cycadites and Pseudocycas. New Phytol. 13:334-40. pls. 3. 1914. , Contributions to the anatomy of Mesozoic conifers. II. Creta- ceous lignites from Clifford, N.J. Bot. Gazette 58:168-77. pls. 12-15. IQI4. , A Jurassic wood from Scotland. New Phytol. 14:205-9. pl. 3. IQI5. pl. 11. 1915. Ho tick, A., Cycads, living and extinct. Jour. N.Y. Bot. Gard. 24:135- 40. 1926. Ho tick, A., and Jerrrey, E. C., Affinities of certain cretaceous plant re- mains commonly referred to the genera Dammara and Brachyphyllum. Amer. Nat. 40:189-215. pls. 1-5. 1906. ———, Studies of cretaceous coniferous remains from Kreischerville, New York. Mem. N.Y. Bot. Garden 3:1-138. pls. 1-29. 1900. Hooker, J. D., On Welwitschia, a new genus of Gnetaceae. Trans. Linn. Soc. London 24:1-48. pls. 1-4. 1863. Hutcuinson, A. H., The male gametophyte of Abies. Bot. Gazette 57: 148-52. figs. 15. 1914. , On the cuticles of some Indian conifers. Bot. Gazette 60: 215-27." [ee —e————E——E——@@mGx