Fecundation in Plants BY DAVID M. MOTHER, Ph. D., Professor of Botany in Indiana University Published by the Carnegie Institution of Washington 1904 msM^mmjB^Sl^j^^m^ http7/wWw.archive.org/detalis/fecuhciaMiiif^ Fecundation in Plants BY DAVID M. MOTHER, Ph. D.. Professor of Botany in Indiana University Published by the Carnegie Institution of Washington 1904 Carnegie Institution of Washington, Publication No. 15. Press of Gibson Bros., Washington, D. C. PREFACE. This volume presents the subject of fecundation in the vegetable kingdom by the discussion of concrete cases, selecting from the great groups of plants certain typical representatives in which the sexual process seems to have been most thoroughly investigated. In the introductory chapter I have discussed typical processes of nuclear division and cell-formation, especially in spore mother-cells, together with a few topics dealing with certain phenomena of the cell and the significance of sexuality. This is considered necessary to a better understanding of sexual reproduction, for problems of sexuality, like problems of evolution, have in late years become reduced to problems of the cell, and, since the nucleus plays by far the most important part in fecundation, I am tempted to say to problems of the nucleus. The processes leading to the development and differentiation of the gametes have been regarded as of prime importance, and they have therefore received emphasis. Whenever the subsequent history of the fecundated egg has been followed to any extent this has been done, as in the Ascomycetes and Floride^^ to show the relation between the real sexual process and the vegetative fusion of nuclei which has been confused with the sexual act, and, as in the Desmids, for the sake of pointing out certain nuclear phenomena that take place during the germination of the zygote with similar phenomena just preceding the sexual act in the Diatoms. Processes which are purely morphological are assumed or dealt with very briefly. In grouping the representative types into the several chapters I have had in mind no particular theory of the evolution of sexuality, but merely the idea of the evolution of the plant kingdom and the corre- sponding differentiation of the sexual organs and cells accompanying this evolution in the groups of plants themselves. The chapters dealing with the lower plants in which the develop- ment of the gametes is not known from a modern cytological standpoint, and in which the behavior of the sexual nuclei in the fusion of the gametes has not been followed — have been made as brief as possible. For a similar reason the mosses and liverworts have been omitted entirely. iv PREFACE. No attempt has been made to discuss the numerous theories bearing upon the subject. Whenever theoretical matters are touched upon the object has been chiefly to suggest probable lines of investigation. I have not hesitated, however, to express my own opinion in all cases in which my special field of study has given me a first-hand knowledge of the subject-matter. To designate the sexual process which consists in the fusion of sex- ually differentiated cells, or gametes, and especially the fusion of their nuclei, the term fecundation has been used instead oi fertilization — fecundation being the equivalent of the German Befruchtung and the Yrenchfecondation. It has been necessary, of course, to copy numerous figures from the papers of other investigators, but in every case due credit is given. In the citation of literature in the text the author is referred to by the year in which his work was published. No attempt has been made to g^ve a complete bibliography, and no doubt many valuable refer- ences have been omitted. The author is indebted to Professors W. Belajeff, H. O. Juel, F. Oltmanns, S. Ikeno, and to Dr. H. Klebahn, Dr. A. H. Trow, Dr. H. Wager, Dr. S. Hirase, and Dr. V. H. Blackman, for re- prints of their papers, from many of which illustrations have been borrowed, and especially to Professor R. A. Harper for helpful suggestions. David M. Mottier. Indiana University, August^ igo2. CONTENTS. Chapter I. — Introduction. Pagk. Nuclear division, .... . 2-30 Karyokinesis in cells of the lower plants in which centrospheres are developed, . 2-10 Dictyota, .... . 2 Erysiphe, . 7 Mitosis in pollen mother-cells, . 11-30 The first or heterotypic mitosis, . 11-26 Resting nucleus and the development of the chromatin spirem II Development of the spindle. 15 Chromosomes, . . 17 Metakinesis, 20 The anaphase, . . 22 The telophase, 23 The nucleolus, . . 25 The second, or homotypic division, 27-31 Cell division, . 31-44 The type of the higher plants. 31 Free cell-formation. 33 Cell-cleavage, . . . • 36 Cell-division in Dictyota and Stypocaulon, . 41 The centrosome and the blepharoplast. 44 The significance of the sexual process and the numerical reduction of the chromosomes, . 49-60 Chapter II. — Fecundation; Motile Isogametes. Ulothrix and Hydrodictyon, ...... 61-65 Copulation of gametes, ....... 65 Ectocarpus, ....... "5 Chapter III.— Fecundation; Non-Motile Isogametes. Spirogyra, ....•••• Sporodinia, ...••••• Closterium and Cosmarium, ...... Diatoms (Rhopalodia, Cocconeis), ..... Basidiobolus, .....••• 67 71 71 73 76 VI CONTENTS. Chapter IV. — Fecundation ; Heterogametes. Sphseroplea, ........ 79 Fucaceae (Fucus, Halidrys), ...... 84 Volvox, ........ 88 OEdogonium, ........ 89 Coleochaete, . . . . . . . 91 Vaucheria, ........ 94 Albugo (Cystopus), ....... 96 Achlya and Saprolegnia, . . . . . .102 Chapter V. — Type of the Ascomycetes and Rhodophyce.(E . Sphaerotheca, . • • • . . . 108 Pyronema, . • • • • . . iir Batrachospermum, ....... 116-119 Dudresnya, ....... 1 19-126 CoUema, ........ 126-128 Chapter VI. — Archegoniates. .Pteridophyta, • • • • .... . . i2g The spermatozoid, ...... 130-136 The egg-cell and fecundation, .... 136-142 Gymnosperms, ....... j^2 Cycas, Zamia, and Ginkgo, . . . . j^2 The male gametophyte and the development of the sperma- tozoids, ....... 142-155 The archegonium, ..... 156-158 Fecundation, ...... 158-163 Pinus, ....... j5. The male and female gametophytes, . . . 163-164 Fecundation, ...... 165-168 Chapter VII. — Angiosperms. The embryo-sac, or female gametophyte, .... 169-174 The male gametophyte, ..... 174-176 The fusion of male and egg-nucleus, .... 176-177 The fate of the second male nucleus in the embryo-sac, . 177-180 Bibliography, 1 81-187 INDEX. Page. Abies ...... 156 Achlya ....... 102-107 Adiantum ...... 136 Albugo ....... 96-100 Aspidium ...... 136 Basidiobolus ...... 76-78 Batrachospermum . "■ • 116-119 Callithamnion ...... I 19-124 Cell-cleavage in Synchitrium discipens 36-38 Pilobolus crystallinus 38-41 Cell-division in higher plants 31-33 Dictyota and Stypocaulon 4'-43 Cell- formation, free, in Erysiphe communis 33-35 Lachnea scutellata . 35 Centrosome, in Dictyota .... 3-7 Erysiphe . . . . 8-10 Centrosome and Blepharoplast 44-49 Cephalotaxis ..... 157' Chara ...... 135-136 Chromosomes in tetraspore mother-cell of Dictyota . 5-6 ascus of Erysiphe 8-1 1 pollen mother-cells of Lilium 17-31 Podophyllum 17-31 Tradescantia 17-31 Significance of numerical reduction 49-60 Closterium ..... 71 Cocconeis ...... 75 Coleochsete ... 91-93 CoUema ...... 126-128 Cosmarium ...... 71, 72 Cycas ...... 142-149, 156, 1 57, 163, 166 Cystopus Csee Albugo). Dasya ...... 124 Diatoms ...... 73-76 Dictyota ..... 2-6, 26 Dudresnya ...... 119-125 Ectocarpus ..... 65, 66 Equisetum ...... 135 Erysiphe ..... 7-10 Fucus ...... 84-88 Ginkgo ...... 149-155. 162, 163, 166 Gloecosiphonia ..... 124 Gnetum ..... . 168, 173 Gymnogramme ..... 130-132 Halidrys ..... 8s Helleborus ..... . 12, 158, 169-171, 173 Hydrodictyon . . • . . 63-65 INDEX. Karyokihesis (see Mitosis). Laboulbeniaceae Larix .... Lilium : Mitosis in pollen mother-cells Development of mitotic spindle in pollen mother-cells Behavior of chromosomes in pollen mother-cells Nucleolus ..... Second or homotypic mitosis in pollen mother-cells Embryo-sac and Fecundation Fate of second male nucleus in embryo-sac Marsilia ...... Mitosis in Dictyota Erysiphe pollen mother-cells Monotropa Nemalion Nucleolus, discussion of CEdogonium Onoclea . Peperomia . Peronospora Picea Pilularia Pinus 'Physcia ...... Podophyllum : Resting nucleus of pollen mother-cell Nature of nuclear membrane Behavior of chromosomes in pollen mother-celt Pteridophyta ..... Pyronema Pythium Rhopalodia gibba Saprolegnia Sphaeroplea . Sphaerotheca Spirogyra Sporodinia Synapsis Tradescantia virginica : Behavior of chromosomes in pollen mother-cell Second or homotypic mitosis in pollen mother-cell Tsuga Tulipa Ulothrix Vaucheria Vicia faba Volvox Zamia Zea mays 133. 130- H9-ISS, 1 Page. 126 158, 170-171 11-30 15-16 17-24 25 27-30 169-177 177-178 134^ 135 2-7 7-1 1 11-29 177 119, 131 25, 26 89-91 '33. 136, 138-141 173 lOI 163 142 156, 163-168 128 II, 12 13. 24 18, 22 129-142 111-116 lOI 73> 75, 76 102, 107 79-84 108-111 26, 67-70, 168 71 13 . 18, 19, 22 27, 29 163, 165, 166, 167 178 61, 62, 65 94.95 25 88 57-161, 163, 166 25, 178 FECUNDATION IN PLANTS. CHAPTER I.— INTRODUCTION. The processes of nuclear division and cell-formation are so closely associated with sexual cells and their development that an adequate understanding of these cells is impossible without a definite and thorough knowledge of the processes involved in their development. Our interpretations of the significance of the sexual process and the phenomena of heredity in all organisms will be more lasting and help- ful as scientific knowledge if these interpretations or doctrines are based upon a well-connected phylogenetic series of the most funda- mental facts. Perhaps no other field of research has been more helpful during the past quarter of a century in enabling the biologist to gain a deeper and more far-reaching knowledge of the physical basis of heredity than the study of mitosis, especially in reproductive cells. The division of the nucleus naturally suggests the division of the cell, or the process by which new cells are formed from a mother- cell, and the study of cell-formation in very recent years, especially among the lower plants, has not only wrought almost a revolution in our knowledge of the processes here involved, but has also furnished new criteria for determining relationships and probable lines of descent. It is deemed necessary, therefore, to introduce the subject of sexual reproduction in plants by a brief presentation of the typical processes of nuclear ahd cell-division in both the lower and higher forms. In doing so these processes will be described in a few of those forms which have been subjected to a critical study by means of the most improved methods and instruments. The processes described will be confined largely, though not exclusively, to spore mother-cells. The division of the nucleus and of the cell presents generally three processes, the development of the karyokinetic spindle, the behavior of the chromatin, and the formation of the cell-plate or new plasma membrane. This division is made merely for the sake of convenience, as it is not implied that three distinct or separate processes are necessarily involved, although the development of the plasma mem- brane in many cases has apparently little or no connection with the 2 '' INTRODUCTION. division of the nucleus. The first two of these processes will be dis- cussed under nuclear division^ while the third will be dealt with in connection with cell-formation. NUCLEAR DIVISION. KARYOKINESIS IN CELLS OF THE LOWER PLANTS IN WHICH CENTROSOxMES AND CENTROSPHERES ARE DEVELOPED. At present there are recognized two types of development of the karyokinetic spindle. In one the spindle arises through the instru- mentality of individualized dynamic centers or centrospheres, as in certain Thallophyta and Liverworts ; in the other, it is developed wholly independently and in the absence of any such centers, as, for example, in the higher plants. We speak of types of spindle develop- ment in this connection also for the sake of convenience, since centro- spheres have not been found in connection with the development of the spindle in all Thallophytes ; but the author does maintain that centrospheres have not been demonstrated to occur in any plant above the Bryophytes, and that in the Angiosperms such structures do not in all probability exist. As illustrating the development of the spindle in which centro- spheres are present, the tetraspore mother-cell in Dictyota dichotoma will be selected from the algae and the mother-cell of the ascus in Erysiphe from the fungi. It is not considered necessary, nor conducive to any better under- standing of the facts presented here, to enter into any lengthy dis- cussion concerning the structure of the firmer framework of the cytoplasm. The consensus of opinion now is that the firmer substance of cytoplasm consists of either a reticulum of fibrillae or of an alveolar or foam structure (Waben of German literature) and that, in many cells, these two structures intergrade into one another. DICTYOTA. The cytoplasm of the tetraspore mother-cell of Dictyota dichotoma during the preparation for nuclear division presents two well-defined portions, the kinoplasm, which is always associated with the nucleus and plays the most important r61e in the karyokinetic process, and the remaining alveolar portion. Numerous chloroplasts are also present. The first indication of mitosis is the appearance, on opposite sides of the nucleus, of two large sharply defined asters of kinoplasmic fibers radiating from a rod-shaped body, which is often slightly bent, lying either close to the nuclear membrane or at some little distance from it (Fig. i , A) . The rod-shaped body is the centrosome^ which NUCLEAR DIVISION. together with the kinoplasmic radiations constitutes the centrosphere. The planes of the longitudinal axes of the centrosomes may be parallel or form various angles with each other. In Fig. i, B, the centrosome at the upper side of the nucleus is seen from the side, the lower from Fig. X. — First mitosis in tetraspore mother-cell of Dictyota dichotonta. A, B, early prophase ; the well-developed centrospheres are on diametrically opposite sides of nuclei. C, the kinoplasmic fibers have begun to enter the nucleus to form the spindle and the chromosomes are being differentiated. D, numerous spindle fibers have entered the nucleus, and the chromosomes are collected in the equa- torial region. the end. Viewed from the pole, the centrosome is always rod-shaped. The kinoplasmic fibers radiate in all directions into the cytoplasm where they pass over into the framework of the same. On the side next the nucleus they may run parallel with its wall for some dis- A INTRODUCTION. tance. Near the nucleus the cytoplasm is more granular, with smallet meshes. It is more nearly a thread-like net-work than alveolar in structure, and appears with differential staining as kinoplasm. This very fine granular thread-work often extends in among the radiations of the centrosphere. The resting nucleus shows a large vacuolated nucleolus and a fine linin-reticulum with rather large meshes, upon which are arranged small and nearly uniform granules, all of which do not react as chromatin. With the advance of karyokinesis, the chromatin begins to collect into larger and somewhat irregular masses that finally become the chromosomes. There is not developed, as in vegetative cells of this plant, a regular and uniform chromatin spirem or ribbon. The nucleolus becomes more vacuolated and soon disappears. The nuclear cavity presents a more granular appearance, the granules staining more densely. The kinoplasmic fibers now penetrate the membrane of the nucleus and enter its cavity, while at the same time the polar radiations seem to diminish in number (Fig. i, C). On entering the cavity some of the fibers proceed in advance of the others. Some pass straight to- ward the center of the nucleus, while others diverge toward the sides. As these fibers approach from opposite sides of the nucleus, they tend to collect the chromosomes into an irregular mass in the equatorial region, where they finally form the nuclear plate (Fig. i, D). Cer- tain of these fibers coming from opposite sides seem to unite at their ends to form the continuous spindle fibers which extend from pole to pole ; others fasten themselves to the chromosomes, and still others diverge toward the nuclear membrane in the equatorial region (Fig. 2, E). In the mature spindle, therefore, the fibers present the following orientation : those radiating from the poles, the continuous spindle fibers extending uninterruptedly from pole to pole, those running from the poles to the chromosomes, and the fibers which diverge from the poles toward the equatorial region and end in the cytoplasm (Fig. 2, F) . The nuclear membrane in the tetraspore mother-cell of Dictyota disappears very gradually during the process of karyokinesis, often persisting at the sides when the spindle is mature (Fig. 2, F). It begins to disappear at the poles as soon as the fibers enter the nuclear cavity, and by the time the anaphase is reached no part of the membrane can be distinctly seen. Thus the spindle, with the exception of the polar radiations, lies within the nuclear cavity, its fibers, however, being largely of cytoplasmic origin. To what extent any nuclear substance contributes to the formation of the spindle is ditficult to determine. On the disappearance of the nucleolus, numerous granules appear in NUCLEAR DIVISION. 5 the nucleus, which stain deeply, closely resembling the chromatin granules. In the meantime the chromosomes increase in size, and it seems reasonable to suppose that the nucleolar substance contributes materially to their growth. The development of the nucleolus in the daughter nucleus and its behavior during the following, or second mitosis, seem to strengthen this theory. The chromosomes, when Fig. 2. — Spindle and telophase of first mitosis in the tetraspore mother-cell of Dictyota dichotoma. E, spindle nearly mature ; nuclear rafembrane has disappeared at poles. F, mature spindle ; the small lumpy chromosomes are regularly arranged in equatorial plate ; nuclear membrane persists at sides. G, dau^ter nuclei still connected by strand of connecting fibers ; at poles of each nucleus is a well- developed centrosphere. arranged in the equatorial plate, appear, especially when crowded to- gether — a phenomenon of frequent occurrence — as rounded lumps (Fig. 2, E, F). A careful study in favorable cases shows clearly that each chromosome is either in the shape of a ring, or so contracted as to leave scarcely any central space, such, for example, as occurs in some higher plants [Podophyllum^ Helleborus) . In such cases each 6 INTRODUCTION. segment or daughter chromosome forms one-half of the ring, or each maybe in the form of a short, thick U (Fig. 2, F). Sixteen chromosomes, the reduced number, are present in the first mitosis. While on the way to the poles the daughter chromosomes sometimes fuse with one another to form large masses.^ This is especially so in the second mitosis. In the construction of the daughter nuclei, one or more larger masses of chromatin are formed by the chromosomes ; a nucleolus appears near the chromatin mass or masses, and a nuclear membrane is laid down (Fig. 2, G). The membrane is unquestionably formed through the agency of the kinoplasmic fibers. The centrosomes increase in size, and the polar radiations are more distinct than in the spindle stage. The connecting fibers usually persist until the nuclear mem- brane is present, but a little later they disappear entirely. The chro- matin mass, gradually becoming less dense, soon disintegrates, and each daughter nucleus passes into the resting condition (Fig. 2, G). From the preceding it will be seen that each daughter nucleus is provided with one centrosome, but in the first mitosis the centrosomes could not be made out until they were on opposite sides of the nucleus and provided with radiations. The question naturally arises : Does the centrosome divide to give rise to the two daughter centrosomes .'' Swingle ('97)? who has traced the persistence of the centrosome through several successive generations of vegetative cells in Stypo- caulon^ one of the Phceophycece, found that a division of the centro- some takes place, and Strasburger ('97) arrives at the same conclusion as regards Fucus. This is the generally accepted view. We shall trace the early development of the spindle in the second mitosis in the tetraspore mother-cell in order to see what evidence is furnished by Dictyota toward the solution of this problem. During the reconstruction of the daughter nucleus (Fig. 3, H) two rod-shaped centrosomes, each with its radiations, were observed close together, and in such a position as to form a wide V, giving the impression that a longitudinal division of the single centrosome had taken place. The manner in which a cluster of radiations is attached to each daughter centrosome seems to lend weight to this conclusion. The daughter centrosomes now separate, moving along the nuclear membrane, but they do not, as in the first mitosis, traverse an angular distance of 180° before the formation of the spindle begins (Fig. 3, I, K) . The development of the spindle is the same as in the first mitosis, as Fig. 3, I, J, K, L, will clearly show. In other brown algse, so far as known (Swingle '97, Strasburger '97) , ' This massing of the chromosomes may not occur in all cases. NUCLEAR DIVISION. the development of the karyokinetic spindle in both vegetative and reproductive cells agrees essentially with that described for Dictyota. In the diatoms the development of the spindle as described by Lauterborn ('96) is singular and without parallel in the plant king- dom. According to this author, the spindle develops directly from the centrosome by a division of the same or by budding. We shall refer to this phenomenon beyond in the section dealing especially with the centrosome. In the red algae the development of the karyokinetic figure is known somewhat in detail only in Corallina officinalis. In this plant, Davis ('98) finds that the spindle arises through the agency Fig. 3. — Second mitosis in tetraspore mother-cell q{ Dictyota. H-K, prophase, showing origin of spindle. L, a nearly mature spindle. of centrospheres which undergo a great change in size during mitosis. The persistence of these bodies was not followed from one cell genera- tion to the next. The paucity of our knowledge of nuclear division in the red algae precludes any further mention of the subject in this group of plants. So far as is known to the author, no centrospheres or centrosomes have been authentically observed in the green algae. ERVSIPHE COMMUNIS, For the fungi, the most accurate and complete account of karyoki- nesis is to be found in the classical work of Harper ('97) on certain Ascomycetes. As an illustration of the process in this group of fungi, which is probably best known cytologically, a brief account of mitosis will be given as described by Harper in the ascus of Erysipke communis.. 8' INTRODUCTION. The ascus of this species offers unusually favorable material for the study of mitosis on account of the clearness with which all details are brought out, and because the three successive nuclear divisions follow each other rapidly, making it possible to trace with unmistakable clearness the persistence of the centrosome from one nuclear genera- tion to the other. Since the spindles lie in different planes, it is pos- sible also to observe, side by side, the same stages at different angles in the same field of the microscope. The following refers especially to the second mitosis in the ascus. W^V;?.-^' Fig. 4. — Mitosis in ascus oi Erysiphe communis. — (After Harper.) A, nucleus in resting stage of second nuclear generation in ascus, the flattened or disk-shaped centro- some closely applied to nuclear membrane. B, early prophase ; the kinoplasmic radiations have been developed about the centrosome. C, D, E, F, successive steps in development of spindle. G, mature spindle, the nuclear membrane still persists at sides. H, end of anaphase ; connecting fibers extend between the daughter nuclei, which are not yet provided with a nuclear membrane. I, daughter nucleus provided with membrane, kinoplasmic radiations present. J, later stage in which the polar radiations have disappeared. Between the successive nuclear divisions in the ascus, the chromatin of the daughter nuclei does not assume the complete resting condition. It consists (Fig. 4, A) of an irregular net with the angles of the meshes somewhat thickened. Generally the net lies tolerably free in the nuclear cavity, and a very distinct nucleolus is present. The centrosphere appears as a flattened disk closely applied to the nuclear membrane, giving the impression as if the two were grown together (Fig. 4, A). The chromatin net appears also attached at this place NUCLEAR DIVISION. 9 and frequently forms a dense mass. These phenomena indicate clearly that chromatin and centrosphere are in dii^ect communication through the nuclear membrane. The first step in the division is characterized by the appearance of a well-developed aster or system of radiations about the centrosome. It seems very probable here that the radiations grow out into the cytoplasm from the centrosome as a center. In the development of the radiations the nucleus probably cooperates. At this stage the chromatin is contracted into a dense net toward the centrosphere and appears in close connection with it. From the chromatin mass several fine achromatic threads extend toward the nuclear membrane (Fig. 4, B). In the next stage observed, the two poles of the spindle have been formed, which lie some distance apart on the nuclear membrane (Fig. 4, C). The polar radiations are well developed, and from each centrosome a cone of spindie fibers extends into the nuclear cavity. The diverging fibers seem to be inserted in the nuclear membrane at points opposite the centrosome. As in Dictyota the two systems of fibers cross each other at nearly righ't angles without in any way uniting. Whether the two centrospheres arose by a division of the primary centrosphere cannot be stated with absolute certainty, since the intermediate stages between B and C, Fig. 4, were not observed, yet from what is known in Stypocaulon and in Dictyota^ it seems reasonable to suppose that the centrosphere may undergo a division in Erysiphe also. The chromatin, at this stage, seems to be reduced in mass to that which will appear in the nuclear plate. It lies distributed in irregular lumps among the fibers opposite the two poles. The nucleolus has now disappeared, or, in some cases, it may remain in the form of a weakly staining residue. The spindle fibers within the nucleus be- come attached to the chromosomes and then contract strongly, bringing the chromosomes into the center of the nuclear cavity (Fig. 4, C, D, E, F). Some of the fibers of the bent spindle appear, at this stage, to extend uninten-uptedly from pole to pole. The continuous fibers are, in all probability, formed by the union of those which are not attached to the chromosomes. The polar radiations now undergo a marked change, becoming shorter and thicker, as if drawn in toward the poles. The majority of the radiations diverge only slightly. They are contracted into bundles or brush-like collections, which stand perpendicular to the surface of the nucleus. Some of these radiations, however, diverge somewhat from the central group, but all the polar radiations are not centered upon a single point. The pole of the spindle is exactly as broad as the base of lO INTRODUCTION. the central group of polar radiations, and, as will be seen from Fig. 4, E, F, G, the impression is that the polar radiations and the spindle contain the same number of fibers, which are continued uninterruptedly through the poles. But the continuity of the fibers is sharply inter- rupted by an achromatic plane at the nuclear membrane, through which the deeply staining (violet, by the Flemming triple stain) fibers pass from nucleus to cytoplasm. Whether the spindle fibers actually end at the nuclear membrane, or whether their substance only stains less densely there, was not determined. However, the phenomenon leaves the impression that the central body consists merely of the bases of the polar radiations closely crowded together. If the centrosome is an individual organ here, it seems that it must consist of a very thin, flat- tened disk, equal in breadth to the blunt end of the spindle. The poles of the spindle now separate farther from each other, whereby the spindle becomes straight. The individual chromosomes, eight in number, which are arranged in the equatorial plate, are sharply defined, and the nucleus has become somewhat elongated (Fig. 4, G). The polar radiations have again become fine elongated fibers, forming regular systems of sun-like radiations. As soon as the daughter chromosomes have reached the poles of the spindle the nuclear membrane disappears (Fig. 4, H). The fibers of the central spindle become now less sharply defined and broken in different places. Their number is also gradually diminished, their substance soon being indistinguishable from the immediately surround- ing cytoplasm. The polar radiations, however, form at this stage a more regular and sharply defined aster, owing to the outer rays bend- ing somewhat backward round the chromosomes (Fig. 4, H). The latter form a dense mass in which the individual elements are no longer to be distinguished. The centrosome is likewise not to be distinguished from the chromatin mass near which it lies. A nuclear membrane is now formed about each daughter nucleus, which appears as a small vesicle with the chromatin mass at the polar side (Fig. 4, I). With the further development of the nuclear membrane the free cavity of the nucleus increases in size. The chromatin mass begins to swell, and is gradually transformed into threads and lumps which are arranged, at first, mostly along the nuclear membrane, but soon become distributed through the nuclear cavity. A nucleolus now appeal's, and with the further growth of the nucleus the chromatin passes over into the netlike framework like that in Fig. 4, J, A. As soon as the nuclear membrane is formed, the polar radiations begin to disappear. In Erysiphe they seem to be transformed into a granular mass (Fig. 4, J). Finally, when the daughter nucleus is MITOSIS IN POLLEN MOTHKR-CELLS. II mature, the centrosphere remains as a much flattened disc closely applied to the nuclear membrane. From the foregoing it is clear that, although differing much in detail, the karyokinetic process in Erysiphe is, in general, similar to that in the brown algae. At our present state of knowledge, it is difficult to explain all the minor differences or to form an estimate of their relative importance. MITOSIS IN POLLEN MOTHER-CELLS. The spore mother-cells of certain Liliacece and other monocotyledo- nous species, as well as a few dicotyledonous plants such as Helleborus and Podophyllum^ have become classical objects for cytological study, and in these genera the mitotic process is now as well understood as in any other angiosperms. The following discussion of the first two nuclear divisions in the spore mother-cells of higher plants is based upon the author's own investigations made upon Lilium martagon^ L. candidum^ Fritillaria persica^ Tradescantia virginica, Helle- borus foetidus and Podophyllum peltatum. THE FIRST OR HETEROTYPIC MITOSIS. RESTING NUCLEUS AND DEVELOPMENT OF CHROMATIN SPIREM. Soon after the last nuclear division in the archesporium, or spore- bearing tissue, which gives rise to the pollen mother-cells, the latter begin that period of growth so characteristic of spore mother-cells pre- viously to the first mitosis. The nucleus is relatively large with a sharply defined membrane, and contains a fine linin network, in which the chromatin granules are held, and one or more nucleoli. The nucleolus may lie in a colorless, spherical cavity, which seems sharply circumscribed. The chromatin appears in larger and smaller granules, which are, as a rule, regularly distributed in the linin thread. The cytoplasm presents a uniform netlike structure (Fig. 5, A). This is the typical structure of a pollen mother-cell. With further growth of the nucleus, the chromatin granules increase in size, probably through the union or aggregation of the smaller granules, while at the same time the linin thread contracts and shortens. In this stage the linin net consists of a complicated spirem or thread with short turns. The chromatin granules have attained a more uni- form size, and lie more regularly distributed in the linin thread (Fig. 5, B). This contraction of the linin thread and fusion of the smaller chromatin granules continues, so that the nuclear thread, which later 12 INTRODUCTION. contains a row of larger granules or disks (the Chrovtatinscheiben of the German literature) of a tolerably uniform size, becomes a hollow spirem whose irregular turns traverse the nuclear cavity (Fig. 5, C). The chromatin disks have usually a jagged or erosed outline, which shows that each disk is composed of smaller granules. The chromatin disks, first carefully described by Strasburger ('82), vary much among themselves in size, and do not always have the same orientation in the linin thread. This fact, together with the twisting of the thread upon its axis, which is a mechanical necessity, gives the impression of a spirem composed of very irregular granules. This is especially notice- FiG. s. — Pollen mother-cell and early prophase of first or heterotypic mitosis. A, F, Podophyllum peltatum. B-E, Helleborus ftetidus . A, typical pollen mother-cell, with nucleus in resting stage, and while the cells are in tissue connection. B, linin net with numerous small chromatin granules. C, spirem in which chromatin disks are of uniform size. D, pieces of chromatin spirem more highly magnified; a, before longitudinal splitting; ^, after longi- tudinal splitting. E, the spirem has split longitudinally; daughter segments show a tendency to separate. F, the chromatin spirem has segmented transversely into chromosomes ; daughter segments twisted about each other. ( All figures represent sections.) able immediately after the longitudinal splitting of the chromatin granules. At this stage the most careful staining is necessary to bring out the chromatin disks clearly, since the linin retains the stain with greater avidity, thereby concealing the former. If the nuclear thread be too densely stained, it will appear more or less homogeneous, in which case the chromatin disks manifest themselves as a succession of enlargements whose granular character is concealed. The chro- matin thread consists, therefore, not of a succession of chromatin disks MITOSIS IN POLLEN MOTHER-CELLS. IJ but of a continuous linin thread in which are held the chromatin disks or granules. In an early stage the nuclear thread shows a marked tendency to con- tract into a ball or mass about the nucleolus. The contraction into a dense ball is regarded by some observers as a perfectly normal occur- rence, to which the name synapsis has been given. My own investiga- tions have convinced me that the contraction of the nuclear thread into a ball is in a large measure due to the reagents, and that synapsis has little or no significance. It indicates probably a very sensitive con- dition of the nuclear thread or net at the stage in which the contraction occurs. Soon after the nuclear net has developed into the spirem, as men- tioned, the chromatin and linin elements split longitudinally (Fig. 5, D, a, 3, E). The daughter spirems remain either closely applied to each other, or, as sometimes happens, they may separate for longer or shorter intervals. They are always twisted upon each other, and, as a consequence, the two parallel rows of disks are not easily seen, especially where the chromatin thread makes short turns. The twist- ing of the daughter spirems upon each other persists after the trans- verse segmentation of the spirem into chromosomes, and in very many cases it is still to be seen during metakinesis (Figs. 6, 7). Very frequently portions of the spirem which run parallel with each other are connected by very fine threads, and, in some cases, as in the pollen mother-cells of Podophyllum^ very delicate cytoplasmic threads seem to penetrate the nuclear membrane and fasten themselves to the chromatin spirem. At this stage also one or more nucleoli, of varying sizes and with a homogeneous or vacuolate structure, are pres- ent. The nuclear membrane, especially in Podophyllum^ does not present from now on the sharp contour of the resting nucleus. It seems to consist merely of a cytoplasmic boundary (Fig. 5, F), and as will be pointed out in a later paragraph, we may conclude that the nuclear membrane consists of an extremely delicate kinoplasmic network, whose meshes in the resting nucleus are so closely arranged that only a sharp line is seen when observed in optical section. As soon, however, as the meshes widen with the increase in size of the nucleus the nuclear membrane loses its sharp contour. It cannot be asserted with absolute certainty that the fine threads extending from the nuclear membrane to the chromatin thread penetrate the membrane and con- tinue into the cytoplasm, but in Podophyllum the evidence seems to be in favor of such a view. At any rate there seems to be an intimate connection maintained between chromatin and cytoplasm. As karyokinesis progresses, the chromatin thread contracts, becom- H INTRODUCTION. ing shorter and thicker, and frequently no trace of the longitudinal splitting can be recognized. There is thus formed the loose, hollow Fig. 6. — Prophase and early stages in development of spindle in^ heterotypic mitosis of pollen mother cell. A, B, Liliunt candidum. C, D, L. ■martagon. A, the kinoplasmic spindle fibers arranged radially about the nucleus, large nucleolus present, and thi chromosomes, each consisting of two rather thick segments twisted about each other, lie along th< nuclear membrane or scattered through nuclear cavity. B, same developmental stage as A ; here the kinoplasmic fibers are disposed partly radially and parti] in form of a weft lying in cytoplasm midway between nucleus and cell-wall. C, the spindle fibers are encroaching upon the nucleus, forming a weft about it; the nuclear membram as such has nearly disappeared ; it seems to have been converted into fibers. D, multipolar spindle complex, in which the chromosomes are irregularly distributed. spirem, which segments by transverse division into the chromo- somes. MITOSIS IN POLLEN MOTHER-CELLS. 1 5 We shall now leave the chromosomes for the present and pass to the development of the spindle. DEVELOPMENT OF THE SPINDLE. The development of the spindle in pollen mother-cells varies some- what in detail in different plants, but it can usually be referred to one type. In all cases, so far as known, it arises as a multipolar structure. As soon as the spirem is segmented into chromosomes, and some- times earlier, the kinoplasmic fibers make their appearance in the cyto- plasm. The arrangement of the kinoplasmic fibers is not quite the same in all cells of the same anther. They may be disposed at first radially about the nucleus (Fig. 6, A), or, as in many cases, may form a weft about the nucleus midway between nuclear membrane and cell- wall (Fig. 6, B). The remaining cytoplasm consists of a fibrillar structure. In this stage the nucleus is filled with a fluid which does not stain, namely, the nuclear sap. The chromosomes are connected with each other and with the nuclear membrane by means of fine fibers, and one or more nucleoli are present. The nucleolus, how- ever, begins to break up at this time, so that one large and several smaller ones may be present. The next step in the development of the spindle may differ slightly in different cells, owing to the orientation of the kinoplasmic fibers. In those cells in which these fibers are disposed radially about the nucleus, the tendency to form poles manifests itself before the disap- pearance of the nuclear membrane. Groups of radiations converge toward various points near the plasma membrane, while others form a weft about the nucleus (Fig. 6, C). A little later the nuclear mem- brane is replaced by this weft, and the fibers begin to enter the nuclear cavity. In some cases well-defined poles (or only a few) are not as yet present. In other cases a greater number of poles are formed, and we have then a very remarkable multipolar complex of kinoplasmic fibers surrounding the nucleus, into wrhich the fibers penetrate from all sides (Fig. 7, E). Gradually more kinoplasmic fibers enter the nuclear cavity until it can no longer be recognized as such (Fig. 6, D). In this complex of spindle fibers the chromosomes are irregularly distributed. They are, however, soon collected together, and to each a bundle of fibers be- comes attached. The chromosomes seem to be aggregated more closely together by a pushing and pulling of the spindle fibei-s. Owing to the irregular arrangement of the chromosomes and the complexity of .the mass of spindle fibers, it is not always possible to determine at this stage the exact manner in which the fibers are fastened to the chro- mosomes (Fig. 7, F). i6 INTRODUCTION. The bipolarity of the multipolar spindle now gradually manifests itself, and the multipolar structure rapidly becomes a typical bipolar spindle in which the chromosomes are arranged in the equatorial plate. Fig. 7. — Heterotypic mitosis in pollen mother-cell (L, fHartagan). Development of spindle continued. E, the weft of spindle fibers forms a multipolar complex. F, a multipolar complex in which bipolarity has begun to manifest itself; the weaker poles seem to be drawn in or together. G, bipolarity is established and chromosomes more regularly arranged in equator. H, mature spindle, showing only 3 of the 12 chromosomes; chromosomes fastened endwise to spindle. This transformation is probably brought about by certain of the larger poles converging toward a common area or point, while others are drawn in (Fig. 7, G). The mature spindle is either truncated at the poles (sometimes broadly so) or pointed, and the chromosomes are MITOSIS IN POLLEN MOTHER-CELLS. 1 7 quite regularly arranged in the equatorial plate. They are usually radially disposed, standing at right angles to the axis of the spindle (Fig. 7, H). The spindle fibers present the following arrangement: to each chromosome are attached two bundles of fibers (one to each daughter segment) which extend to the poles ; other fibers, the central spindle fibers, run uninterruptedly from pole to pole, and still others diverge from the poles toward the cell periphery. This arrangement is commonly found in all cells of the higher plants, whether they be reproductive or vegetative. The spindle does not, as may appear at the first glance, present a system of meridional fibers converging toward the poles, but, as is easily seen from thin sections, the fibers cross and anastomose, giving the impression that the spindle consists of a weft or complex of fibers drawn out in the direction of the poles, which, indeed, it really is. In spore mother-cells of plants, the spindle fibers seem to be gener- ally of cytoplasmic origin, i. c, they appear first in the cytoplasm, forming a weft about the nucleus or radiating from it. In the generative cell of gymnosperms and in the first division following fecundation in these plants, it seems that the fibers or many of them arise from kinoplasm, which is in the nucleus or which entered the same in another form. CHROMOSOMES. As is well known, the chromatin spirem, which has split longitudi- nally in the early prophase, segments by transverse division into twelve chromosomes, the reduced number, or half the number in the vegeta- tive cells of the sporophyte. Each chromosome consists, therefore, of two daughter segments, or daughter chromosomes, which are almost always twisted upon each other (Fig. 7, H ; Fig. 8). After the segmentation of the spirem into chromosomes, these contract, thereby becoming shorter and thicker. Previous to the disappear- ance of the nuclear membrane, they lie near it or are scattered throughout the nuclear cavity (Fig. 6, B). In Lilium^ the daughter chromosomes are, as a rule, closely applied to each other, but in many cases they tend to become separated soon after segmentation, so that various forms of chromosomes result, such as rings, loops, X- and V-shaped forms, depending upon the manner in which the daughter segments are oriented toward each other (Fig. 8, A to K). These various forms persist and may be found in the nuclear plate of the mature spindle. The following will explain the manner in which the more fre- quently occurring forms are brought about in Lilium^ Podophyllum and in many other higher plants : i8 INTRODUCTION. The daughter segments often diverge at one or at both ends (Fig. 8, B, C). In other cases they may be bent and in contact only near the middle (Fig. 8, D). If the daughter segments adhere at the ends, and bend away from each other near the middle, a ring results (Fig. 8, E). Ring-shaped chromosomes may be so bent as to bring the opposite ends near each other, in which case we have a ring partly folded upon itself. This is true in a measure in Fig. 8, E. When the segments forming a ring separate slightly at one end, an open ring is produced. A Y-shaped chromosome will result when the segments are con- tiguous for a part of their length but diverge at one end (Fig. 8, F). Sometimes the daughter segments adhere near the middle but diverge H ■ I J K Fig. 8. — Heterotypic mitosis {Liltum martagon). Diflferent forms of chromosomes. A, B, C, D, chromosomes from prophase. E-K, from equatorial place. E, ring-shaped, F, Y-shaped, and J, typical X-shaped chromosomes. G, H, 1, and K, other forms commonly met with in Lilium. at both ends, so that they may be crossed ; this gives rise to the X- shaped chromosome (Fig. 8, J). Instances are also met with in which the segments of the X-shaped chromosome fuse completely at one end, and the chromosome appears as a continuous rod, folded in such a man- ner that the opposite ends are brought together. In this way loops and incomplete rings are produced (Fig. 8, K). In Fig, 8, G, H, and I are forms of chromosomes that are of frequent occurrence. The orien- tation of the daughter segments toward each other, which results in the different forms of chromosomes described, is, in all probability, of no special importance, since two or more of these forms may be seen in the same nucleus. In Tradescantia^ between the time of the segmentation of the spirem into chromosomes and the mature spindle, the daughter segments often contract into the form of short, thick crescents. These may adhere at MITOSIS IN POLLEN MOTHER-CELLS. I9 the points of the crescents to form ring-like chromosomes (Fig. 9, D, at the right). In the majority of cases, however, they adhere at only one end, and under such circumstances each chromosome consists of two thick and slightly curved pieces placed end to end, and as they are oriented tangentially upon the spindle, reach nearly from pole to pole (Fig. 9, D). The chromosomes in Podophyllum present the same variety of forms found in Lilium and Tradescantia. Here the segments may be in close contact, side by side, or form loops, rings, X's, and Y's. Per- haps the majority of chromosomes in Podophyllum present the form last mentioned for Tradescantia. In Lilium the chromosomes, vs^hen in the nuclear plate, are usually arranged with much regularity about the periphery of the spindle. The majority are fastened to the fibers at the ends, and stand radially to the axis of the spindle (Fig. 7, H). When observed from the pole in this stage, they are seen to radiate like the spokes of a wheel from the central spindle fibers. But all the chromosomes are not so regu- larly oriented upon the spindle, and their manner of attachment to the fibers is also variable. As will be seen in Fig. 8, F-K, they may be fastened to the spindle at some distance from one end or near the mid- dle. Those that are quite regularly ring-shaped are attached near the middle of each segment. In all these cases, the chromosomes are placed tangentially upon the spindle. The X-, Y-, and loop-shaped chromosomes are usually fastened to the spindle as indicated in Fig. 8, F, J, K. Karyokinetic figures are not rare in which two or more of the different forms of chromosomes, with their different orientations and different methods of attachment to the fibers, are found in the same spindle.' The stage of the mature spindle persists some time and evidently • Other interpretations of the chromosomes appearing in the first mitosis have been given by different observers and by the same investigator at different times, owing to the trend of theoretical considerations. One of these, which was announced as early as 1884 by Heuser for Tradescantia virginica (Beobach- tung uber Zellkerntheilung. Bot. Centralblt., 17 : 1884) and which has very recently received support by Strasburger and others (Ueber Reduktionstheilung. Sitzbr. der Konig. Preuss. Akad. der Wiss., 18 : t-28, 1904) is that the two segments of each chromosome appearing in the equatorial plate of the first mitosis are not the result of the longitudinal splitting of the spirem occurring in the early prophase, but are formed by the folding together or approximation of two chromosomes, each consisting of the two daughter segments resulting from the longitudinal splitting. Each chromosome is therefore a bivalent chromosome, and the first or heterotypic mitosis is a qualitative cr reducing division, whereas the second mitosis is equational, the segments separating along the line of the longitudinal split. Strasburger bases his conclusion mainly upon data obtained from studies of the pollen mother-cells of Galtonia candicaus. The figures which he gives in support of this view in the paper cited seem to me to be far from convinc- ing. Moreover, Jules Berghs, in a recent study of the prophase of the heterotypic mitosis in Allium fistulosunt and Lilium lanci/olium (j/tfc/Vjjxwi) (La Cellule, 21: 173-188, 1904), shows clearly, in a careful series of stages, that the two segments of each chromosome are the result of the longitudinal fission and not that of a folding together or approximation of two chromosomes. Unfortunately the papers cited reach me too late for further consideration, as these pages are already in press. 20 INTRODUCTION. represents a slight pause in the process of mitosis. For this reason it is the stage most easily obtained and most frequently observed. METAKINESIS. Up to the stage of the mature spindle, as in Fig. 7, H, each chromosome is seen to consist of two daughter segments oriented in one of the ways described above. As soon, however, as these seg- ments begin to separate in metakinesis, each splits longitudinally in a plane at right angles to the longitudinal splitting which took place in the prophase. In some instances, and when the chromosomes are viewed from the end, each is seen to be composed of four rods, the four granddaughter segments, placed side by side in pairs, forming a tetrad. Fig. 9, A. As a rule the granddaughter segments cannot be definitely recognized until the daughter segments have separated somewhat. Having almost or quite separated, the daughter segments are seen to be in the form of a V, although it never should be for- gotten that V's do not invariably result. As the result of the second longitudinal splitting, each typical V-shaped daughter chromosome consists of two granddaughter segments which adhere or are even fused at the ends to which the spindle fibers are fastened, while the opposite ends diverge (Fig. 9, B). It frequently happens that the opposite ends of the granddaughter segments do not diverge, but lie more or less in contact side by side, so that the retreating daughter chromosomes consist of two applied rods (Fig. 9, F, the middle pairs) . In some cases, as already mentioned, the ends of the granddaughter segments forming the angle of the V fuse, so that the V appears to be one piece formed by bending. The bent or contorted condition of the granddaughter segments during metakinesis is due to the previous twisting of the daughter chromosomes upon each other. If the chromosomes be in the form of rings, as shown in Fig. 8, E, it is evident that the separating daughter chromosomes may also be in the form of a V or U, but such V's and U's will be produced by a bending of the daughter segments. This is true in a great many cases in Lilium and in other plants, among both monocotyledonous and dicotyledonous species. In such cases each U or V is invariably double, as the result of the second longitudinal fission — that is, the granddaughter segments are U-shaped and closely applied to each other (Fig. 9, F, right and left). Sometimes these granddaughter seg- ments may separate slightly, giving the impression of two similar daughter chromosomes lying one just beneath the other. This is one of the several phenomena that have led to erroneous interpretations of the chromosomes. MITOSIS IN POLLEN MOTHER-CELLS. 21 In Fig. 9, C, on the left, is shown a chromosome in metakinesis, which is fastened to the spindle near the middle. Each daughter seg- ment, which is split longitudinally, is in the form of a U-like figure, Fig. 9. — Heterotypic mitosis. Meta- and anaphases. A, B, C, and F, Lilium. D, Iradtscantia. E, Podophyllum. A, metakinesis beginning ; viewed from the end, each chromosome is seen to consist of four rods, due to the second longitudinal splitting, which has taken place at right angles to the first. B, metakinesis accomplished ; ends of granddaughter chromosomes, which are directed toward equator, __ diverge, giving rise to the well-known V-shaped elements; in B all chromosomes arc fastened to spindle fibers at the ends. C, chromosome on left was in form of an incomplete ring ; segments fastened at place of bending ; in this case the U- or V-shaped elements owe their form to a bending ; the chromosome on the right was attached endwise. D, mature spindle of Tradescantia. E, F, anaphase ; the retreating pairs of granddaughter segments are rods hooked at one end, or U's. in which one limb seems a little longer than the other. This chromo- some may originally have been a complete ring, as in Fig. 8, E, in which the segments had separated at one end in advance of the other, ii INTRODUCTION. or it may have had this foinii at an earlier stage. The chromosome at the right in this figure (Fig. 9, C), was attached to the spindle end- wise, and the retreating granddaughter segments will probably form Vs. If the chromosome on the left were rotated 45°, so that the seg- ments would be seen in profile, we might have the picture of two double V's or U's about to separate, for, as shown in the figure, the free ends of the pairs of granddaughter elements tend sometimes to diverge. The two chromosomes in this figure, which belong to the same spindle, show clearly how figures of the same shape may be pro- duced in different ways. In the one on the right the chromosome was probably attached to the spindle by the end, and the V's are formed by the divergence of the free ends, while that on the left was fastened near the middle of each segment, and the V- or U -shape of the retreating segments is the result of a bending. In such chromosomes as Fig. 8, G, H, I, the retreating elements may retain their present form, or they may be bent during metakinesis into U's or V's. When the daughter segments of such chromosomes are separated, they must untwist, and it is reasonable to suppose that the force necessary to separate them when twisted will be sufficient to bend the segments into a U- or V-like figure. THE ANAPHASE. The pairs of granddaughter segments, as they pass toward the poles, are in the form of contiguous, straight, or undulating rods, V's or U's, or, in case one limb of the last two named figures be much longer than the other, as is sometimes observed, the retreating elements will be in the form of hooks. Even in those cases in which both grand- daughter segments are nearly straight or undulating rods of equal length, each is often slightly bent or hooked at the end fastened to the spindle fibers, or the segments may be bent at both ends. The daughter chromosomes in Podophyllu?n and Tradescantia show with great clearness their double character during the anaphase (Fig. 9, E). The granddaughter segments generally lie close side by side, although cases in which they are slightly separated are now and then to be observed. There are in these genera also variations in the forms of the chromosomes which may be explained in the same man- ner as in Lilium. The retreating chi'omosomes and the structure of the spindle suggest that the segments are conveyed to the poles by a pushing and pulling action of the spindle fibers. MITOSIS IN POLLEN MOTHER-CELLS. 2^ THE TELOPHASE. As soon as the daughter chromosomes arrive at the poles, they approach each other very closely, so that, in many cases, the separate individuals cannot be recognized. But very frequently the segments do not become so closely crowded together, and the manner in which the daughter spirem is formed can be followed with accuracy. The formation of the spirem can best be observed when the granddaughter segments arrive at the poles in the form of the familiar V-shaped figures. Generally the ends forming the angles of the V fuse first, unless this has already been accomplished ; then the free ends meet end to end and unite (Fig. lO, G). In this way there is formed a continuous single spirem in which the identity of the individual segments or granddaughter chromosomes is lost. If all the daughter chromosomes were regularly V- or U-shaped the spirem would be regular, consisting of an orderly series of nearly uniform turns ; but the spirem rarely shows such regularity, because the chromosomes vary in size and shape and in the manner in which the granddaughter segments are oriented with respect to each other in the several pairs. During the reconstruction of the daughter nucleus, the chromosomes tend to reticulate, that is, to become irregular and lumpy, so that an irregular skein or net results. This is less pronounced in Liliuin than in many other plants. The fact that pairs of granddaughter segments arrive at the poles in diffei*ent forms, such as V's, double U's, and pairs of parallel rods, shows clearly that in such cases the resulting spirem must be very irreg- ular. The chromosomes are generally so closely crowded together that it is not possible to determine with certainty just how the variously shaped pairs of segments behave. But it is reasonable to suppose that the segments of the double U's and those of contiguous rods must first separate in order to unite end to end, for no case has been clearly made out in Lilium in which a part of the spirem is formed double. The newly formed daughter spirem is close with relatively short turns (Fig. lo, G, H). Between each twa extends the beautiful system of connecting fibers, which represents the central fibers of the spindle. Fibers are also present which extend from each spirem toward the plasma membrane in the direction of the equator. Some of these reach the plasma membrane, while others seem to end blindly in the cytoplasm, or pass over into its thread-work. In Lilhi?n there are no polar radiations. The system of connecting fibers soon becomes barrel-shaped, and the cell-plate makes its appearance in the equatorial region. We shall return to the formation of the cell-plate beyond. 24 INTRODUCTION. The nuclear membranes are not formed about the daughter nuclei In Lilium mart agon until after the division of the cell, at least in many instances. Soon after the division of the cell, however, the nuclear membranes are laid down. In ail plants examined, each appears first Fig. io. — Telophase and daughter nucleus of heterotypic mitosis (^Lilium tnartagon). G, daughter spirem formed by union of granddaughter segments end to end ; each daughter spirem is in the form of a disk from whose edges kinoplasmic fibers extend out in direction of cell-wall ; system of connecting fibers slightly bulged out at middle. H, the cell-plate appears in center of system of connecting fibers. I, J, cell-division is completed, but the daughter nuclei are not yet provided with membranes. K, a daughter nucleus at a later stage with nuclear membrane; chromatin spirem continuous, the free ends having been made by knife in sectioning. as a weft of kinoplasmic fibers, which are undoubtedly dei'ived fi-om the spindle. It is interesting to note that in Lilium and Podophyllum the nuclear membrane appears in the same form in which it disap- peared during the formation of the spindle. The fact that the nuclear membrane arises first as a weft of kinoplasmic fibers is a strong proof that it is of a kinoplasmic nature. MITOSIS IN POLLEN MOTHER-CELLS. 25 The young weft-like nuclear membrane encloses a cavity containing the chromatin and little or no other staining material. With further development the kinoplasmic weft is transformed into the typical nuclear membrane, appearing in section as a sharp line, and the daughter spirem becomes loose and open, [n the mature daughter nucleus the spirem is continuous and of a tolerably uniform thickness. In some cases it is rather regular, consisting of long turns arranged in the form of a wreath (Fig. lo, K), but in the majority of instances the spirem is irregular, with long and short turns so disposed that its course cannot be easily followed. This condition of the spirem is in all .probability due to the variously shaped chromosomes mentioned in a preceding paragraph. THE NUCLEOLUS. In the resting nucleus and during the prophase, one or more nucle- oli are present. These nucleoli take on a deep red or reddish purple color with the Flemming triple stain. They sometimes present a uni- form structure, but, as a rule, the larger nucleoli especially reveal one or more vacuoles. As has been mentioned in a preceding paragraph, the nucleolus very frequently lies within a spherical space which appears in optical section as a colorless court about it. This phe- nomenon is especially striking in vegetative cells of higher plants, such as in root tips of Vicia faba and Zea mays. Experiments seem to show that the colorless space surrounding the nucleolus contains something more than a mere watery fluid which is extracted in dehydration. By subjecting roots of Vtcta, Zea and others to a strong centrifugal force, the author (Mottier, '99) found that the nucleolus together with its surrounding colorless court was thrown out of the nucleus into the cytoplasm. The expelled nucleolus was still surrounded by its colorless court — a fact that seems to show that the colorless substance has a specific gravity much greater than other constituents of the nucleolus, and that it may be provided with its own membrane. This colorless substance may represent unorganized nucleolar matter. Frequently before the nuclear membrane disappears a disorganiza- tion begins by which the nucleolus is broken up into several smaller nucleoli (Fig. 6, C). As the nuclear membrane fades away, and the kinoplasmic fibers enter the nuclear cavity, numerous bodies are found distributed in the cytoplasm which stain exactly as nucleoli, and there is no doubt that these bodies represent nucleolar substance. These extra-nuclear nucleoli were found to be more abundant in Lilium viartago7t. In Lilium candidum there may be none, or only a few 36 INTRODUCTION. small ones, at corresponding stages of mitosis. The piesence or ab- sence of extra-nuclear nucleoli may not depend so much upon the plant, perhaps, as upon the condition or activity of the cell. From the spindle stage of the first to the end of the second division there is no noticeable regularity in the behavior of these bodies. In different cells in the same stage of mitosis they may be present or wholly want- ing. Even after the daughter nuclei are provided with membranes, and a nucleolus is present in each, extra-nuclear nucleoli are to be fre- quently seen in the cytoplasm. The same holds also for the second mitosis. A careful investigation of the behavior of the nucleolus in both Thallophyta and higher plants has shown that the nucleolus appearing in the daugliter nucleus is not one of the extra-nuclear nucleoli which happened to lie near the chromatin, or in such a posi- tion as to be included by the nuclear membrane, but that the nucleolus arises anew in each daughter nucleus. The nucleolus appearing in the daughter nucleus arises usually near or in contact with the chro- matin thread, but it is not implied that the nucleolus represents reserve chromatin. In the higher plants and in those with typical nuclei the morpho- logical evidence furnished by a study of karyokinesis, as well as the evidence of experimental physiology, goes to show that the nucleolus in such plant cells lepresents so much food material which can be drawn upon by the cell according to its needs. Whenever the activity of the cell is more intense, the nucleolar substance tends to become diminished, and it matters not whether the activity is directed toward constructive work or the production of energy. It is true that in some cases the food material furnished by the nucleolus seems to be used in a large measure by the chromatin, for example, in Dictyota^ but in others by other parts of the living substance, as in the growth of the spindle or cell plate. In certain species of Spirogyra (Wisselingh, '98), in which, as it has been claimed by several investigators, the nucleolus furnishes directly one or more chromosomes, greater diffi- culties present themselves. It is not improbable that the nucleolus of such plants as Spirogyra may possess a totally different composition from that of the typical nucleolus, and we may, therefore, speak with propriety of chromatin nucleoli. However the behavior of the nucleolus is not well enough known in the plant kingdom to justify any attempt to harmonize all the facts now known. Applied to the higher plants the above conclusion seems to be very reasonable, since the facts there are almost wholly confirmatory. MITOSIS IN POLLEN MOTHKR-CKLLS. 2*J THE SECOND OR HOMOTYPIC MITOSIS. In the pollen mother-cell of Lilium^ the daughter nucleus does not pass into the complete resting stage, although in some cases the chromatin tends to become reticulated. In the homologous division in the embryo-sac, the daughter nucleus, on the contrary, passes into a structure which approaches closely that of the resting condition. In Tradescantia the chromatin of the daughter nucleus reticulates more than in Lilium while in certain dicotyledonous species, e. g.^ Lirio- dendren and Magnolia (Andrews, 'oi), a complete resting condition is reached. The spindle in Lilium and in all other plants investigated by the author arises also as a multipolar complex of fibers. The develop- ment of the multipolar structure and its transformation into the typical bipolar spindle differ in no essential from that already described for the first mitosis. In Lilium^ it is very evident that the spirem does not segment completely into chromosomes before the disappearance of the nuclear membrane. The spirem does not split longitudinally in this division, since that part of the process was accomplished in the preceding mitosis, but during the transformation of the multipolar into the bipolar spindle the chromatin skein segments into the chromosomes, which are arranged in pairs in the nuclear plate. Within the complex of spindle fibers, the spirem, or pieces of it, provided it has partly segmented, are somewhat crowded together. The various turns are greatly entangled, kinked and knotted, so that the segments cannot be accurately traced out. In only the most favorable cases at this stage can a few segments or parts of the spirem be followed definitely throughout their entire length (Fig. 1 1 , A). The kinked and entangled condition of the skein or its segments is due doubtless to the irregularity of the spirem, for were the turns all of a uniform shape and size a less complicated arrangement would result. The appearance of the chromatin during the development of the spindle suggests that the chromosomes were brought to a more regular arrange- ment in the nuclear plate by a pushing and pulling of the fibers. Judging fi-om the form of certain chromosomes which stand out by themselves, and which can be traced throughout their entire length during the development of the spindle or in the nuclear plate, it seems that the spirem, or a part of it at least, segments into pieces compris- ing the two segments of a chromosome, i. e., the two granddaughter chromosomes of the first division, and that these pieces may correspond to long turns or loops of the spirem (Fig. 1 1, B, C). These loops are 28 INTRODUCTION. fastened to the fibers either at the free ends or at the place of bending. Now, in order that the two segments of such a chromosome may come into contact side by side, as is frequently the case, the parallel parts of the loop need only be brought closely together. This may happen Fig. II. — Second, or homotypic mitosis, in pollen mother-cells {Liliunt). A, multipolar stage of spindle ; chromatin spirem not completely segmented into chromosomes, B, bipolarity established; chromosomes more regularly arranged. C, mature spindle ; chromosomes more regularly disposed in equator, placed radially or tangentially on spindle. D, metakinesis. E, the anaphase. before the spirem is completely segmented, for in many chromosomes the two segments are very closely applied and twisted upon each other before the spindle is mature. In like manner parallel portions of the spirem may come in contact either before or after segmentation. In many other cases, both in the mature spindle and during its develop- MITOSIS IN POLLEN MOTHER-CELLS. 29 ment, the two segments are separated from each other, being in con- tact only at the ends which are attached to the spindle fibers. Under this circumstance one segment may lie tangentially on one side of the equator and the other on the other. Otlier instances are observed also in which the two segments may lie parallel in pairs, but not in contact when arranged in the nuclear plate or at an earlier stage. Such cases as the two last mentioned would seem to indicate that the spirem, or a part of it, is segmented into the granddaughter chromosomes, and that these are then brought together in pairs. It is also probable that pieces of the segmented spirem, which are nearly straight, or only a little curved, may consist of two granddaughter segments, and these are brought side by side by the folding of the piece at or near the middle, so that the free ends are brought into apposition, after which the piece is severed at the point of bending. From a careful study of the second mitosis in the pollen mother-cells of Lilium^ Podophyllum^ Tradescantia and others, the author is inclined to believe that the spirem may segment in the different ways just mentioned. However, the daughter spirem segments transversely into the granddaughter chromosomes, and during the development of the spindle these are arranged more or less in pairs in the nuclear plate (Fig. ii, C). In the nuclear plate, the chromosomes are oriented either radially, obliquely, or tangentially to the major axis of the spindle. The segments may be straight or variously bent, and, in either case, fre- quently twisted upon each other. In Lilium^ the segments are frequently, perhaps in the majority of cases, variously twisted, kinked or knotted, so that they can be followed for only a part of their length. In many cases, the kinked and twisted chromosomes seem to be so contracted as to form lumps. This is true also in Trade- scantia and in numerous other plants. The bent, kinked, and twisted condition of the chromosomes seems to be due to the irregu- larity of the spirem, for it seems probable that, were all the turns of the chromatin skein regular and uniform, the greatly entangled nature of the spirem would not appear during the development of the spindle. We have seen that the identity of the individual chromosomes is lost from observation in the daughter spirem, and the question bear- ing upon the theory of the individuality of the chromosomes, naturally arises as to whether the chromosomes of the second, or homotypic mitosis, are identical with the pairs of granddaughter segments of the anaphase of the preceding, or heterotypic division. In other words, are the two segments of each chromosome, appearing in the nuclear plate of the second nuclear division, sisters ? Or may it be possible that some are sisters, while others are composed of segments from different pairs of granddaughter chromosomes of the first division ? 30 INTRODUCTION. It is generally conceded that the segments of each chromosome are sisters, and it is conceivable that, no matter in what manner or when the daughter spirem may segment during division, the spindle fibers, or those parts of the cell which have to do with the arrangement of the chromosomes in the nuclear plate, are able to bring the sister seg- ments together in pairs. Strasburger, Guignard, and others regard each long loop or turn of the daughter spirem as representing a V or U of the preceding mitosis, and that, consequently, the spirem segments exactly as it was con- structed, i. e., the chromosomes simply separate at the points marking the free ends of the V's and U's. The spirem accordingly breaks up into pieces equal to the length of two segments or two granddaughter chromosomes. It is claimed by Strasburger (1900, pp. 23, 24) that these V's or U's are fastened to the spindle in the same manner as in the first division, namely, at the angles or at the place of bending. Theoretically, there may be little objection to this view. The vast majority of facts, however, show that there is no such regularity in the shape of the chromosomes, or in their manner of attachment to the spindle. We have seen that, in the daughter nucleus, the identity of the individual chromosomes cannot be recognized, and we do not know whether the spirem segments in the same manner in which it was constructed. But if the spirem should segment by transverse division at the points marking the angles of the V-shaped chromosomes instead of at the free ends, then it is clear that the two segments of each chromosome would not be sisters. The result might be that two or more sister chromo- somes would go to the same daughter nucleus, a condition that might furnish a basis for greater vai'iation. We cannot prove either propo- sition, and the author is not disposed to enter into any speculation here upon the subject. The observed facts are these : The identity of the individual chromosomes is lost in the daughter nucleus, and we do not know whether the segments of the respective chromosomes appearing in the nuclear plate of the second mitosis are sisters or not. There is also no basis in fact for the conclusion that one chromosome is heredi- tarily different from another. The first two nuclear divisions in the embryo-sac mother-cell, so far as is known, are quite similar and homologous to those in the pollen mother-cell. In Lilium marlagon^ the species more carefully investi- gated by the author, there is no important difference in the behavior of the chromosomes. It may be mentioned, moreover, that the daughter nuclei resulting from the first mitosis approach more closely the resting condition than in the pollen mother-cell. CELL-DIVISION. 3 1 The question now remains whether in all micro- and macro-spore mother-cells of the higher plants a double longitudinal splitting of the chromatin takes place during the first mitosis and how prevalent such a phenomenon is in both plants and animals. In those plants in which the daughter nucleus passes into the struc- ture of the complete resting stage, it is certainly difficult to understand the significance of the double longitudinal splitting of the chromosomes in the first division. CELL-DIVISION. THE TYPE OF THE HIGHER PLANTS. Modern research has established the very important fact that new cells are formed from uninucleate or multinucleate mother-cells accord- ing to different methods, depending largely upon the manner in which the new plasma membranes differentiating the cells are formed. (I.) Among the higher plants, and some Thallophyta as well, in which cell-division is generally intimately associated with nuclear division, the new plasma membrane or membranes are laid down through the instrumentality of kinoplasmic connecting fibers, extending between the nuclei concerned. (2.) In the ascus of certain Ascomycetes^ where the new cells (spores) are carved out of a common nucleated mass of cytoplasm or mother-cell, the plasma membrane is also formed by kinoplasmic fibers, but these are polar radiations and not connecting fibers. The entire plasma membrane of such cells is new, that of the mother-cell taking no part in the process. This is typical and real free cell- formation. (3.) Another form of cell-division is found among the Myxomycetes and certain Phy corny cetes^ in which the new plasma membranes arise by a process of progressive cleavage, beginning at the surface, with or without any connection with, or aid of, vacuoles. Kinoplasmic con- necting fibers or radiations are in no way connected with this process. This type we may know as cell-cleavage. It resembles the cleavage of animal cells more closely than do the other processes of cell-formation in plants. (4.) There is yet another method of cell-formation typified by Dictyota and Stypocaulon among the brown algae, in which the new plasma membrane seems to be a direct transformation of the meshes or threadwork of the cytoplasm. It is not a cleavage like the last mentioned, nor are any connecting fibers present to take part in the 32 INTRODUCTION. ' formation of the cell-plate. This method is, however, closely related to cleavage. As an illustration of the method of cell-plate formation typical of higher plants, the pollen mother-cells of Lilium furnish excellent material. Here a cell-division follows the first nuclear division. The connecting fibers are well developed, and with suitable fixing and staining the details stand otlt with a clearness unequaled among plants. As we have seen in Fig. lo, G, the daughter spirems are connected by a beautiful system of connecting fibers, which is slightly barrel-shaped at an early stage. The fibers soon show a thickening in the equatorial region, which stains more intensely with gentian violet. The thicken- ings are not granular or lumpy, but rather homogeneous, and are due to the accumulation of kinoplasm, the substance out of which the cell-plate, or plasma membrane, is made. At a little later stage (Fig. ID, H) there appears in the central part of the system of con- necting fibers in the region of the equator a fine homogeneous line, the beginning of the cell-plate. This young cell-plate is evidently in the form of a circular disk, which proceeds in growth uniformly toward the periphery of the cell. The cell-plate is not necessarily formed by the meeting or union of thickened places of the connecting fibers, for in many cases the fibers are too far apart. The kinoplasmic material is brought to the place occupied by the new plasma membrane and there deposited in the form of a fluid substance. With the further growth of the cell-plate the connecting fibers bulge out more and more, being always thicker and more numerous at the outer edge or surface of the system (Fig. lo, H). As the peripheral fibers of the barrel-shaped system bulge out, its longitudinal axis becomes shorter, so that the daughter spirems come eventually to lie in the center of the daughter cells. In Fig. lo, I, the cell-plate is just complete, the peripheral fibers which have reached the plasma membrane of the cell being more numerous there. The cell-plate or plasma membrane is now seen to be double, and it is the author's opinion that the new plasma membrane is formed double. The fact that each daughter or granddaughter cell, when somewhat shrunken at this stage, is seen to possess its own plasma membrane, seems to support this view. Soon after the formation of the plasma membranes, a cell-wall is deposited between them. Until the primordia of the daughter nuclei (Fig. lo, J) are provided with a nuclear membrane, the chromatin spirem is in the form of a circular disk from whose margin radiates a zone of kinoplasmic fibers toward the equatorial edge of the cell. In optical section this zone appears as a bundle of fibers on the right and CELL-DIVISION. 33 left, whose elements diverge, meeting the concave plasma membrane at different points. Other delicate fibers extend from the spirem in all directions toward the plasma membrane. As soon as the nuclear membrane appears these radiating fibers become more uniformly dis- tributed about the nucleus. They undoubtedly take part in the forma- tion of the spindle in the division of the daughter nucleus. FREE CELL-FORMATION. The most beautiful and best known illustration of typical free cell- formation is found in the development of the spores in the ascus of certain Ascomycetes as described by Harper. The delimination of the spores from the cytoplasm in Erysiphe fol- lows immediately after the close of the last of the three successive nuclear divisions which furnish the eight nuclei for the spores. The entire process is accomplished by those kinoplasmic fibers which constitute the polar radiations of the last nuclear division and in a manner quite peculiar to asci. All of the eight nuclei pass through the anaphase at the same time, and, when in the resting condition, cannot be distinguished one from the other, with the exception of those that He close to the wall. The polar radiations persist in connection with those nuclei that form spores, while from those which do not the radiations disappear entirely. The chromatin lies mostly free in the nuclear cavity, but it is always in communication with the nuclear membrane, especially near the centrosphere (Fig. 12, A). As the first indication of cell-formation, the nucleus becomes pointed and develops a beak-like prolongation on the side next to the pole or centrosphere. This point or beak gradually elongates, so that the centrosphere becomes farther removed from the body of the nucleus (Fig. 12, B). As soon as the beak reaches a length which exceeds slightly the diameter of the nucleus, its growth ceases. This beak consists not of a single fiber or thread but of a slender cylindrical tube arising abruptly from a rather broad base. Into the tube there extends quite to the centrosphere a continuation of the chromatin net, by which the latter remains in communication with the centrosphere. In the base of the beak the nuclear network is loose and more open, while in the slender part it is drawn out into a single and twisted thread. As soon as the beak has reached its definitive length the kinoplasmic radiations undergo a remarkable change. The radiations which have a direction similar to that of the beak begin now to bend or grow backward, with the centrosome as a center, toward the nucleus, so that «4 INTRODUCTION. the aster is converted into a hollow cone whose apex is the centro- sphere. Neighboring radiations unite and grow rapidly in length, at the same time bending back toward the nucleus in a manner resem- bling the spray from a fountain. An optical section of this stage is shown in Fig. 12, C. With further growth the kinoplasmic rays give rise to a sort of bell-shaped or half-ellipsoidal structure whose center is occupied by the nucleus and whose pole is formed by the centro- some (Fig. 12, D). Near the centrosome the fibers have already formed a continuous but extremely thin layer, the plasma membrane, separat- ing the cytoplasm of the spore from that of the ascus. At the edge of Fig. 12.— Free cell-formation in ascus oi Erysiphe communis. A, nucleus with centrosphere. B, development of nuclear beak. C, polar radiations extend outward and backward as spray from a fountain. D, formation of plasma membrane from end of beak outward, and continued growth of kinoplasmic fibers backward. E, F, meeting of fibers at opposite end of ellipsoidal spore and establishment of a complete plasma membrane delimiting spore-plasma from remaining plasma of ascus. — (After Harper.) the bell the radiations end as free fibers, continuing their growth, how- ever, in a direction corresponding to the periphery of the ellipsoid (Fig. 12, E). Finally these fibers meet in a point which is directly opposite the centrosome, and unite end to end and laterally. The for- mation of the plasma membrane continues, so that eventually an ellip- soidal or oval cell is delimited from the cytoplasm of the ascus by a complete plasma membrane (Fig. 12, F). At first the plasma membrane is thicker near the centrosome, but later its thickness be- comes uniform throughout. CELL-DIVISION. 35 Fig. 13, I, J, shows several stages of the process just described in two asci of Lachnea scutellata. While this is taking place the nuclear beak becomes smaller and smaller until it is finally reduced to a mere thread in which chromatin and membrane are no longer recognizable. The centrosome remains for a short time as a deeply staining and sharply defined disk adhering to the plasma membrane. Very soon it becomes free from the mem- brane and is drawn back to the somewhat pointed nucleus, wliere it appears as a saddle-like thickening upon the point of the nucleus, or Fig. 13. — Free cell-formation in the ascus. G, H, Erysipke communis. I, J, Lachnea scutellata. G, the plasma membrane is complete ; nuclear beak withdrawn and centrosome saddle-shaped, and closely applied to the nuclear membrane. H, a mature spore with cell-wall ; centrosome closely applied to nuclear membrane at upper side. I, J, portions of two asci showing several steps in process of free cell- formation in situ. — (After Harper.) as a simple disk (Fig. 13, G, H). The nucleus now gradually assumes its original spherical form, the chromatin passing into the structure of the resting stage, while the centrosome remains closely adhering to the nuclear membrane. It will be observed that in the specific case of cell-formation described the plasma membrane is completed before the nucleus has reached the resting stage, but in Lachnea (Harper, 1900) the daughter nuclei of the eight-nucleated stage are completely reconstructed before the beaks are formed. This may be, of course, a case of individual variation and of only secondary importance. 36 INTRODUCTION. CELL-CLEAVAGE. The process of cell-formation by means of a progressive cleavage is best known at present in certain Phycomycetes and Myxomycetes. As a convenient and suitable illustration of this method the process of cleavage leading to spore formation in the sporangium of Synchitrium^ parasitic upon the hog peanut, and of Sforodinia is selected. For our knowledge of cleavage we are again indebted to the researches of Harper ('99). The so-called initial cell of the sporangium of Synchitrium^ when almost fully developed, is large enough to be visible to the unaided eye, and contains a relatively large nucleus (Fig. 14, A). This nucleus divides several times until a large number of nuclei are present, which lie irregularly distributed in the cytoplasm. Cleavage of the cytoplasm now begins. It does not take place by repeated bipartitions, nor by the simultaneous precipitation of a cell- wall about each nucleus. As mentioned in a preceding paragraph, it resembles in a large measure the process in certain animals, as for example, the dividing protoplasm of the germinal disk of the chick, or perhaps more nearly that in certain insect eggs in which a series of nuclear divisions precedes cytoplasmic segmentation.^ The cleavage begins by the formation of furrows on the surface, which grow deeper and deeper in a direction more or less radial. It is progressive and divides the cell into successively smaller portions (Fig. 14, D). The process is described in detail by Harper as follows : These grooves are in reality so narrow as to appear as plates, which grow wider by additions along their inner margins till they intersect, and thus divide the protoplasm into irregular blocks or sometimes pyramids with their bases in the surface of the initial cell (Fig. 14, D, E). Only at the very periphery the separation of the cut surfaces of the protoplasm to form a shallow notch, as it appears in section, reveals the true nature of the process as a pushing in of the free surface to form a deep though extremely narrow constriction. In many cases there is at first no separation of the newly formed surfaces ; they remain closely appressed, up to the periphery of the cell. The groove appears in section, merely as a single line which the Zeiss appochromatic lens 1.40 ap. fails to resolve into two closely appressed surfaces (Fig. 14, B). The position of the line is further emphasized by the arrangement of the vacuoles, which are pushed aside and form in section two more or less regular rows in the plane of the newly formed surfaces on each side of the furrow. Such a line might be taken for a cell-plate which subsequently splits to form the boundaries of the protoplasmic segments or which is metamorphosed into the cellulose walls of the spores. That this line, however, in reality represents from the start two closely appressed surfaces is abundantly shown in many cases. > Hertwig : Die Zelle und die Gewebe, p. 187. CELL-DIVISION. 37 These lines of cleavage are not meridional furrows which divide the cell symmetrically, but they intersect each other at varying angles, marking off the surfaee of the cell by a network of grooves, in which the meshes are of an irregular shape and of unequal dimensions (Fig. H, E). Fig. 14.— Cell-cleavage in Synchiirintn discipens . A, sporangium mother-celt. B, Portion of cell showing two nuclei and two surface cleavage-furrows. C, multinucleate stage, showing progressive cleavage by furrows from surfece. D, median section showing cleavage further advanced. E, section from surface of cell in early stage of cleavage. F, cell after segmentation is completed, showing uninucleate protospores.— (After Harper.) The cleavage is progressive from the surface inward, the furrows deepening in general in a radial direction. Still they may be curved, and are inclined to each other at very varying angles and frequently form intersections at points near the surface of the cell, thus cutting off superficial blocks of protoplasm of varying shapes and sizes (Fig. 14, C), so that we have a central solid mass or 38 INTRODUCTION. cell of protoplasm surrounded by a layer of superficial cells ; in other cases the furrows grow radially inward without intersecting till near the centre, thus form- ing narrow cones and pyramids with their bases outward (Fig. 14, D). With the progress of cleavage the contraction of the protoplasm in Synchitriutn becomes very noticeable, the furrows open widely and the masses tend to become rounded. The cell is thus split up into a number of blocks of varying size and containing a variable number of nuclei. In these large cells or portions of protoplasm cleavage fur- rows show no tendency to orient themselves with reference to the nuclei, but as the process advances and the pieces become smaller the nuclei are seen to be more evenly distributed. Finally, the result is always the separation of the cytoplasm into uninucleate masses or cells (Fig. 14, F). It is interesting to note that the process which, in the beginning, seemed to be independent of the nuclei, is finally directed solely from the standpoint of their distribution. From this process of cleavage in Synchitrium it is at once appar- ent that we have a method of cell-formation which is fundamentally different from either of the two methods described in the preceding pages. Here there ai-e no kinoplasmic fibers developed in connection with the nuclei under whose instrumentality plasma membranes are formed, and, in earlier stages of cleavage in the sporangium, new plasma membranes seem to be developed independently of nuclei, though not in their absence. In certain cases of cell-formation by cleavage, in which very large multinucleate masses of protoplasm are involved, as in the plasmodium of certain Myxomycetes and in sporangia of such Phycomycetes as Piloholus and Sporodinia^ vacuoles play a very important part either directly or indirectly. The first indication of the cleavage which is preparatory to the for- mation of the columella- wall in the sporangium of Pilobolus (Harper, '99) is seen in the gradual appearance of a layer of vacuoles larger than the rest, and lying in the curved surface which marks the outline of the columella : The vacuoles become flattened in their radial axes parallel to the surface of the sporangium, and form thus disk-like openings which tend to fuse at their edges. At the same time a circular cleft is seen to start from the edge of the sporangiophore opening . . . and to develop upward, cutting into the vacuoles, so that they become connected into a continuous furrow (Fig. 15, A). Whether this furrow is continued upward to enclose the whole dome-shaped columella, or whether the vacuoles in the upper portion fuse edge to edge before the cleft reaches them, is difficult to determine. The process is a progressive one, the cleavage being complete in certain portions sooner than in others, and CKLL-DIVISION. 39 at a very late period strands of protoplasm are seen connecting the spore plasma with that in the columella. It is not impossible that many of the apparently disk-shaped vacuoles are sections of curved openings which burrow through the plasma from below upwards. Frequently vacuoles which are distinct in one plane are seen, by focussing up and down, to lie connected. There can be little doubt, however, that a considerable part of cleavage of the columella is accomplished by flattening and lateral fusion of originally ellipsoidal or spheri- cal vacuoles ; that is, the cleavage is not entirely by a furrow from the plasma Fig. 15. — Cell-cleavage in sporangium otPiloboltu erystallinu*. — (After Harper.) A, median section at stage when columella is forming. B, section of spore-plasma from base of sporangium, showing surface cleavage-furrows; a, sporangial wall. C, section of portion of upper part of a sporangium, showing irregular sausage-shaped bodies formed by cleavage of spore-plasma. D, similar to C, but older, showing uninucleate masses (protospores). membrane at the mouth of the sporangiophore, but is at least in part a process of separation by excretion of a liquid into vacuoles and their fusion side by side in situ. These vacuoles are not situated on the extreme boundary of the pro- toplasm adjacent to the large central vacuole, but placed where the dense spore- plasma first becomes characteristically spongy. At the base of the sporangium indeed, they cut through plasma as dense as the densest spore-plasma of the sporangium. Why the cell-wall of the columella could not be deposited on the surface of the central vacuole, as well as on the surface of the small vacuoles, 40 INTRODUCTION. and thus enclose all the protoplasm in the sporangium, is an interesting ques- tion. The necessity is evident that the cleavage should proceed through a tolerably dense plasma, and this is, perhaps, due to the need of two proto- plasmic surfaces in contact in order to form a cell-wall. The fact that the columella is not deposited on the surface of the central vacuole seems to indicate that the limiting layer of a vacuole is not quite a plasma membrane, although it may partake partly of the real nature of one. Although there is much to show that the wall of a vacuole, such as we are dealing with here, and a plasma mem- brane are closely related, yet the author is not quite ready to admit that they are the same. Why two plasma membranes should be in contact in order to form a cell- wall, as suggested by Harper, is not quite clear to the author, since in many cases a single plasma mem- brane will secrete a cell-wall. In the cleavage of the spore-plasma, which begins soon after the columella is complete, vacuoles also take an important part. The cytoplasm becomes somewhat vacuolar, and the numerous nuclei are rather evenly distributed throughout its mass. Cleavage furrows appear now near the base of the sporangium, cutting the surface into irregular polygonal areas (Fig. 15, B). At the same time vacuoles in the interior become angular, appearing three-cornered in section, and their edges cut through the cytoplasm to meet similar cleavage furrows from adjacent vacuoles (Fig. 15, B). In the meantime the surface furrows which have been growing deeper meet and become continuous with the edges of the vacuoles. By pressure of the adja- cent plasma-masses, the surfaces of the vacuoles which were formerly convex become concave, and the vacuoles appear as intercellular spaces between the cleavage-segments. In this manner the spore- plasma is marked out into irregular blocks, apparently without refer- ence to the size or number of nuclei they contain. A continuation ot the process cuts the spore-plasma into oblong rounded sausage-shaped masses containing generally two to four nuclei in a row (Fig. 15, C). These oblong masses now divide transversely to form rounded bodies with one or few nuclei (Fig. 15, D). This completes the primary cleavage by which the spore-plasma has been cut up into smaller units with one or few nuclei. These units are not the spores. They undergo a period of growth and nuclear division before the final cleavage divisions take place by which the mature spores are pro- duced. The last divisions are, however, similar to the first, presenting the simpler process of cleavage or fission. In the sporangium of Pilobolus, we have a cleavage which is of the same type as in Synchitrium., with the exception of the promi- CKLL-DIVISION. 4I nent part taken by the vacuoles in the former. Although the mem- branes of these vacuoles may not, at first, be exactly similar to plasma membranes, they are undoubtedly converted into them. Since we assume that the plasma membrane is largely of a kinoplasmic nature, and attribute to it something of a morphological rank in the cell, it may not be wholly fanciful to suggest that the limiting membrane of a vacuole may be developed into a real plasma membrane, and that this actually takes place in the plants in question. CELL-DIVISION IN DICTYOTA AND STYPOCAULON. There is yet another method of cell-formation which has been observed in certain of the brown algae that differs materially from the process of cleavage already described. There are no kinoplasmic connecting fibers by which a plasma membrane may be formed, nor is it a cleavage such as has been described for certain fungi. The plasma membrane, or cell-plate, seems to be formed directly out of the apparently undifferentiated framework of the cytoplasm. This type of cell-formation has been observed in such Phceophycece as Stypocaulon (Swingle, '97), Fucus (Strasburger, '97), and Dic- tyota (Mottier, 1900). Swingle has followed the development of the cell-plate in great detail in the apical cell of Stypocaulon. Here each division of the nucleus is followed by a cell-division. The bulk of the cytoplasm presents a very beautiful and typical alveolar structure, and the first indication of a cell-plate is seen in certain alveolae, which show a tendency to arrange themselves across the cell in a transverse plane (Fig. 16, B). As soon as this orientation of the alveolae becomes more marked, the transverse alveolar lamellae form a more continuous plane which, in section, appears as a very fine line. During these changes neither an increase in the number of connecting fibers between the nuclei nor any perceptible change whatever in the arrangement of the kinoplasm was to be seen. Only a few fibers or lines of force, indi- cated by the arrangement of the alveoloe of the frothy plasma, extend from the nucleus of the apical cell to the seat of cell-plate formation, and fewer still from the lower nucleus to the same place. It is certain that if there be real fibers, they must be extremely delicate and not numerous enough to lead one to suppose that the cell-plate is laid down by any such process as in the higher plants. The author has found that the development of the plasma membrane in the tetraspore mother-cell of Dictyota (Mottier, 1900) is similar to that of Stypocaulon. Here there is absolutely no visible trace of 42 INTRODUCTION. kinoplasmic connecting fibers between the nuclei, and in the region of the cell-plate the cytoplasm seems undifferentiated. The plasma mem- branes, or cell-plates, which will separate the four spores, are laid down almost simultaneously. In the region where they are to appear the cytoplasm, as elsewhere, except near the nuclei, presents the same visible structure of alveolae, or perhaps a mixture of alveolae and a thread-like network. Rather large and small meshes are intermingled. Fig. i6. — Cell-plate oi Dictyota dichototna and Stypocaulon. A, portion of cell-plate from tetraspore mother-cell of Dictyota, formed apparently by arrangement of alveolar lamellae into a continuous and even plane. B, same from apical cell of Stypocaulon, — (B, after Swingle.) The small-meshed structure is apparently more granular than that with larger meshes. The first visible trace of a cell-plate is manifested by the transverse walls of the alveolae becoming perceptibly thicker and arranging them- selves in such a way as to appear as an uneven or somewhat zigzag line in section (Fig. i6, A). In this cell-plate primordium the walls of both large and small meshes take part. At first certain of the alve- olar lamellae are thinner than others, so that the cell-plate seems CELL-DIVISION. 43 interrupted at these places, but eventually and gradually it attains a uniform thickness. Very soon the cell-plate is a uniform plane, appearing in section as a rather smooth line. The cell-plate is not always laid down everywhere simultaneously, but sometimes it appears at first more marked at the periphery. This seems to depend upon the position of the nuclei. It is evident that in Dictyota no differentiated kinoplasmic connecting fibers can be recog- nized by which the cell-plates are formed. It seems that the appar- ently undifferentiated framework of the cytoplasm, consisting of large and small meshes in the immediate region of the cell-plate, is con- verted into a plasma membrane. The cell-plates are certainly formed under the influence of the nuclei, and kinoplasm in some form enters into the process. The behavior of the cell-plate toward certain stains, particularly gentian violet, and the character and behavior of the cytoplasm in that region, immediately preceding the appearance of the plasma membrane, strongly suggests that the latter is not an actual transformation of the alveolar walls, but that the substance of the cell-plate is deposited by kinoplasm present in the framework of the cytoplasm. The form in which this kinoplasm occurs here is difficult to determine, but it mat- ters very little whether it takes on the form of a fibrous network or of alveolae, or whether it is present merely as a homogeneous fluid. Of the several types of cell formation briefly described in the fore- going pages, the first, or that which is typical for higher plants, occurs generally in all plants from the liverworts up. It obtains also in Chara and Nitella and has been found by Fairchild ('97) in Basi- diobolus. This method doubtless occurs in other algae and fungi. The process of typical free cell-formation, as found in the ascus of the Ascomycetes mentioned, is, so far as known, restricted to this group of fungi A process of free cell-formation has been described by Strasburger in the egg-cell of Ephedra^ but there it differs considerably from that in the ascus, since centrosomes or centrospheres are not present and the kinoplasmic fibers radiate in all directions from each nucleus. The process of cleavage is the method of cell-formation in the plas- modium of Myxomycetes and in certain Phy corny cat es. It is also of undoubted occurrence in many algae and in other fungi. Whether the kind of cell-plate formation described for Stypocaulon and Dictyota occurs outside of the brown algae, future research must determine. The process of constriction characteristic of Cladophora and Spiro- gyra may be looked upon as a kind of cleavage in which the formation 44 INTRODUCTION. of the new cell-wall is gradual and progressive from the old cell-wall inward, instead of being developed simultaneously from a plasma membrane previously formed. Whether in such cases new plasma membranes are formed across the ends of the daughter cells which come in contact with the new transverse cell-wall the author is unable to state. THE CENTROSOME AND THE BLEPHAROPLAST. As illustrations of karyokinesis in which the spindle arises through the agency of centrospheres I have selected thetetraspore mother-cell of Dic- tyota and the ascus of certain Ascomycetes^ because the centrosphere is probably best known in those cells and because the entire develop- ment of the mitotic figure has been followed in great detail. In these plants, as well as in Fucus and certain Sphacelariacece , we have seen that the body which we call a centrosome is one that persists from one cell-generation, or nuclear generation, to another in vegetative and in certain reproductive cells. It seems to be capable of division, and is the centre of radiations that give rise to the karyokinetic spindle. We do not know with absolute certainty that the centrosome divides, although the evidence seems to admit of no other interpretation. In addition to the plants just mentioned, centrospheres have been found in some liverworts, in diatoms, and in certain Rhodophycece. In the diatoms, however, the behavior of the centrosome during karyo- kinesis, as described by Lauterborn ('96), differs widely from the typical cases described in the preceding pages. In species of .Pinnu- laria, Surirella^ and others, Lauterborn finds that the peculiar cen- tral spindle arises from the centrosome by a division or process of budding. "Es scheint mir keinem Zweifel zu unterliegen, dass die Anlage der Centralspindel aus dem Centrosom durch eine Theilung (oder, wenn man lieber will, Knospung) hervorgeht " (1. c, p. 61). In the diatoms in question the original centrosome is a relatively large globular body which is the center of a system of beautiful radia- tions. Soon after the budding off of the primordium of the central spindle, the original centrosome, with its radiations, disappears, and what is taken to be the new centrosomes arise near the poles of the spindle and apparently from it. So far as the author is aware, such a phenomenon has no parallel among plants, and it is impossible to bring the process of spindle- formation in the diatoms, as described by Lauterborn, into line with anything known in other organisms. When we consider the facts alone in the algae and fungi mentioned, we certainly have strong evidence in favor of the doctrine of the genetic THE CENTROSOME AND THE BLEPHAROPLAST. 45 continuity of the centrosomes ; but from the fact that no such organs exist in the higher plants, and that they seem to be wanting in many Thallophyta as well, this view is greatly weakened, if not rendered quite untenable. On the zoological side of the question, the recent researches of Wil- son (1901) on eggs of Toxopcnustes^ which were made to develop parthenogenetically through certain stages by means of chemical stimuli, throw new light upon the subject. In segmenting eggs induced to develop parthenogenetically by means of a treatment with suitable solutions of magnesium chloride, numerous asters (cytasters) often made their appearance in the cytoplasm in addition to the nuclear asters. Similar asters may arise also in non-nucleated fragments of eggs. These " cytasters," just as the segmentation or nuclear asters, may consist of a very distinct centrosome upon which is centered a system of beautiful radiations. The centrosomes divide, and a central spindle is formed between the daughter centrosomes. In fact, the " cytasters " are exactly like the normal cleavage-asters arising in con- nection with the chromatin. As the evidence seems conclusive that the "cytasters" arise de novo, Wilson concludes that centrosomes occurring normally in cells arise also cie novo, and that the doctrine of the genetic continuity of the centrosome is untenable. It is not known whether anything comparable to these " cytasters" ever occurs in a plant egg-cell, which may be made to develop parthe- nogenetically by artificial means, and consequently we cannot accept the conclusion upon this basis as applicable to plants. There are, however, in plants many well established facts which argue strongly against the view that the centrosome or centrosphere is an organ of morphological rank. In 1897, the author made the unqualified statement, to which he still adheres, that centrosomes or centrospheres do not occur in the higher plants, and nearly all research since made along this line has only confirmed this view. We know now that the structures which Guignard so beautifully figured in 1891 for cells of Lilium were the product of preconceived ideas and the misinterpretation of certain facts. There are still a few observers who persist in seeing centro- spheres in the cells of higher plants, in which a score or more of the most competent cytologists, with the aid of the very best methods, have failed to find any such structures. It may be of some interest to note, however, that the centrospheres figured more recently by these obsei-vers are not drawn with the old-time diagrammatic distinctness, and it will probably not be long till these structures will not appear at all in figures illustrating karyokinetic phenomena in Allium cefa and species of Lilium. 46 INTRODUCTION. At the present writing it is the opinion of the author that individu- alized centrosomes or centrospheres do not occur in plants above the liverw^orts, and they are certainly absent in certain species of these [Anthoceros). On the whole, these structures are well established in only a few Thallophyta. As the writer has already stated in a former paper (Mottier, 1900), if we take into consideration only "such plants as Fucus^ Stypocaulon^ Dictyota^ and certain Ascojnycetes^ there are good grounds for the view that the centrosome is an organ of morphological value ; but the evidence furnished by these forms, however convincing it may seem, is not quite sufficient, especially in the light of our knowledge of kary- okinesis in forms in which centrosomes or centrospheres have not been found ; for there is no reason for believing that the spindle fibers in plants devoid of centrosomes are of a different substance from the radiations or spindle fibers developed in connection with an aster. Space will not permit of a discussion of such questions as whether the radiations are outgrowths of the centrosome considered as a mor- phological unit, or constructed out of the kinoplasm by the centrosome, or whether the centrosome is only a denser mass of kinoplasm, formed by the meeting of the polar radiations, and which may persist after the radiations and spindle fibers have disappeared. It may be stated in this connection that in plants there is little to support the view that the radiations are centripetal or centrifugal currents. They do not seem to be currents at all. We understand radiations and spindle fibers to be fine, more or less homogeneous, kinoplasmic threads which are capable of contracting, extending, or becoming changed into a uniform and homogeneous mass. We have now to consider the relation of the centrosome to the blepharoplast^ or cilia-bearer, which is so well known in the sperma- tozoid of the Archegoniates (see Chapter V). Belajeff, Ikeno, and Hirase and a few others regard the blepharo- plast of the fern and certain gymnosperms as the homolog of the centro- some. It seems to the author that such a conclusion is merely a hasty judgment, which does violence to the facts as they are known at present. The development and function of the blepharoplast, as will be seen from the chapter referred to, shows clearly that this structure lacks the more essential distinguishing characteristics of the normal centrosphere, as it is known in the cases most thoroughly investigated. The bleph- aroplast is not the center of- kinoplasmic radiations which form a karyokinetic spindle. So far as has been shown the radiations of the blepharoplast primordia take no part in the formation of the spindle. These primordia do not divide to give rise to new blepharoplasts, but THE CBNTROSOME AND THE BLEPH AROPLAST. 4^ arise de novo. They do not persist through several successive genera- tions of cells, two cell-generations representing the maximum time of their duration. In short, the blepharoplast develops merely the cilia and forms, therefore, the locomotary apparatus of the spermatozoid. No phylogenetic relationship has as yet been shown to exist between blepharoplast^' and centrosome. The fact is that, in those plants in which blepharoplasts occur, there are no centrosomes with which to show any phylogenetic relationship. The main reason, it seems, for regarding the blepharoplast as the homolog of the centrosome is the sole fact that the primordia of the former at a certain period of develop- ment are provided with a system of radiations, giving them the appear- ance of centrospheres. The view concerning the origin and phylogeny of the blepharoplast as advanced by Strasburger is of interest, since it is the only one that seems to take into consideration all the facts. Strasburger derives the blepharoplast from the cilia-bearer of the zoospores and gametes in the algae. In the zoospoi'es of certain algae, e. g".^ Vaucherta^ (Edogo- nium^ and others, the cilia spring from a localized thickening of the plasma membrane (Hautschicht) at the anterior end. In (Edogonium this kinoplasmic thickening is in the shape of a double convex lens, from the edges of which arise the numerous cilia. In the large swarm spore of Vaucheria the nuclei seem to be intimately connected with the formation of the cilia-bearer. The nuclei migrate to the plasma mem- brane and elongate in a direction at right angles to the surface of the spore. The anterior end of each pear-shaped nucleus comes in contact with the plasma membrane. That part of the plasma membrane in contact with the nucleus thickens in the form of a delicate concavo- convex lens, from two points of which, on opposite sides, spring the cilia. The size and shape of the cilia-bearer vary, of course, in different algae. Timberlake ('oi) finds a small body at the base of the cilia in Hydrodictyon^ but it does not seem to be part of the plasma membrane. As Strasburger has pointed out, the " mouth-piece " of swarm spores and gametes is not to be confounded with the cilia-bearer, since the former represents the entire anterior end of the cell free from chloro- phyll. It is true that the cilia-bearer is not well known in the sperma- tozoids of algae, but transitions show that in all probability the sperma- tozoids were derived from male gametes which in every way resembled asexual swarm spores. The spermatozoids of Volvox globator are regarded as a good illustration of this relation, for in structure they occupy an intermediate position between the gametes of algae and the spermatozoids of Chara. In Volvox the two laterally inserted cilia would seem to indicate that the blepharoplast had undergone a lateral 4$ INTRODUCTIOX. displacement, for the entire anterior end of the spermatozoid of Volvox is certainly not blepharoplast. (The very suggestive theory of Stras- burger carries with it a certain degree of probability, yet to what extent it is true further research must determine.) If, however, any genetic relationship exists between centrosome and blepharoplast, the evidence is certainly to be sought in the lower plants. In this connection it is of the greatest importance to know first of all whether, in such algae as the Spkacelariacece^ in which centrosomes are known, any relation exists between the centrosome and cilia-bearer, assuming, of course, that the cilia arise here also from a differentiated body. In Chara and in those Archegoniates^ with blepharoplasts no centrosomes are found, neither is any such body known to take part in the formation of the spindle in such algae as (Edogonium^ and others in which highly developed cilia-bearers occur. Although these facts do not prove anything, yet they lend encouragement to the belief that centrosome and blepharoplast may be homologous structures, or in some degree phylogenetically related. Those who maintain that the cilia-bearers are centrosomes have not, it seems, approached the question from the standpoint just mentioned, but seem to have based their conclusion upon the resemblance between blepharoplast primordia and centrospheres, or upon analogies between the spermatozoids in plants and the spermatozoa of certain animals. Belajeff ('99), who claims that blepharoplasts are homologous with centrosomes, strengthens his view by his observations in spermagenous cells of Marsilia. In the grandmother-cells of the spermatozoids of this plant he finds that the blepharoplast primordia, which lie some distance from the nucleus, divide previous to the division of the nucleus, and between the two separating daughter primordia a small central spindle is developed just as in certain animal cells. From this small amphiaster the karyokinetic figure is developed. This, if true, is the first case on record in plants in which a central spindle is formed between the daughter centrosomes, lying in the cytoplasm some dis- tance removed from the nucleus. In the light of what is now known concerning the development of the spindle in Chara and in the Pieridophyta, the author entertains serious doubts concerning the accuracy of Belajeff's statement. Oster- hout's ('97) studies on the development of the spindle in the spore mother-cells of Equisetum prove beyond all question that centrosomes are not present in that genus. In other Pteridophyta the majority of all investigations, which have been thorough or reasonably exhaustive, shows that centrosomes or centrospheres are absent there also. ' Marchtintia folym«rfka excepted. SIGNIFICANCE OF THE SEXUAL PROCESS. 49 From our present state of knowledge of the development of the blepharoplast there is but one conclusion, it seems to the author, that can be legitimately drawn concerning their origin, namely, that they arise de novo. As regards centrosomes the evidence is more compli- cated and conflicting. Although, in the opinion of the author, the evidence is decidedly against the doctrine of the genetic continuity of the centrosome, yet the proof is not quite conclusive. If centrosomes also arise de novo, then the problem assumes a slightly different aspect, for it is questionable whether we are justified in speaking of homologies between organs that, as such, are without genetic continuity. There is strong evidence, which seems to be increasing from day to day, that it is the fundamental substance known in the plant cell as kinoplasm which is genetically continuous. After a careful considera- tion of the facts, the author is led to the same conclusion concerning the centrosome to which he gave expression in 1900, in a paper on the nuclear division in Dictyota (1. c, p. 178), namely, that it is the kinoplasm which should hold the rank of morphological unit, and that the centrosome should be regai'ded as an individualized part of the same, existing in that form in some organisms and not in others, for reasons that cannot at present be explained. As regards blepharo- plasts, about the only conclusion in harmony with all the facts is that these bodies represent individualized parts of the kinoplasm which arise de novo in certain spermagenous cells, and from which the cilia are developed. SIGNIFICANCE OF THE SEXUAL PROCESS AND THE NUMERI- CAL REDUCTION OF THE CHROMOSOMES. Speaking generally, the phenomena resulting from the sexual process fall into two categories, namely, (i) the transmission of hereditary characters, together with the blending of two lines of descent by the fusion of the sexual nuclei, and (2) the imparting of a growth stimulus to the fecundated egg or to the zygote, by which the energy of growth and division is restored. Correlative with the first category is the reduction in the number of chromosomes. The doctrine of the significance of the numerical reduction of the chromosomes now generally accepted by botanists as a working hypothesis, was first stated in a well organized form and presented formally to botanical science by Strasburger ('94) in a mas- terly essay on the " Periodic Reduction of Chromosomes in Living Organisms." The enunciation of this doctrine marked the beginning of a new epoch in the study of sexuality and in cytological research in plants. 50 INTRODUCTION. The simplest and most primitive organisms known reproduce them- selves asexually, and vv^e are obliged to assume that, from a phylo- genetic standpoint, sexually differentiated organisms were descended from asexual forms. The process of this descent is clearly illustrated by certain of the green algse in which the sexual act consists in the fusion of exactly similar motile gametes. These gametes were un- doubtedly derived from asexual swarm spores, which they closely resemble, except in that they are smaller and often have fewer cilia. In Ulothrix^ for example, and in many of the green algae, the gametes are, so far as is known, smaller and possess only two cilia, while the larger asexual swarm spores bear four cilia. Both sporangia and gametangia are homologous structures, and, so far as is known, the gametes differ only physiologically from the asexual spores. According to Strasburger, to use the language of the translation '} The sexually differentiated plants manifest certain differences in their onto- geny, from which it is possible to infer what was the course along which the phylogenetic differentiation proceeded after sexual differentiation had taken place. The simplest case is that in which the product of fertilization gives rise to an individual similar to those which gave rise to the product of fertilization, and which closes its own life history with the development either of sexual organs or of asexual organs homologous with them. This occurs in many Chlorophycece, where, from the zygospore (the product of the coalescence of similar gametes) or the oospore (the product of the coalescence of dissimilar spermatozoids and ova), a generation is developed which resembles the preced- ing and gives rise either to swarm-spores or to sexual cells homologous with them. Generally, any one sexual generation follows after a number of asexual generations, the relation being, however, dependent on external conditions, so that, as Klebs has shown, the development of a sexual or an asexual generation can be determined by the observer. In such cases there is a homogeneous sequence of generations which does not include any other kind of sequence or alternation beyond the development either of asexual reproductive organs or of sexual organs homologous with them. The asexual reproductive organs are especially concerned with the rapid multiplication of individuals under favorable external conditions ; whilst sexual reproduction is of importance in maintaining the existence of the species under circumstances which are unfavorable to the vegetative existence of the individual. At the same time, sexual reproduction ensures certain advantages arising from the coalescence of distinct sexual cells. In proportion as the asexual mode of reproduction was replaced by the sexual, the numerical conditions of multiplication were maintained either by the development of a number of oospores, as in certain Fucaceae ; or, in addi- tion to the sexual organs, altogether new organs were developed to ensure rapid and vigorous development of new individuals in an asexual manner. This took place in various ways. Either asexual reproductive organs were inter- calated in the life history of the original generation, or an altogether new asexual generation was developed from the product of the sexual act. • English translation, Ann. Bot., 8 : 281-316. SIGNIFICANCE OF THE SEXUAL PROCESS. 5 1 I have quoted thus at length because it seems that this statement of Strasburger is a compact and concise summing up of the phylogenetic development of the process of reproduction and multiplication of indi- viduals among the lower plants. The intercalation of new asexual reproductive organs into the origi- nal generation is strikingly illustrated in many of the fungi, in which the independent individualization of the different stages of development of the sexual generation into special organs of vegetative multiplica- tion, or even into distinct individuals, was carried so far that sexuality seems to have disappeared entirely, as in the higher fungi. On the other hand, in all plants beyond and including the Bryophyta there arose an altogether new generation as the product of the sexual act, whose function is to produce asexually a large number of individuals. The degree of development attained by the new generation in the plants above the Thallophyta differs according to whether its activity was limited to the production of asexual spores alone, or included nutritive functions as well, or whether it became an independent indi- vidual. In the Bryophyta^ especially in some of the simpler liver- worts, the new asexual generation is confined almost exclusively to the production of spores, /. e.^ to the multiplication of the individual, while the original or sexual generation upon which all nutritive func- tion is devolved, together with vegetative multiplication as well, has attained in many cases a cormophytic differentiation. In the Pteri- dophyta and in the higher plants, on the contrary, the center of gravity of phylogenetic evolution is transferred to the new or asexual genera- tion arising from the act of fecundation, and in these plants the asexual generation has attained its highest cormophytic development. Among the Pteridophyta of the present time it is evident that (1. c, p. 283) '*as this evolution took place, the nutritive apparatus of the sexual generation became of less importance, and it became altogether super- fluous from the moment when the asexual generation began to provide its spores with the material necessary for the development of the sexual generation." Along with this evolution there came into existence, as a correlative phylogenetic process, the dimorphic character of the gametophyte, which is characteristic of certain Pteridophyta and of all Spermatophyta. This dimorphism was probably manifested in the chaiacter of the mature gametophyte before any visible trace of it could be recognized in the unicellular stage of the sexual generation, namely, the spore. To illustrate this fact we need only to recall the condition which obtains among certain homosporous Pilicinece^ for example, Onoclea struthiopteris of the Polypodiacece. Here there is no visi- ble evidence of heterospory, yet it is perfectly well known that in every 52 INTRODUCTION. culture of spores some will develop into distinctively male prothallia, bearing only antheridia, while others show a marked tendency to develop into prothallia bearing only archegonia. It is also well known that this tendency toward dimorphism is, in a measure, influ- enced by external conditions, for if spores of Onoclea struthiopteris be sown thickly, and the culture be poorly illuminated and, conse- quently, poorly nourished, the vast majority of the prothallia will be male; but if the spores be sown thinly and well illuminated, a much greater number will become female plants. In all existing forms in which the spores, or unicellular condition of the sexual generation, contain food material for the development of the asexual generation, or its earlier stages, dimorphism is well estab- lished, i. e., those forms are heterosporous, and the conclusion which most naturally follows is that heterospory and the disappearance of the nutritive apparatus of the sexual generation represent correlative phylogenetic processes. Now, during this phylogenetic evolution and, as Strasburger very clearly puts it, — In accordance with the general law which determines the phylogenetic disap- pearance of organs which have become useless, the vegetative parts of the sexual generation became more and more reduced, until little was left but the repro- ductive organs themselves : hence the progressive reduction in the prothallium from the Ferns up to the Phanerogams. This reduction culminated in the complete loss of independent existence by the sexual generation, because it had ceased to be able to nourish itself independently, and [because of] its becoming enclosed by the asexual generation. In consequence of this enclosure of the sexual in the asexual generation, the advantageous rapid multiplication of indi- viduals which the latter originally effected was lost : in order to compensate for this loss, a large number of seeds were produced in the Phanerogams in place of the numerous spores of the Cryptogams ; that is, multiplication is effected now by the product of fertilization instead of by asexual spores. In harmony with this doctrine, an alternation of generations is neces- sary in those plants in which the fecundated egg gives rise to the asexual generation, and the asexual spore to the sexual generation. The development of the plant kingdom, at least so far as sexuality is concerned, seems to show that sexual differentiation was preceded by asexuality, and in those groups in which a true alternation of gen- erations exists the sexual generation is to be regarded as the older and more primitive and as having arisen from an asexual form. In fact, we are able to trace this phylogenetic development step by step, or the evidence at hand, at least, seems to be sufficiently conclusive to justify the general acceptance of the doctrine. Probably the first indi- cation of this development is to be found among such algae as CEdo- SIGNIFICANCE OF THE SEXUAL PROCESS. 53 gonium^ Coleochcete and, as the researches of Oltmanns seem to indicate (See Chapter IV), certain Rhodophycex. From the fecun- dated egg of CEdogonium four swarm-spores are developed, while in Coleochcete a multicellular body is developed, from the cells of which asexual swarm-spores are formed. In both cases the swarm-spores give rise to sexual plants, or the first generation. The product of the fecundated egg in Coleochcete bears a striking resemblance to the sporophyte of such liverworts as Riccia. The fundamental differ- ences lie chiefly in the fact that the covering of the sporophyte in Coleochcete is derived from vegetative branches of the thallus, the oogonium being unicellular, and that the asexual spores are motile, a correlation with the aquatic habit of Coleochcete. In the Rhodo- phycece the cystocarp or cystocarps are the product of the fecundated egg, and the spores give rise to the first generation. This is made all the more probable by the researches of Oltmanns, which go to show that the fusion of the cells of the sporogenous filaments with auxiliary cells is merely a nutritive process. It is of interest to note further that a similar condition is preserved in certain Ascomycetes in which Harper has proved that unquestioned sexuality exists. Such algse as Coleochcete., therefore, seem to point out more or less clearly the phylogenetic road along which the ancestors of the Archegoniates have passed. Research upon the process of fecundation and indirect nuclear division, especially in reproductive cells, during the past twenty years, has given a new insight into the significance of sexuality and the alter- nation of generations in plants. Our knowledge along this line was very materially advanced by the discovery of Van Beneden ('83) that the number of chromosomes is the same in both conjugating nuclei. Further investigations have established the still more important fact that, in both plants and animals, a reduction to one-half of the number of chromosomes in the sexual nuclei preceded the sexual act, and that, as a consequence of the fusion of the male and female nuclei, the number of chromosomes in the fecundated egg is doubled. In all the higher plants it is a well-established fact that the numeri- cal reduction of the chromosomes takes place in the spore mother-cell, and that in the cells of the gametophyte arising from the spore the reduced number persists. In cells of the sporophyte, resulting from the fecundated ^%^,', the increased number obtains until the differentia- tion of the spore mother-cells. It will thus be seen that the funda- mental characteristic of both sexual and asexual generations lies in the number of the chromosomes, and upon this phenomenon rests the sexual differentiation of cells. 54 INTRODUCTION. There is a possibility that this doctrine may not be applicable to cases of apogamy, apospory, and normal parthenogenesis among plants. It has been suggested by Strasburger ('94, p. 300) that the number of chromosomes may become doubled under the influence of correlative processes in an apogamously developed fern which arises as a bud from the prothallium, the nuclei of vv^hose cells contain the reduced number, and for the same reason the reverse may take place in cases of apospory, i. e., the aposporous development of prothallia may be attended vv^ith a correlative reduction in the number of chromo- somes. Until the facts are determined by actual observation, all discussion of this subject must remain a matter of pure speculation. The researches of Juel (1900) upon the normal parthenogenesis of Antennaria alpina are of the highest interest in this connection, as they throw light upon this question so far, at least, as the seed-bearing plants are concerned. In Antennaria alpina^ in which the egg develops parthenogenetically under normal conditions, Juel finds that no reduction in the number of chromosomes takes place in the develop- ment of the embryo-sac, and, consequently, the nucleus of the egg- cell which gives rise to the parthenogenetic embryo contains the same number of chromosomes as the vegetative cells. Contrary to Anten- naria dioica^ in which fecundation regularly occurs, the mother-cell of the embryo-sac of A. alpina develops immediately into the embryo- sac, the heterotypic and homotypic nuclear divisions which follow the appearance of the reduced number of chromosomes being omitted. In cases of normal parthenogenesis among the angiosperms, the facts, so far as they are known, are certainly not at variance with the doctrine of the reduction of the chromosomes as applied to the alter- nation of generations. As has been intimated in preceding paragraphs, the sexual genera- tion has been spoken of as the more primitive condition, and, as will be seen from the following, the reduction in the number of chromo- somes in the spore mothei--cell is regarded by Strasburger as the return of highly organized plants to the original unicellular condition : The morphological cause of the reduction in the number of chromosomes and of their equality in number in the sexual cells is, in my opinion, phylo- genetic. I look upon these facts as indicating a return to the original generation from which, after it had attained sexual differentiation, off"spring was developed having a double number of chromosomes. Thus the reduction by one-half of the number of the chromosomes in the sexual cells is not the outcome of a gradually evolved process of reduction, but rather it is the reappearance of the primitive number of chromosomes as it existed in the nuclei of the genera- tion in which sexual diffierentiation first took place (1. c, p. 288). The phenomenon under consideration is essentially that of the return of the SIGNIFICANCE OF THE SEXUAL PROCESS. 55 most highly organized plants, at the close of their life-cycle, to the unicellular condition : in a word it is the repetition of phylogeny in ontogeny (1. c, P-3"). This theory of reduction must still be regarded as a very helpful working hypothesis, finding its greatest application in the higher plants. In the lower cryptogams the theory is confronted with facts, many of which seem at present to be quite at variance with it. The product of fecundation in the Thallophyta as a rule does not give rise to a definite organism representing the asexual generation, and it is not known at which point in the life-cycle that reduction takes place. It has been suggested that reduction may take place during the germina- tion of the zygote or oospore. Conclusions respecting the time of reduction in the lower cryptogams have been drawn chiefly from the phenomena of certain cell-divisions that seem to be analogous with divisions which follow the reduction in higher organisms, and not from an actual determination of the number of chromosomes. On account of the many difficulties in counting, the number of chromo- somes is known in only a very few algae and fungi, and our knowledge on this subject is so meager with respect to these plants that the few definite facts that have been obtained, although apparently at variance with the theory, may not as yet be considered as offering very serious objections to it. If the reduction in the number of chromosomes signifies what is attributed to it by the theory, it is possible, in the light of facts that have been observed in such algae as Fucus and Dictyota^ that what is considered the sexual generation in the Thallophyta may not be homologous with the gametophyte of higher plants, assuming that homology is based upon the number of chromosomes. Farmer and Williams ('96, '98), and Strasburger ('97) have found that the reduced number of chromosomes in Fucus appears in the oogonium, while in vegetative cells of the* thallus twice that number is present. Stras- burger finds that in the first nuclear division in the oogonium the reduced number appears, fourteen to sixteen having been counted, and this number persists throughout the two succeeding mitoses. In vegetative cells of the thallus, which is regarded as the gametophyte, the number is not far from thirty. In Dictyota I have found the reduced number (sixteen) of chromosomes in the first nuclear division of the tetraspore mother-cell, while in the vegetative cells of the thallus bearing the tetrasporangia about twice that number was counted. Whether in the nuclei of plants arising from tetraspores the reduced ^6 INTRODUCTION. number persists, and whether in the egg-cell this number obtains was not determined.^ As is well known, two views are held concerning the manner in which the reduction in the number of chromosomes is accomplished. One of these views, which has been given prominence by Weismann, holds that the chromosomes are qualitatively different, and that reduc- tion is accomplished during the maturation divisions in animal cells and in the first two divisions taking place in the spore mother-cells of higher plants. For example, in the second maturation division of the animal egg it is maintained that the daughter chromosomes do not arise as a result of a longitudinal splitting, but by a transverse division, or what is known as a qualitative division. The nuclei of the four cells thus resulting, whether representing the egg and its polar bodies or those which develop directly into spermatozoa, are hereditarily different in character, and it is upon this assumption that hereditary variation is based. The other view, which is now very generally accepted by botanists, is that, in plants no qualitative division exists, but the chromosomes of each mitosis arise in every case by a longitudinal splitting. The reduction takes place in the resting nucleus or during the early pro- phase of the first, or heterotypic, mitosis in the spore mother-cell of higher plants. The fact, as shown in preceding paragraphs, that during this first mitosis a double longitudinal splitting of the chromo- somes occurs, probably as a preparation for the two divisions, has led to much confusion, because these divisions were supposed to have been rather the instrument of reduction than a consequence of reduction. Assuming the persistent individuality of the chromosomes, we may conclude on good grounds that the reduction represents the actual and complete fusion of the chromosomes of both parents, which have remained separate in the sporophyte until the formation of the spore mother-cells. There is no visible evidence that a qualitative difference exists between the chromosomes in plants, and our assumption here is that they are hereditarily similar, because of the fact that every indi- rect nuclear division is preceded by a longitudinal splitting of the chromatin. Since the nucleus is the unquestionable bearer of hereditary char- acters, fusion of sexual nuclei in fecundation has for its purpose the blending of two lines of descent and possibly the restoration 'J. Lloyd Williams in a recent paper (Studies in the Dictyotacese, Ann. Bot.,i8: 141-160, 1904) observes facts that seem to point to the conclusion that the plantlets developing from the tetraspores, with their reduced number of chromosomes, may become gametophytes, and that the fecundated egg cells probably develop into tetraspore plants which have been shown to possess the increased number of chromosomes. If this be true, an alternation of generations exists in Dictyota. SIGNIFICANCE OF THE SEXUAL PROCESS. 57 of the power of growth and cell-division. The influence of the hereditary characters of each parent upon each other by their intimate association in the same nucleus seems to be the physical basis of phylogenetic variation, but the manner in which this influence acts to bring about variation, or to impart a more vigorous character to the product of fecundation still remains a matter of speculation. It is well to consider the blending of the two lines of descent as a consequence of fecundation in a relative sense or as a correlative phylogenetic process. In certain of the lower cryptogams, Ulothrix and Basidiobolus for example, in which the gametes arise from adjacent cells of the same filament and in which a sexual differentia- tion is not at all or only scarcely recognizable, there does not seem to be two lines of descent to blend, yet it is conceivable that the sexual character of the nuclei may have been determined before the stage of ontogeny is reached in which the sexual cells manifest themselves as such. If in such forms a reduction in the number of chromosomes occurs, the sexual character of the nuclei is determined at that time. It is well known that among the simpler forms of the algse and fungi, the development of gametes depends to a certain extent upon external conditions, which effect transpiration, atmospheric pressure, food supply, and so forth, yet no one would suppose for one moment that sexuality is the outcome of these external conditions. We have now to touch briefly upon the category of phenomena by which a growth stimulus, or the power of growth and cell-division, is imparted to the product of fecundation. Among many of the lower algae about the only important difference which seems to exist between a gamete and an asexual swarm-spore is the ability of the latter to develop into a normal individual of the adult size. It is true that the iso-gametes of algae, such as Ulothrix., are capable of developing into small dwarf individuals — a fact which indicates that here, at least, the gametes possess the power of independent growth sufficiently to enable the resulting plantlet to develop to a limited extent. As soon, how- ever, as the sexual elements have attained any marked degree of bisexual differentiation in the plant kingdom, the individual gametes are quite incapable of independent development even into the most rudimentary individuals, cases of normal and artificial parthenogenesis excepted. The stimulus to growth and division in bisexual reproductive cells is imparted normally only by the fusion of male and female elements, and the question naturally arises, is this stimulus due to the fusion of the cytoplasm of the male cell with that of the female, or is it due merely to the fusion of the respective nuclei ? Experiments upon arti- 58 INTRODUCTION. ficial parthenogenesis, brought about by the use of chemicals and other stimuli, have thrown some light upon the subject, but in the opinion of the author they are, as yet, far from furnishing an adequate solution of the problem. In Marsilia vestita Nathansohn (1900) fovmd that it was possible to stimulate the egg-cell to a parthenogenetic development by exposing the germinating macrospores to a temperature of 35° C. for 24 hours, and allowing them to continue their development at a temperature of 27° C. As a result about 7 per cent, of the spores gave rise to par- thenogenetic embryos. So far as we know, this is the only case among plants above the Thallophyta in which parthenogenesis has been brought about artificially, and it may be that Marsilia lends itself to this sort of experiment more readily because of the fact that in certain species the tendency toward normal parthenogenesis is strongly mani- fested. In Marsilia drummondii Shaw ('97) found normal parthe- nogenesis to be of frequent occurrence. In these cases of Marsilia the morphological side of the question, especially the behavior of the nucleus, is not known, nor have the number of chromosomes been determined in the cells of the parthenogenetic embryo. On the animal side of the question the experimenter finds, fortu- nately, an abundance of most favorable material in the eggs of sea- urchins and of certain marine worms. The results of several investi- gators (Wilson, Morgan, Loeb, and others) have shown that the eggs of Arbacia and Toxopenustes may be made to develop parthenoge- netically through certain earlier stages by subjecting them for a certain time to a solution of sea-water, whose osmotic power is increased by the addition of a solution of magnesium chloride. The action of the Mg-solution seems to be similar to the growth stimulus imparted to the egg by a spermatozoon in normal fecundation. Equally instructive are the experiments of Winkler (1901) on nucle- ated and enucleated fragments of the egg of Cystosira barbata, one of the Fucacece.^ which were fecundated by the spermatozoids. Both the enucleated fragments and those containing the nuclei developed into small embryo plantlets which were exactly alike and attained about the same stage of development. The development of normally fecundated fragments of egg-cells and that of the entire eggs induced to develop parthenogenetically by chemical or physical stimuli are phenomena which seem to fall into the same category. They show that in all probability the growth stinaulus, or the restoration of the power of division and the blending of hereditary characters are phenomena which in a measure are inde- pendent of each other. Experiments similar to the foregoing have SIGNIFICANCE OF THE SEXUAL PROCESS. 59 their greatest value in the suggestiveness of their results and the new points of view to which these results lead. They do not show that the reactions brought about by these stimuli are the same as those resulting from the union of sexual cells. Although the development of a rudimentary embryo induced by artificial means may proceed in the same manner as the product of normal fecundation, yet the arti- ficial stimulus cannot be looked upon as being equivalent to the sexual process. In the case of the former, we are dealing with a stimulus which merely starts growth, but a mature individual is never developed. The sting of an insect or some similar stimulus may call forth a growth in a leaf of an oak, which results in a gall, a local and limited growth, but never in an oak tree, and we cannot for one moment think of comparing such a stimulus to a sexual process. The author does not agree with those who regard the sexual process merely as a restoration to the egg of the power of growth and division. We are not quite ready to lay aside, as yet, the facts won by twenty years of the most careful morphological research for any chemical or electrical theory of heredity. Our knowledge of sexual reproduction in the plant kingdom indi- cates beyond question that that which is of primary significance in the sexual process is the fusion of the nuclei, and the question still remains, which imparts the growth stimulus, the nucleus or the cjrto- plasm of the sperm ? Or are both necessary ? Strasburger has suggested that the stimulus to grov/th and division is given by the cytoplasm, and especially a particular part of the same, the kinoplasm, brought into the egg by the spermatozoid. Some zoologists have attributed this stimulus to the centrosome of the sperm, but in the plant kingdom no case is definitely known in which a centrosome is brought into the egg by a spermatozoid. The doctrine of Strasburger is perhaps the best that has been proposed, and it seems to have some basis in fact. According to this view the egg is rich in food material, trophoplasm, and poor in kinoplasm, while in the sperm the reverse obtains. The unfecundated egg is incapable of developing, therefore, on account of the lack of energy. This theory, however plausible it may seem, leaves much to be desired. In the first place, it is not known as a fact that the egg is poor in kinoplasm, and that the sperm is correspondingly rich in that substance. In many cases the quantity of cytoplasm of the male cell is so small that it seems almost incredible that it could have such a powerful influence. The spermatozoid of the fern, for example, con- sists of a relatively very small amount of cytoplasm, and the kino- plasmic part of this constitutes an organ of locomotion. Although 6o INTRODUCTION. cytoplasmic band and blepharoplast, or cilia-bearer, enter the egg, yet their function seems to be of secondary importance as compared with that of the nucleus. Again in the higher seed-bearing plants, the generative nuclei are accompanied by only a small portion of cyto- plasm, which cannot be recognized in the embryo-sac, and it seems reasonable that it is merely absorbed as so much food. However, when we remember that in all cases of fecundation at least some cytoplasm accompanies the male nucleus into the egg, there is good ground for the belief that the cytoplasm plays some important r61e, but whether that be anything more than to assist in restoring the power of growth and division must at present remain a question. The behavior of the sexual nuclei during the process of fecundation and the wonderful phenomena of karyokinesis point to the conclusion that the nucleus is the bearer of hereditary characters, and that the blending of these characters in the offspring are largely the result of the fusion of the sexual nuclei. The nuclear fusion is also the basis of all hereditary variation. CHAPTER II.— FECUNDATION; MOTILE ISO- GAMETES. ULOTHRIX AND HYDRODICTYON. There seems to be no question that the simplest and most primitive form of sexuality consists in the union of motile isogametes as found among many of the most primitive algae. The chief difference be- tween the gametes of such forms as Pandorina and Ulothrix^ for example, and their asexual swarm-spores, from which the gametes were undoubtedly derived phylogenetically, seems to be merely phys- iological. Generally speaking, the gamete is incapable of developing into a normal adult individual. It must unite first with another gamete of the same species in order to restore the power of growth and divis- ion necessary to the development into an individual common to the species, and apart from theoretical considerations (I refer to the num- ber of chromosomes which, of course, has not been determined for these lower forms) this is the most fundamental distinction made. Many other well-known forms among the green algae might have been taken as representatives, instead of the two selected, but these have been chosen because the development of the reproductive cells from the mother-cell has been more carefully worked out here, and because the processes in this development are coming to be regarded as more important from a genetic standpoint. In connection with Ulothrix I have selected Hydrodictyon in order to present the cytological processes preparatory to the formation of gametes in uninucleate as well as in multinucleated cells. The cytological development, leading to the formation of gametes and also asexual swarm-spores among the simpler representatives of the green algae, has been investigated by a number of earlier observers, among whom were Alexander Brown, Cohn, Pringsheim, Dodel, Strasburger, Klebs, and lately by Timberlake. The well-known and widely distributed Ulothrix consists of a simple unbranched filament differentiated into base and apex (Fig. 17, A). The cells, except the basal one, which is modified as an organ of attachment, are quite alike. Each contains a single nucleus and a band-shaped chloroplast in the form of an almost complete hollow cylinder. Almost any vegetative cell of the filament save the basal one may, without undergoing any external modification, function as a gametangium. 62 FECUNDATION ; MOTILE ISOGAMETES. The process of cell-formation by which the gametes are devel- oped from the protoplast of the gametangium has been observed and described in some detail by Dodel ('76) and by Strasburger ('92). These authors agree that the gametes arise not by the process of free cell-formation, as understood at the time, but by successive bipartitions of the entire plasmic contents of the cell. According to Strasburger ('92) the process of division in the development of the swarm-spores, which is exactly the same for the gametes, differs from the beginning in a very marked way from the vegetative cell-divisions. At first the cell-contents undergo apparently a sort of rejuvenescence by which the protoplast becomes independent of both the outer and inner plasm D Fig. 17. — Ulothrix zonata. — (After Dodel-Port.) A, young plant. B, Zoosporangia, showing escape of swarm-spores. C, an asexual swarm-spore. D, gametangia, showing gametes and escape of same. E, two gametes. F, G, copulation of gametes. H, zygote. J, zygote after a period of rest. K, germinating zygote whose contents have divided into a number of swarm-spores. membranes. In the first division the granular plasma only is halved. The outer plasma membrane (Hautschicht) is undivided, and the membrane surrounding the vacuole remains unchanged. By further successive divisions these two protoplasts give rise ultimately to the gametes. The process of division is the same whether gametes or asexual swarm-spores result (Fig. 17, D). Strasburger has expressed the opinion that, in the development of the gametes, only one more cell-division is necessary above those required for the zoospores, and this division renders the resulting cells or gametes incapable of further independent development. In what way this last division incapaci- tates the gametes for further independent development was not dis- cussed at the time. That view was probably prompted by Weismann's ULOTHRIX AND HYDRODICTYON. 63 theory of a reduction division of the chromosomes, which at the time received a wider acceptance than at present. In Hydrodictyon Klebs ('91) affirms that the process of cell-forma- tion, giving rise to gametes or asexual swarm-spores, occupies an intermediate position between simultaneous and successive cell-division. From what follows it will be seen that the process is a cleavage similar to that occurring in certain Phycomycetes, but, using the methods that he did, Klebs failed to perceive the true nature of the process. His account in substance is as follows : The first indication of cleavage is manifested in the appearance of numerous small clefts, pointed at the ends, in the plasma layer con- taining the chlorophyll (Fig. 18, A). This can be seen in material cultivated in darkness in a maltose solution, especially after the appli- cation of a weak plasmolysing agent. These clefts soon become longer and more numerous, neighboring ones thereby uniting with each other, so that finally the entire chlorophyll-bearing layer is seg- mented into pieces which are still connected, however, by fine plasmic threads. The cleavage is not confined solely to the chlorophyll- bearing layer, but extends into the colorless plasma in which the nuclei are situated. The plasma membrane and the wall of the vacuole are, on the contrary, unaffected. Previously to and during the cleavage the plasmic layer concerned frequently undergoes a contraction, thus giving rise to colorless spaces, so that this layer appears as a coarse net, as Pringsheim ('71) has described for Bryopsis. These spaces contain also some plasma, and, as the plasma membrane and wall of the vacuole are continuous, the entire cell contents form still a unit, as shown by plasmolys^is. The continuation of the cleavage results in the segmentation of the plasmic contents into numerous bands with irregular and sinuous contour (Fig. 18, B). These bands undergo still further segmentation (Fig. 18, C), until finally the plasmic con- tents are broken up into numerous small pieces, each containing a nucleus, which ultimately separate and develop into gametes (Fig. 18, D). The method of division in these portions referred to in Fig. 18, B, C (Klebs continues), appears to consist in a constriction, progress- ing from one side, but not entirely completed, since the individual parts remain in communication ; yet direct observation shows also that, in the plane of division, a colorless line or furrow is frequently present, which gives the impression that the constriction may proceed from within. The same principle operating in the segmentation of the bands or pieces obtains also in the earlier cleavage of the whole plasmic layer of the cell. There is from beginning to end a progres- sive condensation, but the process that plays the chief r61e is concealed from observation. 64 fecundation; motile isogametes. Using improved methods Timberlake ('oi) in a study of spore- formation in Hydrodictyon uiriculatum Roth., has found that, in the earlier stages of the process, cleavage takes place by means of surface constrictions of the plasma membrane on the outside and the vacuolar membrane on the inside of the protoplasmic layer, as may be seen from Klebs' figures (Fig. i8, B, C). The process is a progressive one, the cleavage furrows cutting out first large irregular multinucleated masses of protoplasm, which are in turn divided into smaller ones, until each Fig. i8.— Cell-cleavage in Hydrodictyon utriculatum.—{Mttt Klebs.) A, cell showing cleavage furrows at early stage in the process ; e, place in protoplasm free from chlorophyll. B, sausage-shaped protoplasts formed in early stage of cleavage. C, two protoplasts similar to those in B, showing manner of further cleavage. D, final result of the cleavage. contains a single nucleus. In this manner the entire protoplast is divided into uninucleated spores or gam- etes, as the case may be. Judging from Strasburger's account of the process in Ulothrix^ it seems probable that cell-formation leading to the development of gametes or swarm-spores is also a cleavage similar to that in Hydrodictyon. In Ulothrix, however, the cells are uninucleate, and a nuclear division must either accompany or precede cell-division. Until the behavior of the nucleus is known, and the process carefully worked out with the aid of more impi'oved methods, the exact nature of the cell-formation in question must remain largely a matter of conjecture. In the light of more recent investigations concerning cell-formation among the lower thallophytes, it is evident that our present knowledge of this process in connection with the development of gametes or asexual zoospores among the algae is very meager and fragmentary. COPULATION OF GAMETES. — ECTOCARPUS. 65 COPULATION OF GAMETES. The gametes of Ulothrix zonata are rounded or oval cells, bearing two cilia at the anterior end (Fig. 17, E). Each contains a nucleus, a red eye-spot, situated about midway between the ends of the cell near the surface, and a chromatophore. According to Strasburger the cilia are developed under the influence of the nucleus and from the anterior, colorless portion or mouth-piece, which consists mostly of kinoplasm. In his later investigation of the subject of swarm cells, Strasburger (1900) finds that the cilia arise from a local kinoplasmic thickening of the plasma membrane at the anterior end. As already mentioned in a preceding paragraph (p. 47), he regards this thicken- ing as the homolog of the blepharoplast of the Archegoniates. In the swarm-spores of Hydrodictyon^ Timberlake finds a small body at the base of the cilia, which, in some cases at least, was not a part of the plasma membrane. The gametes copulate in pairs immediately after they escape from the gametangium (Fig. 17, F, G). It is probable that they may be brought together, or at least held together after coming in contact, by means of a chemotactic stimulus. The stigmatic or eye-spots do not unite, but remain separate and independent in the young zygote (Fig. 17, H). There is no doubt of a nuclear fusion, but how soon this takes place after conjugation is not known, so far as the author is aware. In Hydrodictyon the gametes are small, oval in shape, biciliate, containing one nucleus and, according to Klebs, two pulsating vacuoles. They conjugate in pairs immediately on escaping from the gametan- gium, but I have observed that conjugation may sometimes take place within the mother-cell. If, however, copulation does not follow soon after the gametes are set free, they become incapable of uniting, come to rest and disorganize. Whether this is a rule was not determined. ECTOCARPUS. Among the isogamous Phceophycece the sexual process is doubtless best known in Ectocarpus siliculosus Lyngb. from the investigations of Berthold ('8i), which have been confirmed and extended by Oltmanns (*99) . Ectocarpus is of especial interest in this respect, since it repre- sents a transition from isogamy to heterogamy. In fact, there is in the brown algae, as well as in phylogenetic series of other Thallophyta^ every transition from the type of gametes found in Ectocarpus to that of Fucus. The gametes, although nearly or quite the same size and appearing morphologically alike, are physiologically different, and we may, with much propriety, speak of egg-cells and spermatozoids. 66 FECUNDATION ; MOTILE ISOGAMETES. Both Oltmanns and Berthold agree in the opinion that Ectocarfus siliculosus may be either monoecious or dioecious, for they observed individuals whose gametes would not conjugate with each other, but only with those of another individual. As is well known, the gametes are generally borne in the so-called plurilocular sporangia. The details in the process of nuclear and cell-division in the development of both gametes and asexual swarm-spores have not, as yet, been thoroughly studied. The gametes (Fig. 19, A) are pear-shaped cells with a chro- matophore, nucleus, a reddish brown eye-spot, and two cilia inserted laterally. The cilia are of unequal length, the longer extending for- ward and the shorter backward. The conjugation of the gametes can be most readily followed in a hanging drop, into which both male and female gametes are intro- duced, when the whole process may be observed with the aid of the highest magnifying powers. The female gametes, as a rule, first come to rest, and about each one numerous spermatozoids assemble. If the female gamete comes to rest at the edge of the drop, the male cells cluster about it, attaching themselves apparently by the anterior cilium, giving the familiar picture figured by Berthold (Fig. 19, A). But should the female gamete attach itself to some particle hanging in the arched surface of the drop, this cell then appears as a circular disk surrounded by a wreath of male cells radially disposed. Shortly a male gamete (in exceptional cases two), having attached itself to the female by means of the anterior cilium, approaches the latter appar- ently by the sudden contraction of the same and unites with it, while the remaining male gametes withdraw (Fig. 19, B, C). In a few minutes cytoplasmic union is complete, and within about ten hours after copulation both nuclei have fused (Fig. 19, E, F, G). The chloroplasts do not unite, a fact which is contrary to the peculiar phenomenon described by Overton for Sfirogyra (see page 69). The sexual process in Ulothrix^ Hydrodtctyon, and Ectocarpus may be considered as fairly typical of the lower algae in which fecun- dation consists in the fusion of motile isogametes. In this, probably the simplest and most primitive sexual process, as in the higher plants, it will be seen that fecundation consists in the fusion of the sexual nuclei together with the cytoplasm of the gametes, but the fusion of the nuclei must be regarded as of prime importance. CHAPTER III. -FECUNDATION ISOGAMETES. NON-MOTILE In this chapter will be discussed the sexual process in several forms in which the gametes are non-motile, i. e., they do not escape from the parent plant and move about in the surrounding media, and are either unisexual or show a certain degree of bisexuality, as in Basidio- bolus. The forms used, Spirogyra^ Cosmarium and Closterium among the desmids, certain diatoms and Basidiobolus^ have been chosen solely because the development of the gametes and their union have been most thoroughly investigated in certain species of these genera. Owing to the conflicting results obtained by the several investigators in the much-studied Sporodinta^ the process in this plant, which properly belongs here, will be only incidentally referred to. Fig. 19.— Copulation of gametes in Ectocarpus siltculatus. A, female gamete with numerous male gametes attached, seen from the side. B, C, D, E, successive stages of cytoplasmic fusion.— (After Berthold.) E, F, G, fusion of nucleus.— (After Oltmanns.) SPIROGYRA. Among the algae Spirogyra undoubtedly furnishes the best known illustration of the sexual process in which the gametes are isogamous and non-motile. The process as observed in the living plant has been carefully described long ago by DeBary ('58), Strasburger ('78) and others, and it is now a matter of common observation in almost every botanical laboratory. The nuclear behavior, which cannot be fol- lowed in the living specimen, and which is the most essential part of the process, has received comparatively little attention. Morphologically and physiologically every cell of a Spirogyra filament, except those serving as organs of attachment, is exactly like every other cell, so that the filament may be regarded, in a sense at 68 FECUNDATION ; NON-MOTILE ISOGAMETES. least, as a colony of individuals. Any cell of a filament, save those mentioned, may function as a gamete. In sexual reproduction cells of two filaments lying close side by side send out protuberances toward each other which meet end to end. In the contiguous membranes a circular opening is made by the dissolu- tion of the cellulose walls, through the agency of an enzyme, whereby a continuous canal is formed between the cells (Fig. 20, A). It is highly probable that the conjugating tubes are brought together by the aid, of a chemotactic, directive stimulus. Haberlandt ('90) claims, and his view is shared by Klebs ('96), that the conjugating cells exert a mutual chemical influence upon each other, namely, that a cell will put out a conjugating tube only when influenced by another, probably of a different sex, lying near it. In support of this view, Klebs found that cells of individual filaments cultivated upon agar-gelatin, although having been brought side by side by the folding of the filament, never put out conjugating protuberances. A single male filament, on the contrary, may conjugate with several female filaments whenever their cells lie sufficiently, near one another, but all those cells of the male filament separated some distance from those of the female remain sterile in spite of the tendency to conjugate. The limits of this mutual action of the filaments (Haberlandt, '90) is equal to a distance of two or three diameters of their cells. Slightly beyond this limit the cells may put out short conjugating tubes, but these never reach each other, the stimulus being presumably too weak. Haberlandt states further that the conjugating tubes are not laid down simultaneously, but rather one sends out a protuberance which calls forth the development of the cor- responding tube from the other cell. If the protuberances do not lie exactly opposite, they bend slightly in order to meet each other. A further action of the stimulus is seen when a long male cell copulates with two female cells. Two canals are formed connecting the male with the two female cells, but, of course, only one of the latter receives the gamete. In some species, especially Spirogyra inflata^ according to Klebs, the meeting of the conjugating protuberances is facilitated by a curving or a knee-like bending of the cells, from whose convex sides the protuberances arise. These phenomena are not presented in this connection for the purpose of discussing any special phase of the physiology of the sexual process, but merely to indicate a few features manifested by unisexual elements which show a tolerably well-marked tendency toward bisexuality. When the conjugation canal, joining the gametes, is complete, the turgor in each cell is diminished, so that each protoplast experiences a self-plasmolysis. The contraction usually takes place first in the male SPIROGYRA. 69 gamete, which passes through the canal to unite with the stationary or female gamete (Fig. 20, A). Strasburger ('78) has observed that occasionally the female cell was the first to round up. Haberlandt suggests that the extrusion of water is connected with a mutual stimulus between the cells, for the female gamete contracted only when the male was normal, and, furthermore, the male cell became self-plasmolyzed only when connected with a female cell. The principle underlying the movement of the male gamete through the canal is not well understood. Overton ('88) held that a gelatin- ous substance was secreted, which, upon swelling, forced the proto- plast through the canal. The presence of a mucilaginous substance Fig. 20. — Fusion of gametes in Spiroiyra. A, portion of two conjugating filaments of Spiragyra quinina. — (Aft«r Strasburger.) B, young zygote provided with only a thin wall. C, zygote at a later stage ; the cell-wall is thicker, and the nuclei hav« united, but thcMiucleoli have not fused. has not been demonstrated, however, and it is highly probable that we have to do here with an active plasmic movement operating under the chemotactic stimulus of the two protoplasts. Here fecundation con- sists in the union of the entire plasma of both gametes, though DeBary records the case of Spiroiyra heeriana} in which a small vesicle of plasma is left beyond the partition wall in the conjugation canal. Concerning the behavior of the chlorophyll bands in the zygote, much diversity of opinion exists. DeBary ('58) and Schmitz ('82) observed that in species with one chlorophyll band the two chloro- plasts united in the zygote to form one continuous band. Overton ('88) , on the contrary, asserts that the single band of the female gamete segments at the middle during the fusion of the protoplasts ; the two halves then separate, and each piece unites with the ends of the band ' See Fig. lo, p. 17, Die Naturlichen Pflanzenfamilien, i Theil, z Abtheilung. yo FECUNDATION ; NON-MOTILE ISOGAMETES. furnished by the male gamete. Chmielewskij ('90) finds that in all of the several species examined the chloroplast of the male gamete is dissolved in the zygote, that of the female only remaining. The behavior of the nuclei during fusion cannot be followed with any degree of certainty in the living specimen. As a rule they cannot be seen at all, a fact which led to the view of the earlier observers that the product of union was without a nucleus. One must, therefore, resort to thin and well-stained sections of properly fixed material to observe the details of nuclear fusion. For this purpose I have selected a small-celled species with one chlorophyll band. When the young zygote is provided with a thin cell-wall, the two nuclei, which are exactly alike, judging from their appearance, are seen lying closely applied to each other (Fig. 20, B). Each contains a rather large and distinct nucleolus and the characteristic linin net in which are imbedded small granules that behave toward stains as chromatin granules in resting nuclei of higher plants. In fact, the nuclei of Spirogyra in this condition seem to possess the same structure as the phanerogamic nucleus. The contiguous parts of the nuclear membranes dissolve or disappear as such, and the network of the one unites directly with that of the other, the fusion of the nucleoli following later (Fig. 20, C). Frequently, before complete union of the nuclei, the wall of the zygospore may become much thickened and less easily penetrated by fixing fluids, so that perfect preparations are diflficult to procure. During the development of the zygospore the chloroplasts become vacuolate and the identity of each cannot be made out. . In the preceding paragraphs I have described the nuclear fusion in the zygote as I was able to follow*5t, but for lack of time and suitable material an exhaustive study of the subject was not made, and conse- quently I am not prepared to state whether the peculiar behavior of the nuclei as described by Chmielewskij ('92) for Spirogyra crassa and S. elongata is correct. Chmielewskij states that, as the gametes round up, the nuclear membranes become less distinct, disappearing entirely as the gametes unite. The nuclei now fuse, the fusion being complete by the time the zygote is provided with a thick, dark wall. This fusion takes place during the prophase of division. As soon as fusion is complete the nucleus divides. The daughter nuclei now divide, four nuclei resulting. Two of these then fuse, while the other two divide by direct division and finally disorganize. The fusing nuclei are provided with membranes and are in the resting condition. If the observations of Chmielewskij be true, the process in Spirogyra is without parallel in the plant kingdom, at least so far as the author is aware. SPORODINIA. — CLOSTERIUM AND COSMARIUM. 71 SPORODINIA. Morphologically considered, the sexual process in Sporodinta grandis and in other typical Zygomycetes seems to be similar to that in the Conjugatece^ but in Sporodinia the gametes are multinucleate, and the behavior of the nuclei in the young zygote varies considerably, according to the accounts given by the different obsei-vers. After the cytoplasmic fusion of the gametes, the nuclei of each arrange them- selves into a spherical layer surrounding a globule of oil, and then fuse, producing a hollow sphere full of oil, which L6ger ('95) has called an embryonic sphere {sphere embryonnaire) . These embryonic spheres lie near the poles of the zygote. During the germination of the zygospore the two embryonic spheres fuse. The fused mass reveals numerous nuclei, which pass into the sporajigiferous mycelium and begin to divide. In the azygospore only one embryonic sphere is developed. Wager ('99) regards the union of the nuclei to form the embryonic sphere as the sexual act, and the azygospores are, there- fore, truly sexual, the process of conjugation being of secondary importance. Dangeard ('94, '95) does not accept Lager's interpreta- tion of the embryonic spheres, holding that the fate of the nuclei has not been determined. According to Gruber ('01) no embryonic spheres are to be seen in the newly formed zygote. The numerous nuclei, on the contrary, are uniformly distributed throughout the cytoplasm. After five or six weeks the same condition of things was still found to exist, and what took place finally among the nuclei Gruber was unable to determine. Neither fusion, disorganization nor division of the nuclei was observed even six months after the fusion of the gametes. From what is now known concerning the sexual union of multinu- cleate gametes in other groups of plants, in which the sexual process has been unmistakably followed in every detail, it is very probable that a multiple fusion of the nuclei in pairs obtains also in Sporodinia.^ CLOSTERIUM AND COSMARIUM. In the desmids the process of fecundation agrees essentially with that described by myself for Spirogyra^ except as regards the time of the fusion of the sexual nuclei and the behavior of the chromatophores in the zygospore. During the development of a firm cell- wall about the zygote, according to Klebahn ('91), the chromatophores undergo a marked change, the result of which is the formation of two large rounded balls, which are at first rich in starch and of a yellowish color. The part taken by the four original chromatophores in the » Sec Chapter III, Albugo Bliti, and Chapter IV, Pyrontma. 7a FECUNDATION ; NON-MOTILE ISOGAMETKS. formation of these balls was not determined. The union of the nuclei, which are in the resting stage, does not take place until the germina- tion of the zygote. The behavior of the fusion nucleus, although somewhat beyond the province of our subject, is of such a nature as Fig. 21. — Fusion of sexual nuclei during germination of the zygote in Closteriutn. — (After Klebahn.) A, mature zygote with two large chloroplasts, the two sexual nuclei in contact. B, beginning of germination ; the sexual nuclei have fused while in the resting condition. C, contents escaping from old wall of zygote ; fusion nucleus in prophase of division. D, protoplast free from wall of zygote, fusion nucleus in anaphase of division. E, daughter nuclei reconstructed, division of cell begun. F, spindle stage of second mitosis ; the nuclei lie on opposite sides of cell near periphery. G, second mitosis complete. H, cell-division has taken place; in each daughter cell one of the two nuclei is much smaller and denser; the two large nuclei are provided with membranes. I, the daughter cells have begun to assume form of adult cell ; in each the large nucleus which persists as the nucleus of the cell, has taken a central position ; the smaller one lies near one end of cell. to merit attention, especially in connection with the nuclear behavior previous to the sexual process in the diatoms to be mentioned below. The union of the sexual nuclei in Closterium and Cosmariutn^ CLOSTKRIUM AND COSMARIUM. DIATOMS. 73 according to Klebahn, occurs just prior to the escape of the contents of the zygote from the outer membrane (Fig. 21, A, B) . During the latter process the fusion nucleus often shows signs of approaching karyokinesis (Fig. 21, C). There now follow two karyokinetic divisions in rapid succession, so that each daughter cell may contain two nuclei (for a cell-division may also have taken place) one of which remains as the nucleus of the daughter cell, while the other gradually undergoes disorganization (Fig. 21, D, E, F, G, H, I). (See expla- nation of figure for details.) It will now be seen that the process in the zygote of the desmids differs from that described for Spirogyra by Chmielewskij (see p. 70) : (i) in the fusion of the sexual nuclei in the resting stage ; (2) in that there is no second fusion of two of the four daughter nuclei, but a cell-division, one nucleus going to each of the daughter cells. /T:-^ B C Fig. 22. — Formation of gametes in Rhopalodia gibba. — (After Klebahn.) A, protoplast of cell showing first mitosis ; nucleus in spindle stage. B, second mitosis, each daughter nucleus dividing. C, second mitosis complete, the four nuclei about equal in size. D, part of two conjugating individuals ; the protoplast of the one on left has begun to divide by becom- ing constricted in the middle; two nuclei in each cell are large, other two have become smaller. E, cell-division complete. DIATOMS. In the diatoms the type of isogamous fecundation resulting in the formation of the auxospore recalls the nuclear history subsequent to fecundation in the desmids. As in the case of the desmids we are indebted also to the investigations of Klebahn ('96) and to those of Karsten (1900), for a more accurate knowledge of the nuclear behavior preceding the sexual act. The nuclear activity, which immediately precedes conjugation, is of prime importance here, and it is to this that our attention is especially directed. In Rhopalodia, the form studied by Klebahn, two individuals place themselves side by side, being held together by means of mucilaginous masses. The protoplast of each cell, which contains one nucleus and 74 FECUNDATION ; NON-MOTILE ISOGAMETES. in general two pyrenoids, undergoes a rejuvenescence and finally divides. Prior to this cell-division, however, two successive mitotic divisions of the nucleus^ take place (Fig. 22, A to E). After the first mitosis the daughter nuclei generally move apart toward the ends of the cell whither the pyrenoids also wander (Fig. 22, B). Soon the second mitosis takes place, when four nuclei similar in appearance are present in the protoplast, which may, as yet, show no sign of division (Fig. 22, D). With further progress the protoplast in each individual becomes constricted near the middle and finally divides, two daughter nuclei passing into each daughter cell, which contains one or some- FiG. 23. — Conjugation of gametes in Rhopalodia gibba. — (After Klebahn.) F, conjugating pair seen from valve side ; protoplast of each has divided into two daughter cells or gametes ; each gamete contains, besides the large pyrenoid, a large and a small nucleus. G, cytoplasmic fusion of the two pairs of gametes has taken place; the small nuclei are scarcely recog- nizable. H, a later stage ; the small nuclei have entirely disappeared, while the two functional nuclei in each zygote, which has now changed from a dumbbell to an elongated form, have come nearer together. times two pyrenoids and a chromatophore (Fig. 22, E). A marked change is now manifested in the nuclei. Of the two nuclei in each daughter cell, one increases in size while the other diminishes, becom- ing dense and contracted (Fig. 22, D, E). The next step in the pro- cess is the conjugation of the daughter cells of one individual with those of the opposite one by means of protuberances sent out from the respective cells (Fig. 23, F, G). The large nucleus of each * For details of mitosis see the original paper of Klebahn, '96. DIATOMS. 75 cell, followed by the pyrenoid, passes into the isthmus or connecting portion of the dumbbell-shaped zygote, which soon becomes cylindri- cal or crescent-shaped, and scarcely a trace of the small nuclei are to be seen (Fig. 23, H). During the development of the zygote into an auxospore, the two large functional nuclei assume the structure characteristic of the resting stage (/. e., each presents a granular frame- work and a definite nucleolus) and fuse. The fusion does not take place in every case at a certain developmental stage of the two auxo- spores, but may occur earlier in one than in the other (Fig, 24, I, J). As a rule, however, the fusion is complete when the siliceous valves have begun to develop. The behavior of the small nuclei would seem to indicate that they are utilized as food. A slightly different process, leading to the production of the auxo- spore, is met with in Cocconeis placentula Ehr., as described by Karsten (1900). In this species the protoplasts of the conjugating cells do not divide, and, therefore, only one zygote results. In each cell there is also but one division of the nucleus instead of two as in Rhopalodia. Preparatory to the cytoplasmic union the protoplast of each cell contracts. Each cell is seen to possess two nuclei, one large and one small, so that nuclear division must have taken place at an earlier stage. During the contraction mentioned each protoplast sur- rounds itself with a gelatinous envelope. Near the point of contact of the two individuals the two halves of each shell separate slightly. From the opening in one of the cells, which is regarded as the male gamete, a small papilla protrudes, which grows toward the opening in the female cell, and the gelatinous envelopes are soon in open communi- cation. The entire protoplast of the male cell now passes through this narrow channel into the female cell. The young zygote then increases considerably in size, and begins the formation of a firm cell- wall about itself. Of the four nuclei only the two large ones are now to be seen, the smaller ones having gradually disappeared. The two large functional nuclei, each with a nucleolus, begin to fuse slowly, and, by the time the shell of the zygote is fully formed and the two chromatophores are reduced to one, fusion is complete. From the foregoing it is clear that the nuclear behavior immediately preceding the sexual act in Rhopalodia is strikingly analogous to the process following fecundation in Closterium and Cosmarium. Whether these processes bear any closer relation to each other than mere analogy is a difficult question. It may be suggested that, in the case of the diatoms, we have to do with the development of two perfect gametes in each cell instead of four, a process similar to that in certain Fucacece^ where only part of the egg-cells in the oogonium mature, 76 FECUNDATION ; NON-MOTILE ISOGAMETES. the others being disorganized ; and in the desmids only two out of the four in the germination of the zygote develop into perfect cells. It is not known whether the reduction in the number of chromo- somes, if a reduction actually occurs in either desmids or diatoms, is in any way associated with the nuclear divisions in question, as has been assumed by some authors (see Wilson, "The Cell," p. 198) ; consequently, in the light of our present knowledge, it cannot be said with any certainty that these nuclear divisions represent a preparation for the sexual act, that in the diatoms taking place just before fecun- dation while in the desmids it occurs at the beginning of an ontoge- netic development. BASIDIOBOLUS. A sexual process similar to that in the Conjugatecs is found in Basidiobolus, one of the Phycomycetes. I have selected Basidio- Fio. 34.— Fusion of the sexual nuclei in Rk0f»~ lodia £ih6a.—{A.f\.e.r Klebahn.) I, the two young zygotes or auxospores have elon- gated and begun to assume form of adult; sexual nuclei now in contact. J, middle portion of two auxospores, each with a fusion nucleus. bolus ranarum because of its close re- semblance to certain Mesocarfacece^ especially Mougeotia^ both in structure (the cells possess only one nucleus) and in the sexual process, and because the development of the sexual organs and the fusion of the gametes are well known in detail. Sex- uality in this genus has recently been subjected to a critical study by Fairchild ('97)? whose results form the basis of the following account. Two neighboring cells of a filament send out near the transverse wall a beak-like protuberance, into which the nuclei of the respective cells pass (Fig. 25, A). The nucleus in each of the protuberances now undergoes a karyo- kinetic division, which is followed by the formation of a transverse BASIDIOBOLUS. 11 wall cutting off a small cell at the end of the beak (Fig. 25, B). The manner in which this wall is laid down is worthy of special notice here, since it is formed as in the higher plants, namely, through the instrumentality of the kinoplasmic connecting fibers, appearing at first as a cell-plate. Apart from Chara this is the only instance as yet known among the lower cryptogams in which a cell-plate is thus formed. Immediately the nuclei have entered the beaks, and prior to the prophase of the nuclear division just mentioned, and also before an increase in size of the female gamete, a hole is formed in the transverse wall separating the two gametes. The two daughter nuclei cut off in the ends of the beaks gradually disappear, while the other two pass down deeper into the cytoplasm Fig. 25.— Formation of gametes in Basidiobolut ratutrum Eidam.— {After Fairchiid.) A, two gametes showing the beaks; the nuclei, which are in the beaks, are in the resting condition ; the hole has already formed between the gametes. B, the nuclei have divided and two of the daughter nuclei are cut off in the ends of the beaks, while the other two, which have increased in sire, have passed down near the opening in the transverse wall ; the female gamete has increased greatly in size, the male retaining its former dimensions. of the cells (Fig. 25, B). The male nucleus now passes through this opening and comes in contact with the female nucleus (Fig. 26, C) . During these movements the nuclei attain their original size, and each contains one or more interwoven nuclear threads, in which chromatin granules are situated at rather long intervals. In this condition the two nuclei remain some time before fusing. The entire cytoplasm of the two gametes is utilized in the formation of the young zygospore, which now forms about itself a very thin wall, within which the thick endospore, consisting of several layers, is gradually developed. Owing to the difficulty with which fixing fluids penetrate the thick wall of the zygote the exact time of fusion of the male and female nuclei is not easily determined, but as the zygospore approaches maturity the fusion is complete, so that no trace of male and female 78 FECUNDATION ; NON-MOTILE ISOGAMETES. nuclei can be distinguished (Fig. 26, D). According to Raciborski ('96) the fusion may be delayed until the germination of the zygote. The full significance of the formation of the beaks into which the nuclei wander, the division of the latter, and the cutting off of the small cells which degenerate, can be more fully understood only after the process of sexual reproduction is known in other and related forms. The two small cells cut off in the ends of the beaks may, however, be Fig. 26. — Fusion of sexual nuclei in Basidiobolus ranarum. — (After Fairchild.) C, the sexual nuclei are in contact. D, zygote with fusion nucleus and thick cell-wall. reasonably regarded as degenerate gametes, although it may seem idle to attempt to explain or to bring into line the various peculiar phenom- ena brought out in the several preceding paragraphs that pertain to the desmids, diatoms, Basidiobolus and Spirogyra. In the desmids, diatoms and Basidiobolus^ it is possible that all these phenomena may have resulted independently from similar causes acting during a large part of the phylogenetic history of the respective groups of plants. CHAPTER IV.— FECUNDATION ; HETEROGAMETES. In the preceding chapters we have considered sexual reproduction in certain of those Thallophyta in which no very marked differentia- tion of the gametes has been attained, although in Bctocarpus espe- cially, and even in Sfirogyra and Bastdiobolus, a tendency toward a differentiation into male and female cells is manifested. Nor have we found any modification of the cells bearing the gametes into dif- ferentiated sexual organs, unless the gametangia of such forms as Ectocarpus be so considered, and even then there is no apparent difference between male and female gametangia. As already men- tioned in the introductory chapter, the terms male zx^^ female sexual cells are essentially the expression of a certain fundamental kind of division of labor, and in the developmental history of sexuality in plants we find this division of labor manifested in the gametes them- selves before a corresponding differentiation is apparent in the organs bearing them. SPH^ROPLBA. Among the algae one of the best known and most interesting exam- ples of this fact is illustrated in Sphceroflea annulina. To Ferdinand Cohn ('55) is due the credit of having established the fact of sexual reproduction in this genus, a phenomenon among the algae little known at the time. Later Sphceroplea was studied by Heinricher ('83), Rauwenhoff ('88), Kny ('84) and more recently by Klebahn ('99). Although both Heinricher and Rauwenhoff followed the behavior of the nucleus during certain stages in the development of the sexual cells and in fecundation, yet in many respects their work was incomplete. For a more thorough investigation of this process, however, we are indebted to the researches of Klebahn, who studied the two varieties of the species, S. annulina var. braunii (Keutz) Kirchner and S. annulina var. crassisepta Heinricher. The chief interest in the sexual repro- duction of this plant centers upon the fact that in var. braunii several nuclei are usually present in the egg-cell. The contents of the multinucleate cells of Sphcsroplea present the well-known and characteristic arrangement : In typical cases the cen- tral cavity of each cell is traversed by a row of large vacuoles inter- spersed by smaller ones of varying size. The protoplasm, which forms only a thin layer between the larger vacuoles and the cell-wall, is collected into dense ring-like or band-shaped masses between the So FECUNDATION ; HETEROGAMETES. former. These plasmic rings or diaphragms communicate with each other by plasmic strands or bridges. In the plasmic rings are located the rounded chloroplasts, pyrenoids and the nuclei. Of the latter the number in each ring varies from 3 to 20 in var. draum'i and from i to 4 in var. crassisepta (Fig. 27, A). In those cells in which spermatozoids are developed the nuclei undergo four or five karyokinetic divisions,^ so that ultimately about 300 small nuclei are present in each band (Fig. 28, A to F). During these divisions the pyrenoids disappear, and the chromatophores undergo several divisions and assume a pale, yellowish-brown color. Fig. 87. — Cell-cleavage leading to formation of egg-cells in SpheeropUa braunii. — (After Klebahn.) A, outer view of a protoplasmic ring of a vegetative cell, showing chromatophores, pyrenoids and nuclei. B, portion of an odgonium showing frothy nature of protoplasm and early stages of cleavage. C, small portion of oogonium, showing irregular protoplasts resulting from cleavage, which contain several nuclei and pyrenoids. The plasmic rings up to this time retain their original form. Now the cytoplasm segments into numerous protoplasts, the spermatozoids, in such a manner that each spermatozoid receives only one nucleus (Fig. 29, I, J, K, L). The mature spermatozoids (var. crassisepta) are as a rule spindle-shaped, being smaller at the anterior end, which bears the two cilia. Near the middle lies the veiy small and densely staining nucleus (Fig. 29, L). Kny in his Wandtafel, Lxiir, figui-es four or five yellowish chromatophores in each spermatozoid. The processes leading to the formation of the egg-cells show a marked difference from those taking place in the antheridium. Even ' For details of karyokinesis see Klebahn, '99. SPH^ROPLEA. 8l in the two varieties, as, will be shown, the cleavage is not the same. In var. braunii the ring-like disposition of the protoplasm disappears, while large vacuoles appear, transforming the entire cell-contents into a foamy structure in which larger and smaller strands and masses alternate (Fig. 27, B). In the dense portions of protoplasm nuclei, as well as chromatophores and pyrenoids, are irregularly disposed. Now a cleavage takes place by which the plasmic contents are segmented into irregular protoplasts of varying sizes (Fig. 27, C). These proto- plasts contract (the large vacuoles thereby gradually disappearing) and Fig. a8. — Parts of contents of young antheridia, showing nuclear history preparatory to formation of sperma- tozoids in S. braunii. — (After Klebahn.) A, part of plasmic ring showing two nuclei in prophase of division. B, spindle stage of same mitosis. C, anaphase probably from second mitosis. D, Telophase of a later nuclear division. E, Condition of nuclei between successive mitoses, pyre- noids still present. F, nuclei shortly before formation of spermatozoids ; the pyrenoids have disappeared. round up to form the egg-cells, of which two to four are seen in a cross- section of the cell. Neither shortly before nor during cleavage, according to Klebahn ('99), is there to be observed a division or fusion of the nuclei, so that (contrary to Rauwenhoff who claimed that during the formation of the eggs the number of nuclei was diminished) each egg' may contain, in addition to 2 or more pyrenoids, several nuclei, the number varying from I to 5 (Fig. 29, A to E) . The number of nuclei falling to any egg is largely a matter of chance, since the cleavage planes do not seem to be determined in any way by the number or position of the nuclei in the cytoplasm. * The so-called "giant eggs" are exceptions. 82 FECUNDATION ; HETEROGAMETES. In var. crasstsepta^ whose cells are smaller (narrower) and with fewer nuclei, the process of cleavage differs somewhat. The eggs in this variety contain, as a rule, only one nucleus. When the protoplasm of the oogonium has become frothy, as described for var. braunii^ cleavage planes are formed at right angles to the long axis of the cell, thus separating the contents into a row of short segments.' Here the cleavage follows in such a way that a nucleus will be included in each seg- ment of the cell, although in exceptional cases two nuclei may be included in a segment. In var. braunii we have, therefore, to do with multinu- cleated eggs, while in var. crassisepta each egg-cell is uninucleate. When the egg-cells are mature, small openings are formed in the wall of the oogonium through which numerous spermatozoids enter (Kny, Wandtafel, lxiv). The manner in which the spermatozoids unite with the cytoplasm of the egg was not observed by the authors cited. According to Klebahn ('99) the fecundated egg is readily dis- tinguished by its delicate membrane and by the presence of the sperm nucleus which appears always in sharp contrast to the nuclei of the egg (these resemble vegetative nuclei) as a small, densely staining body about the size of the nucleolus (/. ^., about one micron in diameter) (Fig. 29, A, B). In eggs just fecundated the sperm nucleus lies at the surface beneath the delicate membrane. After a time, the length of which was not determined, the sperm nucleus passes into the interior of the egg, and finally fuses with one of its nuclei (Fig. 29, C, D, E). Before actual fusion the two sexual nuclei remain side by side some time, a phenomenon of very frequent occurrence in the plant kingdom, during which the male nucleus increases in volume, its chromatic sub- stance assuming the form of larger and more distinct granules, until finally the two sexual nuclei can scarcely be distinguished one from the other. The fusion nucleus is easily recognized by its coarsely granular contents, while the other nuclei in the egg appear pale, with a few small granules arranged along the nuclear membrane (Fig. 29, F). From the foregoing it will be seen that in Sph(^roplea annulina var. braunii^ although several nuclei are present in the ^%%^ fecundation consists in the fusion of the spermatozoid nucleus with only one nucleus of the egg-cell. Whether there exists among the several nuclei of the egg any preference in the union with the male nucleus is not known, as there seems to be nothing in the position or appearance of the nuclei which might suggest a preference. The nuclei are irregularly grouped or distributed in the cytoplasm of the egg, and it seems to be purely a matter of chance as to which one will fuse with the sperm nucleus. See Kny'i Wandtafel, lxiy. SPH^ROPLEA. 83 After fusion of the sexual nuclei the oospore develops its character- istic wall (Fig. 29, G, H). Unfortunately Klebahn was unable to trace the fate of the remaining nuclei. Whether they disappear indi- vidually or, after fusion with each other, unite with the fusion nucleus, is a matter of conjecture only. The investigations of Golenken (1900) Fig. 29.— Fecundation of eggs and later development of spermatoroids. A-H, SpkaropUa braunii. I-M, S. crasshe/ita.—I^MtKT Klebahn.) A, egg with 3 nuclei, into which a sperm has just penetrated. B, same stage as A ; egg with 5 nuclei, C, egg with 4 nuclei and 5 pyrenoids ; the sperm nucleus has penetrated farther into egg D, sperm nucleus applied to functional nucleus of egg. E, fusion of two sexual nuclei. F-H, maturation of oospore. I-K, later stages in development of spermatozoids. L, two spermatozoids. M, part of an oogonium showing fecundated eggs and spermatozoids within. seem to throw further light upon the subject. As reported in the Botanisches Centralblatt, 84, p. 284, 1900, this author, who observed the sexual process in a variety of Sphceroflea annulina, which con- tained multinucleate as well as uninucleate eggs, finds that m the multinucleate eggs the nuclei lie near each other close to the surface, and at a spot where the spermatozoids seem to enter. After fecunda- tion the nuclei first distribute themselves regularly within the egg and then finally fuse to form one nucleus. 84 FECUNDATION ; HETEROGAMETES. In var. crassisepta with uninucleate egg-cells the problem is simpler. The observation of the process in this form in connection with var. braunii was fortunate, as it must have served as a control in the interpretation of the phenomena in the multinucleate eggs. If the observations of Klebahn be correct, var. braunii represents the only authentic case among the algas of a normal sexual union of a single male and female nucleus in an egg-cell containing several nuclei of apparently equal morphological value. FUCACEiE. In certain respects the sexual process in Sphceroplea is suggestive of that in the Fucacece. In the latter, however, we have the addi- tional feature that the female gametes or eggs escape into the water, and copulation takes place outside of the oogonium. Probably no other representative of the algae is so favorable for the observation of the external phenomena of the sexual process than is Fucus. The more obvious details of the process have been observed by Thuret, Oltmanns and others, but it is to the recent researches of Far- mer and Williams ('96, '98) that we are indebted for a thorough and comprehensive account of the phenomena to be observed in the living material. The work of these authors supplements also the observations of Strasburger ('97) on the development of the gametes and on the behavior of the sperm-nucleus after it enters the egg. The type of division of the cell and nucleus in the development of the gametes in this group of plants has been fully treated in the intro- ductory chapter, and the escape of the egg-cells from the oogonium is too well known to bear repetition in this place. ^ Since, however, Fucus has figured prominently in recent and much discussed theories bearing upon the significance of the number of the chromosomes in sex and heredity, it is probably not out of place here to state that, in the first nuclear division in the oogonium, the reduced number of chromosomes appears, and that both the nucleus of the egg and the spermatozoid contain this number. In order to observe the behavior of the sexual cells while alive, and to obtain suitable material for the indirect method of study. Farmer and Williams state: Male and female plants were kept in separate dishes, and were covered to prevent drying up. ... On the appearance of the extruded products, the female receptacles were placed in sea-water, and after the complete liberation ot the oospheres a few male branches with ripe antherozoids were first placed ' On the method of the liberation of the sexual cells, see Farmer and Williams, '98, p. 629. FUCACE^. 85 in a capsule of seawater until it became turbid owing to their number. If on examination the antherozoids proved to be active, small quantities were added to the vessel containing the oospheres. ('96, p. 480.) When vigorous antherozoids (1. c, '98, p. 631) are transferred to vessels con- taining healthy oospheres they at once congregate around them, and attaching themselves to the periphery of the eggs, cause the well-known movements by lashing the water with the free cilium. But, as Thuret noticed, fertilization can often be effected without any whirHng movement taking place, and we have observed perfectly normal germination to follow on the addition of apparently inactive antherozoids to the oOspheres. There seems to be a marked difference between the degree of attrac- tion exerted on the antherozoids by the egg-cells under different condi- tions. Thus, when the extruded products have been long exposed to a moist atmosphere, so that all the membranes have become deli- quescent, the spermatozoids are hardly influenced by the oospheres. On the other hand the oospheres which still retain their walls become covered with spermatozoids. The behavior of the spermatozoids in the genus Halidrys is of especial interest in this connection, and I quote again from the same authors (1. c, '98, p. 633) : On watching the behavior of the antherozoids when swimming amongst the oospheres, they are seen to attach themselves to the surface of the eggs by one cilium, whilst they maintain a circular or gyratory movement around their point of attachment. Most often there is a number — a dozen or more — of these groups, each consisting of 4 to 12 antherozoids, distributed over the sur- face of each odsphere. The movement is always in the clockwise direction, and the chromatophore is on the end of the antherozoid remote from the egg. The rate of gyration is fairly rapid, 40 to 50 complete turns being made in a minute. After this has been going on for a while the egg itself evinces a change, swelling somewhat and appearing more transparent than before. Sometimes movements of vacuoles may be discerned, and even the position of the nucleus may change. These alterations ensue as the definite result of the stimulus in some way given by the antherozoids themselves. . . . Sud- denly the antherozoids are seen to leave the &gg like a crowd of startled birds, or else they become quiescent, and these phenomena are immediately followed by a great change in the egg itself, which becomes warty and covered with conical projections. From each papilla a fine thread projects, consisting of a moniliform series of droplets, and the antherozoids may sometimes be observed attached to these threads. After the lapse of a few (3 to 5) minutes the egg resumes its spherical form whilst at the same time its diameter becomes smaller. Still later the fine threads also disappear, whilst the egg regains its original size. As long as the antherozoids are in active motion on the surface of the egg, the latter exhibits a scarcely perceptible rocking movement and is free in the water, but during the events which have just been narrated it 86 FECUNDATION ; HKTEROGAMETES. becomes attached to the surface on which it may be resting. We consider it as certain that the flight of the supernumerary antherozoids marks the moment of actual fertilization, and it seems only possible to interpret the events outside the egg as the results of an excretion from it of some substance which not only exerts on the surrounding antherozoids a negative chemotactic but also a directly injurious effect, for a number of dead sperms may be seen around the fertilized egg. Possibly the bead-like filaments which partly stain like muci- lage, are directly concerned in the process. The facts observed by Farmer and Williams have been given some- what in detail, because they are suggestive of various interesting problems, especially those pertaining to chemotaxis between sexual cells, a province of physiology well worthy of careful investigation, and one which will undoubtedly yield fruitful results. It may be noted that, in the attachment of the spermatozoids to the egg by means of one cilium, and in the sudden withdrawal of the super- numerary sperms as if startled, a certain resemblance exists between Halidrys and Ectocarpus (see p. ^d),^ although these phenomena are less marked in the latter. In the case of normal healthy products, fecundation occurs within a few minutes after the addition of the male cells. The fecundated eggs form a membrane around themselves at once, and behave in a very different manner from those into which no spermatozoids have pene- trated. For example, if the sea-water be gradually drawn off from a mixture of fecundated and non-fecundated eggs, the latter flatten out, their cytoplasm loses its coherence and becomes distributed in all directions, while the former show only local protuberances and burst only at one point. The passage of the sperm-nucleus through the cytoplasm and its fusion with the nucleus of the egg can be followed with anything like accuracy only in thin and properly stained sections. According to Strasburger ('97), the egg of JFucus platycarpus at the time of fecun- dation is globular and p^rovided with only a plasma membrane. The alveoli of its cytoplasm, together with the included chromatophores, are radially disposed about the centrally placed nucleus (Fig. 30, A), an arrangement which seems to facilitate the movement of the sperm to the egg-nucleus. The passage of the sperm through the cytoplasm and its union with the nucleus of the egg take place rapidly, for both Strasburger and Farmer agree that ten minutes after the addition of the spermatozoids to the water containing the eggs the sexual nuclei have united. Strasburger is inclined to the view that th? larger por- tion of the cytoplasm of the spermatozoid on entering the Q^apis, eyelash or cilium ; and irAacTTOS, formed. ' See Introduction, p. 46. PTERIDOPHYTA. 131 Belajeff ('98). Prior to the division of the grandmother-cell of the spermatozoid, /. e., the last cell-division in the spermogenous tissue of the antheridium, which gives rise to the cells that develop directly into the spermatozoids, there appears on opposite sides of the nucleus a small globular body of a homogeneous structure, staining rather densely (Fig. 51, A). These bodies are not provided with any radia- tions. In Onoclea there is, immediately surrounding the nucleus, a region of less granular cytoplasm from which, undoubtedly, the weft of spindle fibers is developed. These bodies, which are the primordia of the blepharoplasts, lie just at the outer edge of this region or weft (Fig. 51, A). In the telophase a blepharoplast primordium lies near the depression of each daughter-nucleus, very near the pole of the spindle (Fig. 51, B, C). Each appears now to be a hollow globular vesicle. Soon after cell-division is completed the development of the daughter-cells directly into spermatozoids begins. The blepharoplast primordium becomes somewhat lens- or crescent-shape in Gymno- grainme,, with the concave side turned toward the nucleus. The nucleus at the same time becomes flattened upon one side and gradu- ally passes into a crescent- or pear-shaped body (Fig. 51, D, E). The blepharoplast has elongated into a thread or band, which follows the convex side of the nucleus and is rather close to it. One end of the band now extends beyond that end of the nucleus which will be anterior in the mature spermatozoid (Fig. 51, F, G). With further development the blepharoplast moves away from the nucleus to a position just beneath the plasma membrane (Fig. 51, H). At this stage the cyto- plasm in Onoclea (Shaw, '98) shows a depression corresponding to the concave side of the nucleus. At about this period in the development in Gymnogramme^ according to Belajeff, the cilia make their appear- ance as outgrowths of the blepharoplast. The nucleus elongates, becoming more slender, and gradually assuming a spiral or corkscrew shape of two or three turns. In the mature spermatozoid (Fig. 52, A) the nucleus is thicker, tapering abruptly, and sometimes to a point, at the posterior end, but gradually forward into a slender anterior end. It is oval in cross section, or, in sonie cases, slightly flattened on the inner side, especially in the thicker posterior part. In mature sperma- tozoids of Onoclea struthiopteris^ fixed and stained on the slide, the cytoplasmic part seems to be in the form of a band which conforms to the spiral course of the nucleus. It is broadest at the anterior end, which extends a short distance, about one or two turns, beyond the anterior end of the nucleus, but it narrows gradually backward, dis- appearing at a point which marks the thickest part of the nucleus 132 ARCHEGOXIATES. (Fig. 52, A). Along the outer edge of the cytoplasmic band extends the blepharoplast as a thread or narrow band from which the cilia arise. The blepharoplast reaches almost or quite to the anterior extremity of the cytoplasmic part, but it cannot be traced farther back than the posterior extremity of the cytoplasmic part, although it may extend some distance farther as a delicate thread closely applied to the nucleus. The blepharoplast is broadest at its anterior end, where it seems to be not perfectly flat, but curved, appearing as a double line, or in cross section as a shallow U. It is, however, very small, so that the exact shape is difficult to determine with certainty. As already stated, it becomes a very delicate thread at the posterior end which is brought Fig. 52. — Two mature spermatozoids drawn from specimens that were fixed and stained upon the slide a few minutes after their escape from the antheridium. A, OnocUa strutkiopteris ; B, Marsilia vestita. close to the nucleus by the narrowing of the cytoplasmic band. It is probably for this reason that it cannot be traced after coming into con- tact with the nucleus. There is nothing to indicate that the blepharo- plast extends to the posterior end of the nucleus. The cilia begin at a short distance from the anterior end, and extend backward about two and one-half or three turns. Their length equals or even exceeds that of the spermatozoid when extended. Judging from Belajeff's figure of a mature spermatozoid, it would seem that the cytoplasm envelops the entire nuclear portion, but in my own preparations, which were made by killing and staining the sper- matozoids upon the slide after they had escaped from the antheridium, no cytoplasmic mantle was seen to surround the posterior part of the nucleus. Thom ('99) states also that the whole nucleus is surrounded by a cytoplasmic envelope. It is possible, of course, that the plasma membrane, or even a thin layer of cytoplasm, may envelop the nuclear portion. The nucleus usually appears homogeneous in structure, but PTERIDOPHYTA. 133 in some cases in which the stain was well washed out the structure appeared coarsely reticulate or granular. This was observed in sper- matozoids of Onoclea struthiopteris that were killed on the slide in chrom-osmic-acetic acid and stained in safranin gentian-violet and orange G. The posterior turns of the spermatozoid embrace the vesicle, which presents a very fine reticulum, and in which coarse granules are held, among them being small starch grains. The author has observed that the vesicle of Onoclea struthiopteris became separated from the spermatozoids a short time after their escape from the antheridium ; for, of the many hundreds fixed and stained upon the slide a few minutes after their escape from the antheridia, relatively few were found with the vesicle adhering. The development of the spermatozoid of Marsilia^ according to Shaw ('98) and Belajeff ('99), differs in certain important details from that of Onoclea. As this process is known in so few of the Pteridophyta^ it is perhaps well to present briefly the facts as they are known in one of the heterosporous forms. At the close of the second from the last division in the spermogenous tissue of Marsilia vestita, or that leading to the great-grandmother- cell of the spermatozoid (the primary spermatocyte of Shaw), there appears at each pole of the spindle, or near it close to the daughter- nucleus, a small body which is called by Shaw a blepharoplastoid . During the resting stage of the nucleus the blepharoplastoid seems to divide. The two halves increase in size and remain together near the nucleus. As soon as the nucleus of the great-grandmother-cell begins to divide, the pair of blepharoplastoids move away from the nucleus and remain at a position in the cytoplasm between one pole of the spindle and the equatorial plane, until the metaphase, or early anaphase, when they disappear. About the same time, or a little later, a small blepharoplast appears near each pole of the spindle. At the close of the division the blepharoplast lies near the nucleus of the grand- mother-cell of the spermatozoid (secondary spermatocyte or sperma- tocyte mother-cell of Shaw). It now divides, and the two daughter blepharoplasts increase in size and separate from each other, at the same time moving away from the nucleus (Fig. 53, A, B). Each takes a position near the pole of the future spindle but always a little to one side of its longitudinal axis. They increase in size and remain apparently unchanged in structure until the anaphase, when each seems to be hollow (Fig. 53, B, C). As soon as the nucleus of the spermatozoid mother-cell (spermatid) 134 ARCHEGONIATES. is formed, a small eccentric body appears in each blepharoplast (Fig. 53, D), then several, so that it appears as if the blepharoplast had broken up into a group of small bodies (Fig. 54, E). Out of these bodies is developed the band, which elongates, and together with the nucleus moves toward the plasma membrane of the cell (Fig. 54, F, G). In cross section the band is broadly U-shaped, but when seen from above it appears as a double line (Fig. 54, H). The band continues to elongate until finally a spiral is formed, which makes five or more turns about the hemispherical half of the cell (Fig. 54, I). The nucleus also elongates, becoming sausage-shaped, and lies in close contact with the larger turns of the blepharoplast. The mature sper- niatozoid in Marsilia is composed, therefore, of a blepharoplast, Fig. 53. — Blepharoplast primordium during division of grandmother-cell of spermatozoid in Marsilia, vestita. — (After Shaw.) A, the two primordia of the blepharoplasts lie in cytoplasm some distance from nucleus. B, they are now on opposite sides of the nucleus but a little to one side of median line. C, the nucleus is iu spindle stage of division; the young blepharoplasts lie near the respective poles of spindle. D, telophase of division; blepharoplast rudiment at pole of each nucleus contains a dense granule. consisting of a funnel-shaped spiral of about ten or more turns, and a sausage- shaped nucleus without a definite visible structure, which is connected with the three larger posterior turns of the blepharoplast (Fig. 52, B). The posterior end of the blepharoplast, which is usually bent in the shape of a hook, extends beyond the nucleus. The rela- tively large vesicle is embraced by the larger posterior tui^ns of the blepharoplast. In Marsilia vestita the author observed that the vesicle remains adhering to the spermatozoid for a longer time than in Onoclea struthiopteris. The vesicle consists of a delicate cytoplasmic reticulum, in which are held large starch and protein granules. The numerous cilia (the spermatozoids were fixed and stained upon the slide) spring from the middle and posterior coils, the two or three anterior coils being free from them. In some cases observed the cilia extended almost to the posterior end of the blepharoplast. As soon as the vesicle drops off, the spermatozoid becomes much elongated, losing its pronounced funnel-shape. PTKRIDOPHYTA. 135 Belajeff ('99) , who has also studied the development of the sperma- tozoid in Marsilia^ agrees with Shaw in so far as the transformation of the primordium of the blepharoplast into the mature cilia-bearing organ is concerned, but, as regards the earlier behavior of the primordia, these observers disagree in certain important particulars. Belajeff, who regards the blepharoplast as a centrosome, finds that in the division of the grandmother-cell of the spermatozoid, the primordia, which lie some distance from the nucleus, divide, and a faint central spindle is formed between the daughter primordia. This structure, he mamtains, gives rise to the karyokinetic spindle just as in some animal Fig. 54. — Transformation of mother-cell into mature spermatozoid in Mars ilia vestita. — (After Shaw.) E, two spermatozoid mother-cells ; each rudiment of blepharoplast has become a group of granules. F, spermatozoid mother-cell ; the blepharoplast {b) is much elongated. c, cytoplasm ; s, starch grains. G, the thread-like blepharoplast and bean-shaped nucleus lie dose to plasma membrane. H, an older suge seen from above; it is app.irent that blepharoplast is a band concave on the outside. ' I, the blepaaroplast and sausage-shaped nucleus {,k) make several spiral turns within the cell close to plasma membrane. cells, and concludes, therefore, that the blepharoplast primordia are centrosomes. The author has already dealt with this matter in the introductory chapter, and a further discussion will not be given here. In Equisetum Belajeff has found that the spermatozoid develops in a manner similar to that of the fern, and there are good reasons for believing that the process of development is much the same in the majority of archegoniates, although our knowledge is yet too meager to warrant any sweeping generalization. It seems fitting in this connection to compare the mature spermato- zoid of the Characece with that of the fern. Belajeff ('94) has shown that in the development of the spermatozoid of Charafcetida the two cilia are borne by a thread-like body which arises in the cytoplasm in a manner similar to the blepharoplast of the fern. The spermatozoid, 136 ARCHKGONIATES. as in the Pteridophyta and gymnosperms^ is a transformation of the entire contents of the cell, and we may with much propriety regard the spermatozoid of Chara and that of the fern as homologous structures. But whether we are dealing with real homologies, or only with striking analogies, is certainly a question concerning which there may be some diversity of opinion. The fate of the spermatozoid of Chara after penetrating the q%% and the union of the two sexual nuclei is practically unknown in detail, and a further discussion of the process of fecundation in the absence of more facts would seem without value, since it is not the purpose to enter here into any discussion of the homologies of the sexual organs of the Characeae with those of the Archegoniates. THE EGG-CELL AND FECUNDATION. In more recent years the process of fecundation has been observed in various genera of the Felicinecs by Campbell, in Onoclea by Shaw, and in Adianium and Aspidium by Thom. The author has followed the process in Onoclea struthlopteris, and his observations confirm those of Shaw, who has traced the behavior of the sexual nuclei in great detail in Onoclea sensibilis. Soon after the division which cuts off the ventral canal-cell, and before the archegonium of Onoclea struthiopteris is full grown, the three central cells contain fine-meshed and densely granular cytoplasm. Their nuclei are in the resting stage. The wall between egg and ventral canal-cell is generally arched slightly downward into the egg- cell. This wall is laid down in this position, at least in many cases, and the concave upper surface of the ^^,, outer,/,, inner prothallial cells; ez, tube cell. B, proximal end of pollen tube capped by exineof spore; two prothallial cells, /i and /,, have rounded off and increased in sire. C, same at later stage of development ; the inner prothallial, or antheridial, cell has divided into the generative cell iind stalk cell {it); >,, first prothallial cell; c, c, primordia of blepharoplast.s ; r, nucleolus of generative cell nucleus. D, later than C ; the blepharoplast primordia {c) have moved away from nucleus. E, proximal end of pollen tube shortly before division of generative cell (.kz) which has increased greatly in size; the large blepharoplasts are provided with beautiful radiations ; the tube nucleus {ezk) has migrated back into proximal end of tube. walls cutting off the prothallial cells," according to Ikeno, are straight, meeting the wall of the pollen spore, while in Zamia Webber finds that these walls, which are only plasma membranes, are arched out into the tube cell. The inner cell (/,) gives rise to the antheridium, and may be known as the antheridial cell. A period of about three months elapses between pollination, which takes place early in July, and fecundation in October. Immediately 144 ARCHEGONIATES. after pollination each spore in the pollen chamber of the macrosporan- gium germinates, the tube cell developing gradually into a branched tube which penetrates the tissue of the nucellus. The tube-nucleus passes into the tube, maintaining a position near the gi-owing region or end as long as the tube continues its growth into the tissue of the nucellus, while the two prothallial cells retain their former position. Contrary to the genus Pinus and other higher Conifers the distal end of the tube does not grow directly toward the archegonia, but later- ally and downward, serving especially as an organ for the absorption of food (Fig. 65, A). The proximal end of the tube, carrying before it the cap of exine, or the remaining outer wall of the spore, finally grows toward the archegonium. The pollen tube has a similar beha- vior in Zamia (Webber, '97) and Ginkgo (Hirase, '98). Soon after the germination of the spore the two prothallial cells increase in size, especially the antheridial cell, which becomes spherical (Fig. 57, B, /2)' Its nucleus is also correspondingly large, and the cytoplasm presents a looser structure. In the meantime the anthe- ridial cell divides, the daughter-nuclei being of equal size. According to Ikeno ('98, p. 172) a wall is not formed between these two nuclei in Cycas revoluta. One of them now increases rapidly in size, so that it occupies nearly the entire cavity of the mother-cell, while the other remains small and is crowded out as a naked nucleus (Fig. 1^7, C, D, j/) . The larger cell is known as the generative cell (Korperzelle of the German literature) and gives rise to two spermatozoids ; the smaller cell is the stalk cell (Fig. 57, C, D, st). As we shall see later Webber finds that the antheridial cell divides regularly into the stalk and generative cells, but the plasma membrane separating the two cells is delicate, and the stalk cell arches over the first prothallial cell in such a manner as to give the appearance of the latter being nearly enclosed by the former (Fig. 60, F, G). It is pos- sible that the same is true also for Cycas. The plasma membrane, being very delicate, may have been overlooked by Ikeno, for the posi- tion of the two cells is such as to make it appear that the stalk nucleus was forced out of the mother-cell. Soon after this stage of development two small bodies appear in the generative cell (body-cell), lying close to the nucleus and on opposite sides (Fig. 57, C, c). Ikeno seems to be of the opinion that the two bodies, which he calls centrosomes, are derived from the nucleus, for the reason that just prior to their appearance outside of the nucleus, objects staining similarly appear within the nucleus. These bodies, which are the primordia of the blepharoplasts, move away from the GYMNOSPERMS. H5 nucleus toward the periphery of the cell (Fig. 57, D, c). With fur- ther growth the generative cell with its nucleus becomes elliptical, their major axis lying parallel with the longitudinal axis of the tube. The two primordia of the blepharoplasts, which lay previously in line parallel with the transverse axis of the tube, are now found in the ends of the generative cell. About each there soon appear beautiful kino- plasmic radiations, giving them a most striking resemblance to centre- spheres with large centrosomes. Later in the period of development, or about the middle of August in Japan, the young blepharoplasts shift their position again, so that their earlier orientation in the gene- rative cell with respect to the axis of the pollen tube is resumed (Fig. 57, E). The generative cell becomes spherical, and the kinoplasmic radiations are very conspicuous. From this time until the end of September, or about one and one- half months, few changes manifest themselves in the generative cell apart from an increase in size. This period in the development is, therefore, a period of growth, which corresponds to a similar period in the development of the archegonium, and at the end of which all elements have reached their maximum size (Fig. 57, E) . The diameter of the generative cell, which contains dense cytoplasm, is about 0.14 mm., and that of the nucleus is about 60 /x. The primordia of the blepharoplasts have also increased considerably in size ; they are about 15 /xin diameter. Apart from the presence of one or more vacuoles, they are rather homogeneous massive bodies. The kinoplasmic radiations are still beautifully developed ; they seem to pass over gradually and insensibly into the alveolar structure of the cytoplasm. About the middle of September the tube nucleus begins to migrate toward the proximal end of the pollen tube, and, by the end of the month, this nucleus, the generative, stalk, and outer prothallial cells are ail in the proximal end, which is capped by the exine of the spore. It may be mentioned here that the migration of the tube nucleus into the proximal end of the pollen tube seems to be a striking confirmation of the doctrine of Haberlandt, namely, that in a growing cell the nucleus generally takes a position near the seat of constructive activity. Since the proximal end of the tube now grows toward the archegonium, and as growth at the distal end ceases, it is to be expected, in harmony with the theory of Haberlandt, that the nucleus which presides over this growth should move toward the region of that activity. Webber has observed the same behavior of the tube nucleus in Zamia. The final processes which now take place in the male gametophyte have to do largely with the development of the two spermatozoids 146 ARCHEGONIATES. from the generative cell. To this phase of development Ikeno has applied the term spermatogenesis. As soon as all the structures mentioned accumulate in the proximal end of the tube, all save the generative cell begin to disorganize and finally disappear. What this disorganization signifies, Ikeno remarks, FiG- 58. — Division of generative cell and further development of blepharoplasts in Cycas revoluta — (After Ikeno.) A, generative cell with nucleus in early prophase of division ; chromatin scattered in masses of granules. B, same with nucleus in late anaphase ; each blepharoplast has separated into a mass of rods from which radiations extend ; they have nothing whatever to do with mitotic spindle. C, blepharoplast of B more highly magnified. D, cell-division is about complete ; the radiations have nearly disappeared from the mass of granules composing blepharoplast. E, two spermatozoid mother-cells, the one on the right in outline; the ciliated blepharoplast has made one turn about the cell ; nuclear beak is in connection with ciliated band. is an open question, but it seems that all of the disorganized elements contribute to the nourishment of the generative cell. The cytoplasm of the generative cell now assumes a coarse, net-like structure, and the nucleus divides (Fig. 58, A, B). The details of this division will not be dwelt upon further than to state that the mitotic spindle arises without the intervention of the ceutrosphere-like GYMNOSPERMS. HI primordia of the blepharoplasts (Fig. 58, B) . This is true for Zamia^ according to Webber, and for Ginkgo^ according to Hirase. At this stage each prinnordium of the blepharoplast is transformed into a group of fine rods about which the radiations, although not so pronounced, are still present (Fig. 58, C). When, however, the daughter chromo- somes have arrived at the poles of the spindle, each blepharoplast has become a mass, or an accumulation, of granules, and the radiations can scarcely be recognized. At the close of nuclear division each daughter-nucleus is homo- geneous, presenting a small number of nucleoli. A cell-plate is formed and the division of the generative cell completed (Fig. 58, D). The next step is characterized by the behavior of thq mass of granules of the young blepharoplast. These are arranged close to the nucleus into a more or less short and broad band whose granular nature is still evident. Seen in profile a number of radiations appear extending out from the band toward the periphery of the cell (Fig. 59, A). These radiations are the developing cilia of the spermatozoid. Whether the cilia are transformed radiations, or arise anew, is a question. Ikeno ('98, p. 180) is inclined to think that the former mode of origin is the more probable. In the meanwhile the nucleus develops a beak which becomes con- nected with the ciliated band (Fig. 59, A). The development of the nuclear beak and the arrangement of the granules into a band take place simultaneously, so that it is not known which phenomenon is of first importance. If the formation of the beak took the initiative, then it would be reasonable to suppose that the direct cooperation of the nucleus in the development of the band is indispensable. In Zamia^ according to Webber, no such nuclear beak occurs in the development of the spermatozoid. Subsequent to this stage in the development of the band its granular nature is no longer recognizable ; it appears as a thin homogeneous thread (Fig. 59, B). The further behavior of the blepharoplast seems to be characteristic of spermatogenesis in Cycas^ Zamia^ and Ginkgo. The ciliated band extends itself in a spiral which ultimately makes five turns around the hemispherical cell, always remaining near its surface just beneath the plasma membrane. During this process the nucleus increases in size and becomes some- what pear-shaped. Its beak, to which is attached apparently one end of the band, increases in length until it almost reaches the surface of the cell (Fig. 58, E, and Fig. 59, B). The free end of the band con- tinues its spiral course around the cell a short distance beneath the plasma membrane. The direction of the spiral is parallel with the 148 ARCHEGONIATES. plane of division of the generative cell. In Fig. 58, E, which repre- sents a median section through the two daughter-cells, the blepharo- plast has made one turn around the cell. The cilia, which at first lay wholly within the cytoplasm, project out through the plasma membrane as the band approaches the surface of the cell. The nuclear beak, which remains in close contact with the band during its earlier develop- ment, finally becomes separated from it (Fig. 59, C). In the mature spermatozoid the blepharoplast, as already stated, makes about five turns around the cell counter clock-wise. As is evident from a median Fig. 59. — Further development of spermatozoid in Cycas revoluia. — (After Ikeno.) A, part of spermatozoid mother-cell showing nuclear beak in contact with granular blepharoplast band. B, spermatozoid mother-cell ; band-shaped blepharoplast longer, one end being applied to nuclear beak. C, later stage ; blepharoplast has made about three turns about the cell ; the nuclear beak seems to have separated from ciliated band. D, nearly ripe spermatozoid in median section. Both nucleus and cytoplasm are lobed on one side as if constricted by blepharoplast, which describes about five turns around the hemispherical cell. section, the mature spermatozoid consists of a large nucleus completely surrounded by a thin layer of cytoplasm, and the blepharoplast lies in a depression or groove (Fig. 59, D). As a result both cytoplasm and nucleus are lobed, thus presenting a wavy contour in section. This phenomenon seems to indicate that during the final increase in size of the nucleus, the blepharoplast acted as a kind of constriction upon the anterior end of the cell. The same is true in both Zamia and Ginkgo. In the mature spermatozoid the cytoplasm which completely surrounds the nucleus is clearly distinguishable. As will be seen for Zamia and GYMNOSPERMS. 149 Ginkgo^ the spermatozoid of Cycas^ as has been pointed out for the fern, is a transformation of the entire mother-cell. The development of the spermatozoid in both Ginkgo and Zamia closely resembles that in Cycas. That in Zamia differs, however, according to Webber, in certain important details, and because of this fact the process in Zamia will be given also in some detail. Webber investigated two species growing in Florida — Zamia Jloridiana and Z. pu7nila. As a rule the mature microspore of Zamia consists of the tube cell and two prothallial cells (Fig. 60, A). Only in exceptional cases were evidences of a third cell observed, but if three prothallial cells are formed in the development of the pollen spore as is claimed for Cycas^ the first is generally absorbed before the spore is mature, leaving only a trace in the form of a dark line. The two prothallial cells are pro- vided with only a plasma membrane. The first prothallial cell is shaped like a plano-convex lens and arches out into the second prothallial cell. The second prothallial cell is attached to the first and arches out into the tube cell (Fig. 60, A, B). This is especially marked during the growth of the pollen tube. The nucleus of the tube cell is larger than those of the prothallial cells, and of the latter the nucleus of the first is larger than that of the second. Very soon in the growth of the pollen tube the second or antheridial cell, together with its nucleus, greatly exceeds the first. The process of pollination, which occurs in Florida in January, brings the pollen grains into the pollen chamber, a cavity in the apex of the nucellus, formed by the disorganization of the tissue of the latter. Webber ('01) states that the passage of the pollen grain through the micropyle is evidently accomplished by suction. A somewhat mucilaginous fluid is secreted by the cells which sur- round the micropyle, and a drop of this fluid is probably protruded at the time of pollination. The fluid disappears later, and during the formation of the pollen chamber a suction is formed by the breaking down of the cells in its formation, so that the fluid, together with the pollen grains that may be held in it, is brought down into the pollen chamber. In a short time after the pollen grains have been brought into the pollen chamber they germinate, the tube bursting out of the exine of the grain at a point opposite the prothallial cells (Fig. 60, B). No matter what the position of the grain may be, the tube always pene- trates the tissue of the nucellus adjacent to the chamber. The tube in Zamia does not branch before entering the nucellar tissue, and only ISO ARCHEGONIATfiS. occasionally afterward (Fig. 6^, A). During the early development of the tube, the prothallial cells increase in size, becoming broader and longer. The first prothallial cell pushes out into the second, which becomes shaped like a concavo-convex lens, and is crescent- shaped in cross-section (Fig. 60, B, C). As stated in a preceding paragraph, the behavior of the tube nucleus is similar to that in Cycas. Fig. 60. — Microspore and development of male gametophyte in Zaniia. — (After Webber.) A, mature pollen grain ; at point of attachment of the two prothallial cells, on left, a dark crescent- shaped line represents a layer in wall of spore, which may be the remains of a third resorbed pro- thallial cell. B, germinating pollen grain, early stage. The two prothallial cells have not yet begun to increase in size. C, later stage of germinating pollen graia ; the tube nucleus has increased in size and passed out into tube ; prothallial cells unchanged. D, proximal end of pollen tube ; the two prothallial cells have increased in size, the first having crowded out into the second in a marked degree. E, proximal end of pollen tube; nucleus of second prothallial cell, antheridial cell, in telophase of divi- sion, lower end of mitotic figure being crowded to one side by the encroaching first prothallial cell. F, prothalliumin proximal end of tube, after division of antheridial cell into stalk and generative cell. G, prothallium in later stage of development after appearance of blepharoplasts ; the double plasma membrane, separating first prothallial cell and stalk cell, shows that there are two distinct and inde- pendent cells of separate origin. A little later the second cell has arched out very greatly, and the increase in size of the first prothallial cell has brought the second, or anthei'idial cell, out beyond the limits of the pollen grain and into the tube (Fig. 60, D). However, the prothallium remains in con- nection with the wall of the pollen spore until the spermatozoids are mature. The next important step in the development is marked by the division of the second prothallial cell into the stalk cell and generative GYMNOSPERMS, 15I cell (body cell) (Fig. 60, E). In this figure the division is in the telophase, the two daughter-nuclei being still connected by the con- necting fibres. Owing to the crescent shape of the cell the spindle lies at an angle to the major axis of the prothalHum, the lower nucleus being crowded to one side by the position of the first prothallial cell, while the upper nucleus occupies a central position in the upper half of the cell, which, when the wall is formed, will become the genera- tive cell (body cell, central cell). The lower nucleus becomes the nucleus of the stalk cell. Fig. 60, F, represents the next stage in which the division is complete. A distinct transverse plasma mem- brane is formed just above the apex of the first prothallial cell which is almost entirely surrounded by the stalk cell. It is clear that should the plasma membrane separating the generative from the stalk cell be very delicate and somewhat obscured, the nucleus of the stalk cell would appear to be forced out to one side. For this reason it seems possible that the plasma membrane separating stalk and generative cells in Cycas was overlooked by Ikeno. In Ginkgo the first prothal- lial cell, which according to Webber is also surrounded by the stalk cell, was considered by Hirase ('98) to be strands of cytoplasm in the second prothallial cell. Miyake ('02), who has also examined Ginkgo^ confirms the observations of Webber. At the stage of Fig. 60, F, according to Webber, the nucleus of the generative cell is 9.79 y- in diameter, that of the stalk cell 7.12 ;/, while the first prothallial cell is 8.9 [i in diameter. The entire prothalHum is 29.37 i^ ^0"g by 16.91 (i wide. Neither during the division of the second prothallial cell into stalk and generative cell nor for some time afterward was anything observed in the cell in connection with the spindle, or elsewhere, that suggested a young blepharoplast. It is not until the generative cell has increased considerably in size that the first traces of the blepharoplasts were recog- nized. At first each blepharoplast consists of a small, deeply staining granule, from which several filaments of kinoplasm radiate, following the meshes of the cytoplasmic reticulum (Fig. 60, G). "The central granule (Webber, '01, p. 31) does not seem to be different in sub- stance from the radiations — stains the same and shows no differentiation of structure. In this stage it is only a half micron in diameter or less, and seems to be scarcely more than the point of the crossing of the filaments of kinoplasm. These granules are located in the cytoplasm about halfway between the nucleus and the cell- wall. Two are formed in each central cell at the same time and apparently inde- pendently. They are commonly located on the opposite sides of 152 ARCHEGONIATES. the nucleus, but, in a number of cases in this stage and in a still later stage, they have been found nearer together, frequently less than 45° apart." The first indication of a differentiation in the blepharoplast as it increases in size is seen in the formation of an outer wall or membrane. The generative cell, v^^hich has remained nearly spherical, increases in Fig. 61. — Prothallium and a dividing generative cell of Zantia. — (After Webber.) A, prothallium in which generaiive Cell has become large and elongated ; the blepharoplasts have taken positions on opposite sides of nucleus, corresponding to longitudinal axis of pollen tube; starch grains have begun to appear in stalk cell. B, division of generative cell, nucleus in anaphase, showing hyaline cytoplasmic areas around poles ; the blepharoplasts, whose outer membranes have separated into pieces or segments, are not con- nected with spindle. size and becomes elliptical or oblong, its major axis nearly coinciding with the longitudinal axis of the pollen tube (Fig. 61, A). The blepharoplasts by this time have taken a position on opposite sides of the nucleus on the line of the major axis of the cell. The kinoplasmic radiations are slightly more prominent than the lamellse or fibrillae of the cytoplasmic reticulum into which they run and disappear (Fig. 6 1 , A) . GYMNOSPKRMS. 153 About the first of April the blepharoplasts have reached nearly one- half the size they finally attain. They are more or less vacuolate, and the kinoplasmic radiations, which have become more abundant, extend in many instances quite to the plasma membrane of the cell. After further growth the generative cell divides into the two cells which develop into the two spermatozoids (Fig. 61, B, and Fig. 62). The blepharoplasts take no part in the division of the nucleus. Al- though their kinoplasmic radiations become fewer, they do not enter into the formation of the spindle, as the latter devel- ops apparently entirely within the nucleus, and is almost mature before the nuclear membrane has dis- appeared. In the spindle stage of this division the blepharoplast is seen to have undergone a noticeable change. It has increased in size and its outer membrane has separated from the con- tents, which are somewhat shrunken. The outer mem- brane has separated into fragments or plates, and appears now as a broken line (Fig. 61, B). The kinoplasmic radiations have almost disappeared. The reticulum of the cytoplasm about the blepharoplast is so arranged as to suggest radiations. It will be remembered that precisely the same phenomenon occurs in Cycas. During the anaphase of division the finer structure of the outer membrane, which still consists of a number of segments, is seen to be made up of numerous small granules placed side by side to form the membrane. The central contents, which stained very densely at an earlier stage, have disappeared, giving place to a delicate hyahne reticulum (Fig. 61, B). Webber suggests that the densely stammg material which resembled nucleoli in its staining qualities was utilized Fig. 62.— Prothallium of Zamia in which the generative cell has divided. The blepharoplasts have separated into granules which are beginning to organize the ciliferous band. The first prothallial cell and stalk cell have become gorged with starch. (The magnification of this figure is only one- half that of A, Fig. 61.)— (After Webber.) 154 ARCHEGONIATKS. as food material in the growth of the blepharoplasts and other parts of the cell. During the telophase the blepharoplast is represented by a more or less irregular or spherical mass of granules, which have evi- dently been derived by the breaking up of the membrane. "It would seem that the outer membrane of the blepharoplast breaks up into numerous segments or granules, which assume a roundish or elliptical form, and through the action of the cytoplasm become crowded to- gether in a mass occupying the position of the original blepharoplast." About the time of the reconstruction of the daughter-nuclei and the formation of the plasma membranes separating the cells, the develop- FiG. 63. — Further development of blepharoplast. — (After Webber). A, two attached spermatozoid mother-cells (spermatids) resulting from division of generative cell; the band of blepharoplast is being formed by fusion of granules. B, fusion of granules to form the band. C, formation of ciliferous band by fusion of granules, more highly magnified. ment of the band, which is to bear the cilia, begins. It appears first as a short, delicate, and deeply staining line extending from the mass of granules toward the nucleus (Fig. 63, A). A little later a similar line or band can be seen on the opposite side of the mass of granules. From Fig. 63, B, it is apparent that the band is developed more or less directly from the granules. The band, which at first is very nar- row, increases appreciably in width (Fig. 63, B, C). The further development of the band with its cilia and the transformation of the daughter-cell into a spermatozoid closely resembles that of Cycas, already discussed at some length in the preceding pages, with the very noteworthy exception that in Zamia there is no nuclear beak formed, GYMNOSPERitS. I55 which is in contact with one end of the blepharoplast in the earlier part of its development (Fig. 63, A). The mature spermatozoid is also quite similar in structure to that of Cycas, consisting of a large nucleus completely surrounded by a layer of cytoplasm in which the ciliferous band, or blepharoplast, is located just beneath the plasma membrane. The blepharoplast is in the form of a helicoid spiral, making about five or six turns counter clock-wise and embracing about one-half of the body of the cell (Fig. 65, B). The spermatozoid, as in the ferns, is a transformation of the entire cell and, therefore, a true spermatozoid. The development of the spermatozoid in Ginkgo according to Hirase ('98) is quite similar to that in Cycas as described by Ikeno. In the generative cell of Ginkgo Webber ('97) and Hirase ('98) find that, when the nucleus becomes strongly flattened or lenticular, a large nucleolus-like body appears on either side of the nucleus between the nuclear membrane and the young blepharoplasts. Other similar but smaller bodies are sometimes present in the cell. Accompanying these two bodies Hirase finds coarsely granular cytoplasm. The bodies in question react toward stains much as do nucleoli, and, since they dis- appear at a later stage, it is probable that they represent merely extra- nuclear nucleolar substance. Miyake ('02) finds that after the division of the generative cell in Ginkgo a cell-wall is formed between the two daughter-cells, and that a distinct and firm wall was always found around the two spermato- zoids. The fact that a wall is or is not formed about the daughter- cells, /. e., the mother-cells of the spermatozoids, does not affect the morphological rank of the spermatozoid. The mature spermatozoid of Zamia is probably the largest male gamete known in the plant kingdom, being plainly visible to the unaided eye. When swimming freely and without pressure it is slightly ovate, nearly round or compressed sphexical (Fig. 65, B). They vary greatly in size, however, ranging in length from 222 to 332 ;u, and in width from 222 to 306 p.. Ikeno describes the spermatozoid of Cycas as being provided with a tail which is merely the elongation of the posterior part of the cyto- plasmic mantle. Measured in sections the length was found to be 160// and the width 70 //. The length of the tail was 80 yu or equal to that of the body. Fujii has shown that the tail attributed to the spermato- zoid of Ginkgo was an artifact, and this statement has been confirmed by Miyake, Since no tail exists in Zamia^ it is probable that that described for Cycas may also have been the result of abnormal conditions. 156 ARCHEGONIATES. THE ARCHEGONIUM. The development of the archegouium in the Cycadacece and in Ginkgo^ which is similar to that of Plnus,, is too well known to require a detailed description in this place. The manner, however, in which the large central cell is nourished during its growth by the immediately surrounding cells of the prothallium is, if Ikeno's observations be cor- rect, a phenomenon of a rather rare occurrence in the Gymnosperms, and merits some special mention. These surrounding cells, which are separated from the central cell by thick cellulose walls, are of a uniform size, each possessing dense cytoplasm and a large nucleus. Before the archegonium is full grown the nuclei of these cells show a fine and distinct threadwork ; but, as this organ approaches maturity, the nuclei, with the exception of the nucleoli, are transformed into homo- FlG. 64. — Three cells from layer of prothallial cells immediately surrounding upper part of central cell of archegonium of Cycas, showing protoplasmic connections between these cells ; in B the bealc of nucleus extends into plasmic bridge. — (After Ikeno.) geneous and diffusely staining bodies. This phenomenon is not confined solely to the cells forming the wall of the archegonium, but it may extend to adjacent cells of the prothallium. This nuclear change takes place only in cells near the upper part of the central cell. Goroschankin has shown that in the Cycadacece fine cytoplasmic connections exist between the central cell of the archegonium and the surrounding cells. Fi'om Ikeno's figures it seems that the cytoplasmic strands in Cycas are relatively large, and that large granular plasmic masses pass over bodily into the central cell (Fig. 64, A, B). Fre- quently the nucleus itself will send out a beak or protuberance toward the nearest plasmic connection. Arnoldi (1900) finds that in several species of Pinus and in Abies the nuclei from the surrounding cells pass into the egg-cell. The prevalence of condensed nuclei in cells surrounding the upper part of the central cell is explained by Ikeno as GYMNOSPERMS. 157 being due to a greater need of food material by this part of the central cell ; for it is here that the greatest activity takes place during the maturing of the egg-cell, which culminates in the formation of the ventral canal-cell. Webber does not find any protoplasmic connections between the egg-cell and those surrounding it in Zamia^ and so far as the author is aware no such protoplasmic connections exist in the higher Gymnosperms. In Cycas the phenomenon described by Ikeno is, if true, probably an adaptation to the rapid transfer of nutritive material from the surrounding cells to the egg-cell. Strasburger ('01, pp. 550-553), in a late publication on the proto- plasmic connections between cells in plants, calls into question the statement that nuclei or nuclear fragments pass bodily through the pits of the surrounding cells into the egg-cell of Gymnosperms as a normal phenomenon, and asserts that it is the result of injury due to pressure or fixing reagents. There seems to be no doubt that in all Gymnosperms in which the egg-cells reach such an enormous size the cells immediately surround- ing the &g^ contribute directly to the nutrition of the latter, but it is not clear why any of the material should pass over bodily into the egg-cell. The final step in the development of the archegonium is the forma- tion of the ventral canal-cell, which takes place immediately preceding fecundation, and consequently this cell persists only a short time (Fig. 67, A). It was probably due to this fact that the presence of a ventral canal-cell was not observed by Warming and Treub. Only a plasma membrane and not a cell-wall is formed separating the ventral canal- cell from the egg. It is not at all improbable that in some cases a plasma membrane may not be formed, and such is reported for Cefh- alotaxis fortuni by Arnoldi (1900). The formation of a plasma membrane is, however, of secondary importance in the formation of the ventral canal-cell, for if the nucleus of the central cell of the archegonium divides karyokinetically, and one of the daughter-nuclei becomes the functional egg-nucleus, the division is certainly to be regarded as the formation of a ventral canal-cell whether a plasma membrane is formed or not. Botanists have sometimes been inclined to refer to the formation of the ventral canal-cell as a maturation process similar to that in the animal egg. Ikeno speaks of this step in the development as the period of maturation (Reifungsperiode), which recalls the formation of the polar bodies in the animal egg, but I do not infer that he considers the two processes homologous. He states, however, that it appears prob- .58 ARCHEGONIATES. able, judging from the karyokinetic figures observed, that the nuclear division leading to the formation of the ventral canal-cell is of the heterotypic type, and takes place essentially as in the first division of the pollen mother-cells of the LiliacecE. This is certainly an error, for in both Gymnosperms and Angiosperms the heterotypic nuclear division occurs in the micro- and macrospore mother-cells and nowhere else in ontogeny. Since the spore mother-cells of the Gymnosperms are homologous with those of the higher plants, we naturally expect to find the heterotypic division in Cycas in the first karyokinesis of the macrospore mother-cell. This is made all the more certain by the researches of Juel (1900), who finds in Larix that the first nuclear division in the macrospore mother-cell is heterotypic. In Larix and Fig. 65. — Upper end of nucellus ; spetmatozoids in pollen tube of Zamia. — (After Webber). A, diagrammatic outline of upper end of nucellus, showing proximal ends of pollen tubes growing down into the cavity just above archegonia ; a, archegonia ; /, prothallium ; pc, pollen chamber ; pt, pol- len tubes; pg, pollen grain. B, two mature spermatozoids in proximal end of pollen tube. in other Gymnosperms the earlier development of the macrospore is precisely the same as in such Angiosperms as Helleborus^ in which the first nuclear division is heterotypic and homologous with the first division in the pollen mother-cell. The formation of the ventral canal-cell may represent some sort of a maturation process, and the conclusion that this cell is an aborted &%% is tempting, but at our present state of knowledge such an infer- ence is scarcely justifiable. FECUNDATION. Soon after its formation the ventral canal-cell disorganizes. The nucleus of the &^^ passes back gradually toward the middle of the cell, at the same time increasing in size. Finally, when the center of the cell is i-eached, the nucleus is usually large, being generally longer than broad, and shows the structure of the resting condition. GYMNOSPKRMS. '59 During the final stages in the development of the spermatozoid the proximal end of the pollen tube, which is still capped by the exine of the spore, grows downward into the prothallial cavity as in Zamia (Fig. 65, A). This cavity in Cycas, according to Ikeno, is filled with a watery fluid derived largely from the archegonia, and in which the spermatozoids swim on escaping from the pollen tube. Webber is of the opinion that in Zamia this fluid is derived largely from the pollen tube. The spermatozoids in Cycas, on escaping from the pollen tube, swim about rapidly, and in a short time penetrate the ^%^. That part of the egg at which a spermatozoid enters is de- pressed, giving the impres- sion that it came against the egg with some force. The nucleus of the spermatozoid now escapes from its cyto- plasmic mantle and migrates toward the nucleus of the egg. The cytoplasm and blepharoplast are left in the upper part of the G%'g as in Zamia (Fig. ^G^ A, B), where they undergo disor- ganization. It frequently happens that several sperm- atozoids reach the ^%%^ but, as a rule, only one penetrates into its interior, the others remaining at the surface. Whether more than one male nucleus ever fuses with the egg-nucleus is not known. When male and female nuclei come in contact they are readily distinguished from each other, the male being smaller, with a more finely granular threadwork. Both are in the resting stage. The male nucleus seems to pi-ess against the female, forming a depression in the latter. In a short time the male nucleus is completely imbedded within the egg-nucleus ; the membrane of the male nucleus disappears, and the two nuclei fuse so completely that the fusion nucleus can scarcely be distinguished from an unfecundated nucleus of the egg. Fig. 66. — Fecundation of egg-cells in Zamia — (After Webber.) A, egg-cell immediately after coming together of male and female nuclei ; the ciliferous 'band of fecundating sper- matozoid lies in upper end of egg ; a second spermato- zoid trying to gain entrance is shown at apex of egg. B, similar to A, but showing longitudinal section of ciliferous band in upper end of egg. l6o ARCHKGONIATES. The processes incident to and accompanying fecundation in Zamia differ only in minor details from those of Cycas. Cei'tain phases of these processes, however, as observed by Webber ('97, I, II, III), are of special interest and importance. They are described as follows ('97, II, p. 18): The proximal ends of the pollen tubes . . . which grow downward through the apical tissue of the nucellus into a cavity formed in the prothallium above the archegonium, have increased in length until the ends almost or quite touch the neck cells of the archegonia, which protrude into the same cavity (Fig. 65, A). It is interesting to note that the pollen tubes when they enter the prothallium cavity, which is filled with air, do not grow at random, but bend slightly outward and grow directly toward the archegonia. . . . The pro- truding tip formed by the old pollen grain is plainly visible with a hand lens, and is evidently the point which first comes into contact with the neck cells of the archegonia. The neck cells are also distended and turgid, and are evi- dently easily broken. If in this stage the end of a pollen tube be touched very lightly with the flat side of a scalpel it bursts, and the antherozoids, together with a drop of the watery contents of the pollen tube are quickly forced out, and the pollen tube immediately shrivels up into a shapeless mass. . . . The pollen tube evidently grows down until the end is forced against the neck cells, when the tube bursts, discharging the mature antherozoids and the watery contents of the tube which supplies a drop of fluid in which the antherozoids can swim. . . . ('97, III, p. 226). As explained in my previous papers, several antherozoids commonly enter each archegonium, two being usually found and sometimes three or four. The entire antherozoid enters unchanged, swimming in between the ruptured neck cells. Only one of the antherozoids is concerned in fecundation, and the others are usually found between the protoplasm and the wall of the archegonium, presenting their original form and appearance, or in some stage of disintegra- tion (Fig. 66, A). Occasionally one of the antherozoids not concerned in fecun- dation pushes for a short distance into the contents of the archegonium, as it is always found in such cases to form a distinct body which stains very differently . . . ('01, p. 65). That one which is utilized in fecundation swims into the protoplasm of the archegonium for a short distance, where it comes to rest and undergoes change. The nucleus slips out of its cytoplasmic sheath and passes on alone from this point to the egg-nucleus, with which it unites. The spiral ciliferous band remains at the apex of the egg-cell in the place where the nucleus left. In very numerous instances, just after fecundation, it has been discovered in this position, and there can be no doubt that this process is the one normally occurring. It shows very plainly and presents nearly the original form of the spermatozoid (Fig. 66, B), but is always stretched out much more than in the normal spermatozoid. . . . The method of escape of the nucleus from the body of the spermatozoid can only be conjectured. It would seem, however, that the rapid boring of the apical or spiral end into the egg-cell may cause too great a pressure on the large body of the spermatozoid, resulting in its bursting and freeing the nucleus , GYMNOSPERMS. l6l while the cilia motion continues probably some time longer, carrying the band farther along and freeing the nucleus from any hindrance by it. The apex of the spiral end of the spermatozoid invariably enters the egg-cell first, and in all of the cases observed where the nucleus has just escaped from the spermatozoid it has been found a short distance behind the spiral of the spermatozoid, as if it had been forced out and left behind. The function of the cytoplasm of the spermatozoid is still in considerable doubt, but that it fuses with the cytoplasm of the egg-cell is certain. Shortly after the nucleus has broken out of the sper- matozoid cell, the thin layer of dense cytoplasm which surrounded it can be seen in a broken, fragmentary form, still somewhat connected with the spiral band. The cytoplasm of the spermatozoid in this stage is very different from that of the egg-cell, being more densely granular and staining more deeply, so that it is easily distinguished. Later, only a coarse granular substance is found inside the spiral coil of the ciliferous band, and it would seem that this is the cytoplasmic matter from the spermatozoid which has mingled with that of the egg-cell. It should be mentioned that the plasma membrane surrounding the spermatozoid has entirely disappeared, no trace of it being visible. It would seem to have fused with some substance of the egg-cell or to have been absorbed in some way. The male nucleus, when it has escaped from the spermatozoid and is observed lying in the cytoplasm at the apex of the egg-cell, is a loose, open structure, seeming to have but little kinoplasmic and chromatin matter. The passage to the nucleus is evidently a rapid one, as few stages have been found between the above and the completion of fecundation. In some instances the path over which the nucleus travelled in reaching the egg-nucleus is discernible by the arrangement of the granules in the cytoplasm, showing the direction of the passage. The egg-nucleus, previous to fecundation, is elliptical and is located slightly below the center of the enormous egg-cell which is about 3 mm. long by 1.5 mm, wide (Fig. 66, A, B), The egg-nucleus is of enormous size, comparatively, being plainly visible to the unaided eye. It is composed of an open, coarse reticulum. So far as the writer has observed there is no depression or '• emp- fangnisshohle " in the upper part of the nucleus where the sperm-nucleus enters, as was found by Ikeno in Cycas. No special attention has been given to this matter, however, and further observation may show it to be present. The male nucleus in entering the egg-nucleus gradually pushes into it as observed by Ikeno in Cycas, and finally becomes entirely surrounded by it. Meanwhile it has changed its structure and become densely granular, differing markedly from the egg-nucleus in this particular. , , . After fecundation is apparently completed the male nucleus appears as a small, nearly round body in the upper portion of the egg-nucleus into which it has penetrated (Fig. f^, B). Further changes in the sexual nuclei were not followed by Webber, and it is not known whether a fusion nucleus is formed in Zamia as described by Ikeno for Cycas. Since the publication of his paper on Cycas, Ikeno ('01) has observed the formation of the ventral canal-cell, the process of fecundation and l62 ARCHEGONIATES. the first division of the fusion nucleus in Ginkgo biloba (Fig. 67, A, B, C, D). These processes agree closely with those in Cycas. In Ginkgo^ however, the male nucleus at the time of fusion is relatively small, being less than one-tenth the size of the female nucleus. As in Cycas and Zamia^ the male nucleus becomes completely imbedded in Fig. 67. — Formation of ventral canal-cell, fusion of sexual nuclei, and the division of the fusion nucleus in Ginkgo. — (After Ikeno.) A, apex of central cell of archegonium showing telophase of nuclear division ; a cell-plate, or plasma membrane, is formed in the connecting fibers. B, egg-nucleus into which a male nucleus {ni) has penetrated. C, fusion nucleus in prophase of division. D, fecundated egg-cell showing fusion nucleus in spindle stage of mitosis ; the mitotic figure lies within limits of the nucleus whose membrane seems to be still intact. the female before the dissolution of its membrane. Both nuclei are in the resting condition at the time of fusion. The spindle of the first karyokinesis following fusion is formed within the nuclear cavity and before its membrane has disappeared (Fig. 67, B, C, D). Nothing is said by Ikeno about being able to distinguish male and female chromatin elements in this division. GYMNOSPERMS. 163 It is interesting to note further that in neither Cycas, Zamia^ nor Ginkgo was the stalk or prothallial cell of the pollen tube found in the egg by any of the observers mentioned. These cells are probably disorganized beyond recognition when the contents of the tube are discharged into the egg. PINUS. THE MALE AND FEMALE GAMETOPHYTES. Apart from the absence of motile spermatozoids and the behavior of the male gametophyte, the process of fecundation in the Coniferales, so far as this is well known, is in general similar to that in Cycas^ Zamia^ and Ginkgo^ and it will be necessary only to point out briefly the more important features of difference. Since the important researches of Strasburger, Goroschankin, and Belajeff upon certain of the higher Gymnosperms, an interesting series of facts has been collected by Dixon ('94), Blackman ('98), Cham- berlain ('99), Murrill (1900), Ferguson ('01), and others. The studies of later observers, who used more improved technique, have been confined principally to the genera Pinus, Picea, and Tsuga, and consequently our knowledge of the sexual process in many other Gymnosperms is sadly wanting. It has been shown by Strasburger ('92) and others that the prothal- lial cell in the ripe microspore of Pinus and other closely related genera is the last one of a series of two or three cells, and that this cell divides, as in Cycas and Ginkgo^ to form the stalk cell and the generative cell of the antheridium (Fig. 68, A, B). The generative cell (body cell) then divides to produce the two non-motile male gametes, each consist- ing of a nucleus surrounded by a specially differentiated mass of cyto- plasm (Fig. 68, C). Contrary to Cycas^ Zamia^ and Ginkgo^ the distal end of the male gametophyte, or pollen tube, grows in a more or less direct line from the pollen chamber down through the nucellus to the archegonium, and while the tube seems to be merely a carrier of the gametes, it can and doubtless does act as an absorber of nutriment as well. The probable need of less food by the male gametophyte of the higher gymnosperms may account for the absence of a specially developed absorbing appa- ratus. This idea is advanced merely as a suggestion and not as an adequate explanation of the difference between the behavior of the tube of Pinus^ for example, and that of Cycas or Ginkgo. Other factors may have been more influential during the phylogenetic develop- ment of these forms. 164 ARCHEGONIATES . The development of the archegonium is the same as in the lower gymnosperms. The ventral canal-cell is separated from the egg merely by a plasma membrane, which is formed by the connecting fibers, as is usual in the higher plants. It persists for a short time only. In Fig. 68. — Pollen grain, end of pollen tube, and fusion nucleus of Pinus itrobtu. — (After Ferguson.) A, mature pollen grain. /^ and/', remains of first and second prothallial cells ; a. c, antheridial cell. B, pollen grain in which antheridial cell has divided, /.c, generative cell ; st.c, stalk cell. C, distal end of pollen tube which is pushing between neck-cells of archegonium ; the male nuclei (x.M.) are of unequal size, v.n., tube nucleus ; st.c, stalk cell ; s.c., cytoplasm of generative cell. D, first mitosis following fecundation. The spindle is formed, but the male and female chromatin spirems are still separate and distinct. Pinus strobus^ according to Ferguson, there are probably instances in which the nucleus of this cell is not reconstructed, and this may be true also in other genera and species. GYMNOSPERMS. 165 FECUNDATION. Goroschanken ('83) observed in Pinus pumilio that both male nuclei pass into the egg-cell, and the same fact was established for Picea vulgaris by Strasburger ('84). Dixon ('94) seems to have been the first to observe that in Pinus sylvestris all four nuclei in the pollen tube, /. e., the two male nuclei, the stalk-cell nucleus, and the tube nucleus pass into the egg-cell of the archegonium. This fact has been confirmed by Blackman ('98) for Pinus sylvestris^ by Murrill (1900) for Tsuga canadensis^ and by Ferguson for Pinus strobus. Accord- ing to Blackman the behavior of the four nuclei in Pinus sylvestris can be easily followed after their entrance into the egg-cell. The two male nuclei around which the cytoplasm of the generative cell can be no longer observed are distinguished by their larger size. In P. strobus one of these nuclei is sometimes larger than the other (Fig. 68, C). The nuclei of the stalk cell and tube are, however, similar, and can scarcely be distinguished from each other. Within the egg one of the two male nuclei moves toward the nucleus of the egg, the other three nuclei remaining near the upper end of the cell. On its way through the cytoplasm of the &^^ the functional male nucleus increases in size, and in some cases in substances stain- ing more readily, but in others the increase in size seems to be due to vacuolation. The nucleus of the egg-cell in Pinus sylvestris at the time of fecundation presents a strikingly peculiar structui-e, which differs from that of the female nucleus in all other plants. After the formation of the ventral canal-cell the female nucleus migrates toward the center of the cell, and, by the time it has reached the middle, it has attained an enormous size, and there is developed within it a rather coarse, uniform, and wide-meshed linin reticulum which persists until a later stage (Fig. 69, A) . Within this linin reticulum the chromatin is distributed in irregular masses of varying size. These masses may be in the form of irregular lumps as if composed of an aggregate of granules, or in shreds or rods with uneven edges. Sometimes they appear globular as small nucleoli. In fact it is quite difficult to distin- guish between some of the small nucleoli and similar chromatin masses, if, indeed, a difference really exists. The quantity of chromatin in the nucleus is proportionally very small. In addition to the linin reticulum there is also present a fine granular substance which appears to be evenly distributed in the nucleus or aggregated along the linin threads. In the former case the nucleus appears more uniformly granular, and its linin reticulum stands out less sharply. The structure of the egg- nucleus in Pinus sylvestris^ as described by Blackman, agrees with 1 66 ARCHEGONIATES. that of my own observations, and from the work of Chamberlain ('99) on Pinus laricio and Murrill (1900) on Tsuga canadensis^ it seems that a similarly constructed nucleus is present in these species. In Pinus strobus (Ferguson, '01) the structure of the egg-nucleus may vary from a most delicate network bearing minute granules to an intei*- rupted, imperfect reticulum composed of large, irregular, diffusely- Fig. 69. — The fusion of the sexual nuclei in Pinus sylvestris. — (After Blackman.) A, egg cell showing male nucleus ( cuvt Fig. 73.— Formation of egg-apparatus and mature embryo-sac in Lilium martagan. A, telophase of third mitosis; the four nuclei, three only shown, form a tetrad; the lower nucleus to the right is the egg-nucleus, the one to left the upper polar nucleus ; plasma membranes delimiting the three cells of egg-apparatus are just formed. B, same stage, perhaps a little later, showing all four nuclei in a plane ; the lower nucleus on left is the upper polar nucleus. C, mature embryo-sac into which the male nuclei have been discharged. «.»., egg-nucleus ; tn.n., male nucleus applied to that of the egg; m.n.', second male nucleus approaching upper polar nucleus; syM., disorganized synergid; />.«., polar nuclei; t.i., trophoplasmic body; ant., antipodal cells. As soon as the end of the pollen tube enters the embryo-sac it opens, discharging the two male gametes and other contents. One of the male nuclei enters the egg-cell and applies itself to the nucleus of the egg, while the other passes on into the cavity of the sac (Fig. 73, C). As soon as the male nuclei have been discharged into the 176 ANGIOSPERMS. embryo-sac and can be distinctly recognized, no trace of the cytoplasm which accompanied them in the tube can be distinguished, so that the exact behavior of this cytoplasm is unknown. Consequently we are concerned hei*e solely with the union of the nuclei. THE FUSION OF MALE AND EGG-NUCLEI, We shall follow first the male nucleus which fuses with that of the egg-cell. It is presumably the first male nucleus which escapes from the pollen tube that unites with the nucleus of the egg, but positive proof on this point is want- ing. In certain spe- cies of Lilium^ and various observers have shown this to be true of many other Angiosperms, the male nucleus, when observed in the egg- cell, is frequently sausage- shaped, worm-like, or S- shaped (Mottier, '97), making one or more spiral - like turns, which is sug- gestive of a worm- like motion, but posi- tive proof of any such movement is want- ing. It applies itself to the nucleus of the ^%g., retaining the form mentioned for some time (Fig. 74, A). The structure of the two sexual nuclei at this stage is accurately shown for Lilium martagon in this figure. The two nuclei are in the resting condition, although the chromatin of the male nucleus is a little more regularly arranged. The male nuclei when in the embryo-sac stain a deeper red, safranin, gentian violet and orange G being used, than the other nuclei of the sac, and for that reason they may be readily recognized. As fusion progresses, the nuclei become quite alike in shape, size and structure (Fig. 74, B). Their membranes gradually disappear at the place of contact, their cavities become one, and the resulting fusion nucleus, which is in the Fig. 74. — Fusion of sexual nuclei. A, vermiform male nucleus applied to egg-nucleus, Lilium martagon. B» egg-cell oi Lilium candidum, showing sexual nuclei in act effusing ; the nuclear membranes have disappeared at place of contact. FATE OF SECOND MALE NUCLEUS IN EMBRYO-SAC. 1 77 resting condition, can scarcely be distinguished from the nucleus of an unfecundated egg. The nucleoli finally unite also. Thfe worm-like or S-shape form of the male nucleus in L ilium ^ first described by the author in 1897 (Mottier, '97, p. 23), has since that time attracted the close attention of students of fecundation generally. Guignard, having observed the same phenomenon in 1899, concluded to designate these vermiform nuclei as antherozoids, evidently attributing to them the power of locomotion. As a matter of fact these nuclei do not possess cilia or any other cytoplasmic organ of loco- motion, nor have the male nuclei in any Angiosperm been found to possess any such structures. Nuclei in many vegetative cells of both plants and animals are known to be able to change their form, and the fact that in the embryo-sac the male nuclei may assume a worm-like shape, which merely suggests a squirming or vermiform motion, is not a sufficient reason for designating them as spermatozoids. So far as is known, all spermatozoids are provided with a cytoplasmic organ of locomotion, existing in the form of a cilium or cilia, and it certainly does not conduce to clearness to apply this term to the male nuclei of the Angiosperms. Strasburger (1900) claims that the vermiform nucleus moves passively in the embryo-sac, basing his opinion upon observa- tions of the embryo-sac of Monotropa in the living condition. A streaming movement was seen in the cytoplasmic strand connecting the egg-cell with the endosperm nucleus, and, in the light of this fact, it is highly probable that the second male nucleus is carried to the endosperm nucleus by that means. THE FATE OF THE SECOND MALE NUCLEUS IN THE EMBRYO-SAC. The fact that one of the male nuclei fuses with a polar nucleus, or with the endosperm nucleus in certain lilies and in species of widely separated families, has also aroused a keen interest among botanists, and has called forth much interesting and suggestive speculation. In 1897 the author called attention to the fact that the second male nucleus in Lilium martagon applied itself to one of the polar nuclei, but the actual fusion was not observed. The plants from which the material was obtained produced few or no seeds that year, and all preparations of embryo-sacs, examined at a time when normally fecundated eggs should have been present, gave only evidence of disorganization, and it was concluded that probably a fusion of the nuclei did not proceed further, which under the circumstances may have been true. Later, other investigators as well as the author have observed this nuclear 178 ANGIOSPERMS. fusion in species of Lilium (Fig. 75, A, B, C). An account of the fusion of one of the male nuclei with the polar nuclei was first pub- lished by Nawaschin ('99) and made known to botanists in general by a reference in the Botanisches Centralblatt. Guignard ('99) in the same year published the results of his obser- vations confirming the statement of Nawaschin. He figured the second vermiform male nucleus in contact with one or both polar nuclei, but none of Guignard's figures showed an actual fusion. Although we are justified in assuming that sexual nuclei, when brought in contact, will fuse, yet the possibility is not excluded that since the sexual nuclei remain side by side for some time before fusion takes place, the causes w^hich have been long known to operate in preventing the formation Fig. 75. — Fusion of second male nucleus with polar nuclei in Lilium )Harta£on. A, an S-shaped male nucleus applied to the upper polar nucleus. B, second male nucleus (shown only in part) and the two polar nuclei close together. C, all three nuclei fusing. of seeds in certain species of Lilium may also prevent the complete fusion of these nuclei after having come in contact. The fusion of a male nucleus with the endosperm nucleus has received different interpretations at the hands of the several investigators. Na- waschin (1900), H. De Vries ('99, 1900) and Correns ('99) evidently see in this fusion a true sexual process, basing their conclusion largely upon the hybrid character of the endosperm of certain varieties of Zea mays. Guignard in his paper upon Tulipa celliana and T. sylveS' iris regards the process as a pseudo-fecundation. From a series of important experiments on the hybridization of several varieties of Zea mays^ Webber (1900) arrives independently at the same conclusion as De Vries, namely, that certain phenomena of xenia are the result of the fusion of one of the male nuclei with the endosperm nucleus. As a result of the crossing, the endosperm, pro- duced in the same embryo-sac with the hybrid embryo sporophyte, FATE OF SECOND MALE NUCLEUS IN EMBRYO-SAC. 1 79 shows certain well-marked characters of the male parent, and accord- ing to the hypothesis of Webber, De Vries, and others, these hybrid characters are transmitted by the male nucleus. In some cases the endosperm does not reveal hybrid characters, but only those of the mother plant, and Webber explains the fact by assuming that in those cases the endosperm nucleus may not have been fecundated. As an explanation of another peculiar feature of xenia in certain varieties of maize, which is shown by a variegated or mosaic endosperm, Webber suggests that probably the second male nucleus may not have united with the endosperm nucleus, but it may have been able to divide in- dependently. If this should occur, there would then be formed in the embryo-sac nuclei of two distinct characters, one group from the division of the endosperm nucleus and one from the sperm nucleus. Or a second hypothesis lies in the probability that the second male nucleus fuses with one of the polar nuclei, and that after fusion the other polar nucleus is repelled and develops independently. In view of the fact that in the sea-urchin (Boveri, '95) the male nucleus is capable of independent division under certain circumstances, these hypotheses are certainly very suggestive, but they have, as yet, among plants no support based upon observation, especially since partheno- genesis is unknown in maize. Before these suggestions can be of much value in explaining the phenomenon, it is necessary to know whether a male nucleus is of itself capable of division in the embryo- sac, and whether one of the polar nuclei without having united with the other or with a sperm nucleus is also capable of independent division. Although the union of a male nucleus with the endosperm nucleus may be conclusively shown to be the cause of hybrid endosperm in maize, yet that fact alone is not sufficient to justify the unqualified conclusion that the fusion represents a real fecundation. Strasburger, in discus- sing this question at some length in the Botanische Zeitung (pp. 293- 316, 1900), argues forcibly against the doctrine of a double sexual process as understood by Nawaschin, and proposes a different interpre- tation of the two sets of nuclear fusions. For the union of the male nucleus and that of the egg-cell which results in an individual sporophyte, the expression generative fecundation is used, while the fusion of the other male nucleus with the endosperm nucleus is designated vegetative fecundation. In the interpretation of Strasburger, the need of genera- tive fecundation by means of sexual nuclei of different origin lies in the equalization of individual variations, which is necessary for the continu- ance of the species, while in vegetative fecundation there is merely the l8o ANGIOSPERMS. manifestation of a growth stimulus. Vegetative fecundation according to this interpretation finds its parallel in such phenomena as described by Klebs ('98, 1900), Loeb ('99, '01) and Nathansohn (1900), in which, by means of physical or chemical stimuli, such as increased tempera- ture or an increase of the osmotic power of the surrounding fluid, unfecundated egg-cells have been made to develop parthenogenetically through certain embryonic stages. According to the view of Stras- burger, therefore, sexual reproduction embraces fundamentally two great and far-reaching factors, namely, the union of hereditary ele- ments and the imparting of a growth stimulus. In the fusion of a male nucleus with the endosperm nucleus, only one of these factors, the stimulus to growth, is manifested, since the interrupted growth of the endosperm is enabled to continue. The result is the same whether the second sperm nucleus unites with the endosperm nucleus or not, and furthermore because the endosperm is not an individual in the sense that the embryo sporophyte is an individual. It is further true that the endosperm nucleus may divide and give rise to several nuclei before the contents of the pollen tube are discharged into the embryo- sac, and in case that no pollen tube reaches the embryo-sac, these same endosperm nuclei never continue their development. It is reasonable to conclude, therefore, that a growth stimulus may be imparted to the endosperm by the act of fecundation in the egg-cell, just as the vegetative tissue of certain parts of the pistil are stimulated to growth by the presence of the pollen tube. Many who agree with Strasburger may probably not consider it necessary or advisable to use the term " vegetative fecundation." The author does not see the necessity of associating the idea of fecundation with this process of nuclear fusion, for the reason that nuclear fusions in vegetative cells do not signify an act of fecundation. In the light of all the known facts, it seems that we have to do here with purely vegetative fusions, and that we are not justified in attributing to such nuclear fusions the idea of sexuality. Although the upper polar nucleus is the sister of the egg-nucleus, it does not necessarily follow that the former is also a female nucleus, since it is certainly not true that the sister cells of egg-cells are even potential gametes. If such an assumption were accepted, then the ventral canal-cell of the arche- goniates might be considered an egg-cell, a doctrine to which the author can not, as yet, subscribe. BIBLIOGRAPHY. Andrews, F. M., 'oi : Karyokitiesis in Magnolia and Liriodendron. Bot. Cen- tralb. Beihefte, ti : Ileft 2, 1901. Arnoldi, W., '00: III. ^"Beitrage zur Morphologic der Gymnospermen. Flora, 87: 46-63, 1900. . IV. Beitrage zur Morphologic der Gymnospermen. Flora, 87 : 194- 203, 1900. Artary, a., '90: Zur Entwickelungsgeschichte dcs Wasscrnetzcs. Moskau, 1890. Bary, a. de, '58: Untersuchungen iiber die Familie der Conjugaten. Leipzig, , '63 : Ueber die Fruchtentwickelung der Ascomyceten. Leipzig, 1863. , '83 : Zu Pringsheims neuen Beobachtung iiber den Befruchtungsact der Gattung Achlya und Saprolegnia. Bot. Ztg., 41 : 38-54, 1883. Bary, de, and Woronin, '70: Beitrage zur Morphologic und Physiologic der Pilze. Dritte Reihe. (Abhandl. der Senkenberg. naturf. Gesellsch. Bd. 7.) Frankfurt a. M., 1870. • , '8i : Beitrage zur Morphologic und Physiologic der Pilze. Vierte Reihe. TAbhandl. der Senkenberg. naturf. Gesellsch., Bd. 12,) Frank- furt a. M., 1881. Bauer, E. , '98 : Zur Frage nach der Sexualitat der Collemacecn. Ber. d. Deutsch. Bot. Gesellsch., 16: 363. 1898. Behrkns, J., '86 : Beitrage zur Kcnntniss der Befruchtungsvorgange bei Fucus vesiculosus. Ber. d. Deutsch. Bot. Gesellsch., 4: 92-103, 1886. ,'90: Einige Beobachtungen iiber die Entwickelung des Oogons und der Oosphare von Vaticheria. Ber. d. Deutsch. Bot. Gesellsch., 8: 314- 318, 1890. Belajeff, W.,'91: Zur Lchre von dem Pollenschlauch der Gymnospermen. Ber. d. Deutsch. Bot. Gesellsch., 9: 280-285, 1891. , '94 : Ueber Bau und Entwickelung der Spermatozoiden der Pflanzen. Flora, 79: 1-48, 1S94. , '97 : Ueber den Nebenkern in spermatogencn Zellen und die Sperma- togenese bei den Farnkrautern. Ber. d. Deutsch. Bot. Gesellsch., 15 : 337-339. ^897. , '97: Ueber die Spermatogenese bei den Schachtelhalmen. Ber. d. Deutsch. Bot. Gesellsch., 15: 339, 1897. , '97 : Ueber die Aehnlichkeit einiger Erscheinungen in der Spermato- genese bei Thieren und Pflanzen. Ber. d. Deutsch. Bot. Gesellsch., 15 : 342, 1897. , '98 : Ueber die Cilienbildner in den spermatogencn Zellen. Ber. d. Deutsch. Bot. Gesellsch., 16: 140, 1898. .'99: Ueber die Centrosomen in den spermatogencn Zellen. Ber. d. Deutsch. Bot. Gesellsch., 17: 199-205, 1899. Beneden, E. van, '83: Rech. sur la maturation de I'ceuf, la f(6condation, et la division ccllulairc. Arch. d. Biol., 4: 1883. Berlese, a. N., '98: Ueber die Befruchtung und Entwickelung der Oosphare bei den Peronosporeen. Jahrb. f- wiss. Bot., 31 : 159-195, 1898. Berthold, G., '81 : Die geschlechtliche Fortpflanzung der eigentlichen Phae- osporeen. Mittheilungen aus der Zoologischcn Station zu Neapel, 2 : 401,1881. (Here the earlier literature.) Bessey, E. a., '01 : Notes on the spermatozoid of Ginkgo. Science, N. S., 13: 255, 1901. Blackman, V. H., '98: The cytological features of fertilization and related phenomena in Pinus sylvestris. Phil. Trans. Royal See. of London, Series B, 190 : 395-426, 1898. Boveri, T., '95 : Ueber die Befruchtung und Entwickclungsfahigkeit kcrnloser Seeigel-Eier, etc. Archiv. f. Entwickclungsmcchanik, 2 : 1895. 181 152 FECUNDATION IN PLANTS. BuLLER, A. H. R., 'oo : Contributions to our knowledge of the spermatozoa ot Ferns. Ann. Bot., 14: 543-582, 1900. Campbell, D. H., '88: The development of Pilularia globulifera L. Ann. Bot., 2: 233-264, 1888. , '92 : On the prothallium and embryo of Osmunda claytoniana L. , and O. ctitnamomea L. Ann. Bot., 6: 49-94, 1892. , '97 : A morphological study of Naias and Zanichellia. Proc. Cali- fornia Acad. Sci., 3d Ser., Bot., i : 1897. , '99 : I. Notes on the structure of the embryo-sac in Sparganium and I^ysichiton. Bot. Gaz., 27: 153-165, 1899. , '99 : II. Studies on the flower and embryo in Spargantum. Proc. Calif. Acad. Sci., 3d Ser., Bot., i : 394, 1899. , 99: III. Die Entwickelung des Embryosackes von Peperomia pellu- cida. Ber. d. Deutsch. Bot. Gesellsch., 17: 352-456, 1899. , '01: The embryo-sac of Peperomia. Ann. Bot., 15: 103-118, 1901. Chamberlain, C. J., '99: Fertilization oi Pinus sylvestris. Bot. Gaz., 27: 268, 1899. Chmielewskij, v., '90: Eine Notiz Uber das Verhalten der Chlorophyllbander in den Zygoten der Spirogyra-Arten. Bot. Ztg., 48: 773-780, 1890. , '92 : Materialen zur Morphologie und Physiologie des Sexualprocesses bei den niederen Pflanzen. (Russian.) Ref. in Bot. Centralbl., p. 264, 1892. CoHN, F., '55 : Ueber die Fortpflanzung von Sphceroplea annulina. Bericht tiber die zur Bekanntmachung geeigneten Verhandl. der K. Pr. Akad. d. Wiss. zu Berlin, 1855, pp. 335-351. CoRRENs, C, '99: Untersuchungen iiber die Xenien bei Zea Mays. Ber. d. Deutsch. Bot. Gesellsch., 17: 401-417, 1899. Dangeard and Leger, '94, '95 : (i) Recherches sur la Structure des Mucorin^es ; (2) La reproduction sexuelle des Mucorindes. Le Botaniste, 4 : 1894, 1895. Dangeard, '94, '95 : Considerations sur les phdnomfenes de reproduction chez les Phycomycfetes. Le Botaniste, 4 : 249, 1894-1895. Darbishire, O. v., '99 : Ueber die Apothecienentwickelung der Flechte Physcia pulverulenie (Schreb.) Nyl. Jahrb. f. wiss. Bot., 34: 329, 1899. Davis, B. M., '96: The Fertilization of Batrachospermum. Ann. Bot., 10: 49, 1896. , '98 : Kerntheilung in der Tetrasporenmutterzelle bei Corallina offici- nalis L. var. Mediterranea. Ber. der Deutsch. Bot. Gesellsch., 16 : 266- 272, 1898. , '00: The fertilization of Albugo Candida. Bot. Gaz., 29: 299, 1900. Dixon, H. H., '94: Fertilization of Pinus sylvestris. Ann. Bot, 8: 21, 1894. Dodel, a., '76: Ulothrix zonata. Jahrb. f. wiss. Bot., 10: 417-550, 1876. Fairchild, D. G., '97 : Ueber Kerntheilung und Befruchtung bei Basidiobolus ranarum Eidam. Jahrb. f. wiss. Bot., 30: 285-295, 1897. Farmer, J. B., and Williams, J. Ll., '96: On the fertilization and the segmenta- tion of the spore in Fucus. Proc. Royal Soc, 60: 1896. Farmer, J. B., '98 : Contributions to our knowledge of the Fucaceae : their life- history and cytology. Phil. Trans. Royal Soc, London, 190: 623- 645, 1898. Ferguson, M. C., '01 : I. The development of the pollen-tube and the division of the generative nucleus in certain species of pines. Ann. Bot., 15: 193-222, 1901. , '01 : II. The development of the egg and fertilization in Pinus strobus. Ann. Bot., 15: 435-479, 1901- Fujii, K., '00 : On the morphology of the spermatozoid of Ginkgo biloba. (The text is in Japanese; the explanation of the figures in English.) Bot. Magazine, Tokyo, 14: 260-266, 1900. GoLENKiN, M., '00: Ueber die Befruchtung bei Spkceroplea annulina und iiber die Structur der Zellkerne bei einigen griinen Algen. Ref. in Bot. Centralbl., 84: 284, 1900. Goroschankin, J., '83 : Zur Kenntniss der Corpuscula bei den Gymnospermen. Bot. Ztg., 41: 825-831, 1883. Grbgoire, v., '99: Les Cinfese Polleniques chez les Liliac^es. La Cellule, 16: 236-296, 1899. BIBLIOGRAPHY. 183 Gruber, E., '01 : Ueber das Verhalten der Zellkerne in den Zygosporen von Sporodinia grandis Link. Ber. d. Deutsch. Bot. GeselUch., 19 : 51-55, 1901. GuiGNARD, L., '99: Le d^veloppement du pollen et la reduction chromatique dans le Naias major. Arch. d'Anat. Microscopique, 2 : 455-509, 1899. . '99= I- Sur les anthdrozoides et la double copulation sexuelle chez les v^gdtaux angiospermes. Revue G^n^rale de Botanique, 11: 129, 1899. , '99 = II' Les ddcouvertes r^centes sur la fdcondation chez les vig6- taux angiospermes. Volume jubilaire du Cinquantenaire de la Soci^td de Biologic, p. 189, 1899. 1 '00 : I. L'appariel sexuel et la double fdcondation dans les Talipes. Ann. d. Sci. Nat. Bot., S^rie VIII, ii : 365. 1900. , '00 : II. Nouvelles recherches sur ia double f^condation chez les vdg^taux angiospermes. Comptes Rendus, Acad. d. Sci., 13K 153,1900. , '01 : La double f^condation dans le MaYs. Journal de Bot., 15: 1901. Haberlandt, G.. '90 : Zur Kenntnissder Konjugation bei Spirogyra. Sitzungs- ber. d. Wiener Akad., Math.-nat. CI., XCIX, Abt. i : 390-400, 1890. Harper, R. A., '95: Die Entwickelungder Peritheciums bei Sp/icerotheca castag- nei. Ber. d. Deutsch. Bot. Gesellsch., 13: 475, 1895. , '96 : Ueber das Verhalten der Kerne bei der Fruchtentwickelung einiger Ascomjceten. Jahrb. f. wiss. Bot., 29: 655, 1896. . '97 : Kerntheilung und freie Zellbildung in Ascus. Jahrb. f. wiss. Bot., 30: 249-284, 1897. 1'99'- Cell-division in Sporangia and Asci. Ann. Bot, 13: 467-534, 1899. , '00 : Sexual reproduction in Pyronema conflnens and the morphology of the Ascocarp. Ann. Bot, 14: 321-400, 1900. Hartog, M., '89: Recherches sur la Structure des Saprolegnides. Comptes Rendus, 108: 687-689. 1889. , '95 : On the cytology of the vegetative and reproductive organs of the SaprolegniecB. Trans. Royal Irish Acad., 30: 1895. , '96: The cytology of Saprolegnia. Ann. Bot., loi : 1896, . '99: The alleged fertilization in the Saprolegniece. Ann. Bot., 13: 447, 1899. Hassenkamp, a., '02 : Ueber die Entwickelung der Cystocarpien bei einigen Florideen. Bot. Ztg., 60: 65-85, 1902. Heinricher, '83: Zur Kenntniss der Algengattung Spharoplea. Ber. d. Deutsch. Bot Gesellsch.. I: 433-450, 18S3. HiRASE, S., '95; Etudes sur la fdcondation et lembryogdnie du Ginkgo biloba. Journ. Coll. Sci. Imp. Univ., Tokyo, VIII : 307-322, 1895. , '97 : Untersuchungen Uber das Verhalten des Pollens von Ginkgo biloba. Bot. Centralbl., 59 : 1897. ) '98 : Etudes sur la fdcondation et I'embryog^nie du Ginkgo biloba. (Second m^moire.) Journ. Coll. Sci., Imp. Univ., Tokyo, 12: 103-149, 1898. Ikeno, S., '96 : The spermatozoids of Cycas revolula (Japanese). Bot. Magazine, Tokyo, 10: 1896. , '97 : Vorlaufige Mittheilung fiber die Spermatozoiden bei Cycas revo' luta. Bot. Centralbl., 59: 1897. , '98: Zur Kenntniss des sog. '* centrosomahnlichen Kdrpers " im Pollenschlauch der Cycadeen. Flora, 85 : 15, 1898. , '98: Untersuchungen fiber die Entwickelung der Geschlechtsorgane und den Vorgang der Befruchtung bei Cycas revoluta. Jahrb. f. wiss. Bot., 32 : 357-379, 1898. , '01 : Contribution a I'dtude de la f^condation chez le Ginkgo biloba. Ann. d. Sci. Bot., S6rie 8, 13: 305-316, 1901. Ikeno and Hirase, '97 : Spermatozoides in Gymnosperms. Ann. Bot, ii : 344, 1897. Johnson, D. S., '00: On the endosperm and embryo of Peperotnta pellucida. Bot. Gaz., 30 : i-io, 1900. JosT, L., '95 : Beitrage zur Kenntniss der Coleochateen. Ber. d. Deutsch. Bot. Gesellsch,, 13: 433-452. 1895. 184 FECUXDATION IN PLANTS. JuEL, H. O., '00: Vergleichende Unlersuchungen iiber typische und parthe- nogenetische Fortpflanzungbei der Gattung Antennaria. Kongl. Veten- skaps-Akademiens Handling:ir, 33: 1-58, 1900 , '00: Beitrage zur Kenntniss der Tetradentheilung. Jahrb. f. wiss. Bot.,35: 626-659, 1900. Karstkn, G., '00 : Die Auxosporenbildungbei der Gattung Cocconets Surirella und Cymatofleiira. Flora, 87 : 253-283, 1900. Klkbahn, H., '88 : Ueber die Zygosporen einiger Conjugaten. Ber. d. Deutsch. Bot. Gesellsch.,6: 160-166, 1888. . '91 : I- Studien iiber Zjgoten. Die Keimung von Closierium und Cosmarium, Jahrb. f. wiss. Bot., 22 : 415, 1891. , '92: II. Studien iiber Zjgoten. Jahrb. f. wiss. Bot., 24: 235, 1892. . '96: Beitrage zur Kenntniss der Auxosporenbildung. Jahrb. f. wiss., Bot., 29 : 595, 1896. , '99 : Die Befruchtung von Spkceroplea annulina Ag. Festschrift fur • Schwendener. 1899. Klbbs, G., '91 : Ueber die Bildung der Fortpflanzungszellen bei Hydrodictyon utriculaium. Bot. Ztg., 49 : 790, 1891. , '96: Die Bedingungen der Fortpflanzung bei einigen Algen und Pil- zen. Jena. 1896. (See also Jahrb. f. wiss. Bot., Bd. 32, 33, 35.) Kny, L., '84: Bot. Wandtafeln. Taf. lxiii-lxiv. Berlin, 18S4. Land, W. J. G., '00: Double fertilization in Composites. Bot. Gaz., 30: 252, 1900. Lautkrborn, R., '96: Untersuchungen iiber Bau, Kerntheilung und Bewegung der Diatomeen. Leipzig, 1896. Leger, M., '95 : Structure et ddveloppement de la zygospore du Sporodinia grandis. Revue G^n. de Bot., 7: 481-486, 1895. LiNDAU, G., '88: Ueber die Anlage und Entwickelung einiger Flechten-apothe- cien. Flora, 71 : i,i,\-i^^^, 1888. , '99: Beitrage zur Kenntniss der Gattung Gyrophora. Festschrift fiir Schwendener. Berlin, 1899. LoEB, J.,'oi: Experiments on artificial parthenogenesis in Annelids (C/its/o/- teris) and the nature of the process of fertilization. Am. Jour, of Physi- ology, 4: 423-459- 1901- LoTSY, J., '99: Contributions to the life-history of the genus Guelum, Ann. du Jardin Bot. de Buitenzorg, 16: 2d Ser. i : 46-110, 1899. MiYAKE, K., '01: The fertilization of Fythium De Baryanum. Ann. Bot., 15: 653-666, 1901. , '02 : The spermatozoid of Ginkgo. Jrnl. Applied Microscopy and Laboratory Methods, 5 : 1773-1780, 1902. Moll, J. W., '93: Observations on karyokinesis in Spirogyra. Verhand. d. koninkl. Akad. van Wetensch. te Amsterdam, Sect. II, Deel i, Nr. 9, 1893. 36 p. MoTTiER, D. M., '97: Beitrage zur Kenntniss der Kerntheilung in den Pollen- mutterzellen einiger Dikotylen und Monokotylen. Jahrb. f. wiss. Bot., 30: 169-294, 1897. • , '98 : Ueber das Verhalten der Kerne bei der Entwickelung des Embryo- sacks und die Vorgange bei der Befruchtung. Jahrb. f. wiss. Bot., 31 : 125-157, 1898. , '00: Nuclear and cell division in Dictyota dichotoma. Ann. Bot., 14: 163-192, 1900. , '03 : The behavior of the chromosomes in the spore mother-cells of higher plants and the homology of the pollen and embryo-sac mother- cells. Bot. Gaz., 35: 250-282, 1903. MuNiER, A., Le nucleole de Spirogyra. La Cellule, 3 : 333, . MURBECK, '01 : Parthenogenetische Embryobildung in der Gattung Alchemilla. Lunds Universitets Arsskrift, 36: 1-40, 1901. MuRRiLL, W. A., '00: The development of the archegonium and fertilization in the hemlock spruce ( 757<^a canadensis C2irr.). Ann. Bot., 14: 5S3, 1900. Nathansohn, a., '00: Ueber Parthenogenesis bei Marsilia und ihre AbhSn- gigkeit von der Temperatur. Ber. d. Deutsch. Bot. Gesellsch., 18: 99- 109, 1900. BIBLIOGRAPHY. 185 Nawaschin, S., '99: I. Neue Beobachtungen iiber Befruchtung bei Fritillaria tenella und Lilium Mariagon. Reference in Bot. Centralb., 77 : 62, 1899. , '99: II. Resultate einer Revision der Befruchtungsvorgange bei Lilium martagon und Fritillaria tenella. Bull, de I'Acad. imp. des Sciences de St. Petersb., 9: 377, 1899. ,'00: Ueber die Befruchtungsvorgange bei einiger Dicotjledonen. Ber. d. Deutsch. Bot. Gesellsch., 18: 224. 1900. Oltmanns, F.,'99: Ueber der Sexualitat der Ectocarpeen. Flora, 86: 1-14, 1899. .'89: Beitrage zur Kenntniss der Fucaceen. Bibliotheca Botanica, Heft 14. 1889. , '95 : Ueber die Entwickelung der Sexualorgane bei Vaucheria. Flora, 80: 3S8-420, 1895. . '98 : Die Entwickelung der Sexualorgane bei Coleochcete fulvinata. Flora, 85 : 1-14, 1895. , '98: Zur Entwickelungsgeschichte der Florideen. Bot. Ztg., 56: 99- 140, 1898. OsTERHOuT, W.J. v., '97: Ueber Entstehung der karyokinetischen Spindel bei Equisetum. Jahrb. f. wiss. Bot. 30: 159-168, 1897. , '96: On the life-historj of Rhabdonia tenera ]. Ag. Ann, Bot., 10: 403, i8()6. ,'00: Befruchtung bei BatracAospermutn. Flora, 87: 109-115, 1900. Overton, C, '88: Ueber den Conjugationsvorgang bei Spirogyra. Ber. d. Deutsch. Bot. Gesellsch., 6: 68-72, 1888. Overton, J. B., '02: Parthenogenesis in Thalictrum purpurascens. Bot. Gaz., 33 : 363-375. 1902. Pfkffer, W., '84: Locomotorische Richtungsbewegungen durch chemische Reize. Untersuchungen aus. d. Bot. Inst, zu Tubingen, i : 1884. Phillips, R. W., '95 : I. On the development of the cystocarp in Rhodomelacece (lihodomelia, Polysiphonia). Ann. Bot., 9: 289, 1895. , '96 : II. Dasya, Chondria, Launencia, and Polysiphonia. Ann. Bot., 10: 1896. , '98. The development of the cystocarp in Rhodymeniales: Delesseri- aceae. Ann. Bot., 12: 173, 1898. Pringshkim, N., '57 : Beitrage zur Morphologic und Systematik der Algen. II. Die Saprolegnieen. Gesammelte Abhandl. 2: 57, 1857. (From Jahrb. f. wiss. Bot., I : 1857. , '58, '60: Beitrage zur Morphologic und Systematik der Algen. Ge- sammelte Abhandl., i : 279, 1895. (Jahrb. f. wiss. Bot., 2 : i860.) , '73 : Weitere Nachtrage zur Morphologic und Systematik der Sapro- legnieen. Gesammelte Abhandl., 2: 177, 1873. (Jahrb. f. wiss. Bot., 9: 191, 1873.) , '82 : Neue Beobachtung iiber den Befruchtungsact der Gattungen Achlya und Safrolegnia. Gesammelte Abhandl., 2 : 167, 1882. Raciborski, M. von, '66: Ueber den Einfluss aussercr Bedingungen auf die Wachstungsweise des Basidiobolus ranartim. Flora, 82 : 107-132, 1896. Rauwenhoff, N. W. p., '88: Recherches sur le Sphceroplea annulina Ag. Archiv. N^erlandaises des sc. exact, et nat, 22 : 91-142, 1888. Sargant, E., '99: On the presence of two vermiform nuclei in the fertilized embryo-sac of Lilium martagon. Proc. Royal Soc. , 65 : 163, 1899. , '00 : Recent work on the results of fertilization in angiosperms. Ann. Bot., 14: 689-712, 1900. ScHMiTZ, F., '79: Untersuchungen iiber die Zellkerne der Thallophyten. Ver- handl. d. Naturhist. Vereins d. Preuss. Rheinlande u. Westfalens, 1897. , '82 : Die Chromatophoren der Algen. Bonn, 1882. , '83 : Untersuchungen iiber die Befruchtung der Florideen. Sitzungs- ber. der Acad, der Wissenschaft zu Berlin, i : 215-258, 1883. Shaw, W. R., '97: Parthenogenesis in Marsilia. Bot. Gaz., 24: 114, 1897. , '98: The fertilization of Onoclea. Ann. Bot, 12 : 261-285, 1898. , '98: Ueber die Blepharoplasten bei Onoclea und Marsilia. Ber. d. Deutsch. Bot. Gesellsch., 16: 177-184, 1898. Stahl, E., '77: Beitrage zur Entwickelungsgeschichte der Flechten. Leipzig, 1887. l86 FECUNDATION IN PLANTS. Stevens, F. L., '99: The compound oosphere of Albugo bliti. Bot. Gaz., 28: 149, 1899. , '01 : Gametogenesis and fertilization in Albugo. Bot. Gaz., 32: 77, 1901. Strasburgkr, E., '69: Die Befruchtung bei den Coniferen. 1869. , '72 : Die Coniferen und die Gnetaceen. 1872. J '78 : Ueber Befruchtung und Zelltheilung. Jena, 1878. , '79: Ueber das Verhalten des Pollens und die Befruchtungsvorgange bei den Gymnospermen. Histol. Beitrage, 4 : 1892. , '79: Die Angiospermen und die Gymnospermen. 1879. , '82 : Ueber den Theilungsvorgang der Zellkerne. 1882. , '84: Neue Untersuchungen iiber den Befruchtungsvorgang bei den Phanerogamen als Grundlage fiir eine Theorie der Zeugung. Jena, 1884. , '92 : Schwarmsporen, Gameten, etc. Histol. Beitrage, Heft 4, 1892. , '94 : The periodic reduction of the number of chromosomes in the life- history of living organisms. Ann. Bot., 8: 281-316, 1894. , '95: Karyokinetische Probleme. Jahrb. f. w. Bot., 28: 151-204, 1895. , '97: Kerntheilung und Befruchtung bei Fuctis. Jahrb. f. w. Bot., 30: 351-374, 1897. , '98: Die pflanzliche Zellhaute. Jahrb. f. w. Bot., 31 : 511-596, 1898. ,'00: Ueber Reductionstheilung, Spindelbildung, Centrosomen und Cilienbildner im Pflanzenreich. Histol. Beitrage, Heft 6, 1900. ,'00: Einige Bemerkungen zur Frage nach der " doppelten Befruch- tung" bei den Angiospermen. Bot. Ztg., 58: 295, 1900. Swingle, W. T., '97 : Zur Kenntniss der Kern- und Zelltheilung bei den Sphace- lariaceen. Jahr. f. wiss. Bot., 30: 297-350, 1897. , '98 : Two new organs of the plant cell. Science, N. S., 7 : 119, 1898. Thaxter, R., '96: Contribution toward a monograph of the Laboulbeniacece. Mem. of Am. Acad, of Arts and Sciences, 12: 187-429. Boston, Decem- ber, 1896. Thom, C, '99: The process of fertilization in Aspidium and Adiantum. Trans. Acad. Sci. St. Louis, 9: 285-314, 1899. Thomas, E. N., '00: I. On the presence of vermiform nuclei in a dicotyledon. Ann. Bot., 14: 318, 1900. , '00: II. Double fertilization in a dicotyledon — Caltha palustris. Ann. Bot., 14: 577, 1900. TiMBERLAKE, H. G., 'oi : Swarmsporc formation in Hydrodictyon utriculatum Roth. Bot. Gaz., 31 ; 203, 1901. Treub, M., '81 : Recherches sur les Cycad^es. Ann. du Jardin Bot. de Buiten- zorg, 2 : 32-53, 1881. Trow, A. H., '95: The karyology of Sapiolegnia. Ann. Bot., 9: 609-652, 1895. , '99 : Observations on the biology and cytology of a new variety of Achlya Americana. Ann. Bot., 13: 131, 1899. , '01 : Observations on the biology and cytology of Pythium uliimum, n. sp. Ann. Bot., 15 : 261-311, 1901. Vries, H. de, '99 : Sur la fdcondation hybride de I'albumen. Comptes rendus d. Acad. d. Sci., 129: 973-975, 1899. Wager, H., '96: On the structure and reproduction in Cysiopus candidus Lev. Ann. Bot., 10: 295-341, 1896. , '99: Sexuality in fungi. Ann. Bot., 13 : 575-597, 1899. , '00: On the fertilization oi Peronospora parasitica. Ann. Bot, 14: 263-297, 1900. Webber, H. J., '97 : I. Peculiar structures occurring in the pollen tube of Zatnia. Bot. Gaz., 23: 453-458, 1897. , II. The development of the antherozoids of Zatnia. Bot. Gaz., 24: 16-22, 1897. , III. Notes on the fecundation of Zatnia and the pollen tube apparatus of Gittkgo. Bot. Gaz., 24: 223-234, 1897. , '00: Xenia, or the immediate effect of pollen in maize. Bull. No. 22, Division of Vegetable Physiology and Pathology, U. S. Dept. Agrl., Sept. 12, 1900. , '01 : Spermatogenesis and fecundation in Zatnia. Bulletin No. 2, Bureau of Plant Industry, U. S. Dept. Agrl., 1901. (Complete bibliog- raphy here.) BIBLIOGRAPHY. 187 WiEGAND, K. M., '00: The development of the embrjo-sac in some monocoty- ledonous plants. Bot. Gaz., 30: 25-46, 1900. WiLLE, N., '94: Ueber die Befruchtung bei Ncmalton multifidum. Ber. d. Deutsch. Bot. Gesellsch., 12 : (57)-(i6o), 1894. Wilson, E. B., '61 : Experimental studies in cytology. I. A cytological study of artificial parthenogenesis in sea-urchin eggs. Archiv. fiir Entwicke- lungsmechanik der Organismen, 12 : 529-588, 1901. Winkler, H., 'oi : Ueber Merogonie und Befruchtung. Jahrb. f. wiss. Bot., 36: 753-756, 1901- WissELiNGH, C. VAN, '98: Ueber den Nucleolus von Spirogyra. Bot. Ztg., 56: 195-226, 1898. UNIVERSITY OF CALIFORNIA BRANCH OF THE COLLEGE OF AGRICULTURE THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW 2 6 193S ■'lln27'400FT •' • iilW22'58 1 ■^ 5m-8 ,'26 LIBRARY, BRANCH OF THE COLLEGE OF AGRICULTURE