V BIOLOGY LIBRARY G VOL. XXXVIII, No. 449 MAY, 1904 THE AMERICAN NATURALIST CONTENTS Page I. The Anatomy of the North American Coniferales together with Certain Exotic Species (continued.} PROFESSOR D. P. PENHALLOW 331 II. Further Instances of Malar Division ..... DR. ALES HRDLICZA 361 in. Studies on the Plant Cell I DR. BRADLEY MOORE DAVIS 367 IV. Notes and Literature: General Biology, Plankton of the Illinos River, Where 397 did Life Begin ? Bermuda, Morgan on Evolution and Adaptation Zoology, Zoological Investigations in the Malay Archipelago, Davison's 399 Anatomy of the Cat, Notes Botany, The Journals, Notes . . . . 402 V- Correspondence . 405 BOSTON, U. S. A. GINN & COMPANY, PUBLISHERS 19 BEACON STREET Ifew York Chicago London, W. C, 70 Fifth Avenue 378-388 Wabash Avenue 9 St. Martin'* Street Enisrtd at tkt Pest-OffUl, Batten, Matt., at Steonti-ClaM Moiltfati*. BIOLOGY The American Naturalist. ASSOCIATE EDITORS : J. A. ALLEN, PH.D., American Museum of Natural History, New York. E. A. ANDREWS, u.T>.,fohns Hopkins University, Baltimore. WILLIAM S. BAYLEY, FH.D., Colby University, Waitrvill* DOUGLAS H. CAMPBELL, PH.D., Stanford University. J. H. COMSTOCK, S.B., Cornell University, Ithaca. WILLIAM M. DAVIS, M.E,, Harvard University, Cambridge. ALES HRDLICKA, M.D., U.S. National Museum, Washington. D. S. JORDAN, LL.D., Stanford University. CHARLES A. KOFOID, PH.D., University of California, Btrkelty. J, G. NEEDHAM, PH.D., Lake Forest University. ARNOLD E. ORTMANN, PH.D., Carnegie Museum, Pittsburg. D. P. PENHALLOW.D.SC..F.R.M.S., McGUl University, Montreal, H. M. RICHARDS, S.D., Columbia University, New York. W. E. RITTER, PH.D., University of California, Berkeley. ISRAEL C. RUSSELL, LL.D., University of Michigan, Ann Arbor. ERWIN F. SMITH, S.D., U.S. Department of Agriculture, Washington. LEONHARD STEJNEGER, LL.D., Smithsonian Institution, Washington, W. TRELEASE, S.D., Missouri Botanical Garden, St. Louis. HENRY B. WARD, PH.D., University of Nebraska, Lincoln. WILLIAM M. WHEELER, PH.D., American Museum of Natural History, New York. THE AMERICAN NATURALIST is an illustrated monthly magazine of Natural History, and will aim to present to its readers the leading facts and discoveries in Anthropology, General Biology, Zoology, Botany, Paleontology, Geology and Physical Geography, and Miner- alogy and Petrography. The contents each month will consist of leading original articles containing accounts and discussions of new discoveries, reports of scientific expeditions, biographical notices of distinguished naturalists, or critical summaries of progress in some line ; and in addition to these there will be briefer articles on various points of interest, editorial comments on scientific questions of the day, critical reviews of recent literature, and a quarterly record of gifts, appointments, retirements, and deaths. All naturalists who have anything interesting to say are invited to send in their contributions, but the editors will endeavor to select for publication only that which is of truly scientific value and at the same time written so as to be intelligible, instructive, and interesting to the general scientific reader. All manuscripts, books for review, exchanges, etc., should be sent to THE AMERICAN NATURALIST, Cambridge, Mass. All business communications should be sent direct to the publishers. Annual subscription, $4.00, net, in advance. Single copies, 86 cents. Foreign subscription, $4.60. GINN & COMPANY, PUBLISHERS. -f >-^ 4 STUDIES ON THE PLANT CELL. I. BRADLEY MOORE DAVIS. INTRODUCTION. is of papers that will follow one American Naturalist. They will in plant cells and the most histories, largely from the point of tudent of developmental processes, has entirely outrun the general several botanical text books and zoologists. We shall attempt to subject in its present state with nt papers ; but this is not to be an ire that is already very large and :d far more satisfactorily several passed through the criticism that y active botanical investigation, ^ason to be proud of the achieve- -esearch upon the morphology and r much of the best work of recent 'his in itself has been a great stim- nese brief accounts which he hopes : to a clearer understanding of the will also serve to contrast the pro- mts with those of the animal cell in several foreign works and in n Development and Inheritance. lly gratified if these papers should /ards investigations on the plant "cemriS!Tl l fl!!Wr^ffl^^^B^ prevalent among botanists. There is a tendency to regard cell studies as a very special field of botanical research with elaborate technique which the average 367 The American Naturalist. BIOi.OGY LIBRARY ASSOCIATE EDITORS: J. A. ALLEN, PH.D., American Museum of Natural History, New York. E. A. ANDREWS, PH.D.,/^J Hopkins University, Baltimore. WILLIAM S. BAYLEY, FH.D., Colby University, mutrvili* DOUGLAS H. CAMPBELL, PH.D., Stanford University. J. H. COMSTOCK, S.B., Cornell University, Ithaca. WILLIAM M. DAVIS, M.E,, Harvard University, Cambridge. ALES HRDLICKA, M.D., U.S. National Museum, Washington. D. S. JORDAN, LL.D., Stanford University. CHARLES A. KOFOID, PH.D., University of California, Btrkelty. J, G. NEEDHAM, PH.D., Lake Forest University. ARNOLD E. ORTMANN, PH.D., Carnegie D. P. PENHALLOW.D.SC..F.R.M.S., MC&I H. M. RICHARDS, S.D., Columbia Universe W. E. RITTER, PH.D., University of Califm ISRAEL C. RUSSELL, LL.D., University o ERWIN F. SMITH, S.D., U. S. Department LEONHARD STEJNEGER, LL.D., Smith. W. TRELEASE, S.D., Missouri Botanical G. HENRY B. WARD, PH.D., University of A WILLIAM M. WHEELER, PH.D., Amer New York. THE AMERICAN NATURALIST is an of Natural History, and will aim to pres facts and discoveries in Anthropology Botany, Paleontology, Geology and Phj alogy and Petrography. The contents leading original articles containing acc< discoveries, reports of scientific expedi distinguished naturalists, or critical su line ; and in addition to these there will points of interest, editorial comments < day, critical reviews of recent literatui gifts, appointments, retirements, and de All naturalists who have anything to send in their contributions, but the < for publication only that which is of tn same time written so as to be intellignx to the general scientific reader. All manuscripts, books for revie\ sent to THE AMERICAN NATURALIST, ( All business communications sh publishers. Annual subscription, $4.00, net, in adve Foreign subscriptio EIBRARY IYCRSISTDF MBRHKY FttOD GINN & COMPANY, PUBLISHERS. A . STUDIES ON THE PLANT CELL. I. BRADLEY MOORE DAVIS. INTRODUCTION. THIS is the first of a series of papers that will follow one another in the pages of the American Naturalist. They will describe the chief structures in plant cells and the most important events in their life histories, largely from the point of view of the morphologist and student of developmental processes. Research upon the plant cell has entirely outrun the general accounts that may be found in several botanical text books and in certain works of prominent zoologists. We shall attempt to give a general survey of the subject in its present state with references to the most important papers ; but this is not to be an exhaustive account of a literature that is already very large and which can probably be treated far more satisfactorily several years from now when it has passed through the criticism that time will give in a field of very active botanical investigation. American botanists have reason to be proud of the achieve- ments of their countrymen in research upon the morphology and physiology of the plant cell, for much of the best work of recent years has come from them. This in itself has been a great stim- ulus to the writer to prepare these brief accounts which he hopes will assis.t the general botanist to a clearer understanding of the progress in this field. They will also serve to contrast the pro- toplasmic activities among plants with those of the animal cell which has been so well treated in several foreign works and in English by Wilson's The Cell in Development and Inheritance. The author will feel especially gratified if these papers should help to change an attitude towards investigations on the plant cell that is unfortunately too prevalent among botanists. There is a tendency to regard cell studies as a very special field of botanical research with elaborate technique which the average 367 368 THE AMERICAN NATURALIST. [VOL. XXXVIII. botanist cannot be expected to master. Those who work in this field are considered as in a department by themselves and are labeled cytologists which is sometimes given as an excuse for knowing little about their results. Cell studies are nothing more than morphological and physiological investigations which are frequently so broad as to break the mould of the narrower mor- phology and physiology of former years. Cell studies must be the foundation of all exhaustive work in morphology and physi- ology. Indeed among the lower plants they constitute almost all there is to morphology and will determine the classification and relationships of great groups. There are no better illustrations of this fact than the effect of Prof. Harper's investigations on the ascus and sporangium upon Bref eld's theory of the origin of the Ascomycetes. And again the results of several investigators upon the multinucleate gametes found among the Phycomycetes and Ascomycetes are of the utmost importance to a correct understanding of the phylogeny of these groups. When students of the plant cell refuse to accept the stamp of cytologist and insist and show that their work is simply fundamental mor- phology and physiology we shall break away from a past that should be outgrown. The material of these papers will be treated under the follow- ing heads. TABLE OF CONTENTS. Introduction. SECTION I. STRUCTURE OF THE PLANT CELL. i. Protoplasmic Contents. ' (a) The Nucleus. (b) The Plastids. (c) Cytoplasm. 1. Plasma Membranes. 2. Trophoplasm. Coenocentra, Nematoplasts, Physodes. 3. Kinoplasm. Centrospheres, Centrosomes, Asters, Filarplasm, Ble- pharoplasts. No. 449-1 STUDIES ON THE PLANT CELL. 369 2. Non-Protoplasmic Contents. (a) Food material and waste products. (b) Vacuoles. 3. The Cell Wall. SECTION II. THE ACTIVITIES OF THE PLANT CELL. 1. Vegetative Activities. 2. Cell Division. (a) The Events of Nuclear Division. 1. Direct Division. 2. Indirect Division (Mitosis). Prophase, Metaphase, Anaphase, Telophase. 3. The Dynamics of Nuclear Division. (b) The Segmentation of Protoplasm. 1. Cleavage by Constriction. 2. Cleavage by Cell Plates. 3. Free Cell Formation. SECTION III. HIGHLY SPECIALIZED PLANT CELLS AND THEIR PECULIARITIES. i. The Zoospore. 2. The Sperm. 3. The Egg. 4. The Spore Mother Cell. 5. The Coenocyte. 6: The Coeno- gamete. SECTION IV. CELL UNIONS AND NUCLEAR FUSIONS IN PLANTS. SECTION V. CELL ACTIVITIES AT CRITICAL PERIODS OF ONTOG- ENY IN PLANTS. i. Garnet ogenesis. 2. Sporogenesis. 3. Reduction of Chromosomes. 4. Apogamy. 5. Apospory. SECTION VI. COMPARATIVE MORPHOLOGY AND PHYSIOLOGY OF THE PLANT CELL. 370 THE AMERICAN NATURALIST. [VOL. XXXVI IK LITERATURE ON THE PLANT CELL. Reference to special papers will be given by the authors name and the date of publication through lists presented at the end of every section. There is no comprehensive treatise devoted to the plant cell but the following general accounts and reviews of the literature are important. 1. Strasburger in the Lehrbuch der Botanik and Pfeffer in his Physiology of Plants present the best general accounts of the structure and activities of the plant cell. 2. Zimmerman in 1893 and '94 (" Beihefte zum Botanischen Centralblatt " vol. 3 and 4), reviewed the literature on the plant cell under the title " Sammel-Referate aus dem Gesammtgebiete der Zellenlehre " and in 1896 collected the literature dealing with the nuclei of plants in a book entitled Die Morphologic und Physiologic des pflanzliclien Zellkernes, Jena, 1896. 3. Dangeard discusses a number of cytological topics in the 6th series of Le Botaniste (1898) with especial reference to his studies on the Chlamydomonadineae. 4. Fischer, Fixirung Farbung und Bau des Protoplasmas Leipzig 1899, presents a critique of the methods of cytological research and the justification of the conclusions based thereon. 5 . The most recent analysis of conspicuous activities of the plant cell is that of Strasburger Ueber Reductionstheilung, Spindelbildnng, Centrosomen und Cilienbildner im Pflanzen- reich, HistologiscJie Beitriigc VI, I9OO. 1 SECTION I. STRUCTURE OF THE PLANT CELL. It is customary to apply the term cell in Botany not alone to the protoplasmic units of organization but also to the enclosing wall that generally surrounds the protoplasm. Indeed these walls alone when entirely emptied of protoplasm in specialized 1 To this list should be added an excellent concise review by Koernicke entitled "Der heutige stand der pflanzlichen zellforschung " Ber. d. dcut. hot. Gesell 21, (66), 1904.^ This article appeared too late to be quoted in the earlier papers of this series. No. 449-1 STUDIES ON THE PLANT CELL. 371 regions of the plant, e. g. vascular and certain supporting and tegumentary tissues, are still called cells. When among the lower forms and at certain periods in the life history of many higher plants the protoplasm is naked (e. g. zoospores, sperms, eggs, etc.), these structures are cells in exactly the sense used by zoologists. We shall consider almost entirely the protoplasmic portion of the plant structure for any extended treatment of the walls would lead us at once into that field of microscopic anatomy termed histology. i. Protoplasmic Contents. The most highly differentiated region of the cell is the nucleus,, a structure remarkably uniform in organization among all plants except the lowest Algae and some very simple Fungi. These more primitive conditions will be considered in Section. VI. Besides the nucleus there are present plastids in all groups except the Fungi. Plastids are likewise specialized protoplasmic elements although much simpler in structure than the nucleus. Nuclei and plastids lie in a protoplasmic matrix called the cyto- plasm. Cytoplasm is more variable in structure and activity than any other region of the cell. Thus three forms of proto- plasm, nucleoplasm, plastidplasm and cytoplasm comprise all the living material of the cell and may be sharply contrasted with the non-protoplasmic contents, mostly food material and waste prod- ucts, which will be considered under a separate head. Definite masses of nucleate protoplasm, with or without plastids are termed protoplasts and such are either unicellular organisms themselves or units of a multicellular structure. (a) The Nucleus. The nucleus is bounded by a delicate membrane that is probably largely or wholly a modification of the surrounding cyto- plasm. The nucleoplasm very rarely completely fills the nuclear membrane, the remaining space being occupied by a fluid known as the nuclear sap. The elements in the resting nucleus consist chiefly -of material that takes the form of a net work so that the 372 THE AMERICAN NATURALIST. [VOL. XXXVI II. effect is that of a much coiled and twisted thread whose loops are united at intervals to form large and small meshes. The ground substance of this thread is called linin and imbedded in it as in a matrix are deeply staining granules of chromatin. Chromatin is regarded as the most important substance in the nucleus, chiefly because of its behavior during nuclear division, and in critical periods of the life history of organisms as at sporogen- esis, gametogenesis and fertilization (to be described in Section V). Just before nuclear division the chromatin becomes organ- ized into bodies named chromosomes which are remarkably uni- form in number and definite in shape for each tissue and period of the plant's life. They will be discussed under " The Events of Nuclear Division " (Section II), and in Sections IV and V. FIG. i. The resting nucleus. , Embryo sac of lily with linin thread and two nucleoli. 6, Root of onion large nucleolus. c, Tetraspore of Corallina showing large chro- matin body and small nucleolus. d, Spirogyra with central body containing chroma- tin, e, Chromatin on linin net work from egg of pine. After Mitzkewitsch and Chamberlain. In the meshes of the linin network or lying freely in the nuclear sap may be found one or more bodies, generally globular in form, called nucleoli. (See Fig. i a and Fig. i b}. The nucleolus is generally regarded as a secretion of the nucleus and it is quite certain that its substance is utilized just previous to and during the period of nuclear division when the spindle is formed. (Strasburger '95 and : oo, p. 125, and from the work of others). The structure is not always homogeneous but may -show in the interior small vesicles or areas of a different con- sistency from the periphery. There is often present also a rather thick outer shell or membrane. Sometimes the chromatin in the nucleus may be gathered into a globular body that resem- bles superficially a nucleolus. Such chromatin bodies are gen- No. 449.] STUDIES ON JTfE PLANT CELL. 373 erally transitory as in Corallina, Davis '98, where the structure (Fig. i c) is only found in the young daughter nucleus and later fragments into many smaller bodies. In Spirogyra however (Moll '94, Mitzkewitsch '98, Van Wisselingh : oo, '02) the chromatin is supposed to be always in a globular mass mixed with nucleolar substance and recalls the conditions in certain Protozoa. These chromatic structures however should never be confused with nucleoli, whose substance is different and which are not permanent in the cell, since they may disappear before or during nuclear division and be formed de novo in each daughter nucleus. The substance of the nucleolus is not well understood. It is frequently impossible to distinguish it from chromatin except when favorably situated in the cell and there is much evidence that it is closely related to that substance. In large nuclei of higher plants the chromatin is sometimes gathered into globular bodies without apparent relation to a linin thread and these are readily mistaken for nucleoli and have been called such, but this loose usage of the term should be avoided. And true nucleoli may be so closely associated with the linin net work as to have the appearance of chromatin. Some of these conditions have been especially described by Cavara, '98. Chamberlain, '99, has made a study of the egg nucleus of the Pine where masses of chromatin may take very irregular forms on the linin threads (Fig. i e] and sometimes resemble small nucleoli. But such conditions should always be sharply distinguished from true nucleoli which are often caught in the meshes of the linin net work and appear to be a part of it when in reality there are no organic attachments. It is certain that nucleoli are of secondary importance in the cell and probably by-products of the general constructive activities of the nucleus. In which case they may be secretions, perhaps closely related to chromatin, or even direct transformations of this substance. It is well known that the nucleus has wonderful constructive powers, when the amount of chromatin and other nuclear substances may be immensely increased, facts that are especially well illustrated at reproductive periods of the plant's life as during sporogenesis and garnet o- genesis. Chromatin is the only substance in the nucleus that is constant 374 THE AMERICAN NATURALIST. [VOL. XXXVI II. in its presence throughout all periods in every cell's history. It passes on from cell to cell through the mechanism of nuclear division without interruption. There are periods of cell history when the nucleus consists only of chromosomes as in the stages of nuclear division called mataphase and anaphase. The other structures of the nucleus have their relation to definite condi- tions that are in part understood. The nuclear membrane probably results from the reaction of the cytoplasm to the secre- tion of nuclear sap among the chromosomes (Lawson, -.03 a). It would then be strictly cytoplasmic in character and similar to the plasma membranes around vacuoles. Nucleoli must be regarded as temporary structures since they generally disappear during nuclear division either dissolving or else passing out into the cytoplasm where they may remain for long periods as deeply staining globules (extra nuclear nucleoli). Linin is believed to be derived from chromatin and in its turn may be transformed into the substance of spindle fibers, which are cytoplasmic, so that chemically it holds a position somewhat intermediate between chromatin and cytoplasm. It seems established that the linin net work is a temporary structure related to the activities of chromatin. (b) The Plastids. These very interesting structures, characteristic of plant cells, have not received the degree of attention that they deserve and much valuable work may be done in the detailed study of their protoplasmic structure and activities at various periods of ontogeny especially through the series of changes that are presented during developmental processes. The primitive types of plastids are relatively large structures, often solitary in the cells, and generally of complex form. These are called chromatophores and are characteristic of many algae especially among the lower groups but are not found above the thallophytes (Anthoceros and Selaginella excepted). The chromatophores of the simplest algae are replaced in most of the higher types of these thallophytes and in all groups above by very much smaller structures, generally discoid in No. 449-] STUDIES OA THE PLANT CELL. 375 form, which are called chloroplasts when green, chromoplasts when the color is other than green or leucoplasts if colorless. These plastids are without doubt derived from the more primitive chromatophores. The colors of chromatophores are various. They are believed always to contain some chlorophyll but this green is frequently so completely masked by other pigments that its presence can only be determined when the additional coloring matters have been extracted. Chloroplasts are universally green except when they may be changing into chromoplasts. Chromoplasts generally take their tint from the predominance of other strong pigments in addition to chlorophyll as phycoerythrin in the red and phyco- phsein in the brown algae. But chromoplasts may be derived from chloroplasts whose green has largely or wholly disap- peared leaving other pigments present as the yellow, xanthophyll, or the orange red, carotin. The remaining plastids, leucoplasts, are devoid of color and are found in embryonic regions such as eggs, growing points, and in the various tissues of seeds, underground organs and other structures where the cells are largely or wholly removed from sunlight. The leucoplasts may become green upon expo- sure to light thus changing into chloroplasts. They are respon- sible for the secretion of reserve starch in many structures (e. g. potato) and in consequence have been called ^myloplasts. Leucoplasts, chloroplasts and chromoplasts are morphologically the same structures. It is well known that they may pass one into the other in the order indicated and that chloroplasts and chromoplasts may lose their color and become leucoplasts. It is generally believed that plastids are not formed de novo. They divide by constriction and thus multiplying are passed on from cell to cell and it is believed from generation to generation. They are therefore usually ranked as permanent organs of the cell. However, it is but fair to call attention to the fact that there are some serious difficulties in the way of a complete acceptation of these views. The protoplasmic structure of the plastids of higher plants is rather simple while that of the chromatophores in algae is more complex since they contain a special organ termed the 376 THE AMERICAN NATURALIST. [VOL. XXXVIIL pyrenoid. The detailed structure of chromatophores was first described by Schmitz ('82) and of plastids by Meyer ('83). The most complete study of plastids however is that of Schimper ('85). The body of the plastid is always denser than the sur- rounding cytoplasm. It has a porous structure that is only visible under high magnification and there are sometimes present very delicate fibrils. The coloring matter, oily in consistency, is held in the pores as minute globules. The plastid may therefore be compared to a very fine-textured sponge saturated with pigment. All of the coloring matter of the plastid may be readily extracted with alcohol leaving the colorless proteid matrix. The pigments of plastids are then in the nature of secretions held in these specialized regions of protoplasm. Chlorophyll is the principal substance and, as has before been said, is almost always present, but the amount is sometimes so small that its green is completely hidden by the color of other pigments. Chlorophyll itself contains greater or less amounts of two other coloring matters that may be readily separated from the pure green, a yellow xanthophyll and an orange red carotin, both substances closely related to chlorophyll. The other pigments, characteristic of the chromatophores in some groups of algae, are however quite distinct from chlorophyll. There is phycocyan, found in the blue* green algae (Cyanophyceae), phycophaein and phycoxanthin, characteristic of the brown (Phaeophyceae) and phycoerythrin of the red (Rhodophyceae). Chloroplasts are found almost universally in green plants above the Thallophytes and are also present in the large group of algae the Siphonales and in the Charales. They are some- times formed very numerously in the cell, reproducing rapidly by fission (see Fig. 2 a 2, 3) and lie in the layer of protoplasm just inside of the plasma membrane. They are sensitive to light and readily shift their position in the cell. Strong illumination results in their retreat from exposed positions to the sidewalls and bottom of the cell where the light is less intense. If the illumination be weak they may all gather on the side most favor- able for the reception of light. These facts are well illustrated by the behavior of the plastids in some of the Siphonales (e. g. No. 449.] STUDIES ON THE PLANT CELL. 377 Botrydium), in the Rhodophyceae (e. g. Polysiphonia) and also in the palisade cells of leaves. Chloroplasts after exposure to light generally contain starch but in some plants this substance is never formed (e. g. Vaucheria, Fig. 2 A i), the nrsFvrsible products of photosynthesis being other substances more of the nature of oil. It is not known whether the starch grain in the PIG. 2. Plastids. a, Chloroplasts: i Vaucheria, with oil globules; 2 Bryopsis ; 3 moss (Funaria), in division and containing starch grains ; 4 Oxalis, with a grain of starch, b, Chromoplasts : i Tropaeolum, epidermal cell from calyx; 2 Fucus, 3 Callithamnion. c, Chromatophores : i Spirogyra, with pyrenoids (/) and caryoicls (c); 2 Hydrodictyon, pyreuoid forming starch; 3 Nemalion; 4 Anthoceros, in divi- sion and containing starch, d, Leucoplasts : i Ph.ijus, pUstid and starch grain at the side of the nucleus ; 2 Iris, from root and containing oil globules ; 3 Iris, in deeper cells of root, with starch grains. After Meyer, Strasburger, P.lla, 1 imber- l.ike and Schimper. chloroplast results from the direct change of some of the pro- teid substance or whether it is a secretion. The conditions are somewhat different when pyrenoids are present in a chro- matophore as will be described presently. The Chloroplasts of higher plants may change color under various conditions and become chromoplasts. Some of the best 378 THE AMERICAN NATURALIST. [VOL. XXXVIII. examples are found in the colored cells of certain floral parts and fruits (Fig. 2, b i). These pigments are generally either xanthophyll (yellowish) or carotin (orange red). Chloroplasts may also turn brown especially in older cells that are losing their contents. The colors of some leaves and flower parts are due not to the plastids but to substances dissolved or otherwise held in the cell sap of the vacuoles. The brilliant coloration of autumn foliage is of this character as well as some of the tints of petals, hairs and other structures. The chromatophores of the higher brown Algae (Phaeophyceae) and most of the red (Rhodophyceae) have the discoid form characteristic of chloro- plasts (Fig. 2 b 2, 3). They might be called phasoplasts and rhodoplasts if one wished to classify plastids according to their color. The structure of chromatophores is frequently complicated by the presence of pyrenoids which may be quite numerous in the body. These structures are denser regions of the chromatophore with a definite boundary. They are proteid in character and are known to vary in size with nutritive conditions and may completely disappear if the cell is starved. They have been regarded as masses of reserve oroteid material but certain func- tions of great importance are also associated with them. The arrangement of starch grains in the chromatophores of many algae is clearly around the pyrenoids as centers. For this reason they have been called amylum centers. Timberlake (:oi) has recently shown in Hydrodictyon that segments are split off from the pyrenoids (see Fig. 2, c 2) and changed directly into starch grains which naturally lie for a time close to the source of their formation and only later become distributed throughout the chromatophore. It is probable that similar conditions will be found in other algae (Conjugales, Protococcales, etc.) and we may soon have a much clearer understanding of the pyrenoid. The indications are that the pyrenoid will prove to be a region of the chromatophore differentiated as a metabolic center, more or less prominent according to conditions of nutrition, and that its most conspicuous activity is the formation of starch by the -direct transformation of portions of its substance. Some other structures besides the pyrenoids have been No. 449.] STUDIES ON THE PLANT CELL. 379 described by Palla ('94) in the chromatophores of several of the Conjugates and have been named caryoids. Caryoids (Fig. 2, c i) are smaller and more numerous than pyrenoids and are distributed irregularly in the chromatophore but chiefly "along the edge. Their function is not known. The leucoplasts complete the list of plastid structures. They are colorless and may be found in underground or other portions of the plant removed from light or where there is little or no photosynthetic activities as in embryo sacs, seeds, growing points, etc. They become impregnated with chlorophyll under condi- tions suitable for photosynthesis thus changing into chloro- plasts. An important function of the leucoplast is the forma- tion of reserve starch in various parts of the plant. The more recent investigations of this process (Meyer, '95, Salter, '98) claim tfyat it is in the nature of a secretion within the substance of the leucoplast. This view is opposed to the older conceptions {Schimper, '81, Eberdt, '91), which regarded the starch grain as formed by the direct change of proteid material in the plastid. In view of Timberlake's ( : 01) studies on the pyrenoid of Hydro- dictyon we may well hesitate to fully accept the views of Meyer and Salter and ask for further investigations of this very difficult subject. In addition to starch leucoplasts may contain proteid crystals and oil globules. The reproduction of plastids and their evolutionary history in ontogeny and phylogeny offers a very attractive field for research. It is well known that plastids multiply by fission and it is generally believed that they never arise de novo but are passed from generation to generation as permanent organs of the cell. The process of division may be very favorably studied in the spore mother-cell of Anthoceros (Fig. 2, c 4). The fission begins (Davis, '99) by a constriction at the surface as though the bounding membrane of cytoplasm exerted pressure upon an elongating structure. There is no evidence that the interior of the chloroplast undergoes any changes that could assist the process further than a possible tendency of the two separating portions to gather their substance together as division proceeds. The conditions suggest that the division is a mechanical separa- tion of material too bulky for the best advantages of the cell, 3 8o THE AMERICAN NA TURAL1ST. [VOL. XXXV III. for the proper balance of protoplasmic elements in narrow confines, a division prompted by the activities of the cytoplasm rather than emanating from within the plastid. The view of the permanence of the plastid as a cell organ has received its strongest support from the classical work of Schimper ('85). We are not prepared to deny it and to assert that the plastid may arise de novo. Yet those who study the cells of embryonic tissues and reproductive phases know that it is extremely difficult to follow the plastids and that these structures require other than the usual methods of cell research to establish their presence. Several writers (Eberdt, Dangeard, Husekand others) have expressed their belief that plastids may arise de novo but no one has thoroughly traced the appearance or disappearance of these structures in any cells. The plastid in phylogeny has never received the attention that it deserves. Beginning with the conditions among the Cyano- phyceae and the lowest Chlorophyceae (which will be further discussed in Section VI) we find the pigment distributed so generally throughout the cell that it is doubtful if the term chromatophore should ever be applied to regions so indefinite in outline. Above these groups the pigment is confined to propor- tionally smaller areas in the cytoplasm and these become chromatophores when their form is clear. The primitive chro- matophores were solitary and filled a large part of the cell. The pyrenoids arose in the chromatophores probably as the result of the influence of metabolic centers upon the protoplasm. It is scarcely possible that a large chromatophore should be absolutely homogeneous throughout ; there would develop one or more centers of metabolic activity and such would exert some influence on the form of the protoplasm. But the large single chromatophore does not seem to be the form best adapted to the work of a cell perhaps, if for no other reason, because it requires a mechanical adjustment of other cell organs to itself and would interfere with the quick circula- tion of material and the general balance of cell activities. It seems possible that mechanical difficulties may have led to the division of large chromatophores and the substitution of numerous small plastids. This change was instituted in the No. 449.] STUDIES ON THE PLANT CELL. 381 higher members of the Phaeophyceae and Rhodophyceae and in the Siphonales, Charales, Cladophoraceae and some smaller groups of the Chlorophyceae. The Conjugales whose chromato- phores are especially elaborate have cells essentially solitary in their life habits and with a very remarkable adjustment of the cell organs to one another to give almost perfect symmetry. With the splitting up of the chromatophore came the loss of the pyrenoid and the final result was the compact plastid so charac- teristic of plants above the thallophytes. I (c) Cytoplasm. There is no region of the plant cell that maintains such varied relations to its environment and performs so many visible activities as the cytoplasm. For this reason the accounts of its structure and behavior have been diverse and there has developed a nomenclature of its parts that is confusing and somewhat difficult to harmonize. Strasburger has for many years (since 1892) employed the term kinoplasm to distinguish an active portion of the cytoplasm (concerned with the formation of spindle fibers and other fibrillae, centrospheres, centrosomes, cilia, plasma membranes, etc.) from more passive nutritive regions which he called tropho- plasm. Kinoplasm corresponds closely to the archoplasm of the animal cell (Boveri, 1888). This classification has been criticised especially by Pfeffer ( : oo) on the ground that it employed names signifying physiological differences when the distinctions as far as we know are those of morphology alone. However the physiological behavior of kinoplasm and tropho- plasm becomes very real to anyone who studies extensively cell activities and the morphological characters serve to emphasize these peculiarities. The truth seems to be that cell studies cannot be pursued from the standpoint of physiology or mor- phology alone but must combine these attitudes. And in the union it is hardly possible or perhaps desirable to construct a terminology with strict regard to either field of study. We shall use the terms kinoplasm and trophoplasm grouping the various cytoplasmic structures under these heads. 382 THE AMERICAN NA TURALIST. [VoL. XXXV III. Cytoplasm has surface contact with three conditions and in each case there is present a delicate plasma membrane, colorless and very finely granular, which is very' different in structure from the cytoplasm within. The first of these three membranes is the outer plasma membrane, which bounding the protoplast, is consequently just inside the cell wall. This membrane is called the " hautschicht " by the German botanists, a word for which we have no exact equivalent, the term ectoplast more nearly expressing the meaning than any other but for several reasons not being very satisfactory. Since this outer plasma membrane lies against a moist cell wall it is virtually surrounded by a film of water. The functions of the cell wall in land plants and its developmental history indicate a close relation to the demands of the outer plasma membrane for a fairly uniform environment of moisture, a matter which will be discussed in the last section of these papers. The second form of plasma membrane surrounds the water vacuoles in the cell. It is very common for the plant cell to have a single large central vacuole containing the cell sap and the membrane around this was named the tonoplast by DeVries in 1885. DeVries believed that this vacuole reproduced itself by fission with each cell division and consequently was a perma- nent organ of the cell. It is, however, now well known that the large central space containing cell sap is not different from other vacuoles, indeed is frequently formed by the flowing together of several small vacuoles as smaller soap bubbles unite in the froth to form a larger one A vacuolar plasma membrane is of course bathed by water since it holds the cell sap and its relation to a moist surface is therefore more evident than in the case of the outer plasma membrane. The third plasma membrane encloses the nuclear sap with the protoplasmic nuclear elements chromatin, linin and the nucleolus. This nuclear membrane was discussed in connection with the nucleus of which it is generally considered a part, but as there stated, the evidence largely indicates that it is cytoplasmic in character, representing a reaction of this protoplasm to the fluid nuclear sap formed around the chromosomes in the daughter nuclei after each division (Lawson :c>3 a ). The nuclear sap No. 449.] STUDIES ON THE PLANT CELL. necessitates the development of a vacuole which becomes bounded by the nuclear membrane. The nuclear membrane in some cases at least differs from a vacuolar membrane in being easily distinguished from the surrounding cytoplasm as a definite film. The structure of all the plasma membranes is much the same as far as the microscope may determine. The protoplasm is dense, colorless and filled with very minute granules (micro- somata). There are no large inclusions such as plastids, parti- cles of food material (starch, proteids, oils, fats, etc.), mineral matter or waste products. These are all held well within the cytoplasm between the outer plasma membrane and the vacuoles. There is good reason to believe that the substance of all plasma membranes is much the same since they perform very similar activities both in relation to the fluids that bathe them and also because their substance in certain cases becomes the proto- plasmic basis of cellulose walls. These resemblances are well established for the outer plasma membrane and that which sur- rounds the vacuoles. Thus, the capillitium of Myxomycetes (Strasburger, '84) is formed from the plasma membranes around the vacuoles after the same method as a cell wall from the outer plasma membrane. And again, during cleavage by constriction (see section II) in the plasmodium and sporangium of the molds (Harper, '99 and : oo, D. Swingle, : 03), vacuoles fuse with cleav- age furrows from the outer plasma membrane to form a common membrane which surrounds each spore mass and secretes a wall, thus showing identity of function and structure. The resem- blances are less conspicuous for the kinoplasm of the nuclear membrane, only appearing indirectly with certain events of cell division (the formation of the cell plate) which will be discussed in the next section of the paper. The evidence indicates that the three plasma membranes are all kinoplasmic in character, a generalization of some importance since it offers explanations of many peculiar cell activities to be described later. Since all plasma membranes have these common characters it may well be questioned whether an elaborate terminology is justified for structures so closely related. The terms ectoplast and tonoplast seem undesirable since they were meant to indi- 384 THE AMERICAN NATURALIST. [VOL. XXXVIII. cate peculiarities of structure and a degree of permanence as cell organs that is not actually present. It seems hardly neces- sary to define the plasma membranes further than by their posi- tion in the cell as the outer, vacuolar and nuclear membranes. All of the cytoplasm bounded by the plasma membranes with the exception of certain conditions to be described later (centre- spheres, centrosomes, asters, filarplasm and blepharoplasts) may be called trophoplasm since it contains structures and substances especially concerned with nutritive functions. Trophoplasm presents an open organization in sharp contrast to the dense kinoplasm. This peculiarity is due in part to numerous small vacuoles which give a spongy appearance to the usual foam like structure and is further complicated by the inclusion of material not strictly a part of the protoplasm in the form of various sized granules. There are sometimes present fibrillae that impart a somewhat fibrous texture. We cannot discuss here the theories of the structure of protoplasm, which has not been so extensively studied in plants as among animals, further than to point out that it varies considerably in different regions of the cell in relation to peculiarities that will be described later. There is sometimes presented very typically the foam structure of Butschli but the introduction of small vacuoles generally gives a spongiose appearance. This subject is critically reviewed by Fischer, '99, and has also been treated in several papers of Strasburger especially in '97. Three well differentiated organs of the cell, probably tropho- plasmic in character, require special mention, viz., coenocentra, nematoplasts and physodes. Coenocentra are very interesting protoplasmic centers found in the oogonia of certain ccenocytic fungi among the Saprolegniales and Peronosporales during oogenesis. They appear just previous to the differentiation of the eggs as small bodies sometimes with delicate radiations (see Fig. 3, a and 8,/), and are found one in each egg origin. They are apt to increase in size as the eggs mature and evidently become the centers of the metabolic activities of the cells, drawing the sexual nuclei into their neighborhood where the latter increase in size (Fig. 3, a 2). The ccenocentrum dis- appears in the ripe oospore and is consequently an evanescent No'. 449.] STUDIES ON THE PLANT CELL. 385 structure. It is probably the morphological expression of a dynamic center in the egg. Ccenocentra have been known for several years and have been given especial attention in the recent investigations of Stevens, '99 and 'or, and the author (Davis, .-03). They will be further considered in our account of Ccenogametes (Section III). Nematoplasts are exceedingly small rod or thread like Fir,. 3. Cytoplasmic structures, a, Coenocentra of Saprolegnia ; i, oogonium, each egg origin with a coenocentrum ; 2, coenocentrum and nucleus from mature egg. b, Nematoplasts from hair of Momordica. c, Nucleus from apical cell of Sphacela- ria, aster with centrosome. d, Nucleus from oogonium of Fucus, aster with centro- sphere. e, Nucleus from germinating spore of Pellia, centrospheres with short cytoplasmic radiations faster like), f, Nucleus from procambium cell of Vicia, kino- I'plasmic caps, g, Pollen mother-cell of Lilium, filarplasm in form of multipolar spindle. A, Development of sperm of Gymnogramme ; i, blepharoplast at side of sperm nucleus; 2, blepharoplast elongating and developing cilia ; 3, mature sperm, blepharoplast and nucleus in parallel bands, cytoplasmic vescicle below. After Zimmermann, Hof, and Belajeff. structures reported by Zimmermann ('93, p. 215) in the cells of hairs of Momordica and the root of Vicia (see Fig. 36). It is probable that organs described by Swingle, '98, and Lagerheim, '99, under the names of vibrioides are the same as or closely 386 THE AMERICAN NATURALIST. [VoL. XXXVI I L related to physodes. Swingle found them in some of the Saprolegniales and certain Rhodophyceae and Lagerheim in Ascoidea. They are probably not uncommon. Nematoplasts may be proteid crystals but there is evidence that they move, bending slowly back and forth, which suggests a higher degree of organization. They should be further studied. Physodes are bladder like structures described by Crato, '92, in certain brown Algae. They contain a highly refractive sub- stance which gives them a very different appearance from vacuoles whose structure they resemble in many respects. Very little is known about the contents of physodes and it may well be questioned whether they are really organs of the cell and not vacuoles set apart to hold some fluids or substances other than cell sap. There are left for us a group of kinoplasmic structures that are especially prominent and sometimes only present during the events of nuclear division and at the times when cilia are formed. They will be discussed in later sections of these papers (Sections II, III, V and VI) and at this time we shall give but a brief statement of their appearances. They are centrospheres, centrosomes, asters, filarplasm and blepharoplasts. Centrospheres are rather large areas of kinoplasm that some- times lie at the poles of nuclear figures and to which are attached the fibrillae that form the spindle and also those that may radiate into the surrounding cytoplasm. If the centre- sphere contains a distinct central body, or if such a small structure be present alone at the poles of the spindle it is called a centrosome. Should either structure be accompanied by definite fibrillar radiations the whole is termed an aster. These latter conditions are sometimes very complex and are the most interesting types of structures. Asters with centrosomes are known for the brown algae in the growing points of Sphacelaria (Fig. 3c), Stypocaulon (Swingle, '97) and the spore mother cell of Dictyota (Mottier, :oo). They are also beautifully shown in certain diatoms (Lauterborn, principal paper '96, Karsten, :oo). Asters with centrospheres and occasionally but not constantly containing centrosome-like bodies are found in the oogonium and germinating eggs of Fucus, see Fig. 3, d (Strasburger, '97*, No. 449.] STUDIES ON THE PLANT CELL. 387 Farmer and Williams, '98). Especially well differentiated asters with centrospheres are present during the mitoses in the ascus, functioning at the end in the peculiar process of free cell formation (Harper, '97). Large centrospheres accompanied by radiations are present during the germination of the spores in certain Hepaticae (Farmer and Reeves, '94, Davis, :oi, Cham- berlain, : 03), but are less conspicuously shown in some and are entirely absent in other phases of the life history. Remarkably large centrospheres with inconspicuous radiations are known in the tetraspore mother cell of Corallina (Davis, '98). Centro- spheres occur in the basidium (Wager, '94, Maire, : 02). Cen- trosomes have been reported during the mitoses in the sporangium of Hydrodictyon (Timberlake, :O2). Centrosomes have also been described in other types of the thallophytes but we are justified in asking for further work on these bodies since they are generally without raditions and may not have at all the significance indicated. Neither asters, centrospheres or centrosomes seem to be normally present in groups above the bryophytes, nuclear division taking place in these plants by methods, not found in other organisms, which will be described in succeeding sections. Vegetative and embryonic tissues of plants above the thallo- phytes present very different conditions from those described in the foregoing paragraph. The centrosphere is replaced by a less definite structure in the form of a kinoplasmic cap which appears at the ends of the dividing nucleus and determines the poles of the spindle (see Fig. 3, /). They have been described in the cells of vegetative points of several pteridophytes and spermatophytes by Rosen, '93, Hof, '98, and Nemec, '99 and : 01, and in the seta and late divisions in the germinating spore of the liverwort Pellia (Davis, :oi). The most highly developed conditions of spindle formation are found in the spore mother cells of the bryophytes, pterido- phytes and spermatophytes. Here the nucleus becomes surrounded by a weft of fibrillae which form a kinoplasmic envelope probably derived in part from the nuclear membrane. The fibrillae are at first quite independent of one another or of common centers. Most of the fibrillas enter into the spindle 388 THE AMERICAN NATURALIST. [VOL. XXXVIII. which may in the beginning have several poles (see Fig. 3,^), but these generally swing at last into a common axis so that the spindle finally becomes essentially bipolar. The term filarplasm is applied to this free fibrillar condition of kinoplasm without organized centers. Filarplasm is peculiar to plant cells and its remarkable activities in connection with multipolar spindles have only been found in groups above the thallophytes. Centrospheres, centrosomes and asters among the lower plants resemble in general the same structures in the animal cell. But filarplasm presents a higher form of kinoplasmic structure with perhaps the most complex activities known in the process of spindle formation. We shall consider them especially in Section III when treating the spore mother cell. The blepharoplasts are in some respects the most complex structures derived from kinoplasm. They are most conspicuous in the sperm cells of higher plants (spermatophytes and pteridophytes) but they are undoubtedly present in lower forms and probably in zoospores. The blepharoplast develops cilia as delicate fibrillae from its surface. The origin and homol- ogies of the blepharoplast are uncertain. In some forms they resemble centrosomes at the poles of the last nuclear figures in sperm tissue. But in other cases they are entirely independent of such spindles, a character which cannot be brought into harmony with the activities of centrosomes. They finally lie one at the side of each sperm nucleus, see Fig. 3, Ji, and with the development of the sperm they follow the spiral twist, when present, as a parallel band (Fig. 3, //, 2 and 3). This structure will receive detailed treatment in our account of the sperm (Section III). 2. Non Protoplasmic Contents. It is not possible to distinguish with certainty all the non- living material of a cell from its protoplasm. We have at one extreme cells from which the protoplasm has almost or wholly disappeared and which are either entirely empty or set apart solely as receptacles for various substances, sometimes waste products and sometimes food materials. In contrast with this No. 449.] STUDIES ON THE PLANT CELL. 389 condition are the cells filled with cytoplasm so homogeneous in structure that only the most delicate granules (microsomata) can be distinguished in the clear substance. Waste products such as mineral matter, resins, certain oils, solutions of tannin and various poisons, such as the alkaloides, may be easily recognized. Most food substances such as starch, proteid grains (aleurone), albumin crystals, oils, fats, etc., are readily separated from the protoplasm in which they lie. But the difficulties are much greater with the smaller particles of proteid material, which are frequently such minute granules as to approach the microsomata in size. These may give to the protoplasm a granular consistency that breaks up the foam or spongiose structure characteristic of the pure condition. These granules are undoubtedly in most cases substances intimately concerned with the metabolism of the cell and are members of the chains of constructive and destructive processes that charac- terize life phenomena. The other non protoplasmic structures of cells are vacuoles^ which are essentially bubbles of fluid lying in the denser proto- plasmic medium and surrounded by plasma membranes. The watery fluid of vacuoles contains various substances in solution, carbohydrates such as the sugars glucoses and inulin, mineral salts, asparagin, tannin, alkaloids, etc., and occasionally oil and not infrequently crystals. Vacuoles may be formed in large numbers in protoplasm. They tend to run together as do bubbles in a froth and in this way the large central vacuole becomes established in the cell, gathering to itself many smaller vacuoles until the protoplasm is forced to lie as a relatively thin layer next the cell wall. The fluid in the central vacuole (cell sap) is generally thinner and more watery than that in the smaller vacuoles. The latter are apt to be more rich in albumen which may be transformed into proteid grains as is especially well illustrated in the secretion of aleurone. Cell sap may be colored by pigments in solution and the tints of flowers are largely due to this cause alone or to the effects of its color in combination with various plastids in the cell. It is possible that physodes, described among the cytoplasmic structures, are in reality vacuoles filled with substances other than cell sap, which are not as yet understood. 390 THE AMERICAN NATURALIST. [VoL. XXXVIII. 3. The- Cell Wall. Many of the chief peculiarities of plant organization and activities are due to the presence of the cell wall, its influence on structure and mode of life. The cell wall is not an excretion from the cell like a mineral shell but is formed by the direct change of portions of the protoplasm. The regions concerned may be the outer plasma membrane, the vacuolar plasma mem- brane or the substance that makes up the spindle fibers which form the cell plate. These structures are all kinoplasmic in character and have to do with the formation of cell walls in various ways which will be described in Section II under the topic " The Segmentation of the protoplasm." The transforma- tion of finely granular films of kinoplasm into cellulose is not well understood but there is an evident solution of the granules (microsomata) and the change of the resultant substance into the cell wall. As a chemical process this change means the replacement of molecules of an albuminous nature by those of a carbohydrate substance. The most complete account of the cell wall is that of Strasburger, '98. Cell walls are chiefly composed of cellulose, but other sub- stances are always present, modifying the structure in various ways to give widely different properties. These modifications are generally due to infiltrations of foreign substances but some- times cell walls become incrusted with mineral deposits. The group of cellulose compounds is very large and it is extremely difficult to identify the various substances in structures so small as the cell walls. For a detailed treatment of the chemistry of the cellulose group the reader is referred to Cross and Bevans, '95, and for a general account to Pfeffer, : oo, p. 480-485 . There are microchemical tests for cellulose that give good reactions for most tissues but which cannot be relied upon for some walls (as in fungi and many algae) yet it is well understood that the cell walls of these organisms are from the biological point of view essentially the same as for other plants. The cell walls of some fungi are very largely composed of chitin. Several substances known to be present in cell walls give them marked characteristics. Their association with the cellu- No. 449-1 STUDIES ON THE PLANT CELL. 391 lose is so intimate as to resist very severe treatment and there- fore these cell walls are essentially cellulose groups modified chiefly in their physical properties by the presence of foreign substances. The most conspicuous modifications of this charac- ter are lignification, suberization and cutinization. Lignified walls are permeable to water and gases. Several substances have been separated from the cellulose of lignified walls, among them lignone, coniferin, vanillin, etc. Suberized and cutinized walls are largely but probably never wholly impervious to water and gases ; the one is infiltrated with suberin and the other with cutin, substances that resemble one other very closely. Even walls that appear to be pure cellulose have other sub- stance united with them, the most important being pectose and callose. Cell walls frequently become gelatinous or mucilaginous, when the outer layers swell and lose their form or they may be transformed into gums. These changes are well illustrated in the coats of seeds and fruits and among the algae and fungi. The cells of algae frequently secrete gelatinous envelopes or sheaths of substances so closely related to cellulose that were they condensed they would form a firm cell wall. The cell wall may grow in two directions by methods quite different from one another. There is first surface growth which results in a stretching of the cellulose membrane (growth by intussusception) . And second there may be growth in thick- ness by the formation of successive layers of cellulose inside of one another, giving the wall a striated structure (growth by apposition). The second type of growth is chiefly interesting since it makes possible many peculiarities of structure, because the newly formed layers may not be deposited uniformly inside the primary wall. In some cells the secondary thickenings have the form of rings or spirals or a reticulate structure. The reticulate condition passes insensibly into the pitted cell in which the secondary layers cover the greater part of the surface leav- ing the primary wall only exposed at the pits. Further dis- cussion of these cells falls more within the range of histology than the purposes of this paper. The cell wall offers a very interesting field of research among the thallophytes and especially in the lower groups where we 392 THE AMERICAN NATURALIST. [VOL. XXXVIII. may expect to find these envelopes in a fairly primitive con- dition and may be able to establish the steps in the origin and differentiation of this very important accessory structure to the plant cell. ( To be continued.") LITERATURE CITED FOR SECTION I "THE PLANT CELL." CAVARA. '98. Intorno ad alcune strutture nucleari. Atti. dell. Inst. hot. Univ. di Pavia, II, 5, 1898. CHAMBERLAIN. '99. Oogenesis in Pinus laricio. Bot. Gaz. 27, 268, 1899. '03. Mitosis in Pellia. Bot. Gaz. 36, 27, 1903. CRATO. '92. Die Physode, ein Organ des Zelllenleibes. Ber. d. deut. hot. Gesell. 10, 295. CROSS AND BEVAN. '96. Cellulose, an outline of the chemistry of the structural elements of plants. 1895. DAVIS. '98. Kerntheilung in der Tetrasporenmutterzelle bei Corallina officinalis L. var. mediterranea. Ber. d. deut. bot. Gesell. 16, 266, 1898. '99. The spore mother cell of Anthoceros. Bot. Gaz. 28, 89, 1899. :01. Nuclear studies on Pellia. Ann. of Bot. 15, 147, 1901. :03. Oogenesis in Saprolegnia. Bot. Gaz. 35, 233 and 320, 1903. DEVRIES. '85. Plasmolytische Studien iiber die Wand der Vacuolen Jahrb. f . wiss. Bot. 16, 465, 1885. EBERDT. '91. Beitrage zur Entstehungsgeschichte der Starke. Jahrb. f. wiss. Bot. 22, 293, 1891. FARMER AND REEVES. '94. On the occurrence of centrospheres in Pellia epiphylla, Nees. Ann. of Bot. 8, 219, 1894. FARMER AND WILLIAMS. '98. Contributions to our knowledge of the Fucaceae ; their life history and cytology. Phih Trans. Roy. Soc. 190, 623, 1898. FISCHER. '99. Fixirung, Farbung und Bau des Protoplasmas. Leipzig, 1899. No. 449-] STUDIES ON THE PLANT CELL. 393 HARPER. '97. Kerntheilung und freie Zellbildung im Ascus. Jahrb. f. wiss. Bot- 30, 249, 1897. '99. Cell division in sporangia and asci. Ann. of Bot. 13, 467, 1899. :00a. Cell and nuclear division in Fuligo varians. Bot. Gaz. 30, 217,, 1900. HOP. '98. Histologische Studien an Vegatationspunkten. Bot. Centb. 76, 65 1 KARSTEN. :00. Die Auxosporenbildung der Gattungen Cocconeis, Surirella und Cymatopleura. Flora 87, 253, 1900. LAGERHEIM. '99. Ueber ein neues vorkommen von Vibrioiden in der Pflanzenzelle. K. Svenska. Vet. Akad. Forhand. No. 6, 1899. LAUTERBORN. '96. Untersuchungen iiber Bau, Kerntheilung und Bewegung der Diatomeen. Leipzig 1896. LAWSON. : 03a. On the relationship of the nuclear membrane to the protoplast. Bot. Gaz. 35, 305, 1903. MAIRE. :02. Recherches cytologique et taxonomique sur les Basidiomycetes. Bull. d. 1. Soc. Mycol.d. France. 18, 1902. MEYER. '83. Ueber Krystalloide der Trophoplasten und iiber die Chromoplasten der Angiospermen. Bot. Zeit. 41, 489, 503, 525, 1883. '95. Untersuchungen iiber die Starkerkorner. Jena 1895. MlTZKEWITSCH. '98. Ueber die Kerntheilung bei Spirogyra. Flora 85, 81, 1898. MOLL. '94. Observations sur la caryocinese chez les Spirogyra. Arch. Neer. d. Sci. exactes et naturelle 28, 1894. MOTTIER. :OO. Nuclear and cell division in Dictyota dichotoma. Ann. of Bot. 14,. 163, 1900. NEMEC. '99c. Ueber die karyokinetische Kernthielung in der Wurzelspitze von A Ilium cepa. Jahrb. f. wiss. Bot. 33, 313, 1899. : 01. Ueber centrosomahnliche Gebilde in vegetativen Zellen der Gefass- pflanzen. Ber. d. deut. bot. Gesell. 19, 301, 1901. PALLA. '94. Ueber ein neues Organ der Conjugaten Zelle. Ber. d. deut.. bot. Gesell. 12, 153, 1894. PFEFFER. : 00. The physiology of plants. Clarendon Press 1900. 394 THE AMERICAN NATURALIST. [VOL. XXXVIII. ROSEN. '95. Beitrage zur Kenntniss der Pflanzenzellen Cohn's Beitr. z. Biol. d. Pflan. 7, 225, 1895. S ALTER. '98. Zur naheren Kenntniss der Starkekorner. Jahrb. f. wiss. Bot. 32, 117, 1898. SARGANTS. '97. The formation of sexual nuclei in Lilium Martagon, II, Sperma- togenesis. Ann. of Bot. 11, 187, 1897. SCHIMPER. '81. Untersuchungen liber das Wachstum der Starkekorner. Bot. Zeit. 39, 185, 1881. '85. Untersuchungen iiber die Chlorophyllkorper und die in ihnen homologen Gebilde. Jahrb. f. wiss. bot. 16, 1885. SCHMITZ. '82. Die Chromatophoren der Algen. Bonn. 1882. STEVENS. '99. The compound oosphere of Albugo Bliti. Bot. Gaz. 28, 149, 1899. 'Olb. Gametogenesis and fertilization in Albugo. Bot. Gaz. 32, 77, 1901. STRASBURGER. '84. Zur Entwickelungsgeschichte der Sporangien von Trichia fallax- Bot. Zeit. 42, 305, 1884. '95. Karyokinetische Probleme. Jahrb. f. wiss. Bot. 28, 151, 1895. '97a. Kerntheilung und Befruchtung bei Fucus. Jahrb. f. wiss. Bot. 30, 351, 1897- '97b. Ueber Cytoplasmastructuren, Kern und Zelltheilung. Jahab. f. wiss. Bot. 30, 375, 1897. '98. Die pflanzlichen Zellhaute. Jahrb. f. wiss. Bot. 31, 511, 1898. : 00. Ueber Reductionstheilung, Spindelbildung, Centrosomen und Cilienbildner im Pflanzenreich. Hist. Beit. 6, 1900. SWINGLE, W. T. '97. Zur Kenntniss der Kern und Zelltheilung bei den Sphacelariaceen. Jahrb. f. wiss. Bot. 30, 297, 1897. '98. Two new organs of the plant cell. Bot. Gaz. 25, no, 1898. SWINGLE, D. : 03. Formation of spores in the sporangia of Rhizopus nigricans and Phycomyces nitens. Bu. Plant Ind. U. S. Dept. Agri. Bull. 37, 1903. TlMBERLAKE. :01. Starch-formation in Hydrodictyon utriculatum. Ann. of Bot. 15, 619, 1901. : 02. Development and structure of the swarmspores of Hydrodictyon. Trans. Wis. Acad. of Sci. Arts and Letters 13, 486, 1902. No. 449.] STUDIES ON THE PLANT CELL. 395 VAX WlSSELINGH. :00. Ueber Kerntheilung bei Spirogyra II. Flora 87, 355, 1900. :02. Untersuchungen iiber Spirogyra IV. Bot. Zeit. 60 HC^ioo2 WAGER. '94. On the presence of centrospheres in fungi. Ann. of Bot. 8, 321, 1894. ZlMMERMANN. '93 and '94. Sammel-Referate aus dem Gesammtgebiete der Zellenlehre. Bei. z. bot. Centb. 3 und 4, 1893-94. ( To be continued.) VOL. XXXVIII, NO. 450 JUNE, 1904 THE AMERICAN NATURALIST A MONTHLY JOURNAL DEVOTED TO THE NATURAL SCIENCES IN THEIR WIDEST SENSE CONTENTS Page I. Charles Emerson Beecher DR. R. T. JACKSON 407 n. Variation in the Bay Flowers of the Common Cone Flower (Kudbeckia hirta) F. C. LUCAS 427 in- Studies on the Plant Cell.- II DE. BEADLEY MOOE DAVIS 431 IV. Notes and Literature : Zoology, Notes on Recent Fish Literature Botany, 471 Notes 474 V- Publications Eeceived 479 BOSTON, U. S. A. GINN & COMPANY, PUBLISHERS ag BEACON STREET New York Chicago London, W. C, 70 Fifth Avenue 378-388 Wabash Avenue 9 St. Martin'* Street Snttrtd at tfu Post-Officf, Xttttm, Mint., at Simd-Claa Mail Matter. The American Naturalist. ASSOCIATE EDITORS: J. A. ALLEN, PH.D., American Museum of Natural History, Ntw York. E. A. ANDREWS, PH.D ., Johns Hopkins University, Baltimore. WILLIAM S. BAYLEY, FH.D., Colby University, Waieruiili. DOUGLAS H. CAMPBELL, PH.D., Stanford University. ]. H. COMSTOCK, S.B., Cornell University, Ithaca. WILLIAM M. DAVIS, M.E., Harvard University, Cambridge. ALES HRDLICKA, M.D., U.S. National Museum, Washington. D. S. JORDAN, LL.D., Stanford University. CHARLES A. KOFOID, PH.D., University of California, Berkeley. ]. G. NEEDHAM, PH.D., Lake Forest University. ARNOLD E. ORTMANN, PH.D., Carnegie Museum, Pittsburg. D. P. PENHALLOW,D.Sc.,F.R.M.S., Me Gill University, Montreal H. M. RICHARDS, S.D., Columbia University, New York. W. E. RITTER, PH.D., University of California, Berkeley. ISRAEL C. RUSSELL, LL.D., University of Michigan, Ann Arbor. ERWIN F. SMITH, S.D., U.S. Department of Agriculture, Washingtor*. LEONHARD STEJNEGER, LL.D., Smithsonian Institution, Washington, W. TRELEASE, S.D., Missouri Botanical Garden, St. Louis. HENRY B. WARD, PH.D., University of Nebraska, Lincoln. WILLIAM M. WHEELER, PH.D., American Museum of Natural History, New York. THE AMERICAN NATURALIST is an illustrated monthly magazine of Natural History, and will aim to present to its readers the leading facts and discoveries in Anthropology, General Biology, Zoology, Botany, Paleontology, Geology and Physical Geography, and Miner- alogy and Petrography. The contents each month will consist of leading original articles containing accounts and discussions of new discoveries, reports of scientific expeditions, biographical notices of distinguished naturalists, or critical summaries of progress in some line ; and in addition to these there will be briefer articles on various points of interest, editorial comments on scientific questions of the day, critical reviews of recent, literature, and a quarterly record of gifts, appointments, retirements, and deaths. All naturalists who have anything interesting to say are invited to send in their contributions, but the editors will endeavor to select for publication only that which is of truly scientific value and at the same time written so as to be intelligible, instructive, and interesting to the general scientific reader. All manuscripts, books for review, exchanges, etc., should be sent to THE AMERICAN NATURALIST, Cambridge, Mass. All business communications should be sent direct to the publishers. Annual subscription, $4.00, net, in advance. Single copies, 85 cents. Foreign subscription, $4.60. GINN & COMPANY, PUBLISHERS. STUDIES ON THE PLANT CELL. II. BRADLEY MOORE DAVIS. THE ACTIVITIES OF THE PLANT CELL. i. Vegetative Activities. EVERY cell passes through a history whose events repeat in a broad way activities that have become established in the organ- ism by the experience of its ancestors. The most important of these events is nuclear division, which is accompanied in most plants by cell division, the important exceptions being certain groups whose protoplasm is multinucleate throughout all, or almost all, vegetative conditions (e. g., coenocytic Algae and Fungi, plasmodia and multinucleate cells in various tissues). Protoplasm, whose nuclei can no longer divide, becomes inca- pable of reproducing itself and must take a dependent position in the organism, where the length of its life will be determined by the good fortune of its environment and its vitality. Such protoplasm becomes strictly vegetative in its functions, and while these activities may be very highly specialized and of the utmost importance to the organism as a whole, nevertheless such a cell has lost certain of the constructive, and in consequence repro- ductive, possibilities characteristic of living matter. The most evident and important of these constructive activities have to do with the increase of nuclear material (chiefly chromatin), which leads to its distribution through nuclear division, and the devel- opment of a complicated mechanism (the spindle) to effect this result. As Weismann first pointed out, from the standpoint of cell studies, there is a stream of germ plasm flowing with every spe- cies, protoplasm relatively fixed in its characteristics and poten- tially immortal. The chief peculiarities of germ plasm are its reproductive powers and the generalized structure that enables it 43' 432 THE AMERICAN NATURALIST. [VOL. XXXVIII. to turn to any form of activity possible to the species. Portions of the germ plasm are constantly being detached from the main stream and relegated to more or less special duties. Such pro- toplasm becomes the body plasm, or soma, of the individual. Specialized body plasm generally loses very shortly the reproduc- tive possibilities of germ plasm, ,and in consequence must finally die, for its nicely adjusted dependence upon surrounding cells cannot last forever. Yet it has been one of the surprises of biological science that specialized tissues may keep for a very long time the reproductive qualities of germ plasm. Investiga- tions on regeneration in particular have brought these facts con- spicuously to the front. As an extreme example among plants, it is known that even the epidermal tissues of leaves and scales of certain ferns (Palisa, : oo) may sometimes retain the funda- mental qualities of germ plasm and reproduce the plant. There are no visible characters that separate body plasm from germ plasm, excepting, of course, when body plasm begins to show signs of degeneration. Germ plasm may only be distin- guished by its potentialities of growth and reproduction, potenti- alities that cannot be accurately determined because the stimulus to development is, in the last analysis, an external one and the conditions which govern it may be so intricate as to escape close scrutiny. Germ plasm is found in its most generalized condition in the cells of growing points, in embryonic and meristematic regions, and in the reproductive tissues. These tissues are well recog- nized as the most favorable for cell studies because they present most clearly the details of protoplasmic activities. Almost all that we know of cell activities have come from investigations of such regions. One of the first signs of that specialization which transforms germ plasm to body plasm is the slowing up and final end of nuclear and cell division. With this change come a great variety of modifications (peculiarities of cell wall, plastids, cyto- plasmic activities, etc.) which may be readily associated with the particular work of that tissue. The vegetative activities of germ plasm are chiefly those of growth, which in the end mean reproduction, the embryonic cells No. 450.] . STUDIES ON THE PLANT CE~LL. 433 drawing upon food that has been prepared for them and is either stored in special structures (as seeds, spores, bulbs, etc.), or manu- factured in differentiated organs 'or tissues (leaves, chlorophyll bearing tissue, phlcem, etc.). The vegetative activities of body plasm are far more specific than those of germ plasm. Their tissues have particular and highly developed activities, some deal- ing chiefly with photosynthetic processes, some (phlcem) distrib- uting the organized food over the plant body, some storing the food in large quantities. Besides these there are mechanical functions performed by highly differentiated tissues, even though largely composed of empty cells, as the vascular tissue, support- ing tissues, and the external protective integuments. It is not our purpose to discuss any of these vegetative activ- ities in detail, but only to distinguish as sharply as possible the characteristics of germ plasm with its generalized activities from the specialized body plasm. These generalized characters, as before stated, are constructive activities which mean growth and lead to nuclear and cell division. It is probable that any tissue which presents them has regenerative powers that under the proper environment might be expected to reproduce parts or the entire organism. Germ plasm is distributed more widely through- out the organism than is generally supposed, and many highly specialized tissues still retain the spark of regenerative possi- bilities. The significance of these conditions is not generally appreciated, perhaps because the environmental conditions of regeneration are little understood and are exceedingly hard to adjust experimentally. There is presented here a very attractive field of botanical investigation, a union of cell studies with the more gross anatomical methods of experimental morphology. 2. Cell Division. Cell division takes place only after periods of growth that have led to a multiplication of nuclei and in the tissues of plants above the thallophytes is very generally a part of the history of each mitosis. This is because of the structure called the cell plate which is essentially an organ of cell division. But the thallophytes present other methods of cell division which bear no especial relation to nuclear activities, and in certain groups of 434 THE AMERICAN NATURALIST. [Vou XXXV I II. the thallophytes nuclear division may proceed through the entire vegetative life of the organism without any segmentation of the protoplasm which only takes place during the reproductive phase of spore formation. But fundamentally protoplasmic segmenta- tion depends on increase in the amount of protoplasm which demands the multiplication of nuclei so that nuclear division always precedes cell division, and we shall consider the events in that order. (a) Events of Nuclear Division. i. Direct Division. The nucleus divides after one or two methods, either directly by constriction or fragmentation, or indirectly (mitosis) when there is present a fibrillar apparatus called the spindle. Direct division is the only form present in the simplest plants and phy- logenetically must have preceded the elaborate mechanism de- manded for indirect division. This topic will be given especial attention in Section VI '. Direct division is also present in cer- tain specialized cells and tissues of higher plants. These are generally old cells or tissues that are far removed from the gen- eralized structure and potentialities of germ plasm. Yet some- times direct and indirect division occur in the same cell, e. g., Valonia (Fairchild, '94), and such forms might be made the subject of very interesting investigations. In some cases the phenomenon ,of direct nuclear division accompanies pathological conditions or the degeneration of cells and may take the form of extensive fragmentation. It would be outside of our purpose to discuss such phenomena which is obviously abnormal, and the primitive forms of nuclear division will be taken up later (Sec- tion VI). It is possible that direct division in higher plants is in a sense a reversion to early ancestral conditions, a reversion that only comes on when for some reason the normal activities of the germ cell are in abeyance or have ceased. 2. Indirect Division (Mitosis). Indirect nuclear division, mitosis or karyokinesis, is character- ized by a mechanism which varies greatly among plants in its No. 450.] STUDIES ON THE PLANT CELL. 435 method of development. The characteristic appearance of this apparatus is a spindle like figure formed of fibrillae. The poles of the spindle may be occupied by centrosomes or centrospheres or they may be entirely free from such organized kinoplasmic bodies. The essential structures of the spindle are sets of con- tracting fibers which separate the chromosomes into two groups drawing them to the poles of the spindle where the daughter nuclei are organized. But besides these fibers there are gen- erally present other fibrillae which complicate the nuclear figure. Some of these extend from pole to pole (spindle fibers) others lie outside of the spindle and end freely in the cytoplasm or attach themselves to chromosomes (mantle fibers), and if centro- somes or centrospheres be present there are likely to be fibers radiating from these centers to form asters. The events of mitosis are generally grouped into four periods : (a) Prophase, to include the formation of the spindle and prep- aration of the chromosome.s ; (b) Metaphase, the separation of the daughter chromosomes ; (c) Anaphase, the gathering of the daughter chromosomes into two groups which pass to the poles of the spindle ; (d) Telophase, the organization of the daughter nuclei. It is almost needless to say that these periods merge so gradually one into the other that sharp lines cannot be drawn between them. The activities during prophase are especially variable. Prophase. There are two types of spindles in plants, ( i ) those that are formed within the nuclear membrane and (2) those whose fibers originate largely or wholly from kinoplasm outside of the nucleus. Intranuclear spindles have been reported in a number of groups of the thallophytes. They seem to be the rule in the mitoses of oogenesis in the Peronosporales (Wager, '96, :oo, Stevens, '99, :oi and : 02, Davis, :oo, Miyake, :oi, Trow, :oi, Rosenberg, :O3). They are present in Saprolegnia, Fig. 5a (Davis, 103). Fairchild ('94) reports them for Valonia. Farmer and Williams ('98, p. 625) state that the spindle of Ascophyllum is largely intranuclear. Harper (: oo) has not described them for the Myxomycetes, but very little is known about the prophases of mitosis in that group and their presence is quite probable. Timberlake (:O2) is not positive whether the 436 THE AMERICAN NATURALIST. [VOL. XXXVIII. spindles of Hydrodictyon are intranuclear or not ; they lie in a clear space which, however, may be a vacuole rather than the outline of a nuclear cavity. It seems probable in such a type that the vacuole is really the nuclear cavity whose plasma membrane (nuclear membrane) becomes less clearly defined. The development of the spindle is very difficult to follow among these lower forms because it is so small. Stevens (103) found an exceptionally favorable type in Synchytrium and came to the conclusion that the spindle developed from the threads of the spirem (limn) entirely within and independent of the nuclear membrane. Very remarkable intranuclear spindles have been described in the central cell of the pollen tube of Cycas (Ikeno, '98 b) and Zamia, Fig. 5d (Webber, :oi). Murrill (: oo) found them in the mitosis following the fusion of gamete -nuclei in the egg of Tsuga, Ferguson (:oib) at the same period for pine, and Coker (: 03) in Taxodium. They are also reported by Strasburger (: oo) in the cells of young anthers and nucelli of the lily and in grow- ing points (Viscum) and possibly may be found quite generally in cells weak in kinoplasmic cytoplasm. The development of the spindles in the above forms has not been studied in detail, but the fibers are probably derived from the linin. We are given a clue to the process by the events of spindle formation in the spore mother cell of Passiflora (Williams, '99). In this angiosperm the nuclear cavity becomes filled with a fibrillar network developed from the linin, the nuclear wall becomes transformed into a mesh connecting the intranuclear fibers with a surrounding cytoplasmic reticulum. The fibers in the central region of this net work develop the spindle which is consequently very largely of intranuclear origin. Among the thallophytes the poles of intranuclear spindles are frequently occupied by deeply staining bodies which have been called centrosomes ; but these structures can hardly be homol- ogous with the well-known centrosomes of other thallophytes, e. g., Stypocaulon (Swingle, '97) and Dictyota (Mottier, : oo). They are probably merely temporary accumulations of material with no morphological significance. Spindles that arise from fibers external to the nucleus (extra No. 450.] STUDIES ON THE PLAWT~1?ELL. 437 nuclear spindles) are of two main types : (i) those associated with centrosomes, centrospheres or kinoplasmic caps, and (2) those composed of independent fibrillae developed as a mesh around the nucleus. The latter condition is especially character- istic of the spore mother cell and is perhaps the highest type of spindle formation known for either animals or plants. It is very interesting to trace the relations of this highest condition to the lower types through certain lines of evolution to be discussed in Section VI. Spindles with centrosomes are known in Sphacelaria, Stypo- caulon (Swingle, '97), Dictyota, Fig. 4 a (Mottier, : oo), the zoo- sporangium of Hydrodictyon (Timberlake, : 02), in certain diatoms (Lauterborn, principal paper '96, Karsten, : oo) and in the basidium (Wager, '94 and Maire, :O2). The best accounts of the behavior of the centrosomes are given by Swingle and Mottier. Indeed there is much doubt about the history and significance of the bodies in the other forms, although the con- stancy of their presence at the poles of the spindles indicates that they are really centrosomes. The conditions in the diatoms are especially complicated ; an account of Lauterborn's work has been published in English by Rowley, :o3. In Stypocaulon, Sphacelaria (Fig. 3 c, Section I) and Dictyota (Fig. 4 a) the cells studied have permanent asters which lie at the side of the nucleus and which divide just previous to the mitosis and sep- arate so that they come to lie on opposite sides of the nucleus. Fibers develop from the centrosomes on the sides nearest the nucleus and elongating push against the nuclear membrane and finally enter the nuclear cavity to form the spindle. Spindles with centrospheres are well known in Fucus (Farmer and Williams, '96, '98, Strasburger, '9/a), Corallina, Fig. 5 c, (Davis, '98), in the ascus, Fig. 5 b (Harper, '97 and '99), and in the germinating spore of Pellia, Fig. 4 c (Farmer and Reeves, '94, Davis, :oi, Chamberlain, :O3). Centrospheres have been reported in other forms but the types mentioned above have received the most careful study. It is probable that the centro- sphere is but a larger, more generalized kinoplasmic center than the centrosome, a protoplasmic region whose dynamic activities do not focus so sharply as in the latter structure. There are 438 THE AMERICAN NATURALIST. [VOL. XXXVIII. bodies, as in the basidium, which stand intermediate in size between centrosomes and centrospheres and are probably only called the former because they are very distinct in outline. Centrospheres in Fucus (Fig. 3 d, Section I), Corallina (Fig. 4 b} and Pellia (Fig. 3 e, Section I, Fig. 4 c) are formed de novo for each mitosis by an accumulation of kinoplasm at the poles of the elongating nucleus. The centrospheres in the ascus divide before each of the three successive mitoses and finally remain, one for each nucleus, to instigate the peculiar process of free cell formation characteristic of the ascus. Centrospheres are frequently the centers of asters which, however, are usually not as sharply denned as those with centrosomes, possibly because the fibers are not grouped with the same degree of symmetry as is shown around controsomes. Spindle fibers from centrospheres develop in precisely the same manner as from centrosomes, i. e. by the growth of the fibrillae into the nuclear cavity through the dissolving nuclear membrane. The activity is well shown in the oogonium of Fucus, and Farmer ('98, p. 638) believes "that the intranuclear part of the spindle is differentiated out of nuclear material that is unused for chromosome formation." The entrance of spindle fibers from centrospheres at the ends of a nucleus has been observed by myself in Corallina, Fig. 4 b (Davis, '98). The germinating spores of Pellia, Fig. 4 c (Davis, :oi, Chamberlain, : 03) furnish especially good illustrations of the entrance of spin- dle fibers into the nuclear cavity and the development of the spindle in this form is coincident with the dissolution of the nucleus which, according to Strasburger's theory ('95), indicates that the latter structure contributes material for the growth of spindle fibers. In connection with the centrosphere mention should be made of the blepharoplasts of the cycads and Ginko which are remark- able bodies with radiating fibers. They have been considered by some as asters with centrosomes, but it is known that they take no part in spindle formation or other mitotic phenomena in these forms, and consequently need not be considered at this time. They will be treated in some detail in the account of the sperm (Section III). No. 450.] STUDIES ON THE PLANT CELL. 439 Kinoplasmic caps which form spindles are probably an evolu- tion from the type of centrosphere that is developed de novo with each mitosis as in Pellia. Such centrospheres by -becoming less definite in form and lacking radiating fibers would be called kinoplasmic caps. Indeed the centrosphere so evident in the early cell divisions of the germinating spore of Pellia becomes a kinoplasmic cap in the later mitoses of the older gametophyte (Davis, :oi). Spindles developed from kinoplasmic caps are characteristic of FIG. 4. Prophases of Mitosis, a. Dictyota ; late prophase in spore mother cell, fibers from the two asters with centrosomes have entered nuclear cavity to organize the spindle, chro- mosomes gathering to form the nuclear plate, b, Corallina, early prophase in tetra spore mother cell; two centrospheres, the fibers for.n one having entered the nuclear cavity, chromosomes shown, c , Pellia, nucleus in germinating spore ; spindle fibers from ill defined centrospheres entering nuclear cavity, chromosomes and a nucleolus present, d, Gladiolus, first mitosis in pollen mother cell ; a multipolar spindle, nuclear wall breaking down at one side and fibrillse entering the nuclear cavity, chromosomes and a nucleolus present. After Mottier and Lawson. the mitoses in vegetative tissues, meristematic and other embry- onic regions. They have been especially studied in higher plants by several investigators and for a large number of forms, those most completely described being Psilotum (Rosen, '95), Equise- tum, Allium and Solanum (Nemec, 'Q8a and '98b, *99b and '99c), Pteris, Ephedra and Vicia, (Fig. 3 /, Section I) (Hof, '98) and Allium (McComb, :oo). The polar caps first appear as accumu- lations of kinoplasm on opposite sides of the nucleus which generally elongates. The protoplasm is granular and although 440 THE AMERICAN NATURALIST. [VOL. XXXVIII. central bodies have been reported most investigators are agreed that they are only granules without regularity or special signifi- cance. They are no longer believed to be centrosomes. Fibrillae are developed from the kinoplasmic caps and grow out against the nuclear membrane and finally enter the nuclear cavity to form the spindle. A large part of the substance of the kinoplasmic cap is transformed into these spindle fibers. Papers by Schaffner ('98) on Allium and Fulmer ('98) on the seedling of the pine are the last attempts to bring the centro- some into the history of spindle formation in vegetative tissues of higher plants. But their results cannot stand against the accumulation of studies which indicate that centrosomes are not present in the cells of any plant above the thallophytes with the possible exception of the mysterious blepharoplast and certain structures appearing in some phases in the life history of Hepaticae. Centrospheres are unquestionably present in the Hepaticoe and centrosomes have also been reported. The centrospheres are, however, so generalized as to approach the kinoplasmic caps in structure and development and it seems quite possible that they are the forerunners of this manifesta- tion of kinoplasm. The so-called centrosomes of the liverworts do not exhibit the specialized structure or behavior of cen- trosomes among the thallophytes and it is probable that they are only smaller and somewhat more clearly defined centro- spheres. These structures in the Hepaticae seem to hold an intermediate relation between the definite kinoplasmic bodies (asters, centrosomes and centrospheres) of the thallophytes and the remarkable kinoplasmic activities in higher plants which reach their highest expression in the processes of spindle forma- tion in the spore mother .cell. These topics will be treated in Section VI. Structures resembling kinoplasmic caps have been reported in several other tissues than those noted above. Thus Murrill (:oo) finds in the formation of the ventral canal cell of Tsuga a dense fibrous accumulation beneath the nucleus which develops one pole of the spindle in essentially the same manner as other polar caps. The other pole of the spindle in this case appears to be formed differently for the fibers seem to be intranuclear. No. 450.] STUDIES ON THE PLANT CELL. 441 It would be interesting if two types of spindle formation were present at opposite pole$ of the same nucleus and further inves- tigation of this subject is much to be desired. The mitoses in the central cell of Pinus (Ferguson, :oib, Chamberlain, '99, and Blackman, '98) and Picea (Miyake : O3a) show spindle for- mation from accumulations of fibrillae outside of the nucleus but without conspicuous polar caps. Still more striking than the irregular spindle of Murrill in Tsuga, described above, is Miss Ferguson's (:oia) account of the mitosis in the generative cell of the pollen grain of Pinus. The spindle here begins to develop as a cap-like accumulation of kinoplasm below the nucleus. The fibers enter the nuclear cavity and in cooperation with a nuclear reticulum form a system of fibers that extend through the nuclear cavity to the inner side of the nuclear membrane beyond. This portion of the nuclear membrane persists until after metaphase so that one pole of the spindle is found wholly within the nucleus while the other is external and of unques- tioned cytoplasmic origin. Coker, : 03, regards the spindle which differentiates the nucleus of the ventral canal cell in Taxoclium as almost wholly of nuclear origin and the chromosomes as derived largely from the nucleolus. There are evidently some interesting complications in this form which deserve further study. It should be noted that whenever spindles are formed in con- nection with centrosomes, centrospheres or kinoplasmic caps that the fibers have a definite region of attachment from which they extend into the nuclear cavity. Such regions constitute a sort of anchorage for the spindle fibers. In this respect the physi- ological side of the process of spindle formation in these forms is quite similar to that of the animal kingdom and in sharp contrast to other methods that are found in higher plants, which will now be considered. When spindles are formed after the second method, i. <") As was stated in Section I, the nuclear membrane probably represents the reaction of the granular kinoplasm to a fluid secretion around the chromosomes which becomes the nuclear sap (La wson, :O3a). However, the nuclear membrane is generally a definitely organized film, much more sharply defined than vacuolar membranes. The development of the linin network is not well understood. It is readily seen that the chromosomes become joined end to end and sometimes elon- gate. The amount of chromatin diminishes as the linin substance appears, but it is not certain whether the chromatin is changed directly into linin, or whether the latter substance is a secretion: The best evidence rather favors the former view. Nucleoli are also believed to hold a very close chemical relation to chromatin. It is uncertain whether or not the chromosomes lose their organic identity in the daughter nuclei. Investigations on this problem are surrounded by many difficulties. It has been claimed by Guignard ('99) for Naias and Strasburger (: oo) for several forms that the chromosomes may be followed with cer- tainty through the period between the first and second mitosis in the spore mother cell. But other investigators have not been able to trace the chromosomes after telophase and are inclined to believe that the chromosome completely loses its identity in the resting nucleus. One of the last investigations of Lilium (Mottier, : 03) argues strongly for the latter view, and all who have followed nuclei from one mitosis into another know that the resting nucleus with its linin network and the granular chromatin present conditions that generally make the recognition of chromosomes impossible with the instruments and technique at our command, but this does not prove that they may not be present. The theory of the permanence of the chromosome has met with much favor because it is argued that otherwise how could the number be maintained so regularly through immense num- bers of mitoses. But it can hardly be said that the doctrine is established. It has also found favor because all the events of No. 450.] STUDIES ON THE PLANT CELL. 449 mitosis emphasize the importance of the chromosomes which are really the only enduring structures in the nucleus and have led to their being considered as the probable bearers of heredi- tary qualities. 3. The Dynamics of Nuclear Division. Mitotic phenomena in certain plant cells present evidence that has very direct bearing on some of the theories that deal with mechanical and dynamical explanations of nuclear division. The methods of spindle formation and the various forms of kinoplas- mic structures (centrosomes, centrospheres and kinoplasmic caps) which generally in plants seem not to be permanent organs of the cells all tend to support Strasburger's conception of kinoplasm, which is an outgrowth and application to plants of Boveri's well known theory of archoplasm. The centrosome theory is supported by very few investiga- tions in Botany, the most notable being that of Swingle ('97), for Stypocaulon, who believes that the centrosome divides with the aster and is maintained as a permanent organ throughout successive cell divisions. Other examples of similar conditions may be found among the thallophytes which, after all, have received very little attention, and such types as Dictyota and the diatoms offer excellent subjects for studies covering a series of cell divisions. But in contrast to Stypocaulon it should be noted that the conspicuous centrospheres of Fucus and Cor- allina disappear with each mitosis to be formed anew, and the same conditions obtain in the germinating spores of liverworts (Pellia). There seems to be no place for the centrosome in spindle formation as presented in the spore mother cells of all groups above the thallophytes (see Sec. III). Neither does mitosis in the vegetative tissues of these groups, characterized as it is by the presence of kinoplasmic caps, conform to the program of the centrosome theory. The morphological manifestations of kinoplasm are so various that we are driven to a very general conception of its organiza- tion. Kinoplasm runs through cycles in which the structure passes from a granular condition to a fibrillar and then back again 450 THE AMERICAN NATURALIST. [VOL. XXXV II I. to the granular state. By the granular state we mean one in which no fibrillae seem to be present, but instead the microsomata are densely and homogeneously massed. It is possible that such microsomata form a closely packed network, but no such struc- ture is visible under the microscope. The first appearance of kinoplasm at prophase of mitosis is frequently the granular condition. This state is illustrated by such accumulations as centrospheres and kinoplasmic caps and by the granular zone that has been reported around the nuclei of some spore mother cells. Granular kinoplasm becomes fibrillar probably by the arrange- ment of the microsomata into a reticulum from which fibers extend freely into the surrounding cytoplasm. These fibers undoubtedly elongate during prophase, extending in various directions. Some press against the nuclear membrane and when this breaks down grow rapidly into the nuclear cavity. Of these a portion extend from pole to pole and form the central spindle. Others attach themselves to the chromosomes and lie either among the central fibers or somewhat outside of the spindle (mantle fibers). Still others may extend freely into the cyto- plasm as astral rays from the pole of the spindle, a very com- mon condition when centrosomes or centrospheres are present. A contraction of the fibrillae, beginning with metaphase, is just as characteristic of mitosis as their elongation during prophase. The fibers attached to the chromosomes draw the latter to the poles of the spindle. The central fibers in higher plants draw away from the poles and give their substance to the cell plate. The substance of contracted mantle fibers, with other kinoplasm at the poles of the spindle, probably become distributed around the group of daughter chromosomes so that they finally lie sur- rounded by a sphere of kinoplasm. It does not seem as if we knew much more about the struc- ture and activities of kinoplasm during mitosis than is indicated in this cycle of change from a granular condition through a fibrillar state back to the granular condition, with a period when the fibers elongate and another when they contract. This with few exceptions is the history for every mitosis. The exceptions deal with peculiar conditions or structures. Thus, for example, No. 450.] STUDIES ON THE PLANT CELL. 45 I the astral rays of the centrospheres in the ascus instead of con- tracting to a center or disappearing in the cytoplasm after the last mitosis grow around the nucleus and cut out a portion of the cytoplasm to form the spores, thus contributing their sub- stance to a plasma membrane. There is little doubt that kinoplasmic fibrillae actually exist as structural elements in the protoplasm. Their growth and move- ment in the cytoplasm and nuclear cavity, their multiplication and shifting arrangements as the spindle develops, and their contraction to the poles of the spindle or to a cell plate give these fibers an individuality that cannot be explained on the theory that they merely represent lines of force or paths of dynamic stimuli. They apparently perform all the activities mentioned above by virtue of their own structural organization which is that of rows of microsomata and in this organization resemble and are probably closely related to cilia. There is an excellent discussion of this subject by Allen, : 03, p. 302, etc. Some authors believe that there is a streaming movement in the astral rays (Chamberlain, : 03, for Pellia) either towards or away from the pole of the spindle. This view is founded on the granular appearance of the radiations which are sometimes very thick in Pellia and enlarge at the points where they join the centrospheres or the outer plasma membrane. It is not alto- gether clear that the larger of these structures are quite the same as spindle fibers since they seem to be actually strands of cytoplasm rather than fibrillae. It is probably safe to assume that the forms which kinoplasm takes have relation to dynamic activities, but it is not easy to define these. Thus centrosomes, centrospheres and kinoplasmic caps may well be the centers from which dynamic stimuli extend, and they may be the focal points of other energies. These problems have been very little investigated among plants. It is obvious that differentiated regions of kinoplasm have important physical relations to other portions of the protoplasm, one of the most important being the anchorage which they give to fibrillae, thereby largely governing the direction of such strains as come about through the contraction of these structures in the later periods of mitosis. 452 THE AMERICAN NATURALIST. [VOL. XXXVIII. But the essential characteristics of kinoplasm stand out sharply from whatever point the phenomena of mitosis is viewed, and in this protoplasm with its power of forming contractile fibers is vested some of the most conspicuous activities of nuclear division as well as the important powers given plasma membranes in relation to the segmentation of protoplasm to be considered presently. The dynamic activities concerned with the spindle present only half the story of mitosis. The other important events occur inside of the nucleus. One of these is the dissolution of a por- tion or the whole of the nucleolus which takes place as the spindle develops and we have already given the views of Stras- burger ('95 and : oo), supported by the studies of other investi- gators, that its substance in certain instances furnishes material for the development of the spindle. But the chief events in the interior of the nucleus deal with the accumulation of chromatin on the spirem thread which with the disappearance of the linin indicates that the latter substance may become converted into the former. The splitting of the spirem ribbon longitudinally is of the utmost significance for thereby is made possible an exact and homogeneous distribution of the chromatic material in the nucleus. We do not know how the spirem ribbon splits nor have we as yet any evidence of the origin and evolution of this peculiar activity. (b) Segmentation of the Protoplasm. Mitosis in the uninucleate cells of plants is generally followed by immediate cell division, which takes place in groups above the thallophytes through the formation and cleavage of the cell plate in the equatorial region of the spindle between the daughter nuclei. Among thallophytes, as so far studied, cell division is chiefly through cleavage by constriction. There are forms among the thallophytes and also in the spermatophytes whose nuclei gather about themselves a portion of the cytoplasm, wherein they lie, which becomes cut out of the general mass by a cell wall. This is free cell formation. Multinucleate masses of protoplasm, such as plasmodia and No. 450.] STUDIES ON THE PLANT CELL. 453 portions of coenocytes, generally divide extensively at repro- ductive periods and always through cleavage by constriction with, however, the frequent cooperation of vacuoles which heip~to cut the protoplasm in the same manner as the cleavage furrows. Cleavage by constriction is undoubtedly the most primitive type ; free cell formation and cleavage by cell plates being special and very highly developed protoplasmic activities. i . Cleavage by Constriction. A simple example of cleavage by constriction is presented by such an alga as Cladophora. The process consists in the build- ing out of a ring of cellulose from the side wall into the cell cavity. The outer plasma membrane forms a fold, thus placing the two surfaces opposite one another (see Fig. 8 a), and the wall is laid down between these. Spirogyra forms its wall in precisely the same manner as Cladophora with this peculiarity, that the new wall finally cuts through the protoplasmic strands that connect the daughter nuclei. These strands are said to con- tain spindle fibers (Van Wisselingh, : 02) which may contribute to the plasma membranes forming the cell wall, as it is completed. Another illustration of cleavage by constriction is presented in the formation of gametes of moulds (Sporodinia) and the abstric- tion of conidia (Erysipheae), both processes having been studied by Harper, '99, p. 506. In these cases a cleavage furrow pro- ceeds from the surface inward and divides the protoplasm. The partition wall of cellulose is formed later between the two free plasma surfaces. The only differences between the processes above described are that in the first forms the cleavage proceeds more slowly and the wall follows the furrow as it progresses in the interior of the cell, while in the latter types cleavage is com- plete before the plasma membranes develop the wall. Cell division in the red Algae (Rhodophyceae) is also a process of constriction similar to Cladophora, but the wall is not generally formed entirely across the filament so that adjacent cells remain connected by thick strands of protoplasm. These processes become much more complicated when large masses of multinucleate protoplasm are divided up into many 454 THE AMERICAN NATURALIST. [VOL. XXXVIII. smaller bodies as during spore formation among the Myxomy- cetes and Mucorales. Very complete studies have been made of these conditions by Harper, '99 and :ooa. In the slime mould (Fuligo) cleavage begins by furrows on the external surface which " cut down at all angles into the homogeneous proto- plasm." The direction of the cleavage furrows is further com- plicated by the fact that many of them start from the bottom and sides of deep folds. All of the furrows may bend and secondary cleavage planes strike off from them which in time unite with one another until the protoplasm is divided progres- sively into very many small masses (see Fig. 8 b~) that finally round themselves off and secrete walls, becoming spores, some- times with one nucleus and sometimes with several. Cleavage in the sporangium of Synchytrium and the moulds, as described by Harper, '99, is in general similar to that in the plasmodium with, however, the additional feature that lines or planes of vacuoles are often utilized to assist a cleavage furrow in effecting the segmentation of the protoplasm. The separa- tion of the spore plasm of the sporangium of Pilobolus from the filament below begins with a cleavage furrow from the exterior ; but this furrow follows and makes use of a curved plane of flattened vacuoles with the result that a dome shaped cleft is developed and two plasma membranes are presented face to face, which form the columella wall between them. The segmenta- tion of the spore plasm in Pilobolus is affected somewhat similarly through the cooperation of cleavage furrows from the exterior with vacuoles which cut into the protoplasm at various angles to meet one another and the cleavage furrows. The bodies first formed in the sporangium of Pilobolus are not the final spores. Harper suggests that they may correspond to the zoospores of Saprolegnia. They are generally uninucleate and begin immediately a period of growth within the sporangium characterized by extensive nuclear multiplication and several divisions of the protoplasmic body by constriction. Harper finds that the spore plasm of Sporodinia is separated from the filament below by a dome-shaped plane of flattened vacuoles which fuse together and, unlike Pilobolus, cut their way to the surface of the sporangium. Thus the cleavage is deter- No. 450.] STUDIES ON THE PLANT CELL. 455 mined entirely by the activity of vacuoles. Spore formation, however, is accomplished by cleavage furrows which progress from the exterior inwards and, without the aid of conspicuous vacuoles, cut out multinucleate masses of protoplasm which become the spores. Dean Swingle (: 03) has extended the studies of Harper on spore formation in the molds to Rhizopus and Phycomyces. He confirms Harper's account of the general processes of cleavage by furrows cooperating with vacuoles, and notes the following characteristics in the types studied. In Rhizopus the position of the columella is determined by a dome-shaped series of flat- tened vacuoles which fuse and meet a cleft that extends upward from the outer plasma membrane at the base of the sporangium. The spores are formed in Rhizopus by branching systems of curved furrows that cut the protoplasm into multinucleate masses, and in Phycomyces by angular vacuoles that develop into furrows which extend in various directions and unite with one another and with clefts from the region of the columella. Other excellent illustrations of cleavage by constriction are presented in the sporangia of such types as Hydrodictyon, Clado- phora and Saprolegnia. Timberlake (: 02) has given an account of Hydrodictyon, and the events are also fairly well understood for Saprolegnia. Segmentation begins in Hydrodictyon by the development of cleavage furrows in the outer plasma membrane, which cut into the protoplasmic layer at right angles to the surface and meet similar furrows that make their way from the large central vacuole outward. These cleavage planes spread lat- erally, uniting with one another, until the protoplasm is all divided into uninucleate masses which become the zoospores (Fig. 8 c). In Saprolegnia (see Davis, 103, for general account) conspicuous cleavage furrows develop from the central vacuole and make their way to the exterior, finally breaking through the outer plasma membrane. When this takes place there is an immediate escape of cell sap, which was under pressure, and a shrinkage of the sporangium so that the zoospore origins appear to fuse, but this is not really the case, for cleavage is continued and the zoospores soon separate. A physiological explanation of cleavage by constriction must 456 THE AMERICAN NATURALIST. [VOL. XXXVIII. consider two sets of factors. There is an evident contraction of the protoplasm in many examples because water is given off. The shrinkage of the surface would undoubtedly form furrows, but, as Harper has pointed out, these furrows do not develop in an accidental manner. Non-nucleated masses of protoplasm are never separated from the nucleated, but the segmentation pro- ceeds after a system by which the final products contain only one nucleus or at most a limited number. So it is probable that the nuclei are the ultimate centers controlling the segmentation which at its commencement may be quite irregular. This explanation of sporogenesis in the plasmodium and the spo- rangium is not altogether satisfactory for the cell division of Cladophora, the abstriction of conidia or the development of the gametes of a mould. In these examples the cleavage begins at definite regions of the plasma membrane, so that the stimulus must be local, and the direction of the plane has a definite relation to the axis of the plant. It is important to note (see Harper, : oo, p. 240-249) how inadequate are some of the well-known theories of the segmen- tation of protoplasm as explanations of cleavage by constriction. Hofmeister's law ('67) that cell division is across the axis of growth obviously cannot be applied to the irregular segmentation in the plasmodium and sporangium, nor is Sachs' well-known law of growth in vegetative points adequate. Sachs, '94, and in the Lectures on the Physiology of Plants, chap. XXVII, conceives a growing point of a higher plant or an embryonic structure as a mass of protoplasm whose cell walls are determined by principles of rectangular intersection of perpendicular planes. The outer form of the structure determines the angles of periclines and anticlines and the transversals conform to these. There is not the slightest hint of such an order in the distribution of cleavage planes in the multinucleate masses of protoplasm just described and Sachs' law in so far fails of general application whether or not it be satisfactory for the conditions with which he especially deals. There are also explanations of cell division, applicable to the tissues of many higher organisms, based on the position of the nuclear figure in the cell, which determines the position of the cell plate but these theories cannot handle the events in the No. 450.] STUDIES ON THE PLANT CELL. 457 plasmodium or sporangium where the cleavage planes are formed without regard to the time of nuclear division or the position of mitotic figures. 2. Cleavage by Cell Plates. Cleavage of the protoplasm by means of the cell plate is almost universal in cell division of plants above the thallophytes. It is one of the peculiarities of plant cells, having been found in comparatively f e w Animals and there represented rather imper- fectly by the so-called^ mid-body. The general events of the process have been known since Treubs' studies of 1878, and were clearly described by Strasburger in 1880. Timberlake, : oo, in a recent paper gives an historical review of the subject. When, after the metaphase of mitosis, the two sets of daugh- ter chromosomes separate from one another there is left between them the spindle, made up of the central fibers. The first appearance of the cell plate is a line of granules in the equatorial region of this spindle where the nuclear plate formerly lay. But several important events proceed this condition. The con- necting central fibers begin to thicken, first near the daughter nuclei, and then gradually towards the equatorial region of the spindle. The number of fibers may increase greatly, probably by the separation of bundles of fibrillae composing the spindle into independent elements (Timberlake, : oo, p. 94). But there is evidence that new fibrillae are sometimes formed from the vicinity of the daughter nuclei, some of which may enter the spindle and cooperate with the connecting fibers. In certain forms (e. g. Allium) there is' an accumulation of a stainable sub- stance between the connecting fibers in the equatorial region of the spindle. The reaction of this substance to stains indicates a carbohydrate composition. The cell plate really begins with the thickening of the con- necting fibers in the equatorial plane of the spindle. In some forms these thickenings are elongated bodies, in others mere granules. The earlier writers (Treub, '78, Zacharias, '88) did not believe that they came from the spindle fibers, but there seems to be now no doubt of their origin from these elements, 458 THE AMERICAN NATURALIST. [VOL. XXXVIII. which contract and thicken as the plate develops. The bodies composing the cell plate finally lie in a plane extending the entire width of the spindle (Fig. 8 d) and they then broaden and come in contact with one another to form a continuous membrane, which, as has been said, may lie in a matrix of car- bohydrate material. The cell plate grows rapidly as the central spindle fibers shorten and contribute their substance to the structure. During this contraction the surrounding cytoplasm FIG. 8. Segmentation of the Protoplasm, a, b, c, cleavage by constriction, d, cleavage by cell plate, e, f, g, free Cell Formation. , cell division in Cladophora. b, cleavage of spore plasm in Fuligo. c, spore formation in Hydrodictyon. d, first division of spore mother-cell in Pellia. e, spore formation in ascus, i and 2 (Erysiphse) astral fibers cutting out cytoplasm around nuclei, 3 portion of ascus with developing spores (Lachnea). f, obgonium of Albugo, egg surrounded by membrane pierced by antheridial tube, ccenocen- trum and female gamete nucleus within, g, egg of Ephedra with four embryo cells. After Strasburger Harper and Timberiake. enters the region between the barrel shaped group of fibers and the daughter nuclei (Fig. 8 d}. It is probable that the cell plate is composed entirely of the substance of spindle fibers and in consequence is kinoplasmic in character. The cell plate widens with the accretion of material from the central spindle, which in some cases is assisted by the radiating fibers that, lying outside of the spindle, contract and add their material to the edge of the plate. The cell plate thus extends laterally and finally reaches No. 450.] STUDIES ON THE PLANT CELL. 459 the neighboring cell walls, fusing with the outer plasma mem- brane. There are certain mitoses, as in some spore mother cells and in the embryo sac (see Section III) where the c^l -plates are absorbed into the cytoplasm leaving the' original cell with two or more nuclei and without partition walls. It is uncertain whether the edge of the plate is ever extended by the develop- ment of additional peripheral fibrillae (Timberlake, : oo, p. 161) from the daughter nuclei. Cell division is accomplished by the splitting of the cell plate (Strasburger, '98) into two plasma membranes. The division generally begins in the center and the cleft progresses towards the periphery until it reaches the cell wall. During the process the thickened rod shaped portions of the spindle fibers are pulled apart. There are thus left two kinoplasmic membranes opposite one another and continuous with the outer plasma membrane surrounding the daughter cells. The cause of this cleavage is not apparent, but there are reasons for believing that the split is essentially a thin vacuole which, starting near the center, cuts its way through the cell plate to the periphery after a manner very similar to the behavior of vacuoles during the cleavage of the plasmodium and in the sporangia of certain moulds. And there may be shown in this activity a relationship of cleavage by cell plate to some of the events of cleavage by constriction. After division is complete there follows the formation of a cell wall between the two cell surfaces after the method usual to plasma membranes. The new cell wall generally begins in the oldest portion of the cell plate where the cleft first appeared and is gradually built out peripherally until it reaches the side walls. The first indica- tion of the wall is the appearance in the cleft of a stainable carbohydrate substance which resembles the material that was primarily present between the fibers of the central spindle and which disappears with the formation of the cell plate. This material is probably the basis of the first deposits on the surface of the two plasma membranes, but the nature of the final sub- stance is exceedingly various. A cell wall may be formed that is homogeneous throughout but often the thickened wall presents three regions, two layers of a cellulose basis formed by the 460 THE AMERICAN NATURALIST. [VOL. XXXVI II. respective plasma membranes and between them the so-called middle lamella. The middle lamella has been the subject of much discussion. It is not the remains of the cell plate as was once supposed. Neither is it exactly a cement between two cell walls. Its his- tory is undoubtedly various, for the composition shows much plasticity. The origin of the middle lamella at the surface of a plasma membrane indicates a morphology similar to a cell wall, but the substance, pectic in character, shows transformations far removed from the cellulose compounds that are formed later and which give thickness to the cell wall. Allen (:oi) discusses the subject in detail. The origin of the cell plate is a subject of interest which will be further discussed in Section VI. There are some types, espe- cially among the thallophytes, where a cell plate is present, but apparently in a somewhat undeveloped and rudimentary condi- tion. These forms suggest transitional conditions between cleav- age by constriction with the aid of vacuoles, so general among the thallophytes, and cleavage by the cell plate, characteristic of higher groups. The most interesting examples are Anthoceros, Chara, Basidiobolus, Pelvetia, Fucus, and Sphacelaria. Cell plates are formed with each of the two successive mitoses in the spore mother cell of Anthoceros (Van Hook, : oo ; Davis, :oi, p. 158), but the structure in some species is exceedingly small (e. g., A. Icevis} and can scarcely extend more than one- tenth of the distance across the cell. It is larger in other forms, as in the one studied by Van Hook ; but even there the nuclear figure of the second mitosis is only one-third of the width of the cell. The protoplasm divides simultaneously in the four spores with the characteristic arrangement. If this division were deter- mined entirely by cell plates there would be required an exten- sive development of fibrillae, of which there is no evidence in the cell. But their place seems to.be taken by numerous delicate strands of cytoplasm which connect the four protoplasmic masses, each of which contains a large chromatophore and an accom- panying nucleus. A film is formed in the intermediate region, and this marks the position of the cell wall. It is, of course, quite certain that the two cell plates of the second mitosis are No. 450.] STUDIES ON THE PLANT CELL. 461 a part of this membrane and may start its development, but the final structure must contain very much more material than could possibly be contributed by the sparsely developed spindle fibers . Thus, although the splitting of the cell plate may start the proc- ess of segmentation, its final course and end is probably deter- mined by cleavage through vacuoles, thus utilizing a method characteristic of the thallophytes. Chara appears to have a fairly well developed cell plate (Deb- ski, '97) which extends almost entirely across the cell, presenting very exceptional conditions among the thallophytes. This pecu- liarity is in keeping with other characters of the spindle, which begins its development outside of the nuclear membrane and, lacking centrosomes, resembles the nuclear figures of higher plants. It is possible that nuclear studies upon Chara through- out ontogeny might show a variation that would be very signifi- cant for the evolutionary problems concerned with the structure of protoplasm. Fairchild ('97) reports a cell plate for Basidiobolus when the beak cells are cut off from the gametes. The structure, as fig- ured and described, is not, however, conspicuous. He points out general resemblances between cell division in this form and in the Conjugales, where, as Van Wisselingh (: 02) described later for Spirogyra, spindle fibers connect the daughter nuclei .and may cooperate towards the end of cell division with a cleav- age furrow from the side of the cell. The conditions in the Fucales are not altogether clear. Both Strasburger ('97a) and Farmer and Williams ('98) report that the central spindle disappears in Fucus without the formation of the cell plate and that the wall is developed between the daugh- ter nuclei in a region of granular cytoplasm. However, in Pel- vetia some of the radiating fibrillae from opposite sides of the daughter nuclei bend around these structures and end in the new wall. It is not plain that they contribute much it anything to its formation in the way of substance, but it would seem prob- able that they hold a directive relation to the structure (Farmer and Williams, '98). The Sphacelariaceae seem to be somewhat similar to the Fuca- les in their methods of cell division. The beautiful figures of 462 THE AMERICAN NATURALIST. [VOL. XXXVIII. Swingle ('97) for Stypocaulon give details of the region of the cytoplasm that forms the partition wall between the daughter nuclei. There is a zone of fine meshed protoplasm between much larger vacuoles. It is possible that some very long fibrillae may connect the daughter nuclei with this zone, but they do not form a cell plate. Consequently the wall must be developed in this delicate alveolar layer, which probably splits along some plane of vacuoles. The process of cleavage is then really related to such activities of vacuoles as occur in the sporangium of the Mucorales and in the plasmodium. But the position of the alve- olar layer may be determined by the fibrillae, since it is always situated nearest to the smaller of the two daughter nuclei. It seems likely that the process of cleavage in the Fucales will be found to be similar to Stypocaulon when the details of struc- ture in the internuclear cytoplasm is known. So this group, with others, is likely to furnish conditions in which spindle fibers may determine the position of the cell wall and exert a directive influ- ence upon it without actually laying down a cell plate. As has been pointed out, the splitting of the cell plate is probably a cleavage along a very thin flat vacuole, so that the process in its essential characters is the same as cleavage through a series of vacuoles. Thus cleavage by tlie cell plate is possibly an out- growth from that phase of cleavage by constriction in which the extensive fusion of vacuoles determines the planes of separation. The important advance lies in the new factors, introduced through the activities of fibrillae, which become very conspicuous as actual contributors of material to the kinoplasmic film which is laid down as the cell plate. This function of the fibrillae probably developed slowly from conditions such as those in Stypocaulon and Pelvetia, where their influence upon the position of the cell wall, if any at all, can scarcely be more than directive. 3. Free Cell Formation. Whenever a nucleus becomes the center around which cyto- plasm is gathered and separated from the rest of the cell con- tents, so that the new cell lies freely in the protoplasm of the old, this is free cell formation. Illustrations are presented by No. 450.] STUDIES ON THE PLANT CELL. 463 the spores of an ascus, the oospore of the Peronosporales, the embryo cells of Ephedra, and probably other gymnosperms, and in some cases seemed to be exemplified in the conditions pre- sented by the egg and synergids and the antipodals of the embryo sac. Spore formation in the ascus is known through the studies of Harper ('97 and '99). After the final divisions in the ascus the nuclei lie in the cytoplasm, each with an aster at its side (Fig. 8 e, 3). A delicate prolongation carries the aster with its cen- trosphere away from the main body of the nucleus (e, i). The rays of the aster now bend over and grow around the nucleus, presenting an umbrella-like figure (e, 2). They finally meet on the opposite side, and thereby cut out a portion of the cytoplasm which is included in the spore. The . substance of the aster fibers forms the basis of a kinoplasmic film which becomes the plasma membrane of the ascospore and develops the spore wall externally after the usual method. This peculiar activity of an aster is unparalleled in plant or animal cells. Oogenesis in the Peronosporales has been described in some detail by several authors, but the process has not generally been called free cell formation. Yet at the end of the process the oospore, enveloped by periplasm, lies free in the oogonium. In the beginning the ooplasm gathers in the center of the oogonium as a denser alveolar region around that peculiar protoplasmic body (generally present) the ccenocentrum. This accumulation forces the vacuoles, together with most of the nuclei, to the periphery, where they lie in a sort of protoplasmic froth next the cell wall and constitute the periplasm. The spore wall de- velops at the boundary of the ooplasm, so that it lies close to the large vacuoles (Fig. 8/) in the periplasm. There must be an accumulation of kinoplasm, perhaps from the plasma membranes of numerous vacuoles, to form a delicate layer between the two regions of the oogonium. This layer of kinoplasm probably splits along the line of vacuoles between the ooplasm and peri- plasm, for the primary walls are certainly established between two plasma membranes, because the secondary layers are added to it from both sides. Nuclei in division frequently lie very close to the boundary of the ooplasm, but there is no evidence 464 THE AMERICAN NATURALIST. [VOL. XXXVIII. that the kinoplasmic membrane has any relation to these mitotic figures. That is to say, there are no fibrillae to contribute sub- stance to the membrane, and its development must be concerned with vacuoles alone. In this respect the process recalls the part played by vacuoles in the plasmodium and in certain sporangia during cleavage by constriction. Free cell formation after the method in the egg of Ephedra (Strasburger, '79), which is also likely to be found among other gymnosperms, takes place during the differentiation of the em- bryo cells. The cytoplasm collects around each nucleus, forming a sphere (Fig. 8 g), and a wall is developed on the outside of this body. Details of the process are not known, and it is not clear whether the position of the membrane is determined by the vacuoles that must border upon this region or whether there are fibers radiating from the nucleus which might lay down a cell plate around the denser protoplasm ; but the evidence favors the former possibility. Somewhat similar conditions are presented in the egg appa- ratus of many embryo sacs. In certain forms (e. g., the lily so well described by Mottier, '98) the egg nucleus and synergids are thickly invested by radiating fibers, and these, together with the cell plates, may readily determine the position of the plasma membrane that forms the cell wall. But fibers do not seem to be conspicuously present in the egg apparatus of many other embryo sacs (Excellent illustrations can be found among the Ranunculaceae). In these cases the protoplasm collects around the nuclei as dense areas bordered by vacuolar cytoplasm, and it is possible that the vacuoles by fusing with one another cut out these respective regions and thus determine the plasma mem- branes of the egg and synergids. Such processes would extend the activities of vacuoles, which accompany cleavage by constric- tion in the thallophytes, to the highest groups of plants. It is curious that with all of the work upon the embryo sac we should know less about the segmentation of the protoplasm around the synergid, antipodal, and segmentation nuclei in this structure than in the sporangia of the molds, the ascus, or dur- ing spore formation in the Myxomycetes. ( To be continued]. No. 450.] STUDIES ON THE PLANT CELL. 465 LITERATURE CITED FOR SECTION II. ALLEN. :01. On the origin and nature of the middle lamella. Bot. Gaz. 32, I. ALLEN. : 03. The early stages of spindle formation in the pollen mother cells of Larix. Ann. of Bot. 17, 281. BELAJEFF. '94. Zur Kenntniss der Karyokinese bei der Pflanzen. Flora, 79, 430. BLACKMAN. '98. The cytological features of fertilization and related phenomena in Pinus silvestris L. Phil. Trans. Roy. Soc. of London, 190, 395. CHAMBERLAIN. '99. Oogenesis in Pinus laricio. Bot. Gaz. 27, 268. CHAMBERLAIN. : 03. Mitosis in Pellia. Bot. Gaz. 26, 29. COKER. : 03. On the Gametophytes and Embryo of Taxodium. Bot. Gaz. 36, i and 1 14. DAVIS. '98. Kerntheilung in der Tetrasporenmutterzelle bei Corallina officinalis L. var. mediterranea. Ber. d. deut. bot. Gesell. 16, 266. DAVIS. '99- The spore mother-cell of Anthoceros. Bot. Gaz. 28, 89. DAVIS. :00. The fertilization of Albugo Candida. Bot. Gaz. 29, 297. DAVIS. :01. Nuclear studies on Pellia. Ann. of Bot. 15, 147. DAVIS. :03. Oogenesis in Saprolegnia. Bot. Gaz. 35, 233 and 320. DEBSKI. '97. Beobachtungen iiber Kerntheilung bei Chara fragilis. Jahrb. f. wiss. Bot. 30, 227. FAIRCHILD. '94. Ein Beitrag zur Kenntniss der Kerntheilung bei Valonia utricularia. Ber. d. deut. bot. Gesell. 12, 331. FAIRCHILD. '97. Ueber Kerntheilung und Befruchtung bei Basidiobolus ranarum Eidam. Jahrb. f. wiss. Bot. 30, 285. FARMER AND REEVES. '94. On the occurrence of centrospheres in Pellia epiphylla Nees. Ann. of Bot. 8, 219. , FARMER AND WILLIAMS. '96. On the fertilization and segmentation of the spores in Fucus. Ann. of Bot. 10, 479. 466 THE AMERICAN NATURALIST. [VOL. XXXVIII. FARMER AND WILLIAMS. '98. Contributions to our knowledge of the Fucaceae ; their life history and cytology. Phil. Trans. Roy. Soc. of London, 190, 623. FERGUSON. :01. The development of the pollen tube and division of the generative nucleus in certain species of Pinus. Ann. of Bot. 15, 193. FERGUSON. :01. The development of the egg and fertilization in Pinus strobus. Ann. of Bot. 15, 435. FULMER. '98. Cell division in Pine seedlings. Bot. Gaz. 26, 239. GUIGNARD. '99. Le deVeloppement du pollen et la reduction chromatique dans le Naias major. Arch. d. Anat. micro. 2, 455. HARPER. '97. Kerntheilung und freie Zellbildung im Ascus. Jahrb. f. wiss. Bot. 30, 249. HARPER. '99. Cell division in sporangia and asci. Ann. of Bot. 13, 467. HARPER. :00a. Cell and nuclear division in Fuligo varians. Bot. Gaz. 30, 217. HOP. '98. Histologische Studien an Vegetationspunkten. Bot. Centb. 76, 65. HOFMEISTER. '67. Die Lehre von der Pflanzenzelle. Leipzig. IKENO. '98. Untersuchungen iiber die Entwickelung der Geschlechtsorgane und dem Vorgang der Befruchtung bei Cycas revoluta. Jahrb. f . wiss. Bot. 32, 557. JUEL. '97. Die Kerntheilung in den Pollenmutterzellen von Hemerocallis fulva und die bei denselben auftretenden Unregelmassigkeiten. Jahrb. f. wiss. Bot. 30, 205. KARSTEN. :OO. Die Auxosporenbildung der Gattungen Cocconeis, Surirella und Cymatopleura. Flora, 87, 253. LAUTERBORN. '96- Untersuchungen iiber Bau, Kerntheilung und Bewegung der Diato- meen. Leipzig. MAIRE. :02. Recherches cytologique et taxonomique sur les Basidiomycetes. Bull. d. 1. Soc. mycol. d. France, 18. McCoMB. :00- The development of the karyokinetic spindle in vegetative cells of higher plants. Bull. Tor. Bot. Club, 27, 451. CH UNIVERB.TY No. 450.] STUDIES ON THE PLANT CELL. 467 MlVAKE. :01- The fertilization of Pythium de Baryanum. Ann. of Bot. 15, 653. MlYAKE. :03rt. On the development of the sexual organs of Picea excelsa. Ann. of Bot. 17, 351. MOORE. :03. The mitoses in the spore mother-cell of Pallavicinia. Bot. Gaz. 36, 384- MOTTIER. '97. Beitrage zur Kenntniss der Kerntheilung in den Pollenmutterzellen einiger Dikotylen und Monokotylen. Jahrb. f. wiss. Bot. 30, 169. MOTTIER. '98. Ueber das Verhalten der Kerne bei der Entwickelung des Embryo- sacks und die Vorgange bei der Befruchtung. Jahrb. f. wiss. Bot. 31, 125. MOTTIER. :00- Nuclear and cell division in Dictyota dichotoma. Ann. of Bot. 14, 163. MOTTIER. : 03. The behavior of the chromosomes in the spore mother-cells of higher plants and the homology of the pollen and embryo-sac mother- cell. Bot. Gaz. 35, 250. MURRILL. :00- The development of the archegonium and fertilization in the hem- lock spruce (Tsuga Canadensis Carr.) Ann. of Bot. 14, 583. NEMEC. '98ff. Ueber die ausbildung der achromatische Kerntheilungsfigur im vegetativen und fortpflanzungs Gewebe der hoheren Pflanzen. Bot. Centb. 74, i. NEMEC. '98. Ueber das Centrosoma der tierischen Zellen und die homodyna- men Organe bei den Pflanzen. Anat. Anzeig. 14, 569. NEMEC. '99*. Ueber Kern und Zelltheilung bei Solanum tuberosum. Flora, 86, 214. NEMEC. '99c. Ueber die karyokinetische Kerntheilung in der Wurzelspitze von Allium cepa. Jahrb. f. wiss. Bot. 33, 313. OSTERHOUT. '97. Ueber die Entstehung der karyokinetischen Spindel bei Equi- setum. Jahrb. f. wiss. Bot. 30, 1 59. PALISA. :00. Die Entwickelungsgeschichte der Regenerationsknospen welche an den Grundstiicken isolirte Wedel von Cystopteris Arten ent- stehen. Ber. d. deut. bot. Gesell. 18, 398. 468 THE AMERICAN NATURALIST. [VOL. XXXVIII. ROSEN. '95. Beitrage zur Kenntniss der Pflanzenzellen. Cohns Beitr. z. Biol. d. Pflan. 7, 225. ROSENBERG. :03. Ueber die Befruchtung von Plasmopara alpina. Bihang. till. k. svenska vet-akad. Handlingar. 28. ROWLEY. :03. Some points in the structure and life history of diatoms. Jour. Quekett Mic. Club, II, 2, 417. SACHS. '94. Physiologische Notizen VIII, Mechanomorphosen und Phylogenie. Flora 78, 215. SCHAFFNER. '98. Karyokinesis in the root tips of Allium cepa. Bot. Gaz. 26, 225. STEVENS, F. L. '99. The compound oosphere of Albugo Bliti. Bot. Gaz. 28, 149. STEVENS, F. L. :01#. Gametogenesis and fertilization in Albugo. Bot. Gaz. 32, 77.- STEVENS, F. L. :02- Studies in the fertilization of the Phycomycetes, Sclerospora. Bot. Gaz. 34, 420. STEVENS AND STEVENS. : 03. Mitosis of the primary nucleus in Synchytrium decipiens. Bot. Gaz. 35, 405. STRASBURGER. '79. Die Angiospermen und die Gymnospermen. Jena. STRASBURGER. '80. Zellbildung und Zelltheilung. Jena. STRASBURGER. '95. Karyokinetische Probleme. Jahrb. f. wiss. Bot. 28, 151. STRASBURGER. '97- Kerntheilung und Befruchtung bei Fucus. Jahrb. f. wiss. Bot. 3, 35i- STRASBURGER. '98. Die pflanzlichen Zellhaute. Jahrb. f. wiss. Bot. 31, 511. STRASBURGER. :00. Ueber Reductionstheilung, Spindelbildung, -Centrosomen und Cili- enbildner im Pflanzenreich. Hist. Beitr. 6. SWINGLE, D. :03. Formation of the spores in sporangia of Rhizopus nigricans and of Phycomyces nitens. Bu. Plant Ind. U. S. Dept. Agri. Bull. 37. SWINGLE, W. T. '97. Zur Kenntniss der Kern und Zelltheilung bei den Sphacelariaceen. Jahrb. f. wiss. Bot. 30, 297. No. 450.] STUDIES ON THE PLANT CELL. 469 TlMBERLAKE. : 00. The development and f unction f of the" cell plate in higher plants. Bot. Gaz. 30, 73. TlMBERLAKE. : 02. Development and structure of the swarm spores of Hydrodictyon. Trans, wiss. Acad. of Sci., Arts and Let. 13, 486. TREUB. '78. Quelques recherches sur la r61e du noyau dans la division des cel- lules vegetales. Amsterdam. TROW. :01. Biology and cytology of Pythium ultimum. Ann. of Bot. 15. 269. VAN HOOK. : 00. Notes on the division of the cell and nucleus in liverworts. Bot. Gaz. 30, 394. VAN WISSELINGH. :02. Untersuchungen iiber Spirogyra, IV Beitrag. Bot. Zeit. 60, 115. WAGER. '94. On the presence of centrospheres in fungi. Ann. 'of Bot. 8, 321. WAGER. '96 On the structure and reproduction of Cystopus candidus Lev. Ann. of Bot. 10, 295. WAGER. :00. On the fertilization of Peronospora parasitica. Ann. of Bot. 14, 263. WEBBER :01. Spermatogenesis and fecundation of Zamia. Bu. of Plant Ind. U. S. Dept. of Agri. Bull. 2. WILLIAMS, CLARA L. '99. The origin of the karyokinetic spindle in Passiflora crerulea. Proc. Cal. Acad. Sci. Bot. Ill, i, 189. ZACHARIAS. '88. Ueber Kern und Zelltheilung. Bot. Zeit. 46, 33 and 51. VOL. XXXVIII, Nos. 451-452 JULY-AUGUST, 1904 THE AMERICAN NATURALIST A MONTHLY JOURNAL DEVOTED TO THE NATURAL SCIENCES IN THEIR WIDEST SENSE CONTENTS Page I. Proceedings of the American Society of Zoologists 485 H. The Anatomy of the Coniferales (Continued) PROF. D. P. PENHALLOW 523 III- A List of Bermudian Birds seen during July and August, 1903. HAROLD BOWDITCH 565 IV. Neritina virginica Variety Minor PROFESSOR M. M. METCALF 565 V. Studies of the Plant Cell. Ill . DR. B. M. DAVIS 571 VI. Notes and Literature : Zoology, Dodge's General Zoology, Coues' Key to 595 North American Birds, Boulenger on the Classification of Bony Fishes, Notes on Recent Fish Literature Palaeontology, Eastman's Transla- 605 tion of Zittel, Vol. II Botany, a new Book on Ferns, Porter's Flora 608 of Pennsylvania, The Journals. BOSTON, U. S. A. GINN & COMPANY, PUBLISHERS New York Chicago London, W. C, 70 Fifth Avenue 378-388 Wabash Avenue 9 St. Martin'* 8trct Entered at the Post-Office, Boston, Mats., as Second-Clast Mail Matifr. The American Naturalist. ASSOCIATE EDITORS: J. A. ALLEN, PH.D., American Museum of Natural History, New York. E. A. ANDREWS, Ptt.D.,foAns Hopkins University, Baltimore. WILLIAM S. BAYLEY, FK.D., Colby University, ffiaierzritfc. DOUGLAS H. CAMPBELL, PH.D., Stanford University. J. H. COMSTOCK, S.B., Cornell University, Ithaca. WILLIAM M. DAVIS, M.E., Harvard University, Cambridge. ALES HRDLICKA, M.D., U.S. National Museum, Washington. D. S. JORDAN, LL.D., Stanford University. CHARLES A. KOFOID, PH.D., University of California, Berkeley. J. G. NEEDHAM, PH.D., Lake Forest University. ARNOLD E. ORTMANN, PH.D., Carnegie Museum, Pittsburg. D. P. PENHALLOW,D.Sc.,F.R.M.S., Me Gill University, Montreal. H. M. RICHARDS, S.D., Columbia University, New York. W. E. RITTER, PH.D., University of California, Berkeley. ISRAEL C. RUSSELL, LL.D., University of Michigan, Ann Arbor. ERWIN F. SMITH, S.D., U.S. Department of Agriculture, Washington. LEONHARD STEJNEGER, LL.D., Smithsonian Institution, Washington. W. TRELEASE, S.D., Missouri Botanical Garden, St. Louis. HENRY B. WARD, PH.D., University of Nebraska, Lincoln. WILLIAM M. WHEELER, PH. D., American Museum of Natural History, New York. THE AMERICAN NATURALIST is an illustrated monthly magazine of Natural History, and will aim to present to its readers the leading facts and discoveries in Anthropology, General Biology, Zoology, Botany, Paleontology, Geology and Physical Geography, and Miner- alogy and Petrography. The contents each month will consist of leading original articles containing accounts and discussions of new discoveries, reports of scientific expeditions, biographical notices of distinguished naturalists, or critical summaries of progress in some line ; and in addition to these there will be briefer articles on various points of interest, editorial comments on scientific questions of the day, critical reviews of recent literature, and a quarterly record of gifts, appointments, retirements, and deaths. All naturalists who have anything interesting to say are invited to send in their contributions, but the editors will endeavor to select for publication only that which is of truly scientific value and at the same time written so as to be intelligible, instructive, and interesting to the general scientific reader. All manuscripts, books for review, exchanges, etc., should be sent to THE AMERICAN NATURALIST, Cambridge, Mass. All business communications should be sent direct to the publishers. Annual subscription, $4.00, net,' in advance. Single copies, 85 cents. Foreign subscription, $4.60. GINN & COMPANY, PUBLISHERS. r STUDIES ON THE PLANT CELL. III. BRADLEY MOORE DAVIS. SECTION III. HIGHLY SPECIALIZED PLANT CELLS AND THEIR PECULIARITIES. VERY much of our knowledge of the structure and behavior of protoplasm in plants has been derived from the study of cer- tain cells whose organization has reached an exceptionally ad- vanced degree of differentiation. The peculiarities of these cells are obvious and have proved of great interest but we have as yet scarcely made a beginning in the study which must trace and relate these characteristics of the most complex products of cellular evolution in plants to their more simple progenitors. This section will describe in some detail the structure and protoplasmic activities of the following six highly specialized cells : . i, The Zoospore ; 2, The Sperm ; 3, The Egg ; 4, The Spore Mother-Cell ; 5, The Coenocyte ; 6, The Coenogamete. i. The Zoospore. Zoospores are interesting not only for their own peculiarities but also because they are well known to be the progenitors of the sexual cells or gametes which become later differentiated into the egg and sperm. Comparative studies upon three cells so closely related and yet so diverse in their extremes of struc- ture are sure to yield important results. The zoospore is generally an uninucleate cell, colorless in the Fungi, but containing a chromatophore or plastids in all other groups of thallophytes. There are usually two or four cilia attached to the anterior pointed end which is free from coloring matter and at this region one may expect to find a red pigment spot. Some zoospores are exceptional for special peculiarities, as those of Vaucheria which are multinucleate, each nucleus 572 THE AMERICAN NATURALIST. [VOL. XXXVIII. being accompanied by a pair of cilia, or those of CEdogonium whose colorless forward end bears a crown of numerous cilia. The zoospore stands among the higher forms for a type of motile organism that is very close to the bottom of the assemblage of groups and developmental lines which make up the Algae. The forms most closely related to the zoospore are in the family Chlamydomonadeae of the Volvocales. But at this general low level of the plant kingdom there are several groups whose members pass most of their .lives in motile conditions (Volvo- cales, Flagellates and Peridinales) and the cells of all of these types resemble zoospores to a greater or less degree in their structure and habits, so that this condition represents a wide- spread and well defined stage of evolutionary development. Therefore when zoospores are formed in the life history of some higher plant they represent a return on the part of the organism for a short time to the structure and mode of life of an ancestry perhaps related in some way to the groups that still have the motile habits throughout most of their existence. For these reasons close comparisons in structure between the zoospore and motile Algae will be interesting and should help to explain the peculiarities of these cells. These peculiarities chiefly concern the organ that forms the cilia (blepharoplast), which becomes very complex in the sperm, and the pigment spot. Unfortunately studies upon these problems have been few and we are not prepared to make a general statement of the condi- tions. The most recent investigation on the structure of the zoospore is that of Timberlake (: 02), but Strasburger has written extensively on the subject, especially in the Histologische Bei- trage ('92 and : oo). The later paper (:oo, p. 177-215) reviews the entire subject of cilia formation. Dangeard has presented an account of the Chlamydomonadeae, '99, and in :oi described especially Polytoma, comparing its structure with that of the animal spermatozoan. Polytoma (see Fig. 9 a) is a colorless organism but its cell structure and life history place it unquestionably among the Chlamydomonadeae. The two cilia arise from a small body (blepharoplast) situated at the extremity of the cell. A delicate Nos. 451-452.] STUDIES ON THE PLANT CELL. 573 thread-like structure, which Dangeard calls the rhizoplast, extends from the blepharoplast into the cytoplasm and sometimes ends at the side of the nucleus in a granule (condyle). The-c-ilia grow out from the blepharoplast. This apparatus is not known to bear any relation to centrosomes or to the kinoplasm of nuclear figures present at the time of spore formation. But it should be noted that the blepharoplast is situated directly under if not actually in the outer plasma membrane, which is kino- plasmic. The filamentous connection between blepharoplast and nucleus is probably important, especially since it has also been found in zoospores (Timberlake, : 02, for Hydrodictyon) but we do not even know its developmental history much less its function. Further study will be necessary to make clear possi- ble relations to kinoplasm around the nucleus or to centrosomes. Consequently Dangeard' s comparison of Polytoma to the animal spermatozoon is not convincing for it seems to be established for the spermatozoon that portions of the middle piece at least and the flagellum are derived from a true centrosome. Indeed from the meager evidence now at hand the blepharoplast of Polytoma is as likely to be a structure differentiated from the plasma membrane as to have any relation to the nucleus. But detailed studies on sporogenesis may discover a history more in harmony with that of Hydrodictyon. We have summarized a portion of Timberlake's (:O2) account of sporogenesis for Hydrodictyon in the previous section under the head of " Cleavage by constric- tion." We shall consider now certain details. Small spherical bodies are found at the poles of the spindles during nuclear divi- sion in the mother-cell. They are undoubtedly accumulations of kino- plasm and perhaps stand for centro- somes. However they have no polar radiations nor could they be followed between mitoses when the nuclei were in resting conditions. FIG. 9. The Zoospore. a, Polytoma; 6, Hydrodictyon ; c, Development in Oedo- gonium. (a, after Dangeard : 01 ; b, Tim- berlake : 02 ; c, Strasburger'92.) It is not probable therefore 574 THE AMERICAN NATURALIST. [VOL. XXXVIII. that these structures are permanent in the cell. After nuclear multiplication is ended segmentation proceeds until the nucleate masses of protoplasm separate from one another as zoospores. Then a body may be found lying in contact with the plasma membrane and bearing a pair of cilia (Fig. 9$). This basal body (blepharoplast) by its reaction to stains seems to be entirely distinct from the plasma membrane and is connected with the nucleus by very delicate threads. There is a time just previous to the differentiation of the zoospores when the nuclei lie very close to the cleavage furrow that finally separates the adjacent zoospore origins. A granule may sometimes be observed close to these nuclei and it is possible that this is the first appear- ance of the basal body (blepharoplast). If this should prove correct the structure may have a direct relation to the kinoplasm around the nucleus, a relation that is afterwards maintained through the two or three delicate fibers that connect these structures. Thus the blepharoplast if not directly derived from a centrosome may at least have its origin from the same region of kinoplasm. However these possibilities are mere speculations and the investigation of these points is very much to be desired in a number of algal and fungal types. We are now brought to the views of Strasburger as expressed in his writings of '92 and : oo. His investigations have been chiefly on Vaucheria, Cladophora and CEdogonium. In all of these forms the cilia come from a body (blepharoplast) which he believes to arise from the outer plasma membrane (Haut- schicht). The nucleus lies close to the plasma membrane at the time when the blepharoplast is formed and may determine its development there as a dynamic center, but the blepharoplast is not a centrosome according to Strasburger. It is of course kinoplasmic since it develops from the plasma membrane and this would accord with its activities as a cilia forming organ. The blepharoplast is extraordinarily large in CEdogonium (see Fig. 9 c) and develops a ring of numerous cilia on the exterior while at the same time fibrillar rays grow back into the cyto- plasm and probably help to give a compact organization to the zoospore. This structure is very suggestive of the centrosphere and aster that cuts out the ascospore (see Section II, Free Cell Nos. 451-452.] STUDIES ON THE PLANT CELL. 575 Formation) and in spite of Strasburger's conclusions that it is derived entirely from the plasma membrane we are justified in asking for a fuller description of its development. There is the possibility of a different origin wherein the nucleus may play an important part which, in the light of Timberlake's studies on Hydrodictyon, suggests that Strasburger may not have discov- ered the earliest beginning of the blepharoplast in CEdogonium. And the same doubts apply to Cladophora and Vaucheria. There is thus considerable divergence in the views of the origin and nature of the blepharoplast in zoospores, Strasburger believing that they are developed as a specialized region of the plasma membrane with no relation to centrosomes, and Timber- lake holding that the structure in Hydrodictyon is not a part of the plasma membrane but comes from the interior of the proto- plasm. The problem is also involved with conditions in the sperm, where there is likewise a difference of opinion as to the homologies of the blepharoplast but an undoubted origin at least in the pteridophytes and gymnosperms from the interior of the cell. We should naturally expect the blepharoplasts of zoospores and sperms to be homologous and consequently the problem is of great theoretical interest and will be taken up again in our discussion of the sperm. Its solution demands a most thorough study of the development of some of the larger zoospores as in CEdogonium and certain species of the Confer- valesand Volvocales. The pigment spot is almost universally present in zoospores and is also characteristic of the cells of many motile organisms as in the Volvocales and Flagellates while occasionally found in other groups. The structure has been called an eye spot from its fancied resemblance to the simple eyes of certain Crustacea (Cyclops, etc.) but this term is unsatisfactory since it is not established that the pigment spot is primarily a receptive organ for light or warmth ; but even should it prove to be thus sensi- tive (which is very probable) thereby orienting the cell with respect to the direction of incoming rays, that is not a function comparable to sight. The coloring matter of the pigment spot is held as a single globule or as -a collection of numerous small granules in meshes 5 7 6 THE AMERICA N NA TURALIST. [VOL. XXXVIII. of the protoplasm. It is frequently associated with a plastid. The pigment may be readily broken down and dissolved out by such reagents as alcohol and ether. In chemical composition it is very close to haematochrome and thus may be related to chlorophyll or a derivative of that substance. The cytoplasm around the pigment spot is undifferentiated and when the color- ing matter is removed it is very difficult and sometimes impos- sible to find the situation of the structure. Consequently the pigment spot can hardly be considered a protoplasmic organ since it is merely an accumulation of coloring matter at some point in the cell. Strasburger (:oo, p. 193) states that the pigment spot of certain zoospores (Cladophora, etc.) is formed in the plasma membrane but this is not true of many other motile cells (Flagellata) and there is no doubt that in some cells (e. g. the gametes of Cutleria) the pigment spot is a portion of a plastid. The literature upon the structure and function of pigment spots is reviewed by Zimmermann (Beitrage z. hot. Centralb. Bd. 4, p. 159, 1894) and since then Wager ('99) has presented a detailed study of Euglena. 2. The Sperm. The sperm is unquestionably derived from the zoospore through primitive types of gametes which were identical with zoospores in all essentials of morphology. I have described the origin and evolution of sexual cells of plants in two recent papers (Popular Science Monthly, Nov. 1901, p. 66 and Feb. 1902, p. 300). We should expect the simplest forms of sperms to have the characters of zoospores and this is the fact. The sperms of the Algae, as a rule, have the same number of cilia (usually two) as their ancestral asexual zoospores. They gener- ally contain a chromatophore, although sometimes much reduced, and there is present the pigment spot. The cilia are attached at the pointed end or at the side, arising from colorless pro- toplasm that sometimes contains the pigment spot while the chromatophore, when present, and the nucleus lie at some distance from this region of the cell. The sperms of bryophytes and pteridophytes are much attenuated in form and lack the Nos. 451-452.] STUDIES ON THE PLANT CELL. 577 pigment spot and chromatophore. Those of the bryophytes and the Lycopodinese are biciliate while other pteridophytes have multiceliate sperms the cilia being distributed on a band (blepharoplast) which lies along one side of the spiral structure. A large portion of the spiral in these sperms is composed of nuclear substance and much of the remaining cytoplasm with granules and vacuolar inclusions may frequently be found in a vesicle attached to the larger end of the spiral. The only motile sperm among the Fungi is that of Mono- blepharis. The male cells of other Fungi are non-motile bodies (spermatia) generally formed from the ends of delicate filaments which are found in special organs called spermagonia. Spermagonia have been described in the Uredinales, the lichens and in the Laboulbeniaceae but their function is only clearly established for the last two groups. They are very highly differentiated in the Laboulbeniaceae and comprise several types of structure. Another type of male cell, found in certain groups of the Phycomycetes and Ascomycetes, is the ccenogamete (to be described presently) which is however not the homologue of the sperm but of the mother-cell or antheridium that develops such structures. Sperms of the red Algae (Rhodophyceae) are likewise non-motile and they are invariably formed singly in small cells at the ends of filaments. These non motile sperms of Fungi and red Algae are exceedingly small uninucleate bodies without further complexity of structure as far as is known. We shall not attempt to discuss the earlier literature that treats of the structure and development of the plant sperm. In 1894 Belajeff published a German translation of a paper written two years before in Russian which presents the views of previous investigators and to this the reader is referred for such historical references. At that time various opinions were held respecting the organization of the sperm, some writers (Campbell, Guignard and others) believing that it was chiefly or wholly nuclear in origin, while another group (Zacharias, '87, Belajeff, Strasburger, '92, etc.) thought that the cytoplasm shared very largely in its structure. Belajeff ('94a) from studies among the Characeae showed with especial clearness that the cytoplasm was an important constituent of this sperm since the nuclear 578 THE AMERICAN NATURALIST. [VOL. XXXVIII. material occupied a restricted region in the middle of the spiral structure. This was the first of a series of investigations which have given especial attention to cytoplasmic activities during spermatogenesis and placed the entire subject in a new light. The year 1 897 brought forth almost simultaneously three' short papers by Webber ('9/a, '9/b, '9/c) and Belajeff ('97a, '9/b, '9/c) respectively. Webber had studied the development of the motile sperms of Zamia and Ginko, Belajeff certain forms of the Filicineae and P^quisetineae. These were of the nature of pre- liminary announcements and both authors published later more detailed descriptions and discussions. The discoveries of motile sperms in Ginko by Hirase and of Cycas by Ikeno were announced in several short papers during the years 1896 and '97 but without descriptions of their development. This litera- ture together with later papers of Ikeno, Shaw, Belajeff, Hirase, and Fujii is reviewed in Webber's last contribution (:oi) and also in Strasburger's discussion of " Cilienbildner " (: oo, p. 177) to which the reader is referred for the most complete treatments of spermatogenesis in plants yet published. The cycads and Ginko are the most favorable subjects known for studies in spermatogenesis. Detailed accounts of the cycads are given by Ikeno ('98b) for Gycas and by Webber (:oi) for Zamia, these forms agreeing with one another in all essentials. Two sperms are developed from the daughter cells (spermatids) following the division of the so-called body cell in the pollen tube. The process really begins in the body cell with the appearance of the blepharoplasts. Their development has been followed with especial attention in Zamia. They are formed de novo in the cytoplasm at some distance from the nucleus and while the latter is in the resting condition. They appear inde- pendently of one another, generally on opposite sides of the nucleus but sometimes much nearer together (Fig. io). Each is a large deeply staining body with numerous radiations extending into the cytoplasm. The blepharoplasts then increase in size and, moving farther away from the nucleus, take positions exactly opposite to one another. The nucleus of the body cell now divides, its spindle being clearly intranuclear (Fig. 5 d] and consequently holding no visible relation to the blepharoplasts. Nos. 451-452.] STUDIES ON THE PLANT CELL. 579 which lie at a considerable distance from the structure (Fig. 10 b}. The latter cannot then be said to occupy the position of centro- somes in relation to this spindle. Meanwhile important changes, which are best known for Zamia, take place in the blepharoplast. In this type the structure forms a hollow sphere which breaks up into segments and finally into granules as mitosis proceeds. The radiations disappear without holding any apparent relation to the spindle. During telophase each of the two blepharoplasts FIG. 10. Spermatogenesis in Cycas. a, Body cell in pollen tube with two blepharoplasts ; s, stalk cell; /, prothallial cell; b, anaphase of mitosis in the body cell the spindle lying between the two blepharoplasts which have begun to form cilia ; c, Blepharoplast elongat- ing, in contact with a process from the nucleus; d, end of blepharoplast attached to the nucleus at a later stage of development ; e, sperm showing section of the flattened spiral blepharoplast with cilia projecting beyond the cell. (After Ikeno, '98.) < appears as a mass of granules at some distance from the daugh- ter nuclei which are to become the sperm nuclei. As a result of this division the spermatids (sperm mother-cells) are differen- tiated. At the close of the mitosis the blepharoplast enters upon its functions of forming in the spermatid a cilia bearing band which is to lie as a spiral around the sperm. The granules first extend as a delicate deeply stained line towards the nucleus and then in the opposite direction. The nucleus in Cycas puts forth a papilla (Fig. io^r) which meets this line of granules and remains attached to it for some time. The line thickens into a 580 THE AMERICAN NATURALIST. [VOL. XXXVIII. band which lengthens and finally takes the form of a spiral of five or six turns which becomes more or less closely applied to the plasma membrane (Fig. 10 e, blepharoplast in section). The cilia develop as protuberances from the outer surface of the band (Fig. loc and d} and grow through the plasma membrane to the exterior of the cell. The nucleus in the meantime has increased in size until it occupies the greater part of the top shaped sperm (Fig. loe). The history of spermatogencsis in Ginko is strikingly parallel to that of the cycads. The chief features were first described by Webber ('97c) and in greater detail by Hirase ('98). The two blepharoplasts appear de novo on opposite sides of the nucleus in the body cell. They show the same high state of differentia- tion as those of the cycads, being large and the center of a number of prominent radiations. Ginko however presents a peculiarity not reported in the previous group. A large spheri- cal body lies between each blepharoplast and the nucleus in an area of granular cytoplasm. This structure stains deeply like the globules of nucleolar substance which are frequently found in the cytoplasm after nuclear division. They are probably accumulations of a somewhat similar material at these points in the cell to be utilized at later periods of spermatogenesis, since they decrease in size as the sperms mature. The spindle in the body cell is formed between the blepharoplasts but its poles lie at some distance from and are entirely independent of these structures. During this mitosis the spherical bodies pass to one side of the spindle so that the daughter nuclei (sperm nuclei) finally take the position formerly occupied by them. The blepharoplast becomes granular and begins to lengthen into a band, one end of which becomes attached to the nucleus that puts forth a small papilla towards the blepharoplast. The band elongates and takes the form of a spiral which makes several turns around one end of the cell just under the plasma mem- brane. Cilia then develop along this band as in the cycads. The earlier accounts, describing a short tail on the sperm were founded upon material that was not altogether normal and have been corrected by Webber and Fujii. The mature sperms have essentially the same form as those of Zamia and Cycas. Nos. 451-452.] STUDIES ON THE PLANT CELL. 581 There has been some discussion on the morphology of these motile sperms of the gymnosperms. The claim has been made that they are ciliated spermatids (sperm mother-cells) and there- fore different from the sperms of pteridophytes which are formed inside of mother-cells that upon their escape are left behind as empty cysts. However a close analysis of their struc- ture will show that the sperms in both groups have an identical protoplasmic organization. There is a nucleus and a greater or less amount of cytoplasm in which the blepharoplast lies and the entire structure is surrounded by a plasma membrane. Any differences in the processes of spermatogenesis can only concern the greater or less development of a cellulose membrane around the spermatids. It may be true that this cellulose membrane is entirely absent in Cycas and Zamia, but if present it would be merely a shell like envelope around the sperm and cannot affect its morphological unity and agreement with the sperms of pteridophytes. A comparative study of the composition and formation of the walls enclosing sperm nuclei in the sperma- tophytes is much needed to carefully distinguish between plasma membranes and the cellulose secretions that may be developed by them. While the cycads and Ginko have very much the largest sperms known and are consequently extremely favorable for an examination of spermatogenesis nevertheless some surprisingly detailed studies have been made among the Filicineae and Equise- tineae. Following his preliminary announcements ('9/a, '9/b, '9/c), Belajeff published in '98 an account of spermatogenesis in Gymnogramme and Equisetum. These forms present histories parallel to each other and to the cycads. Two deeply staining bodies (blepharoplasts) appear on opposite sides of each nucleus previous to the final mitosis in the antheridium which differen- tiates the spermatids. Consequently each spermatid receives a blepharoplast which lies close beside the nucleus. The bleph- aroplast begins to elongate and is followed by the nucleus so that both structures form two parallel bands which take a spiral form. (Illustrated in Fig. 3^ of Section I.) The rest of the cytoplasm remains as a vesicle which comes to lie, at the larger end of the sperm. - The cilia of Equisetum could be traced to 582 THE AMERICAN NATURALIST. [VOL. XXXVI I L definite granules in the band as it develops from the compact spherical blepharoplast. There appeared almost simultaneously with the foregoing con- tribution of Belajeff a paper by Shaw ('QBb) on Onoclea and Marsilia. Shaw investigated the cell divisions preceding the formation of the spermatids in Marsilia and discovered some very interesting conditions. The two blepharoplasts which are found in the mother cell of the spermatid are foreshadowed by smaller bodies which appear at the poles of the spindle in the two previous mitoses. The first of these structures was called a blepharoplastoid. The blepharoplastoid first appears besides the daughter nucleus after the third mitosis previous to the dif- ferentiation of the spermatids. There is therefore one for each nucleus of the grandmother cell of the spermatid. This bleph- aroplastoid divides but the halves remain close together and the pair passes to one side of the cell. With the next mitosis (the second previous to the differentiation of the spermatids) two new structures are formed at the poles of the spindle and from these the blepharoplasts arise. They accompany each daughter nucleus after this mitosis into the mother-cell of the spermatid. Then each divides and the two blepharoplasts pass to opposite sides of the nucleus which prepares for the final mitosis of the series. This division gives a daughter nucleus to each blepharoplast and the spermatid is thus organized. The later history of the spermatid as it changes into the sperm is identical with Belajeff 's results. Belajeff ('99) followed Shaw's account of Marsilia with a study of the same form and came to very different conclusions which have to do chiefly with his belief that the blepharoplast is a cen- trosome, a view that will presently be considered in connection with the opinions of Strasburger and others. Belajeff found centrosome like bodies (blepharoplastoids of Shaw) at the poles of spindles in various mitoses preceding the formation of the spermatids with their unquestioned blepharoplasts. He is not willing to concede that these centrosome like structures pass into the cytoplasm to disappear there as Shaw states for the blepharoplastoids. He also found the blepharoplasts at the poles of the spindles, which was not observed by Shaw, and holds that they have a part in spindle formation. Nos. 451-452.] STUDIES ON THE PLANT CELL. 583. We are now prepared to take a general survey of the proc- esses of spermatogenesis to harmonize as much as possible the conflicting opinions respecting the homologies of the blepharo- plast. Strasburger ( : oo, pp. 177-215) has critically reviewed the subject and his conclusions are of great interest. He em- phasizes the kinoplasmic character of the blepharoplast, whether it be a differentiated region of the plasma membrane (as he believes for the zoospores of Cladophora, CEdogonium, etc.) or a special development in the interior of the cytoplasm (pterido- phytes and gymnosperms). Strasburger thinks that all kino- plasmic structures, be they centrospheres, centrosomes or blepharoplasts, hold a very close physiological relation to the substance of the nucleolus and that their appearance and size is largely the result of nuclear activities. Accordingly the bleph- aroplast might occupy the position of a centrosome without being genetically related to that structure, and in fact centro- somes or centrospheres are to be considered more as products of the cells' activities than as self perpetuating permanent organs. There is abundant evidence that the last possibility is the fact in many forms both plants and animals. Since centrosomes are not fpund at other periods of the life history of gymnosperms and pteridophytes, Strasburger concludes that the blepharo- plasts cannot be genetically related (homologous) with such a structure. Ikeno and Hirase from their earliest writings have considered the blepharoplast to be a centrosome. Ikeno ('QSa) held that the blepharoplast corresponded with the middle piece of the animal spermatozoon. Hirase ('94 and '97) although noting for Ginko that the blepharoplasts did not divide and took no part in spin- dle formation nevertheless called them attractive spheres. The conclusions of Shaw ('98) and Belajeff ('99) for the same type (Marsilia) have just been summarized and present very different points of view. Belajeff believes that the blepharoplast of Mar- silia holds the same relation to the poles of the spindles as a centrosome. But Belajeff's conception of the centrosome ('99, p. 204) is that of a morphological and dynamic center which may or may not be easily demonstrated according to the amount of stainable substance present. From these discussions it is ^584 THE AMERICAN NATURALIST. [VOL. XXXVIII. evident that final judgment cannot be passed until certain ques- tions of fact are established by reinvestigations. Shaw and Belajeff cannot both be wholly correct in their observations and interpretations and much depends upon the exactness of future studies upon Marsilia, other pteridophytes, and in the bryo- phytes. The problems are also related to the processes of zoospore formation among the thallophytes. With respect to the bryophytes Ikeno ( : 03) has recently published an account of spermatogenesis in Marchantia poly- morpJia. He reports for the mitoses in the antheridium, prelim- inary to the differentiation of the sperm mother-cells, that a centrosome appears at the side of each nucleus and divides, the two daughter bodies passing to opposite sides of the nucleus and becoming the poles of the spindle. He gives evidence that the daughter centrosomes sometimes divide again when at the poles of the spindle in anaphase. The centrosome cannot be found at the side of the daughter nucleus after the mitosis is com- pleted but it appears when the nucleus is ready for the next divi- sion. Ikeno's explanation of the reappearance of the centrosome is unusual. He believes that the centrosome is formed within the interior of each nucleus as a deeply staining body among the linin threads. This body moves to the nuclear membrane and is thrust out into the cytoplasm through a protuberance from the nucleus. It then lies outside of the nucleus and becomes the functioning centrosome, dividing to form two centrosomes that separate to preside over the poles of the spindle. After the final mitoses in the spermatogeneous tissue the centrosomes remain to become the blepharoplasts of the sperms. Each blepharoplast passes to the plasma membrane of its sperm cell and develops two cilia. There is formed at this time another deeply staining body in the cytoplasm considered by Ikeno equiv- alent to a " Nebenkorper." The nucleus begins to elongate and the "Nebenkorper" takes a position between it and the blepharoplast and in this manner the much attenuated sperm is organized from the mother-cell. Ikeno considers the blepharoplast of Marchantia to be actu- ally a centrosome as shown by its behavior during mitosis. His account therefore in the main supports Belajeff 's interpretation Nos. 451452.] STUDIES ON THE PLANT CELL. 585 of the blepharoplastoids of Shaw which as just described are regarded by the latter author as centrosomes. Both Belajeff and Ikeno are inclined to use the term centrosome with a looseness that is unusual since the first accounts of this structure gave to it a place in the cell which is not strictly followed in these authors' descriptions of spermatogenesis. Ikeno's account of the intranuclear origin of the centrosome is extraordinary. Intranuclear centrosomes have been reported in several animal forms but they do not leave the nucleus in the manner described by Ikeno. On the whole the writer is more in sympathy with the views of Webber (:oi,pp. 70 to 81), Strasburger and Shaw than those of the other authors. Assuming that the observations upon the cycads and Ginko are correct, Webber is certainly justified in emphasizing the striking fact that the blepharoplasts are completely independent of the spindle in the body cell and that they are formed de novo at a distance from its nucleus. These are peculiarities which, if established generally through- out spermatogenesis in plants, will remove the processes entirely from the activities of centrosomes in certain thallophytes (e. g. Stypocaulon, Dictyota) and in many animal cells. It is certainly to be expected that a centrosome when present will always hold an intimate relation to spindle formation during mitosis. It need not be a permanent organ in cell genesis and an ever increasing number of investigations indicate that it frequently is not. Therefore many authors hold that the centrosome is rather the morphological expression of a dynamic center than a protoplas- mic structure with an individuality comparable to the organs of a cell. But these universal characteristics of centrosomes are apparently not present in the blepharoplasts of the gymnosperms nor, according to Shaw, in the pteridophytes (Marsilia). But then the observations of Belajeff and Ikeno are not in accord with those of Shaw and it is possible that studies in zoospore formation and gametogenesis among the thallophytes may pre- sent the subject in new lights. For as shown in our discussion of the zoospore it is not clear whether the blepharoplasts in those cells are always derived in the same manner. We have Strasburger's view that the 586 THE AMERICAN NATURALIST. [VOL. XXXVIII. structures are thickenings of the outer plasma membrane (hautschicht) and opposed to this Timberlake's account for Hydro- dictyon in which the blepharoplast is considered as a structure independent of the plasma membrane although lying in contact with it. It must be apparent that the results of Timberlake are in essential agreement with the events of spermatogenesis in the pteridophytes and gymnosperms while those of Stras- burger introduce new elements in giving to the plasma mem- brane the functions of forming a blepharoplast. The process of spore formation in the ascus must also be considered in this con- nection for in that sporangium a centrosphere associated with each nucleus develops numerous fibrillse that resemble so much a cluster of cilia as to suggest at once a blepharoplast-like struc- ture, but this centrosphere of course is an important factor in spindle formation during the mitoses in the ascus. Indeed we may well ask for further studies in spermatogenesis and zoospore formation before we can expect a solution of the problem of the blepharoplast. Comparisons have been made between the sperms of animals and plants, and some authors (e. g. Wilson : oo, p. 175, Belajeff '9/c) consider the two cells in essential agreement as to structure and development. However these views rest on the assumption that the blepharoplast is truly the homologue of a centrosome. It seems to be established that the locomotor apparatus of the animal spermatozoon is derived chiefly from one or more centro- somes, generally with the co-operation of archoplasm (idiozome, Nebenkern) present in some form near the nucleus. It is true that in plants the locomotor apparatus is derived from kinoplasm which as we pointed out in Sections I and II corresponds closely to the archoplasm of Boveri, but this is very far from implying that structures formed by the archoplasm and kinoplasm respec- tively need be homologous. Indeed both archoplasm and kino- plasm are distinguished by their physiological activities rather than by their morphological manifestations which are too various to allow of close genetic relationships. Therefore it seems far from established that spermatogenesis in plants is along the same lines as in animals, especially since the weight of evidence at present indicates that the blepharoplast is not a centrosome. Nos. 451-452.] STUDIES ON THE PLANT CELL. 587 There are numerous problems connected with the physiology of the sperm that bear directly upon its protoplasmic structure. Some of these will be treated in Section IV in connectiofTwith processes of fertilization. But at this time it is well to call attention to the intimate association that sometimes exists between the nucleus and blepharoplast. These structures come into actual contact in Cycas and Ginko through a process put forth from the nucleus. It should also be remembered that Timberlake and Dangeard found the blepharoplasts in the zoospores of Hydrodictyon and in the cells of Polytoma con- nected with the nucleus by one or two fibers. The nuclear beak that bears the aster in the ascus suggests a similar relationship. These conditions indicate that the activities of locomotion may depend vitally upon the nucleus. 3. The Egg. The subject of fertilization is reserved for the next section (Section IV) of this series and the present account will deal only with the structure of the unfertilized egg. As the sperm is derived from a motile gamete identical with the zoospore, so the egg has had a similar origin. We have traced the steps in this evolutionary process among the algae in a former paper (Popular Science Monthly, Feb. 1903, p. 300). The first indi- cation of a differentiation in the sex of primitive gametes is one of size. The male gametes tend to become smaller while the female contains a greatly increased amount of cytoplasm. One of the important factors determining this differentiation is the number of nuclear divisions which take place in the cells that produce respectively eggs or sperms. There are generally a great many more mitoses in antheridia than in oogonia and con- sequently a given amount of protoplasm must be very much divided to provide each nucleus with its quota of cytoplasm. The tendency of oogenesis on the contrary is to conserve the protoplasm for relatively few nuclei, provided for several eggs or for a single nucleus in a solitary egg, with the result that the egg cell is generally richly supplied with protoplasm. Such proc- esses result in large cells with a prominent chromatophore or 588 THE AMERICAN NATURALIST. [VOL. XXXVIII. numerous plastids and not infrequently a considerable amount of food material. The primitive female gametes were provided with cilia like the male, but with their increase in size came a sluggishness of movement which resulted in much shorter peri- ods of motility on the part of these sexual cells. There are some algae (Ectocarpus siliculosus, Cutleria, Aphanochaete) whose motile female gametes come to rest shortly after their escape from the oogonia and are fertilized as quiescent cells by the active sperms. These female gametes at the time of fertili- zation behave physiologically like eggs although their develop- ment shows a morphology identical with the sperm. When such female gametes dispense with cilia entirely they become eggs. The absence of cilia does away with very much of the com- plexity which we have just described for sperms. There is no trace of the blepharoplast in the egg and no indication of the activities associated with this structure, so conspicuous in sper- matogenesis. The large motile female gametes of such Algae as Bryopsis, Cutleria, Aphanochaete and certain species of Chlamy- domonas and Ectocarpus will probably show some interesting conditions when the details of their cell structure and develop- ment are known, for some of these types are likely to throw light on the relation which the blepharoplast bears to other structures in the cell. The eggs of all plants (Fungi excepted) are believed to be richly stocked with plastids in sharp contrast to the sperms which are entirely destitute of these structures in all groups above the algae. The plastids in the eggs of Algae contain the pigments characteristic of the respective groups giving these cells a very rich coloration and sometimes an elaborate internal structure since these plastids or the single chromatbphore generally main- tain a symmetrical relation to the nucleus. Leucoplasts (see Fig. \\a) have been found in the eggs of angiosperms (Schimper, '85) but detailed studies on the cytoplasm of such cells in spermatophytes, pteridophytes and bryophytes are greatly to be desired to determine the history of plastids dur- ing the development of these germ cells and at later periods after fertilization. Nos. 451-452.] STUDIES ON THE PLANT CELL. 589 The distribution of the plastids in the eggs of Algae may be so general that the entire cell is colored as in Fucus, Volvox and Sphaeroplea. Or, the plastids may be largely or wholly-wkh- drawn from some portion of the egg. It is usual for eggs retained within the parent cell (oogonium) to present a colorless area of protoplasm that becomes the point at which the sperm fuses with the egg. Such a hyaline region is called the recep- tive spot and is generally situated (see Fig. lib] at the side of the egg nearest the pore or opening in the oogonium through which the sperms enter. Excellent illustrations are presented among the Algae in Vaucheria (Oltmanns, '95), CEdogonium (Pringsheim, '58, Klebahn, '92) and Coleochaete (Pringsheim, '60, Oltmanns, '98). It has been suggested that the receptive spot is related to the clear ciliated end of the ancestral motile gamete and zoospore but the structures have not been critically compared to determine the precise character of their proto- plasmic structure and development. The receptive spot in some forms (Vaucheria, CEdogonium, Fig. 1 1 b") lies directly under the opening that is formed in the oogonium and its protoplasm is probably concerned with the fermentative action that destroys the wall at that point. The red Algae (Rhodophyceas) do not have eggs although in their sexual evolution they are at the level of heterogamy. The female gamete (carpogonium with its trichogyne) is a cell homol- ogous with an oogonium and its protoplasmic contents corre- spond to an egg, but the protoplast never withdraws from the cell wall to lie freely as a naked mass of protoplasm within the structure. But the general agreement of the carpogonium and trichogyne with the oogonium and its neck like extension in Coleochaete seems to determine without doubt the homologies of the former. There are very few eggs among the fungi that are strictly comparable to those of the Algae. Monoblepharis (Thaxter '95a) however unquestionably furnishes such an example. But the eggs of the Saprolegniales and Peronosporales are probably in the author's opinion not directly derived from those of Algae. They are either a peculiar form of sexual cell called the cceno- gamete (Davis :oo and 103) or closely related to this structure 590 THE AMERICAN NATURALIST. [VOL. XXXVIII. which will be given a separate treatment in this section. The ccenogamete is the homologue of a multinucleate gametangium but its evolutionary tendencies seem to be towards such a reduction in the number of nuclei that in the highest expression of its sexual differentiation the female cell contains a single nucleus and has the general form of an egg. But this process of sexual evolution is entirely independent of the well known lines of development in the Algae (Davis, Popular Science MontJdy, Feb. 1903). The female sexual cell of the Ascomycetes (called the ascogonium or archicarp) is probably in most forms the homologue of a gametangium. These subjects will be treated in our account of the ccenogamete. The egg in the archegonium of bryophytes and pteridophytes is generally reported to have a clearer region on the side nearest the neck and this is called th'e receptive spot. It is reported by Campbell in his investigations on Pilularia ('88), Iscetes ('91), Osmunda ('92a), Marsilia ('92b), and Marattia ('94), by Shaw in Onoclea ('98) by Thorn in Aspidium and Adiantum ('99) and by Lyon in Selaginella (:oi). The receptive spot is generally believed to be a portion of the egg differentiated to receive the sperm. It is an open question whether this area is morpholog- ically the homologue of the receptive spot in the eggs of algae and the clear area at the ciliated end of motile gametes and zoospores. The problem demands a detailed study of the finer protoplasmic structure to determine whether or not it is kino- plasmic in character. The nucleus is generally situated near the center of the egg and the portions of the cell farthest away from the neck of the archegonium contain coarsely granulate protoplasm which is evidently trophoplasmic, i. e., much of its substance is of the nature of food material and the products of metabolism. The leucoplasts would be supposed to lie in this region of the cell but we know nothing of their presence and behavior in the egg of bryophytes and pteridophytes. The eggs of gymnosperms generally speaking present sharp contrasts to those of pteridophytes. They are very large, prob- ably the largest uninucleate cells in the plant kingdom, and consequently very attractive for cell studies and some of the best work on the events of the maturation and fertilization of plant Nos. 451-452.] STUDIES ON THE PLANT CELL. 59 1 eggs has been done on this group (to be treated in Section IV). Passing over earlier investigations that described accurately the general structure of the egg of gymnosperms we shall consider the results of a number of comparatively recent papers that treat especially the pine, spruce (Picea), hemlock (Tsuga), fir (Abies), cycads, Ginko, Gnetum, Taxodium, etc. Oogenesjs and fertilization in the pine has been the subject of several extensive studies the chief being papers by Dixon ('94), Blackman ('98), Chamberlain ('99) and Ferguson (:oib). The protoplasm of the egg is at first vacuolate but later takes on a denser structure which becomes very puzzling because of numer- -/. s FIG. ii. The Egg. a, Daphne, showing leucoplasts ; b, oedogonium, showing receptive spot; c, pine, with numerous proteid vacuoles ; d, embryo sac of the lily, gamete nuclei fusing, remains of one Synergid (s) shown, (a, after Schim- per, '85; b, Klebahn, '92; c, Ferguson, :oi.) ous granular inclusions and masses of amorphous material which together with fibers present a very complex texture. The fibers are sometimes collected in fascicles and they may form a sort of weft at the periphery of the egg or radiate out from the nucleus which is generally surrounded by a kinoplasmic sheath. The complexity is greatly increased as the egg grows older by the development of remarkable structures called proteid vacuoles (See Fig. 1 1^) which have been especially described by Blackman and Ferguson. The number of proteid vacuoles is exceedingly variable in the egg but they sometimes fill three fourths of the structure. They are spaces in the cytoplasmic reticulum filled 592 THE AMERICAN NATURALIST. [VOL. XXXVI II. with granules and irregular masses of a proteid nature some of which stain like nucleoli. The proteid vacuolee were considered nuclei by earlier writers (Hofmeister and Goroschankin) and recently this view has been revived by Arnold (: oob) who describes the migration of large numbers of nuclei from the cells of the jacket surrounding the egg into that structure. These results have not been confirmed by Ferguson who agrees with the interpretation of other writers that the resemblance of the proteid vacuolesto nuclei is superficial. Miss Ferguson believes that the material of the proteid vacuoles is derived in part from the nucleoli in the cells of the jacket and from those in the egg. A vacuole is reported (Ferguson) at the end of the egg nearest the neck of the archegonium and this is regarded as a sort of receptive spot since the pollen tube discharges its contents into this cavity. The egg nucleus is very large and its contents are not arranged with the regularity generally present in resting nuclei. There are numerous bodies which Chamberlain believes to be chromatic in composition that look very much like nucleoli and have been so designated by that writer. But there is gener- ally one large unquestioned nucleolus and besides this many smaller nucleoli are reported by Ferguson as held in the linin reticulum. Then portions of the linin frequently take irregular forms and stain heavily. There is also present besides the linin, chromatin and nucleoli much granular material (metaplasm), especially in the nuclei of younger eggs, which probably holds some relation to the chromatin although it may readily be dis- tinguished at certain times from that substance. Recent accounts of the spruce and fir, by Miyake (: O3a and : O3b) describe conditions very much as in the pine. The egg of the spruce (Picea) is apparently not so fibrous in structure but proteid vacuoles give it a coarse granular structure. He finds no evidence in support of Arnoldi's (: oob) peculiar views that the proteid vacuoles are derived from nuclei that have passed into the egg from cells of the sheath. They are simply masses of nutritive material. There is some doubt whether the vacuoles present at the end of the egg really represent a differentiated receptive spot. The egg of the fir (Abies) conforms in all essentials to the structure in the pine and spruce. There are numerous proteid vacuoles. Nos. 451-452.] STUDIES ON THE PLANT CELL. 593 It is probable that the eggs of other conifers will be found to present much the same protoplasmic structure and activities as those of the pine. Thus Murrill (: oo) describes for thehejmlock spruce (Tsuga) a vacuolar receptive spot and figures masses of food material very much like the proteid vacuoles. The general features of the egg of Cephalotaxus (Arnoldi, : ooa), Thuja (Land, : 02), Podocarpus (Coker, : 02), Taxodium (Coker, '03) have been recently described and those of Abies, Larix and Taxus are familiar from older writers but the pine remains as the type of conifer in which the events of oogenesis are best known as regards the details of protoplasmic activities. Besides the pine we have had some very complete investiga- tions on cycads and Ginko (Hirase, '98, Ikeno, 'gSb and :oi, Webber, :oi). In some respects these types and especially the cycads seem to be the most favorable of all the gymnosperms for the study of gametes and the processes of fertilization (to be described in Section IV). The cytoplasm of the egg is com- paratively homogeneous in structure so that the cell is relieved from the complicated fibrous structure and proteid vacuoles present in the pine. Ikeno ('98b) finds that the egg of Cycas develops a crater like depression just before and at the time of the fusion of the sperm thus presenting a rather highly special- ized receptive spot. We know almost nothing of the detailed structure of the egg in the Gnetales. Ephedra (Strasburger, '72) develops arche- gonia much like those of other gymnosperms and we should not expect their eggs to be materially different even in details. But the conditions in Tumboa (Welwitschia) are peculiar and approach more closely those of angiosperms where the egg nucleus is scarcely differentiated from neighboring nuclei lying freely in the protoplasm at one end of the embryo sac. The eggs of Tumboa (Strasburger, '72) are merely cells of the prothallus that push out small projections to meet the pollen tubes. Gnetum presents a further simplification or reduction since the female nuclei lie freely in the protoplasm at one end of the embryo sac. In Gnetum gnemon the lower half of the embryo sac is filled with a tissue (Lotsy '99) but in several other species studied by Karsten ('92, '93) no cell walls are found in the entire sac until after fertilization. 594 THE AMERICAN NATURALIST. [VOL. XXXVI II. The angiosperms present no especial advance over Gnetum in the organization of the egg except that this structure is generally reduced to a single female nucleus and the cytoplasm immedi- ately around it (see Fig. u d). This egg nucleus flanked by two companions (synergids) and the accompanying protoplasm compose the egg apparatus whose morphology is still a matter of dispute. It is possible that the synergids may stand for portions of a reduced archegonium, but the two nuclei bear such close relations to the egg and polar nucleus that it seems very probable that they are homologous with these structures which have clearly defined sexual potentialities. In spite of the numerous studies on embryo sacs in various groups of angio- sperms we do not yet know precisely how the cytoplasm becomes gathered around the egg nucleus and the synergids. The spindles that are formed between these nuclei in some types (e. g., Lilium) have been supposed to lay down walls by means of cell plates. But there are other forms in which the proto- plasm seems to separate along planes of vacuoles without rela- tion to spindle fibers. ( To be continued.) VOL. XXXVIII, No. 454. OCTOBER, 1904 THE AMERICAN NATURALIST A MONTHLY JOURNAL DEVOTED TO THE NATURAL SCIENCES IN THEIR WIDEST SENSE CONTENTS Page I. The Anatomy of the Coniferales (concluded) PKOF. D. P. PENHAILOW 691 II. Studieiof the Plant Cell.- IV DE. B. M. DAVIS 728 III- The Affinities of the Ophioglossacese and Marsiliacese PROF- D. H CAMPBELL 761 BOSTON, U. S. A. GINN & COMPANY, PUBLISHERS 9 BEACON STREET New York Chicago London, VT. C, 70 Fifth Ayenue 371-311 Wabaah Avenue 9 St. Martin'* Street Snttrtd / tht Pii-Offtft, Btttnt, Mint., at SetnJ-Cl*lt Mail tf*4trr. The American Naturalist. ASSOCIATE EDITORS: J. A. ALLEN, PH.D., American Museum of Natural History, New York. E. A. ANDREWS, ?n.D., Johns Hopkins University, Baltimore. WILLIAM S. BAYLEY, FK.D., Colby University, Waiemilii. DOUGLAS H. CAMPBELL, PH.D., Stanford University. J. H. COMSTOCK, S.B., Cornell University, Ithaca. WILLIAM M. DAVIS, M.E., Harvard University, Cambridge. ALES HRDLICKA, M.D., U.S. National Museum, Washington, D. S. JORDAN, LL.D., Stanford University. CHARLES A. KOFOID, PH.D., University of California, Berkelty, J. G. NEEDHAM, PH.D., Lake Forest University. ARNOLD E. ORTMANN, PH.D., Carnegie Museum, Pittsburg. D. P. PENHALLOW,D.Sc.,F.R.M.S., Me Gill University, Montreal. H. M. RICHARDS, S.D., Columbia University, New York. W. E. RITTER, PH.D., University of California, Berkeley. ISRAEL C. RUSSELL, LL.D., University of Michigan, Ann Arbor. ERWIN F. SMITH, S.D., U. S. Department of Agriculture, Washington. LEONHARD STEJNEGER, LL.D., Smithsonian Institution, Washington. W. TRELEASE, S.D., Missouri Botanical Garden, St. Louis. HENRY B. WARD, PH.D., University of Nebraska, Lincoln. WILLIAM M. WHEELER, PH.D., American Museum of Natural History, New York. THE AMERICAN NATURALIST is an illustrated monthly magazine of Natural History, and will aim to present to its readers the leading facts and discoveries in Anthropology, General Biology, Zoology, Botany, Paleontology, Geology and Physical Geography, and Miner- alogy and Petrography. The contents each month will consist of leading original articles containing accounts and discussions of new discoveries, reports of scientific expeditions, biographical notices of distinguished naturalists, or critical summaries of progress in some line ; and in addition to these there will be briefer articles on various points of interest, editorial comments on scientific questions of the day, critical reviews of recent literature, and a quarterly record of gifts, appointments, retirements, and deaths. All naturalists who have anything interesting to say are invited to send in their contributions, but the editors will endeavor to select for publication only that which is of truly scientific value and at the same time written so as to be intelligible, instructive, and interesting to the general scientific reader. All manuscripts, books for review, exchanges, etc., should be sent to THE AMERICAN NATURALIST, Cambridge, Mass. All business communications should be sent direct to the publishers. Annual subscription, $4.00, net, in advance. Single copies, 86 oenta. Foreign subscription, $4.60. GINN & COMPANY, PUBLISHERS. STUDIES ON THE PLANT CELL. IV. BRADLEY MOORE DAVIS. SECTION III. HIGHLY SPECIALIZED PLANT CELLS AND THEIR PECULIARITIES (Continued}. 4. The Spore Mother-cell. THE spore mother-cell and its homologues the pollen mother- cell and certain embryo-sacs have furnished some of the most interesting subjects for cell studies in the plant kingdom. Sporogenesis in all plants above the thallophytes seems to be a period when nuclear structures are especially clearly differ- entiated and when the mechanism of mitosis reaches the highest degree of complexity. These intricate conditions are only equalled by processes in the development of the female game- tophyte of some angiosperms, and during endosperm formation, also in the events of spermatogenesis and with the segmentation of the egg nucleus of certain gymnosperms. Sporogenesis is one of the critical periods in the life history of a higher plant since it is the time when the asexual genera- tion (sporophyte) passes over to the sexual (gametophyte). This provides certain important features such as the reduction phe- nomena concerned with chromosomes and greatly adds to the interest in these cells. These matters will receive special atten- tion in Section V, but they must be borne in mind to appreciate fully the significance of many events of spore formation. The general history of the spore mother-cell may be described as follows : It is the product of the last mitosis in the repro- ductive tissue called the archesporium. This mitosis always presents the number of chromosomes characteristic of the sporophyte which is double the number found in the game- tophyte. Therefore the nucleus that passes into the spore mother-cell has the sporophyte number of chromosomes. Two 725 726 THE AMERICAN NATURALIST. [VoL. XXXVIII. mitoses occur successively in the spore mother -cell in all forms. The first mitosis presents half the number of chromosomes found in the last nuclear division in the archesporium and is consequently the reduced or gametophyte number. The reduc- tion of the chromosomes then takes place during the period of rest between the last mitosis in the archesporium and the first in the spore mother-cell. There are two mitoses in the spore mother-cell. In some forms these are exactly alike and present essentially the same characters as the usual typical mitoses of plants. But among the spermatophytes there are likely to be peculiarities in the arrangement and distribution of the chromo- somes. In consequence the first mitosis may be heterotypic and the second homotypic in contrast to the normal typical con- ditions. The description and explanation of these characters will be reserved for the groups that illustrate them the best. They have nothing to do with qualitative reduction phenomena as was formerly supposed. There is sometimes a well defined period of rest after the first mitosis with the formation of a wall between the two daughter nuclei, but frequently the second mitosis follows immediately after the first so that the spore mother-cell comes to contain four daughter nuclei. Cell walls may then be formed between these nuclei simultaneously so that the resultant spores are dis- posed in a radially symmetrical arrangement that is termed tripartite. These cell divisions are almost universally present in the spore mother-cell, the only exceptions being certain sperma- tophytes whose megaspore mother-cells develop directly into embryo sacs, the two mitoses (heterotypic and homotypic) being included within these structures and forming a part of the game- tophyte history. Why the number of spores should generally be four is unexplained. There does not seem to be .any physio- logical significance in the number or other reasons why it should not be more or less. Indeed it is somewhat variable in the spermatophytes for microspore or pollen mother-cells form two and three pollen grains in certain types and five, six and seven have been found in others, while much larger numbers have been occasionally reported. In no case is the microspore mother-cell known to develop directly into a pollen grain, al- No. 454-] STUDIES ON THE PLANT CELL. 727 though the megaspore mother-cell regularly becomes an embryo sac in some forms (e.g., Lilium). But an increasing number of observations indicate that the megaspore mother-cell generally develops two, three or four potential megaspores although nor- mally only one of these becomes an embryo sac. The interest in the, protoplasmic activities of sporogenesis lie chiefly in the elaborate methods of spindle formation and mech- anism of mitosis, in the organization and distribution of the chromosomes, in the functions and activities of the nucleolus, and in the organization of the cell plate and development of the cell wall. There is a very extensive literature on the spore mother-cell some of which, however, merely treats the broad features noted in studies of a general morphological character on the development of sporophylls or floral structures. We shall only attempt to consider the most important contributions, and for convenience will begin our treatment with the Hepaticas and conclude with the spermatophytes where the conditions are the most complex. The Hepaticae or liverworts furnish some remarkable spore mother-cells, and are now the subject of considerable interest and some discussion. They were first brought conspicuously to the attention of botanists by a paper of Farmer ('94) on Pal- lavicinia decipiens. Farmer described a remarkable series of events in this type. The nucleus of the spore mother-cell became surrounded before division by dense protoplasm that extended into the four lobes of the cell in the form of a four- rayed star which he called a " quadripolar spindle." After its development four chromatic droplets appeared in th,e nucleus to indicate its approaching division. These chromatic droplets became four chromosomes which by division were doubled in number. The eight rod shaped chromosomes moved in pairs towards the four lobes of the spore mother-cell. There was a further division of each chromosome, making sixteen in all, and the four groups of four each passed simultaneously to the poles of the ''quadripolar spindle" which persisted to the end. It should be noted that the striking peculiarities of Farmer's account lie in the division of the four primary chromosomes into sixteen, and in their simultaneous distribution through a " quadripolar 728 THE AMERICAN NATURALIST. [VOL. XXXVIII. spindle " to form at once four daughter nuclei. These events are unparalleled, as far as the writer is aware, in the plant or animal kingdom, and consequently the account deserves especial attention. A four-rayed figure around the nucleus is not surpris- ing because the spore mother-cell of the Jungermanniales is four lobed, and its centrally placed nucleus lies in a restricted area. But the simultaneous distribution of quadrupled chromosomes to form four daughter nuclei is a process whose establishment would be of fundamental significance. Farmer also described a centrosome at each pole of the " quadripolar spindle." Farmer ('95, b, and c ) followed his paper on Pallavicinia with studies on other liverworts. He reported the " quadripolar spindle" in the early stages of mitosis in several of the Junger- manniales, but did not find the quadrupling and simultaneous distribution of the chromosomes as in Pallavicinia. The " quad- ripolar spindle " when present was a temporary structure replaced later by the bipolar spindles of two successive mitoses with a longer or shorter interval between. Farmer considers the " quadripolar spindle " of these forms as transitional between that of Pallavicinia and the normal bipolar spindle. The Ric- ciales, Marchantiales and Anthocerotales present two successive mitoses after the usual manner in the spore mother-cell. The writer has described the events of sporogenesis in Pellia (one of the Jungermanniales) in a paper covering the nuclear activities at several periods in its life history (Davis, :oi), and confirmed much of Farmer's account of the mitoses in this spore mother-cell. These are two in number and successive, with a very well defined resting period between the first and the second. There is a four-rayed figure present during the prophase of the first mitosis, and this seems to correspond to Farmer's " quad- ripolar spindle." The nucleus lying in the center of the four lobed spore mother-cell becomes invested by a kinoplasmic sheath which develops -a fibrillar structure. Many of these fibrillae extend into the lobes of the spore mother-cell because the nucleus is confined to a narrow space in the constricted cen- tral region of the cell and the lobes offer the only possible relief for the crbwded conditions. However,^the four-rayed structure is not present when the chromosomes are ready for distribution, No. 454-] STUDIES ON THE PLANT CELL. 729 but there is found instead one large, broad poled spindle. (See Fig. 5 e.) A cell wall is formed between the two daughter nuclei (Fig. 8 d] which divide again after a very short period of rest, the two spindles lying at right angles to one another. The poles of the spindles are rather blunt, and there are no centro- somes or centrospheres in either mitosis. The four-rayed struc- ture of prophase must be regarded as preliminary to spindle formation because the chromosomes are not ready for distribu- tion, and when that period arrives the structure has been re- placed by the true spindle of the first mitosis. These facts led me to question Farmer's account of mitotic phenomena in Palla- vicinia and his conception of the " quadri polar spindle," and I suggested that this structure might prove to be a phenomenon of prophase, a view to which Farmer (:oi) has taken exception in a criticism of my results. Recent investigations of Moore (: 03) on Pallavicinia are flatly contradictory to the conclusions of Farmer for Pallavicinia decipiens and support my suggestions. Moore finds that there are two mitoses in the spore mother-cell of Pallavicinia lyellii, the second (Fig. 12 c, d} following immediately upon the first {Fig. 12 b], each with bipolar spindles and without centrosomes. The chromosomes, eight in number, appear in the usual way with each mitosis (Fig. 12 c, d). There is no " quadripolar spindle" in Farmer's sense, no quadrupling and simultaneous distribution of the chromosomes. The prophases preceding the first mitosis present a tetrahedral form as is shown in Fig. 12 a. This is accentuated by the fibrillae which gather at the points to make a four-rayed structure extending into the lobes of the spore mother-cell. This condition is identical with similar stages in Pellia and in other leafy liverworts, and is a feature to be expected from the fact that the spindle fibers develop chiefly or wholly externally to the nuclear membrane in a rather crowded region of the cell. The nucleus at this time is unquestionably in prophase as shown by the undifferentiated chromosomes and because this stage passes immediately into a bipolar spindle of the normal type (Fig. 12 b). It seems very probable that Farmer was mistaken in his conclusions for Pallavicinia decipi- ens, and that the mitoses in the spore mother-cell of this form 730 THE AMERICAN NATURALIST. [VoL. XXXVIII. are not different in any essentials from those of other plants. FIG. 12. Spore mother-cells of Hepatic.-e. a, b, c, d, Pallavicinia lyellii. a, Prophase ; the fibrillae gathered on four sides of the nucleus which has a tetrahedral form pointing into the four lobes of the spore mother-cell; the nuclear membrane has not yet broken down; similar stages of prophase were probably considered by Farmer as quadripolar spindles. b, metaphase of the first mitosis ; the spindle in all respects a normal bipolar structure without centrospheres. c, Metaphase of the second mitosis; one spindle shown in side view, the other, almost perpendicular to the first, presents the eight chromosomes at the nuclear plate, d, anaphase of the second mitosis: one spindle viewed from the side, the other from one end shows the group of eight grand-daughter chromosomes. e,f, g, antho- ceros laevis, h, i, a larger species from Italy, e, prophase; one pole of spindle developed. f, just after metaphase of the first mitosis; eight chromosomes; blunt poled spindle with- out centrospheres. g t metaphase of second mitosis; very small spindle, h, cell p.ate forming in the spindle between two nuclei, z", two nuclei at the side of their respective chromatophores and the cell plate between, after the second mitosis ; a third chromato- phore shown with strands of protoplasm connecting it with other regions of the cell, (a, b, c, d, after Moore, :os ; h, i, after Van Hook, : oo.) The " quadripolar spindle " proves to be nothing more than a condition of prophase. Besides Pellia and Pallavicinia, which are the most thoroughly studied of the lower liverworts, we know the processes of sporo- No. 454-] STUDIES ON THE PLANT CELL. 731 genesis in the highest type, Anthoceros (Davis, '99). This form is exceedingly attractive for such investigations because the spore mother-cells may be found in all conditions upon the same sporophyte. However, the small size of the nuclei and spindles is a disadvantage. Just previous to the first mitosis the nucleus becomes surrounded by a mesh of delicate fibrillae (kinoplasmic). Later the nucleus takes an angular form, and the fibrillse are found conspicuously at the prominent poles (Fig. 12 c). The nuclear membrane breaks down and the fibers become arranged to form a bipolar spindle (Fig. 12 /) without centrosomes or centrospheres. There is a short period of rest after the first mitosis, but no wall is formed between the two daughter nuclei. The small spindles of the second mitosis (Fig. 12 g) are like- wise bipolar. They lie at right angles to one another and the cell plates that are laid down determine, in part, the position of the walls that are formed between the four granddaughter nuclei and which divide the spore in a tripartite manner. These cell plates are very small (Fig. 12 Ji and i ), but they have been observed in a favorable species of Anthoceros by Van Hook ( : oo). It is not clear how these plates become extended to the wall of the spore mother-cell unless (as suggested in Sec. II) their edges make use of planes of vacuoles when the protoplasm separates to develop the cleft between the four daughter cells. The poles of the spindles in Anthoceros are flattened and entirely free from structures that might be considered centrosomes. Other interesting events of sporogenesis in Anthoceros are the division of the chromatophores and the nuclear condition termed synapsis. The young spore mother-cell contains a single large chromatophore. This increases greatly in size and becomes filled with starch grains. The chromatophore divides succes- sively into two and then four portions which arrange themselves symmetrically in the cell with the nucleus in the center. The mitoses then follow and the four daughter nuclei are distributed, one for each chromatophore in the cell. This provision of four chromatophores long before the mitoses in the cell seems very remarkable (Davis, '99, p. 94 and 95). Synapsis is a condition very common in the nucleus of spore mother-cells before divi- sion. The chromatic material becomes gathered into a compact 732 THE AMERICAN NATURALIST. [VOL. XXXVIII. mass besides the nucleolus. The significance of synapsis is not clear, but the subject will be discussed in Section VI. How- ever, there is good evidence from Anthoceros that the phenome- non is a normal event and not an artefact, because synapsis is always found at a certain period of sporogenesis, and nuclei in neighboring spore mother-cells a little older or younger present their chromatic material with the usual arrangement (Davis, '99, p. 96 and 97). To summarize the conditions in the spore mother-cells of the Hepaticae, all conclusions, in the author's opinion, indicate : (i) That the spindles develop from a surrounding weft of fibrillae without the assistance of centrosomes. (2) That the mitoses are always two in number and successive with the same number of chromosomes for each division. (3) That the cell walls may be formed successively as in Pellia and some other of the Jun- germanniales or simultaneously, to give tetrahedral spores, as in Anthoceros, types of the Marchantiales and Ricciales, Pallavi- cinia and some companion forms in the Jungermanniales. It will be interesting to note the essential agreement in these matters between the Hepaticse and the higher plants. Nothing is known of the nuclear activities during sporogenesis in the other great division of the bryophytes, the mosses (Musci). The spore mother-cells in this group are always small and unattractive for cell studies but the Sphagnales appear to be rather the most promising for such investigations, which are greatly to be desired. The pteridophytes have furnished some important contribu- tions to our knowledge of the spore mother-cell. There is first the paper of Osterhout ('97) on spindle formation in Equisetum, which was one of a group of three contributions (Mottier, '97, Juel, '97) that did much to dispose of a then prevalent belief that the development of the spindle in higher plants was con- trolled by centrosomes. This investigation was followed by a study of Smith (: oo) on spindle formation in Osmunda. Calkins ('97) and W. C. Stevens ('98^) considered especially the formation and reduction of chromosomes in several of the ferns, and arrived at contradictory conclusions. Strasburger (: oo, p. 76 to 79) has reviewed these results in relation to studies of his own on Osmunda. No. 454.] STUDIES ON THE PLANT CELL. 733 Osterhout's ('97) account of spindle formation in Equisetum is noteworthy. He found that the nucleus of the spore mother- cell became surrounded by a web of delicate fibrillse, ^vhich, extending radially into the surrounding cytoplasm (Fig. I3#), were later (Fig. 13^) gathered into numerous pointed bundles or cones. After the dissolution of the nuclear membrane these FIG. 13. Spore mother-cells of Pteridophytes. a, 6, c, Equisetum limosum. a, prophase of first mitosis; the radially disposed fibrilla? are gathering together into cones, b, prophase, older than; the nuclear membrane has broken down and the fibrillae have entered the nuclear cavity; the cones lie in two groups opposite one another, c, just before meta- phase ; the fibrillar cones are nearer together and the chromosomes have gathered to form the nuclear plate, d, e,f, g, Osmunda regalis. d, very early prophase of the first mito- sis; nucleus in the spirem stage surrounded by a granular and fibrillar zone of kinoplasm. e, prophase, somewhat older than d\ fibrillar kinoplasm showing polarity, f, still older; chromosomes formed; one pole of spindle developed, g, metaphase; a tri-polar spindle. (a, b, c, after Osterhout, '97; d, e,f,g, Smith, :oo.) cones arranged themselves side by side in two sets to form the spindle of metaphase (Fig. 13^). The spindle is then from the outset multipolar, and even though some of the cones unite when they become grouped around a common axis, nevertheless the poles of the spindle at metaphase show their composite nature in the absence of a common focal point for the fibrillas. There are no centrosomes at the poles and no reason for their presence at any stage in the process of spindle formation. Smith's (: oo) study of Osmunda presents an important con- firmation of Osterhout's conclusions that the spindle in pterido- 734 THE AMERICAN NATURALIST. [VOL. XXXVIII. phytes developed without centrosomes, while illustrating a proc- ess of spindle formation along somewhat different lines. Smith distinguished a zone of kinoplasm around the nucleus previous to spindle formation. This zone became granular, and then the granules arranged themselves in rows to form fibrillae (Fig. 13 d\ which, however, did not extend into the cytoplasm radially, but lay generally parallel to one another, so that the spindle appeared bipolar from the beginning (Fig. 13 2, shortly after metaphase of the first mitosis (heterotypic). 3, meta- phase of the second mitosis (homotypic). c, d, Scilla Sibirica. ci, megaspore mother- cell. C2, after the first mitosis, rj, after the second mitosis, the lower cell of the pair to become the embryo sac. C4, after the second mitosis, the upper cell of the pair to become the embryo sac. di, anaphase of the first mitosis (heterotypic). dz, anaphase of the sec- ond mitosis (homotypic). e, Liliiim martagon ; portion of embryo sac mother-sell, nucleus surrounded by a felt of fibrillae. /, Lilium candidum\ embryo sac mother-cell, nucleus surrounded by radiating fibrillffi. g, h, i, Lilium martagon. g, late prophase of first mitosis in embryo sac mother-cell, a multipolar spindle, h, anaphase of first mitosis (heterotypic). i, anaphase of second mitosis (homotypic). (a, b, c, d, after Schniewind- Thies : 01 ; e,f, g, h, i, Mottier '97.) spore or pollen mother-cells. Schniewind-Thies (:oi) figures very completely the mitoses in Galtonia. The first mitosis in the megaspore mother-cell (Fig. 150) is heterotypic because the No. 454-3 STUDIES ON THE PLANT CELL. 743 chromosomes (Fig. i$b, i, 2) show clearly the V-shaped forms characteristic of this division. The second mitosis (Fig. 15 b, 3) is homotypic. The lowest cell of the group of four (Fig". 15 a, 3) becomes the embryo-sac and the mitoses that take place within it as the female gametophyte develops are all typical. This account illustrates a simple history in megaspore mother- cell development and is considered the first of three types in a classification proposed by Schniewind-Thies (:oi). The second type of development is one in which two mega- spores are generally developed from a mother-cell and one of these becomes the functional embryo-sac. Schniewind-Thies presents an excellent illustration of this type in Scilla. The first mitosis in the megaspore mother-cell (Fig. 1 5 c] is hetero- typic (Fig. 15^, i)and results in two cells (Fig. I5 c I Poirault and Raciborski, '95 ; Sapin-Trouffy, '96 ; Maire, : oo a, b, c, : 02 ; Holden and Harper, : 03) that the aecidiospores and the mycelium derived from them and pre- ceding the development of the uredospores and teleutospores contain pairs of nuclei which divide in such a manner (conjugate division) that the nuclei of the pair are derived through two unbroken lines of succession for a long vegetative period and always maintain complete independence of one another. Every young teleutospore and basidium contains such a pair of nuclei which shortly fuse so that the mature structure is uninucleate. Dangeard and Sapin-Trouffy have from the first regarded the nuclear fusion within the teleutospore, whether of rust or smut, as a sexual act and the ripe teleutospore a fertilized egg, regard- less of the fact that its morphology was not that of any known sexual organs. Dangeard ('94-'95 c ; : oo) likewise considered the nuclear fusions in the basidium as sexual. Raciborski ('96) suggested that the series of conjugate mitoses leading to the nuclear fusions in the teleutospore represented a vegetative phase intercalated between the beginning of a sexual act and its finish in the teleutospore. His explanation, in the light of the recent paper of Blackman (:O4a), was nearest the truth. Maire (: 02) presents the most extensive account of the nuclear struc- ture in the higher Basidiomycetes previous to and during the formation of the basidia. He held that the fusion of the paired nuclei (synkaryon) in the basidium was not the whole act of fertilization which must begin with the formation of the paired nuclei. Maire (: 02, p. 189) gave some suggestions as to how and where the paired nuclei arose but neither he nor any of the authors mentioned above knew clearly their origin. Blackman (: O4a) has made the most important contribution to the subject of fertilization and alternation of generation in the Uredinales, showing clearly that the paired nuclei appear in the life history of Phragmidium violaceum and Gymnosporangium No. 460.] STUDIES ON PLANT CELL. V. 247 clavariceforme just before the development of the aecidium. They arise in Phragmidium by the migration of a nucleus from an adjacent cell into an element (the fertile cell) which represents a female sexual organ. The morphology of the female organ is not clear but there are suggestions of a structure similar to the procarps of the Rhodophyceae and Laboulbeniales. The fertile cell, after receiving its second nucleus, develops a chain of aecidiospores, the two nuclei becoming so closely associated in the paired condition that they divide simultaneously (conjugate mitosis) from now on until the teleutospores are formed. Thus the cells of all mycelium beginning with the secidiospore con- tain paired nuclei up to the development of the teleutospores, including of course the uredospores when present. This period of the life history may be considered as representing a sporophyte generation, especially since the total of chromatin in the pair of nuclei is double the amount when the nuclei are solitary. The sporophyte phase ends with the fusion of the pair of nuclei in each cell of the teleutospores and in the reduction phenomena that take place with the germination of the teleutospore, includ- ing the formation of the promycelium. The sporidia developed by the promycelium are uninucleate and the cells of the mycelium derived from them are uninucleate up to the production of the aecidium. This constitutes the gametophyte phase of the life history. The spermogonia by their morphology seem to be male organs, now functionless. In such of the Uredinales as have no aecidium, as also in the higher Basidiomycetes and the Ustilaginales, it is probable that both sexual organs are suppressed since no trace of such struc- tures has been found. However, we may expect to discover periods in all of these forms when paired nuclei come into the life history and after a series of conjugate divisions fuse in the teleutospore or basidium. Such pairs of nuclei, as stated before, are known in the Ustilaginales (Dangeard, '93) and in a number of forms of the Uredinales and the nuclear fusions have been followed in the teleutospore. H olden and Harper (: 03) have given an especially clear account of the paired nuclei in the mycelium and uredospores of Coleosporium together with their fusion in the teleutospore. Maire (: 02) describes the paired 248 THE AMERICAN NATURALIST. [VoL. XXXIX. t nuclei (synkaryons) and their fusion in the basidium in a large number of Hymenomycetes and Gasteromycetes. Evidence is thus accumulating that the cells in the mycelium of higher Basidiomycetes (Hymenomycetes and Gasteromycetes) are binucleate for extended periods previous to the formation of basidia where nuclear fusions always take place. Binucleate cells in the higher Basidiomycetes were first reported by Maire (: ooa ; : oob), in the tissue preliminary to spore formation. He also con- firmed Dangeard ('g^-'g^c) in his view that only two nuclei unite in the basidium contrary to accounts of Rosen ('93) and Wager ('99, p. 586) which described a succession of fusions involving sometimes as many as six or eight nuclei. Harper (:O2) has given for Hypochnus one of the most complete accounts of the behavior of paired nuclei previous to and during the development of the basidium. The cells of the mycelium of this simple Hymenomycete were found to be binucleate as far back as they were studied which included all of the conspicuous vegetative structure. Only a single pair of nuclei enters the basidium and fuses. Harper's results are then in agreement with the extended observations of Maire (:O2) as are also the detailed studies of Ruhland (:oi) on a number of forms and Bambeke (:O3). Taken together they seem to show clearly that the mycelium, for long periods preliminary to the formation of basidia, contains paired nuclei and that the basidia receive each a single pair, which nuclei fuse. There is thus an exact correspondence between the life histories of the Ustilaginales, Uredinales, and higher Basidiomycetes with respect to the period of paired nuclei and their fusion in the teleutospore or basidium. Dangeard called the fusion in the basidium a sexual act and the structure an oospore regardless of the morphological difficulties of such a conception. Maire (: 02, p. 202) states that the origin of the paired nuclei is the only phenomenon strictly comparable to fertilization and Blackman's studies support this view. Ruh- land (:oi) regards the conditions as a deviation from the normal type of sexuality calling it " intracellular karyogamy." The origin of the paired nuclei is not known for any higher Basidi- omycete and the discovery of this period and determination of the events leading to the change from uninucleate mycelium to No. 460.] STUDIES ON PLANT CELL. V. 249 binucleate is one of the most interesting problems in this field 'of botany. This is the point where we should expect to find the remains of sexual organs, if any are present in _th higher Basidiomycetes, but it is not likely that they will be found. It seems more probable that the mycelium with the paired nuclei (perhaps sporophytic in character) arises apogamously with a complete suppression of the sexual organs in agreement with such of the Uredinales as have no aecidium and the Ustilaginales. Blackman's explanation of the history of the paired nuclei in Phragmidium is full of interest. As stated before, he regards the fertile cell which develops a chain of aecidiospores, "as a female reproductive cell which undergoes a process of fertiliza- tion by a union with an adjacent cell of the mycelium and its reception therefrom of a nucleus. The mycelium then which arises with the aecidiospore is sporophytic in character and so remains until the fusion of the pairs of nuclei in the teleuto- spores. The male organs of the rusts are the spermogonia and the male gametes the spermatia which are of course now func- tionless so that the "process of fertilization" is through the introduction into the female cell of a nucleus which is not phy- logenetically a male sexual element. Blackman's (104 a, pp. 349-353; :O4b) conception of the process as an act of ferti- lization involves some principles which will be briefly outlined. Blackman believes for Phragmidium " that the primitive normal process of fertilization by means of spermatia has been replaced by fertilization of the female cell through the nucleus of an ordinary vegetative cell " and regards the process as very similar to the phenomenon reported in the apogamous development of ferns by Farmer, Moore, and Digby (: 03), which will be consid- ered presently. Blackman points out that normal processes of fertilization such as we have included under the head of " sexual cell unions and nuclear fusions " do not involve in many forms (probably all types with a sporophyte generation) an immediate union of the chromatin of the sexual nuclei which is known to remain distinct during the first cleavage mitosis in a number of types (e, g., Pinus and some other gymnosperms). So there is nothing in the delayed fusion of the paired nuclei up to the teleutospore that is seriously against his explanation of the "fer- 250 THE AMERICAN NATURALIST. [VOL. XXXIX. tilization " of the female cell of the Uredinales. Indeed, we may expect to find that the actual fusion of paternal and mater- nal chromatin does not take place in the higher plants until the end of the sporophyte generation in the spore mother cell, as zoologists have concluded that such union occurs just previous to gametogenesis in animals. But is Blackman justified in regard- ing the phenomenon substituted for the activities of ancestral sexual organs in Phragmidium, now functionless, as a sexual act and is it desirable to apply the term fertilization to the phe- nomenon ? Blackman (:O4b, p. 153) speaks of the introduction of a nucleus into the fertile cell of the Uredinales and the phenome- non in the apogamous development of the fern after the account of Farmer, Moore, and Digby (:O3) as " reduced forms of ferti- lization." It may be questioned whether the use of the term fertilization is fully justified by the events under discussion. We are all likely to agree with these authors that the physiological aspects of the phenomena in the cases under consideration are similar to sexual acts. But, by the writer, the act of fertiliza- tion is always considered in phylogenetic relations and strictly limited to the union of sexually differentiated cells, which are defined by their morphology through principles of homology. Whenever on~e or both of the gametes are suppressed in a life history and a succeeding generation develops of the sort that normally follows a sexual act, then such a development is apoga- mous and the phenomena always introduce features which are foreign to the processes of normal fertilization and the funda- mental principles of sexuality. Perhaps the most important characteristic of sexuality from an evolutionary standpoint is the fusion of gametes of unrelated parentage, for in the mingling of diverse protoplasm lie two factors: (i) a physiological stimulus to development, and (2) an increased probability of inherited variation which in new combi- nations will appear to the advantage of the species. Blackman's " reduced forms of fertilization" which I should prefer to con- sider apart from normal fertilization as examples of apogamy, and have so classed in this treatment, do satisfy the physiologi- cal requirements of a sexual act in that a form of nuclear fusion No. 460.] STUDIES ON PLANT CELL. V. 25 1 is substituted for the union of gamete nuclei but the phylogenetic and evolutionary aspects of sexuality are disregarded. Also, the nuclei that fuse are sometimes very closely related, which is a condition generally avoided in sexual processes except where peculiarities of habit make close inbreeding necessary. It is true that large groups, such as the Basidiomycetes, perhaps certain regions of the Ascomycetes, some Phycomycetes, and some forms of the higher plants and algae seem to have given up normal sexual processes but there is much evidence that in many cases this loss of sexuality is associated with a certain degree of segregation and with peculiarities of life conditions apart from the normal activities of all organisms or quite different from the ancestral stock. The groups are likely to be distinguished by highly specialized life habits of a sort that make it impossible for inherited sexual organs to function, either through mechan- ical difficulties or because one or both degenerate. It seems to me much clearer to regard all illustrations of Blackman's ''reduced forms of fertilization" under the general term of apogamy even though it may be clear that they are physio- logical substitutes for sexual acts and to reserve the term fertil- ization for the union of gametes which can always be clearly identified through morphology in ontogeny and phylogeny. The success of a group even though ancestral sexual processes may be suppressed does not enter into a problem which is at bottom a morphological one. Success is relative and we really have no means of estimating its degree save by actual experiment. It is not likely that any biologist would claim that sexual degenera- tion is advantageous to any species although the organic world is full of forms which have dispensed with sexuality and still hold their places. These are the reasons why I have grouped cell unions and nuclear fusions as sexual and asexual on a mor- phological basis founded on phylogenetic principles and why in Section V, we shall devote some attention to the substitutes for sexuality under the head of apogamy. The Ascomycetes present a phenomenon of nuclear fusion within the ascus which may properly be considered at this time since there is a certain resemblance to the nuclear fusions in the teleutospore and basidium. Dangeard ('94-'95b) gave the 252 THE AMERICAN NATURALIST. [VOL. XXXIX. first account of this phenomenon describing it for several forms. The mother cell of an ascus sometimes terminates a hypha but more commonly is situated a little back from the end at a point where the hypha bends abruptly like a knee. The mother cell contains two nuclei, closely related to each other, that unite, after which the fusion nucleus divides to form the ascospores. Dangeard considered this fusion to be a sexual act and the product an oospore which germinates immediately to form the ascus. He regards the ascus as a sporangium, and equivalent to the promycelium which he calls a conidiophore. Dangeard is not willing to accept any of the evidence that the ascocarp ever results from a sexual act or that sexual organs either func- tional or abortive are present at any stage in the life history of Ascomycetes. Sexuality, according to him, is reduced to the fusion within the ascus alone. He (Dangeard, '96 '9/a, b ; : oo) discredits the work of Harper on Sphaerotheca, Erysiphe, and Pyronema and the older accounts of De Bary and his pupils on sexual organs of the Ascomycetes. A series of short papers in Le Botaniste (: 03, Fas. i) presents Dangeard's last attack on the work of Harper and a reafnrmation of his peculiar views. Harper's description of sexual processes in Sphaerotheca ('95 ; '96) Erysiphe ('96), and Pyronema (: oob) are so convincing that, together with our knowledge of sexual organs in the lichens, Laboulbeniales, and Gymnoascales, we must accept the old view of De Bary that the ascocarp represents a development (probably sporophytic) from a sexual phase even though it may be established that there is much apogamy in the Ascomycetes. Harper gives the clearest account of the nuclear fusion in the ascus of any author without, however, committing himself to speculations on its significance. The subject is well sum- marized in his paper on Pyronema (: oob, pp. 363, 394). He finds in Erysiphe, Pyronema, and some other forms that the ascus is always developed from a penultimate cell of a hypha which bends sharply so that this cell appears to lie at the tip. There are two nuclei at the end of the ascogenous hypha and these divide simultaneously in a very characteristic manner so that the young ascus receives two of the resultant four nuclei, but each is derived from a different one of the original pair and No. 460.] STUDIES ON PLANT CELL. V. 253 consequently they are not sisters. The two nuclei in the ascus then fuse. The origin of the original pair is not known. No satisfactory explanation of this fusion in the ascus has been advanced. The conditions in the Ascomycetes are not the same as in the Basidiomycetes. There is no series of paired nuclei in the ascogenous hyphse and no evidence of a delayed fusion of gamete nuclei following a sexual act nor of nuclear fusions associated with the apogamous development of a sporo- phyte generation. On the contrary, a sexual act with the fusion of gamete nuclei has been clearly established in some forms preliminary to the development of the ascocarp and the nuclear union in the ascus is plainly a supplementary phenom- enon. Wager and Harper point out analogies to the account of Chmielewski ('9Ob) for Spirogyra, considered in a previous part of this section, which described a double nuclear fusion in the zygospore. Thus the primary, sexually formed nucleus of the zygospore is reported to divide into four secondary nuclei, two of which break down while the remaining two unite forming the second and final fusion nucleus of the spore. It is hard to see how these second nuclear fusions can be sexual and Groom ('98) is perhaps correct in considering them superimposed on the sex- ual act, but their physiological significance is not clear. Some recent papers support in general Harper's investigations on the ascus. Guilliermond (: 043 ; : O4b) describes the devel- opment of the ascus and ascospores in a number of forms. In an unnamed species of Peziza he found, however, that the ascus developed from the terminal cell of the ascogenous hypha which received two nuclei (that fuse) of the four that are found at the tip. Maire (: O3a ; : O3b) has reported a similar history for Galactinia succosa. Both Maire and Guilliermond note the resemblance of these conditions to the nuclear associations in the young basidium and Maire does not hesitate to consider the two nuclei in the tip of the ascogenous hypha as much reduced synkaryons, (paired nuclei) appearing for a very short period just previous to the nuclear fusions in the ascus. Maire fol- lows Dangeard in denying the sexual processes described by Harper in the Ascomycetes and would allign the events in the ascus with those in the basidium. Guilliermond agrees with 254 THE AMERICAN NATURALIST. [VOL. XXXIX. Harper that the number of chromosomes presented in the mitoses within the ascus is large (8, 12, 16, in various species) as against Dangeard and Maire who have claimed that the number is uniformly 4. Guilliermond's account of spore forma- tion in the ascus supports that of Harper (described in Section II) in all essentials and gives especial attention to the structure of the epiplasm and its inclusions. In summary : the significance of the nuclear fusions in the ascus seems very much of a mystery. If they could be associ- ated with an apogamous development of the ascocarp we should have conditions analogous to those in the Basidiomycetes but following a sexual act as it does in Sphaerotheca, Erysiphe, and Pyronema we find a phenomenon whose raison d' etre is not apparent. However, we do not know the history of the nuclei preceding the group of four at the end of the ascogenous hypha and perhaps it may be discovered that events at this period are concerned with nuclear reduction at the end of a sporophyte generation. One of the most interesting announcements of recent months is that in a preliminary note of Farmer, Moore, and Digby (: 03) on the nuclear history preceding the apogamous development of a species of Nephrodium. They found that the cells of the prothallus at the point where the sporophyte arose became binucleate by the migration of nuclei from neighboring cells. The two nuclei might remain separate for some time or fuse at once. The authors speak of the whole process "as a kind of irregular fertilization" and Blackman considers it analogous to the entrance of the nucleus into the fertile cell of Phragmidium and the establishment of the paired nuclei in the Uredinales. As we discussed the phenomenon in that connection I consid- ered the use of the term fertilization unfortunate since it included processes which however similar physiologically held no relation morphologically and phylogenetically to normal sexual processes. As stated then, it seems to me much clearer to regard all such apogamous phenomena apart from sexual proc- esses, pointing out as far as possible physiological resemblances but recognizing the wide gap in morphology established by the past evolutionary history of the plant. The interest in the phe- No. 460.] STUDIES ON PLANT CELL V. 255 nomena does not become less by this treatment which certainly avoids much confusion of expression. There is left for consideration one other group_ol nuclear fusions which may have sexual significance although such is not obvious, namely the fusions of polar nuclei in the embryo sac of angiosperms and the triple unions of the above with a second sperm nucleus which is often called "double fertiliza- tion." Several excellent reviews of this subject have appeared, notably by Strasburger (: oob), Sargant (: oo), Coulter and Chamberlain (: 03), Mottier (:O4a, b), and Guerin (:O4). The explanation of this phenomenon is likely to rest finally upon morphological analysis but at present we are uncertain of the homologies of the polar nuclei and the part they play in the evolutionary history of the endosperm. The most striking theory of the endosperm was proposed by LeMonnier ('87) who suggested that the fusion of the polar nuclei gave origin to a second embryo modified to nourish the normal embryo. One of the polar nuclei is always closely related to the egg nucleus so that in the triple fusions (the sperm with two polar nuclei) we have conditions very close to normal fertilization, the dis- cordant element being not the sperm nucleus but the antipodal polar nucleus. The triple fusions would seem at first thought to be rather favorable to LeMonnier's theory although it is plain that with such a diverse mixture of chromatin from thre.e nuclei the resultant structure can scarcely be called a sporophyte embryo from the very grotesqueness of its make-up. Miss Sargant considers the fusion of the second sperm with the micropylar nucleus as sexual in character but so complicated by the introduction of the antipodal polar nucleus that the result is a bizarre structure not strictly comparable to a normal embryo. In the final solution of this problem we must know whether in phylogeny the sperm and micropylar polar nucleus fused first and the antipodal entered into the process later or whether the polar nuclei began the habit and the second sperm nucleus was drawn afterwards into the activities. Should the first possibility be established the sexual nature of the process would seem clear while in the second the events would be of the nature of asexual nuclear fusions. While we know very little 256 THE AMERICAN NATURALIST. [VOL. XXXIX. of the origin and evolution of the endosperm in angiosperms there is some evidence in favor of the second possibility. Strasburger (: oob) holds that the double and triple nuclear fusions in the embryo sac are not true sexual acts even though they may involve an important principle of fertilization, namely, a stimulus to growth. According to him, sexual processes pre- sent two distinct features which he designates as " generative fertilization" and " vegetative fertilization." Generative fertili- zation deals with the mingling of ancestral hereditary substances in the nuclei and establishes the basis for such characters as hold the species true to its past or introduce new qualities as variations into the germ plasm. Vegetative fertilization brings to the fusion nucleus simply a stimulus to growth such as may be given to unfertilized eggs by changes in their physical and chemical environment. We might apply this classification to many of the examples of asexual nuclear fusions which we have discussed, as in the apogamous development of the fern and the origin of the paired nuclei in the rusts, and they have the ele- ments of vegetative fertilization in Strasburger's sense. But such distinctions are very subtle and it seems rather doubtful whether they add much to the clearness of our conceptions. The growth stimulus of " vegetative fertilization " is always an accompaniment of ''generative fertilization" and would be expected of any cell unions or nuclear fusions. The pecul- iarities of sex lie in the phylogenetic features of the phenomena, /. e., in the union of differentiated gametes with their long evolu- tionary history and not in the mere fusion of any nuclei at any time. From this point of view the double fusions of polar nuclei or the triple fusions, when a sperm nucleus becomes involved in the phenomenon, are of very doubtful sexual nature since no phylogenetic connections have been established with the normal sexual processes of the spermatophytes. Indeed, there are many irregularities in the process of endosperm formation which com- plicate the discussion and make it very difficult to trace relation- ships. Thus nuclear fusions are described in the late stages of endosperm formation when several of the free nuclei become included in the same cell area by the formation of the cell walls No. 460.] STUDIES ON PLANT CELL. V, 257 (Corydalis, Strasburger, '80 ; Tischler, : oo ; Canna, Humphrey, '96). Such nuclei are known to unite two or more and some- times several together within the cells, forming fusion nuclei with a large and variable number of chromosomes. In Peperomia and Gunnera the endosperm nucleus results from the fusion of several free nuclei and a number of instances are recorded in which no fusion of the polar nuclei takes place, but the endo- sperm is derived from the division of one or both. Such irregu- larities, which will probably be greatly increased in number as investigations proceed, indicate that the double and triple fusions preceding the differentiation of the endosperm nucleus are not of phylogenetic importance but are more likely to be special developments in relation to peculiarities of seed forma- tion among the angiosperms rather than of a sexual nature. However, the triple fusions, when a sperm enters into the composition of the endosperm nucleus, seem to furnish a cytological explanation of the phenomenon of xenia and thus come into very close physiological relations to sexual processes. In xenia we find the effects of hybridization expressed immedi- ately outside of the embryo in the endosperm of the seeds. If paternal chromatin has entered into the composition of the endo- sperm nucleus or should the sperm nucleus by itself give rise to a series of endosperm nuclei the appearance of paternal char- acters would be expected. This explanation of xenia was worked out independently by DeVries, Correns, and Webber, the last author having published a particularly clear and full account of the phenomenon (Webber, :oo). Even though the relation of xenia to hybridization is apparent, it is nevertheless clear that we are dealing with an exceptional process only possible because of the unusual conditions within the embryo sac which allow a second sperm nucleus to enter into the activities of seed forma- tion and it is certainly not established that these activities have any phylogenetic relations to past sexual processes. Some interesting studies of Nemec (: 02-: 03 ; : 04) upon asexual nuclear fusions may open the way for explanations of some of the examples which we have considered as asexual in the latter portion of this paper. Nemec found that mitosis in the root tip of Pisum sativum could be checked during anaphase 258 THE AMERICAN NATURALIST. [VoL. XXXIX. by treating the material with chloral hydrate so that no walls were formed between the daughter nuclei, which remained in the common mother cell and presently fused with one another. The fusion nucleus presented a^ double number of chromosomes (twice that of the normal sporophyte) in succeeding mitoses which became reduced in a few hours so that later divisions showed the number characteristic of the sporophyte. Nemec regards nuclear fusions and reduction phenomena as self regulat- ing processes which follow the vital cell fusions characteristic of fertilization. The latter (cell fusions) are then the essential phenomena of sex and nuclear activities follow automatically. Reduction phenomena are atavistic in character. Nemec con- siders these results in serious conflict with Strasburger's ('94) theory of the periodic reduction of the chromosomes, believing that the number of chromosomes is not so likely to give the characters of the respective sporophyte and gametophyte gener- ations as other factors. Nemec's contribution is chiefly of interest to us in the present connection as showing that nuclear fusions may result from dis- turbances of the normal environment very far removed from the conditions that produce sexual cells. And this emphasizes our contention that sexual processes must be judged through phylo- genetic analysis and not by physiological resemblances. Thus the nuclear fusions in the ascus, in the basidium, preceding apog- amous development of the fern, and perhaps the union of polar nuclei in the embryo sac may be involved with special physio- logical conditions although they resemble outwardly sexual processes and are sometimes a substitute for these. But never- theless they are asexual nuclear fusions lacking that funda- mental character of sexuality, the result of sexual evolution, namely, a fixed position in a life cycle established by phylogeny and expressed by the classic phrase " ontogeny repeats phylog- eny." They are departures from the normal life history either apogamous in character or concerned with some other peculiarity of the plants' existence. No. 460.] STUDIES ON PLANT CELL. V. 259 LITERATURE CITED IN SECTION IV, "THE PLANT CELL." BAMBEKE. : 03. Sur 1'evolution nucle*aire et la sporulation chez Hydnangium carneum Wallr. Mem. d. 1'Acad. Roy. Sci., Litt., Beaux Arts d. Belgique, vol. 54. BARKER. : 01. A Conjugating Yeast. Phil. Trans. Roy. Soc. London, vol. 194, p. 467. BERTHOLD. '81. Die geschlechtliche Fortpflanzung der eigentlichen Phaeospo reen Mitt. a. d. Zool. Sta. Neapel, vol. 2, p. 401. BLACKMAN. '98. The Cytological Features of Fertilization and related Phenomena in Pinus sylvestris L. Phil. Trans. Roy. Soc. London, vol. 190, P- 395- BLACKMAN. : O4a. On the Fertilization, Alternation of Generations and General Cytol- ogy of the Uredineas. Annals of Bot., vol. 18, p. 323. BLACKMAN. : 04b. On the Relation of Fertilization to " Apogamy " and " Partheno- genesis." New Phytologist, vol. 3, p. 149. CHAMBERLAIN. '99. Oogenesis in Pinus laricio. Bot. Gazette, vol. 27, p. 268. CHMIELEWSKI. '90a. Eine Notiz iiber das Verhalten der Chlorophyllbander in den Zygoten der Spirogyren-Arten. Bot. Zeit., vol. 48, p. 773. CHMIELEWSKI. '90b. Materialien zur Morphologic und Physiologic des Sexualprocessen bei den niederen Pflanzen. (Russian.) Review, Bot. Centralbl., vol. 50, p. 264. COKER. : 03. On the Gametophytes and Embryo of Taxodium. Bot. Gazette, vol. 36, p. 114. COULTER and CHAMBERLAIN. :03. Morphology of the Angiosperms. New York, 1903. DANGEARD. '93. La reproduction sexuelle des Ustilagine"es. Compt. Rendus, vol. 117, p. 496. DANGEARD. '94-'95a. La reproduction sexuelle de rEntyloma Glaucii, Le Bot- aniste, vol. 4, p. 12. 260 THE AMERICAN NATURALIST. [VOL. XXXIX. DANGEARD. '94-'95b. La reproduction sexuelle des Ascomycetes. Le Botaniste, vol. 4, p. 21. DANGEARD. '94-'95c. Memoire sur la reproduction sexuelle des Basidiomycetes. Le Botaniste, vol. 4, p. 1 19. DANGEARD. '96-'97a. La reproduction sexuelle dans le Sph&rotheca Castagnei. Le Botaniste, vol. 5, p. 27. DANGEARD. '96-'97b. Second me'moire sur la reproduction sexuelle des Ascomycetes. Le Botaniste, vol. 5, p. 245. DANGEARD. : 00. La reproduction sexuelle des champignons. Le Botaniste, vol. 7, p. 89. DANGEARD et SAPIN-TROUFFY. '93. Une pseudofe'condation chez les Uredine"es. Compt. Rendus, vol. 117, p. 267. DAVIS. '96b. Development of the Cystocarp of Champia parvula. Bot. Gazette, vol. 21, p. 109. DAVIS. :00. The Fertilization of Albugo Candida. Bot. Gazette, vol. 29, p. 297. DAVIS. :01. The Origin of Sex in Plants. Pop. Sci. Month., vol. 60, p. 66. DAVIS. : 03. The Evolution of Sex in Plants. Pop. Sci. Month., vol. 62, p. 300. DAVIS. :04. Oogenesis in Vaucheria. Bot. Gazette, vol. 38, p. 81. DE BARY and STRASBURGER. '77. Acetabularia mediterranea. Bot. Zeit., vol. 35, p. 713. FAIRCHILD. '97. Ueber Kerntheilung und Befruchtung bei Basidiobolus ranarum Eidam. Jahrb. f. wiss. Bot., vol. 30, p. 285. FARMER, MOORE, and DIGBY. :03. On the Cytology of Apogamy and Apospory. I. Preliminary Note on Apogamy. Proc. Roy. Soc. London, vol. 71, p. 453. FARMER and WILLIAMS. ''98. Contributions to our Knowledge of the Fucaceae ; their Life History and Cytology. Phil. Trans. Roy. Soc. London, vol. 190, p. 623. FEDERLEY. : 03-: 04. Die Copulation der Conidien bei Ustilago Tra^opogonis pratensis Pers. Finska Veten. Selsk. Forhandl., vol. 46, no. 2. No. 460.] STUDIES ON PLANT CELL. V. 26 1 FERGUSON. : 04. Contributions to the Knowledge of the Life History of Pinus with Special Reference to Sporogenesis, the Development of the Gametophytes and Fertilization. Proc. Wash. Acad. Sci., vol. 6, p. i. GARDINER. '88. On the Continuity of Protoplasm through the Walls of Vegetable Cells. Arb. d. bot. Inst. Wiirzb., vol. 3, p. 52. GARDINER. ' 98. The Histology of the Cell Wall with Special Reference to the Mode of Connections of the Cells. Proc. Roy. Soc. London, vol. 62, p. 100. GARDINER. : 00. The Genesis and Development of the Wall and Connecting Threads in the Plant Cell. Proc. Roy. Soc. London, vol. 66, p. 186. GOLENKIN. '99. Ueber die Befruchtung bei Sphceroplea annulina und iiber die Structur der Zellkerne bei einigen griinen Algen. Bull. Soc. Imp. Nat. Moscow, 1899, p. 343. GOROSCHANKIN. '83. Zur Kenntniss der Corpuscula bei den Gymnospermen. Bot. Zeit, vol. 41, p. 825. GROOM. '98. On the Fusion of Nuclei among Plants. Trans, and Proc. Bot. Soc. Edinburgh, 1898. GUERIN. : O4. Les connaissances actuelles sur la fe"condation chez les phanerogams. Paris, 1904. GUILLIERMOND. :03. Recherches cytologiques sur les levure. Rev. Gen. d. Bot., vol. 15, pp. 49, 104, 1 66. GUILLIERMOND. : 04a. Contribution a I'e'tude de la formation des asques et de 1'epiplasme des Ascomycetes. Rev. Gen. d. Bot, vol. 16, p. 49. .GUILLIERMOND. :04b. Recherches sur la karyokinese chez les Ascomycetes. Rev. Gen. d. Bot., vol. 1 6, p. 129. HARPER. '95. Die Entwickelung des Peritheciums bei Spkccrotkeca Castagnei. Ber. deutsch. bot. Gesellsch., vol. 13, p. 475. HARPER. '96. Ueber das Verhalten der Kerne bei der Fruchtentwickelung einiger Ascomyceten. Jahrb. f. wiss. Bot., vol. 29, p. 655. HARPER. '99a. Nuclear Phenomena in certain Stages of Development of , the Smuts. Trans. Wise. Acad. Sci., Arts, Lett, vol. 12, p. 475. 262 THE AMERICAN NATURALIST. [VOL. XXXIX. HARPER. :02. Binucleate Cells in certain Hymenomycetes. Bot. Gazette, vol. 33. P- I- HARPER. :00. Sexual Reproduction in Pyronema confluens and the Morphology of the Ascocarp. Annals of Bot., vol. 14, p. 321. HAUPTFLEISCH. '88. Zellmembran und Hiillgallerte der Desmidiaceen. Mitt. a. d. naturwiss. Ver. f. Neu-Vor. u. Riigen, 1888. HAUPTFLEISCH. '95. Die Ortsbewegung der Bacillariaceen. Mitt. a. d. naturwiss. Ver. f. Neu-Vor. u. Riigen, vol. 27. HILL. :01. The Histology of the Sieve-tubes of Pinus. Annals of Bot., vol. 15, P- 575- HOLDEN and HARPER. : 03. Nuclear Division and Nuclear Fusion in Coleosporium sonchi- arvensis Lev. Trans. Wise. Acad. Sci., Arts, Lett., vol. 14, p. 63. HUMPHREY. '96. The Development of the Seed in Scitamineae. Annals of Bot., vol. 10, p. i. IKENO. '98b. Untersuchungen iiber die Entwickelung der Geschlechtsorgane und den Vorgang der Befruchtung bei Cycas revoluta. Jahrb. f. wiss. Bot, vol. 32, p. 557. IKENO. : 01. Contribution a l'e"tude de la fe"condation chez le Ginkgo biloba. Ann. Sci. Nat., Bot, ser. 8, vol. 13, p. 305. IKENO. :04. Blepharoplasten im Pflanzenreich. Biol. Centralbl., vol. 24, p. 21 1. KARSTEN. : 00. Die Auxosporenbildung der Gattungen Cocconeis, Surirella, und Cymatopleura. Flora, vol. 87, p. 253. KlENITZ-GERLOFF. '91. Die Protoplasmaverbindungen zwischen benachbarten Gewebsele- menten in der Pflanze. Bot. Zeit, vol. 49, pp. 17, 33, 49, 65. KlENITZ-GERLOFF. : 02. Neue Studien iiber Piasmodesmen. Ber. deutsch. hot. Gesellsch.. vol. 20, p. 93. KING. : 03. Observations on the Cytology of Araiospora pitlchra Thaxter. Proc. Boston Soc. Nat. Hist., vol. 31, p. 211. KLEBAHN. '91. Studien iiber Zygoten. I. Die Keimung von Closterium und Cosmarium. Jahrb. f. wiss. Bot, vol. 22, p. 415. No. 460.] STUDIES ON PLANT CELL. V. 263 KLEBAHN. '92. Studien iiber Zygoten. II. Die Befruchtung von GLdogonium Boscii. Jahrb. f. wiss. Bot., vol. 24, p. 235. KLEBAHN. '96. Beitrage zur Kenntniss der Auxosporenbildung. I. Rhopalodia gibba (Ehrenb.) O. Miiller. Jahrb. f. wiss. Bot, vol. 29, p. 595. KLEBAHN. '99. Die Befruchtung von Sphceroplea annulina Ag. Festsch. f. Sch- wendener, p. 81, 1899. KLEBS. '96. Die Bedingungen der Fortpflanzung bei einigen Algen und Pilzen. Jena, 1896. KNY. : 04. Studien iiber intercellularen Protoplasma. Ber. deutsch. bot. Gesellsch., vol. 22, pp. 29, 347. KOENICKE. :01. Ueber Ortsveranderungen von Kernen. Sitz. niederrh. Gesellsch. f. Nat. u. Heilk. z. Bonn, 1901, p. 14. KOENICKE. : 04. Der heutige Stand der pflanzlichen Zellforchung. Ber. deut. bot. Gesellsch., vol. 21, p. 66. KOHL. '91. Protoplasmaverbindungen bei Algen. Ber. deutsch. bot. Gesellsch., vol. 9, p. 9. KOHL. '97. Die Protoplasmaverbindungen der Spaltoffnungsschliesszellen und der Moosblattzellen. Bot. Centralbl., vol. 72, p. 257. KOHL. : 00. Dimorphism . der Plasmaverbindungen. Ber. deutsch. bot. Gesellsch., vol. 18, p. 364. KOHL. :02. Beitrage zur Kenntniss der Plasmaverbindungen in den Pflanzen. Beiheft z. Bot. Centralbl., vol. 12, p. 343. KRUCH. '90. Appunti sullo sviluppo degli organi sessuali e sulla fecondazione della Riella clausonis Let. Malpighia, vol. 4, p. 403. KUCKUCK. '98. Ueber die Paarung von Schwarmsporen bei Scytosiphon. Ber. deutsch. bot. Gesellsch., vol. 16, p. 35. KUHLA. : 00. Die Plasmaverbindungen bei Viscum album. Bot. Zeit, vol. 58, p. 29. LAND. : 02. A Morphological Study of Thuja. Bot. Gazette, vol. 34, p. 249. 264 THE AMERICAN NATURALIST. [VOL. XXXIX. LAW SON. : 04a. The Gametophytes, Archegonia, Fertilization and Embryo of Sequoia sempervirens. Annals of Bot, vol. 18, p. i. LAWSON. : 04b. The Gametophytes, Fertilization and Embryo of Cryptomeria japonica. Annals of Bot., vol. 18, p. 417. LEMONNIER. '87. Sur la valeur morphologique de 1'albumen chez les angiospermes. Journ. d. Bot., vol. i, p. 140. MAIRE. : OOa. Sur la cytologie des Hymenomycetes. Compt. Rendus, vol. 131, p. 121. MAIRE. : OOb. Sur la cytologie des Gasteromycetes. Compt. Rendus, vol. 131, p. 1246. MAIRE. : OOc. L'eVolution nucle"aire chez les Endophyllum. Journ. d. Bot., vol. 14, p. 80. MAIRE. : 02. Recherches cytologiques et taxonomiques sur les Basidiomycetes. Bull. Soc. Mycol. d. France, vol. 18. MAIRE. : 03a. Recherches cytologiques sur le Galactinia succosa. Compt. Rendus de 1'Acad. Sci. Paris, Nov. 9, 1903. .MAIRE. : 03b. La formation des asques chez les Pezizes et Involution nucldaire des Ascomycetes. Compt. Rendus d. Se*an. d. 1. Soc. Biol., vol. 55, p. 1401. MEYER. '96. Die Plasmaverbindungen und die Membranen von Volvox globator, auteus und tertius mit Rucksicht auf die tierischen Zellen. Bot. Zeit., vol. 54, p. 187. MEYER. : 02. Die Plasmaverbindungen und die Fusionen der Pilze der Flori- deenreihe. Bot. Zeit., vol. 60, p. 139. MlCHNIEWICZ. : 04. Ueber Plasmodesmen in den Kotyledonen von Lupinus-Arten und ihre Beziehung zum interzellularen Plasma. Oester. bot. Zeitsch., vol. 54, p. 165. MIEHE. : 01. Ueber die Wanderungen des pflanzlichen Zellkernes. Flora, vol. 88, p. 105. MlYAKE. : 01. The Fertilization of Pythium deBaryanuin. Annals of Bot., vol. *5>P- 653- No. 460.] STUDIES ON PLANT CELL V. 265 MlYAKE. : 03a. On the Development of the Sexual Organs in 'Picea excelsa. Annals of Bot., vol. 17, p. 351. MIYAKE. :03b. Contribution to the Fertilization and Embryogeny of Abies bal- samea. Beiheft z. Bot. Centralbl., vol. 14, p. 134. MOENKHAUS. : 04. The Development of the Hybrids between Fundulus heteroclitus and Menedia notata with especial Reference to the Behavior of the Maternal and Paternal Chromosomes. Amer. Journ. of Anat., vol. 3, p. 29. MONTGOMERY. :01. A Study of the Germ Cells of Metazoa. Trans. Amer. Phil. Soc., vol. 20, p. 154. MOTTIER. '98. Ueber das Verhalten der Kerne bei der Entwickelung des Embryo- sacks und die Vorgange bei der Befruchtung. Jahrb. f. wiss. Bot., vol. 31, p. 125. MOTTIER. : 04a. The Development of the Spermatozoid in Chara. Annals of Bot., vol. 1 8, p. 245. MOTTIER. :04b. Fecundation in Plants. Carnegie Inst. Washington, 1904. MULLER. '98-'99. Kammern und Poren in der Zellwand der Bacillariaceen. I, II. Ber. deutsch. bot. Gesellsch., vol. 16, p. 386 ; vol. 17, p. 432. MURRILL. : 00. The Development of the Archegonium and Fertilization in the Hemlock Spruce (Tsuga Canadensis Carr.) Annals of Bot., vol. 14, p. 583. NEMEC. :02-:03. Ueber ungeschlechtliche Kernverschmeltzungen. I, II, III. Sitz. kon. bohm. Gesellsch. Wiss., 1902-03. NEMEC. :04. Ueber die Einwirkung des Chloralhydrats auf die Kern- und Zelltheilung. Jahrb. f. wiss. Bot, vol. 39, p. 645. OLTMANNS. '95. Ueber die Entwickelung der Sexualorgane bei Vaucheria. Flora, vol. 80, p. 388. OLTMANNS. '98a. Die Entwickelung der Sexualorgane bei Coleochate pulvinata. Flora, vol. 85, p. i. OLTMANNS. '98b. Zur Entwickelungsgeschichte der Florideen. Bot. Zeit., vol. 56, p. 99. 266 THE AMERICAN NATURALIST. [VOL. XXXIX. OLTMANNS. '99. Ueber der Sexualitat der Ectocarpeen. Flora, vol. 86, p. 86. OSTERHOUT. :00. Befruchtung bei Batrachospermum. Flora, vol. 87, p. 109. POIRAULT and RACIBORSKI. '95. Sur le noyau des Ure'dine'es. Journ. d. Bot., vol. 9, p. 318. RACIBORSKI. '96. Ueber den Einfluss ausserer Bedingungen auf die Wachstumsweise des Basidiobolus ranarum. Flora, vol. 82, p.. 107. REINHARDT. '92. Das Wachstum der Pilzhyphen. Jahrb. f. wiss. Bot., vol. 23, p. 479. ROBERTSON. : 04. Studies in the Morphology of Torreya californica Torrey. II. The Sexual Organs and Fertilization. New Phytologist, vol. 3, p. 205. ROSEN. '93. Studien iiber die Kerne und die Membranbildung bei Myxomyceten und Pilzen. Cohn's Beitr. z. Biol. d. Pflan., vol. 6, p. 237. ROSTAFINSKI Und WORONIN. '77. Ueber Botrydium granulatum. Bot. Zeit, vol. 35, p. 649. RUHLAND. :01. Zur Kenntniss der intracellularen Karyogamie bei den Basidio- myceten. Bot. Zeit., vol. 59, p. 187. Russow. '83. Ueber den Zusammenhang der Protoplasmakorper benachbarter Zellen. Sitz. dorpater naturf. Gesellsch., 1883. SAPPIN-TROUFFY. '96. Recherches histologiques sur la famille des Uredine'es. Le Botan- iste, vol. 5, p. 59. SARGANT. : 00. Recent Work on the Results of Fertilization in Angiosperms. Annals of Bot, vol. 14, p. 689. SCHMIDLE. '99. Einiges iiber die Befruchtung, Keimung und Haarinsertion von Batrachospermum. Bot. Zeit., vol. 57, p. 125. SCHRAMMEN. :02. Ueber die Einwirkung von Temperaturen auf die Zellen des Vege- tationspunktes des Sprosses von Vicia foba. Verhand. d. natur- hist. Vereins d. preuss. Rheinlande, vol. 59. SCHUTT. '99. Centrifugales Dickenwachstum der Membran und extramembran- oses Plasma. Jahrb. f. wiss. Bot., vol. 33, p. 594. No. 460.] STUDIES ON PLANT CELL. V. 267 SCHUTT. : OOa. Die Erklarung des centrifugalen Dickenwachstum der Membran. Bot. Zeit., vol. 59, p. 245. SCHUTT. : OOb. Centrifugale und simultane Membranverdickungen. Jahrb. f. wiss. Bot, vol. 35, p. 470. SHAW. '98a. The Fertilization of Onoclea. Annals of Bot., vol. 12, p. 261. SMITH. : OO. The Haustoria of the Erysipheae. Bot. Gazette, vol. 29, p. 153. SMITH, ISABEL. : 04. The Nutrition of the Egg in Zamia. Bot. Gazette, vol. 37, p. 346. STEVENS. '99. The Compound Oosphere of Albugo blitz. Bot. Gazette, vol. 28, p. 149. STEVENS. :01b. Gametogenesis and Fertilization in Albugo. Bot. Gazette, vol. 32, p. 77- STRASBURGER. '80. Zellbildung und Zelltheilung. Jena, 1880. STRASBURGER. '82. Ueber den Bau und Wachsthum der Zellhaiite. Jena, 1882. STRASBURGER. '94. The Periodic Reduction of the Number of Chromosomes in the Life History of Living Organisms. Annals of Bot., vol. 8, p. 281. STRASBURGER. '97a. Kerntheilung und Befruchtung bei Fucus. Jahrb. f. wiss. Bot., vol. 30, p. 351. STRASBURGER. :00b. Einige Bemerkungen zur Frage nach der doppleten Befruchtung bei den Angiospermen. Bot. Zeit, vol. 58, p. 293. STRASBURGER. : 01. Ueber Plasmaverbindungen pflanzlichen Zellen. Jahrb. f . wiss. Bot, vol. 36, p. 493 SUTTON. : 02. On the Morphology of the Chromosome Group in Brachystola magna. Biol. Bull., vol. 4, p. 24. SUTTON. ': 03. The Chromosomes in Heredity. Biol. Bull., vol. 4, p. 231. TANGL. '79-'81. Ueber offene Communicationen zwischen den Zellen des Endo- sperms einiger Samen. Jahrb. f. wiss. Bot, vol. 12, p. 170. TERNETZ. : 00. Protoplasmabewegung und Fruchtkorperbildung bei Ascophanus carneiis. Jahrb. f. wiss. Bot., vol. 35, p. 273. 268 THE AMERICAN NATURALIST. [VOL. XXXIX. THAXTER. '96. Contribution towards a Monograph of the Laboulbeniaceae. Mem. Amer. Acad. Arts and Sci., vol. 12. THOM. '99. The Process of Fertilization in Aspidium and Adiantum. Trans. Acad. Sci. St. Louis, vol. 9, p. 285. TISCHLER. : 00. Untersuchungen iiber die Entwickelung des Endosperm und der Samenschale von Corydalis cava. Verh. naturhist.-med. Ver. Heidelberg, vol. 6, p,. 351. TOWNSEND. '97. Der Einfluss des Zellkerns auf die Bildung der Zellhaut. Jahrb. f. wiss. Bot, vol. 30, p. 484. TROW. : 01. Biology and Cytology of Pythium ultimum n. sp. Annals of Bot., vol. 15, p. 269. TROW. : 04. On Fertilization in the Saprolegnieae. Annals of Bot, vol. 18, p. 541. WAGER. '96. On the Structure and Reproduction of Cystopus candidus Lev. Annals of Bot, vol. 10, p. 295. WAGER. '99. The Sexuality of Fungi. Annals of Bot., vol. 8, p. 575. WAGER. : 00. On the Fertilization of Peronospora parasitica. Annals of Bot.. vol. 14, p. 263. WEBBER. : 00. Xenia, or the Immediate Effect of Pollen in Maize. Bull. 22, Div. Veg. Path, and Phys., U. S. Dept. of Agric. WEBBER. : 01. Spermatogenesis and Fecundation of Zamia. Bull. 2, Bureau Plant Ind., U. S. Dept. of Agric. WILLIAMS. :04b. Studies in the Dictyotaceae. II. The Cytology of the Gameto- phyte Generation. Annals of Bot., vol. 18, p. 183. WILSON. :00. The Cell in Development and Inheritance. New York, 1900. WOLFE. :04. Cytological Studies on Nemalion. Annals of Bot, vol. 18, p. 607. WOYCICKI. '99. On Fertilization in the Coniferas. (Russian.) Review, Bot. Zeit, vol. 58, p. 39, 1900. VOL. XXXIX, No. 463 JULY, THE AMERICAN NATURALIST A MONTHLY JOURNAL DEVOTED TO THE NATURAL SCIENCES IN THEIR WIDEST SENSE CONTENTS Page I. Eestoration of the Titanothere Megacerops . PROFESSOR R. 8. LULL 419 II. Synopses of North American Invertebrates. XXI. The Nemerteans. Part I PROFESSOR W. R. COE 425 III. Studies on the Plant Cell. -VI DR. B. M. DAVIS 449 IV. Notes and Literature: Zoology, McMurrich's Human Embryology, The Arthropods and Coelenterates of the Maldive and Laccadive Archipel- agoes, Townsend's Birds of Essex County, Massachusetts, Notes . . 501 V- Correspondence : A Biological Station in Greenland, Fleas and Disease . 505 BOSTON, U. S. A. GINN & COMPANY, PUBLISHERS 29 BEACON STREET New York Chicago London, W. C 70 Fifth Avenue 378-388 Wabash Avenue 9 St. Martin's Street Entered at the Pott-Office, Boston, Mast., as Second-Class Mail Matter. The American Naturalist. ASSOCIATE EDITORS: J. A. ALLEN, PH.D., American Museum of Natural History, New York. E. A. ANDREWS, ?H.D.,J0Ans Hopkins University, Baltimore. WILLIAM S. BAYLEY, PH.D., Colby University, Waterville. DOUGLAS H. CAMPBELL, PH.D., Stanford University. J. H. COMSTOCK, S.B.,- Cornell University, Ithaca. WILLIAM M. DAVIS, M.E., Harvard University, Cambridge. ALES HRDLICKA, M.D., U.S. National Museum, Washington. D. S. JORDAN, LL.D., Stanford University. CHARLES A. KOFOID, PH.D., University of California, Berkeley. J. G. NEEDHAM, PH.D., Lake Forest University. ARNOLD E. ORTMANN, PH.D., Carnegie Museum, Pittsburg. D. P. PENHALLOW,D.Sc.,F.R.M.S., Me Gill University, Montreal. H. M. RICHARDS, S.D., Columbia University, New York. W. E. RITTER, PH.D., University of California, Berkeley. ISRAEL C. RUSSELL, LL.D., University of Michigan, Ann Arbor. ERWIN F. SMITH, S.D., U.S. Department of Agriculture, Washington* LEONHARD STEJNEGER, LL.D., Smithsonian Institution, Washington. W. TRELEASE, S.D., Missouri Botanical Garden, St. Louis. HENRY B. WARD, PH.D., University of Nebraska, Lincoln. WILLIAM M. WHEELER, PH.D., American Museum of Natural History, New York. THE AMERICAN NATURALIST is an illustrated monthly magazine of Natural History, and will aim to present to its readers the leading facts and discoveries in Anthropologj', General Biology, Zoology, Botany, Paleontology, Geology and Physical Geography, and Miner- alogy and Petrography. The contents each month will consist of leading original articles containing accounts and discussions of new discoveries, reports of scientific expeditions, biographical notices of distinguished naturalists, or critical summaries of progress in some line ; and in addition to these there will be briefer articles on various points of interest, editorial comments on scientific questions of the day, critical reviews of recent literature, and a quarterly record of gifts, appointments, retirements, and deaths. All naturalists who have anything interesting to say are invited to send in their contributions, but the editors will endeavor to select for publication only that which is of truly scientific value and at the same time written so as to be intelligible, instructive, and interesting to the general scientific reader. All manuscripts, books for review, exchanges, etc., should be sent to THE AMERICAN NATURALIST, Cambridge, Mass. All business communications should be sent -direct to the publishers. Annual subscription, $4.00, net, in advance. Single copies, 85 cents. Foreign subscription, $4.60. GINN & COMPANY, PUBLISHERS STUDIES ON THE PLANT CELL. VI. BRADLEY MOORE DAVIS. SECTION V. CELL ACTIVITIES AT CRITICAL PERIODS OF ONTOGENY IN PLANTS. WE shall discuss in this paper the behavior of the protoplasm at a number of critical periods in the life history of plants when the organism passes from one phase to another of a fundamen- tally different character. At such times great changes take place in the potentialities of the cells which inaugurate the new developments, changes that are generally most conspicuously shown in the structure of the nucleus. Some of the most inter- esting events of cell and nuclear history take place at these times, as would be expected from the importance of the phe- nomena. We shall treat the material under the following heads : (i) Gametogenesis, (2) Fertilization, (3) Sporogenesis, (4) Re- duction of the Chromosomes, (5) Apogamy, (6) Apospory, (7) Hybridization, (8) Xenia. i . GAMETOGENESIS. The events of gametogenesis are clearly known for the higher plants but there is some confusion and almost no detailed infor- mation in the accounts of the thallophytes where the nuclei are very small and the details of the mitoses preceding the forma- tion of sexual cells exceedingly difficult of study. There is complete agreement among all investigators that the mitoses which precede the differentiation of gamete riuclei in spermatophytes, pteridophytes, and bryophytes are typical karyo- kinetic figures not differing essentially in the behavior of the chromosomes from the mitoses generally characteristic of the gametophyte generation. This information is based upon a 450 THE AMERICAN NATURALIST. [VOL. XXXIX. large number of studies of nuclear figures in antheridia and archegonia, the generative cell of the pollen tube and micropylar region of the embryo-sac. There are no reduction phenomena in these higher groups at the period of gametogenesis. The subject is complicated in some types of spermatophytes where the gametophyte phase is so reduced that the mitoses which precede gametogenesis may follow immediately upon the two mitoses characteristic of sporogenesis or be separated from them by only one or two divisions. For example, it is known in several types of the lily family (Lilium, Tulipa, Fritillaria, Erythronium, etc.) that the two mitoses of sporogenesis (hetero- typic and homotypic) are included in the embryo-sac and become a part of that gametophyte history. The third and final mitosis in this history differentiates the egg in the micropylar end of the embryo-sac and is a typical nuclear division. This subject was treated in some detail in Section III of these "Studies" (Amzr. Nat., vol. 38, pp. 741-745, 1904). When the mitoses of sporogenesis are not included within the embryo-sac we find almost without exception three typical mitoses preceding the differentiation of the egg in the angiosperms and a very large number in the gymnosperms, and of course in the pteridophytes and bryophytes the whole vegetative period of the gametophyte which is generally green and self-supporting. There are from two to three mitoses in the pollen grain and male gametophyte of the angiosperms before the development of the sperm nuclei and a somewhat larger and more variable number among the gymnosperms. It is necessary at the outset to understand clearly what are the events of gametogenesis in spermatophytes because several authors have carried the phenomena of sporo- genesis over into the period of gametogenesis, where it can have no proper place in exact morphology. Such papers will be treated in connection with " Sporogenesis " and " Reduction of the Chromosomes," for they concern primarily these phe- nomena alone. Gametogenesis must be considered at present chiefly from our knowledge of the conditions in the higher plants as they furnish almost the only detailed information that we have on the subject. Upon this as a basis we are justified in suggesting No. 463.] STUDIES ON PLANT CELL. VI. 45 1 possibilities in the thallophytes which must remain as specula- tions until investigations have advanced much farther in this difficult field of cell study. The basis of any theories at present must be phylogenetic, a principle that has not been followed in some of the work upon the thallophytes. Gametogenesis in plants is full of interest because of the sharp differences from the processes of spermatogenesis and oogenesis in animals. In animals the period of gametogenesis is one of unusual activity. After the germ cells are differenti- ated there follows a period of cell growth, with the peculiar activity termed synapsis, during which the number of chromo- somes is reduced to one half the number characteristic of the species. The germ cells emerge from the growth periods as primary spermatocytes or oocytes which give rise respectively by two successive mitoses to four spermatids or to an egg with its accompanying polar bodies. The gametes have one half the number of chromosomes characteristic of the species, so that the period of gametogenesis is one of chromosome reduction. The character of this process of reduction will be considered when we take up the analogous phenomena in plants after the discussion of sporogenesis. Gametogenesis in plants is in strik- ing contrast to that in animals. In all higher groups (those above the thallophytes) we know that the gametes have the same number of chromosomes as the vegetative cells of the parent plant (gametophyte). There is no reduction of the chro- mosomes at the time of gametogenesis, that phenomenon taking place at the end of the sporophyte generation with sporogenesis. Also, there are no peculiarities of the mitoses immediately preceding gametogenesis excepting such as concern the devel- opment of cilia-bearing organs (blepharoplasts) or slight pecul- iarities in the form or size of the spindles, for such nuclear figures are frequently different in these particulars from the mitoses in vegetative cells of the gametophyte. The differences concern chiefly the structure of the sperm, and have been de- scribed in our account of that structure (Amer. Nat., vol. 38, July and August, p. 576, 1904). To Strasburger above all others should be given the credit of making clear these important characteristics of gametogene- 452 THE AMERICAN NATURALIST. [VOL. XXXIX. sis in plants. Strasburger's paper of 1894 on " The Periodic Reduction of the Number of Chromosomes in the Life His- tory of Living Organisms" (Annals of Bot., vol. 8, p. 281) was the first elaborate presentation of the principles of gametogene- sis and reduction phenomena in plants and has become classical as the foundation of the present attitude in botanical science and the basis and stimulus of a large amount of confirmatory research. The matter really crystallized after the discovery that the sporophyte generation of the higher plants possessed nuclei with twice the number of chromosomes characteristic of the gametophyte and that the reduction took place in the spore mother-cell just previous to sporogenesis. These facts were gradually established by a number of investi- gations beginning with Strasburger ('84, '88) and Guignard ('84, '85). Guignard ('91) presented the first complete count of the number of chromosomes in the life history of a plant (Li/inm martagoii), determining the reduction period to be in the spore mother-cell, and Overton ('93 a and b) independently reached the same conclusions for the same plant and extended the knowl- edge of the chromosome count in gametophyte and sporophyte to a number of other types. Overton 's paper was important in its suggestiveness for extended research among the higher cryp- togams. Other investigations followed shortly in the gymno- sperms, pteridophytes, and liverworts, all supporting the view that the nuclei of the sporophyte generation, following the fusion of gamete nuclei* had double the number of chromosomes char- acteristic of the gametophyte and that the reduction phenomena occurred at the end of the sporophyte generation in the spore mother-cell. The significance of reduction phenomena at sporo- genesis must be phylogenetic since it represents a return of the organism at this time to the ancestral gametophyte condition. The details of this literature belong to the account of " Sporo- genesis " and " Reduction of the Chromosomes," and will be taken up later. But it is necessary to present the outline at this time to make clear the important fact that no reduction of the chromosomes takes place during gametogenesis in all groups above the thallophytes. The theories of gametogenesis among the thallophytes rest No. 463.] STUDIES ON PLANT CELL. VI. 453 upon information which in point of completeness falls very far short of our knowledge of the groups above. Indeed, no forms have been studied with the detail that is known in higher groups chiefly for the reason that the investigator is forced to deal with very small nuclei and mitotic figures whose chromosomes are exceedingly minute and because of various technical difficulties. The theories in general fall into two groups : (i) those which have an obvious basis in attempts to reconcile events with the processes of gametogenesis in animals, and (2) those proceeding from the view that for phylogenetic reasons the periods and phe- nomena of gametogenesis in the lower plants should correspond with those of the higher. We may pass over with a few words the early crude attempts to establish structures for plants comparable to the polar bodies of animals. For example at the conclusion of oogenesis in some algae (e. g., Vaucheria, CEdogonium) a globule of slime is exuded with the opening of the oogonium. It was suggested that such material is thrown off from the egg but we now know that it is not protoplasmic in character but is apparently derived from a softening of the cell wall. Then the ventral canal cell has been compared to a polar body but it seems clear now that all of the canal cells are homologous and a part of what was form- erly an extensive gametogenous tissue within the archegonium. Then the small group of cells cut off below the oogonium of the Charales and the fragmented nuclear material in the trichogyne of the red algae have been compared to substance thrown off from the egg but without any knowledge of the nuclear struc- ture. Finally the nuclear degeneration which is a very conspic- uous feature of oogenesis in certain groups whose oogonia are multinucleate (Peronosporales, Saprolegniales, Pelvetia, etc.) has been considered related to reduction phenomena. But the nuclei in all of these forms bear every evidence of being in each type homologous structures whose large numbers have a phylo- genetic raison d'etre and the extensive degeneration is associated with the principles of sexual evolution which tend to conserve protoplasm for the good of a lesser number of gamete nuclei even to the sacrifice of others that are potentially equivalent. We will now consider the few instances among the thallo- 454 THE A At ERIC AN NATURALIST. [Vou XXXIX. phytes in which a reduction of the chromosomes is reported just previous to or during gametogenesis. The best known case is Fucus since this type has been studied by three investigators : Farmer and Williams ('98) and Strasburger ('97a). They agree in describing the nuclear figure that differentiates the oogonium from the stalk cell as exhibiting a large number of chromosomes (28 or 30) while the three mitoses within the oogonium, which give rise to the eight eggs, present only one half that number (14 or 15). Apparently there is a reduction by one half just before the mitoses in the oogonium. Since there is no sporo- phyte generation in Fucus it is of course difficult to compare these conditions with those in higher plants, but, as will be explained later, there are some reasons why we should not expect to find reduction phenomena at gametogenesis in any thallophyte. Reduction phenomena at gametogenesis have also been sug- gested for various types of the Peronosporales and Saproleg- niales but not, however, in exactly the same way as in Fucus. There are always, as far as is known, one or two mitoses within the oogonium before the gamete nuclei are organized and it has been held that these are reduction divisions by Rosenberg for the Peronosporales and by Trow for the Saprolegniales. Rosenberg (: O3b) described for the oogonium of Plasmopara a condition of synapsis in the nuclei preceding the two mitoses and compared this sequence with the events of sporogenesis in higher plants in which the two divisions within the spore mother-cell are pre- ceded by a period of synapsis. Rosenberg did not determine the number of chromosomes in the vegetative nuclei so that he has no positive evidence of reduction in the oogonia. With respect to the two mitoses and the preliminary synapsis I have already pointed out in criticism of Rosenberg's studies (Bot. Gaz., vol. 36, p. 154, 1903) that the number of mitoses is variable in the oogonia of the Peronosporales and Saprolegniales and apparently entirely absent in the species of Vaucheria studied by myself (Davis, :O4a). Also, the phenomenon of synapsis, which is easily recognized in the large nuclei of the spore mother-cell, would be difficult to establish in the small nuclei within the oogonia of the forms mentioned above. Nuclei can be found No. 463.] STUDIES ON PLANT CELL. VI. 455 in a number of structures with their contents somewhat massed at one side or in the center but such conditions must not be confused with the remarkable process of synapsis Jn_the spore mother-cell. Among all the excellent studies of gametogenesis in the Peronosporales I cannot find any clear evidence of a re- duction of the chromosomes at gametogenesis. Quite different is the account that Trow (104) brings forward to support his view of chromosome reduction during gametogen- esis in the Saprolegniales. Trow describes two mitoses in the oogonium of Achlya debaryana : in the first the number of chro- mosomes is eight which becomes reduced to four in the second. Trow's account of a second mitosis in Achlya is very different in a number of particulars from the results of all investigations on gametogenesis in the Peronosporales and Saprolegniales. Two centrosomes with radiations are said to appear at the poles of the spindle at anaphase, structures which were not present in the first mitosis. Some of these asters become the center of the egg origins and are later accompanied by deeply staining material constituting a body which Trow terms an ovocentrum and which perhaps corresponds to a coenocentrum. Relatively few of the nuclei in the oogonium are said to pass through this second mitosis and some of their products, with the accom- panying asters, break down. The remainder become the func- tional gamete nuclei of the eggs. There are many complex activities described by Trow in connection with the appearance of the asters during the second mitosis and also at the side of the sperm nuclei which are said to enter the oogonium, events that cannot be correlated with the processes of gametogenesis and fertilization as we understand them for the Peronosporales. They are treated briefly in a review by myself (Bot. Gas., vol. 39, p. 6 1, 1905), where, however, I misunderstood a distinction that Trow draws between the aster and the ovocentrum (see an answer by Trow, Bot. Gaz., vol. 39, p. 300, 1905). My impres- sion is that either Trow has been mistaken in his interpretations or that there are present events which must entirely change our conception of gametogenesis in the Saprolegniales and Perono- sporales, but which are not fully explained by Trow's paper. Let us now think of gametogenesis among the thallophytes 456 THE AMERICAN NATURALIST. [VOL. XXXIX. with reference to what we know of the process in higher groups and the principles of the origin and evolution of sex and the sporophyte among the lower. It seems clear that the sporo- phyte generation is characterized by a double number of chromo- somes as a result of the fusion of gamete nuclei at fertilization. We must then lay the fundamental inception or origin of the sporophyte to the stimulus of the sexual act. That is, the sexu- ally formed fusion cell must have different potentialities from the germ plasm of the parent gametophyte and it cannot pro- duce a gametophyte again until these potentialities are worked off and the protoplasm returns to the dead level of the ancestral stock (the gametophyte). By the potentialities of the sporo- phyte plasm we mean primarily a greater energy or growth stimulus which must express itself differently from the gameto- phyte. Morphologically we can only distinguish sporophyte plasm from gametophyte plasm by the double number of the chromosomes but of course the complexities of the sexual act would make great differences in the chemical structure of the t\vo. The divergences in the history of the gametophyte and sporophyte, as shown throughout ontogeny and phylogeny, are but the final expressions of the different potentialities of the protoplasm in each generation. The -morphological forms of expression of the sporophyte are extraordinarily various and in the long evolutionary history of this generation have developed great structural differentiation but with every life history the sporophyte has the same beginning (fertilization, with the doub- ling of the chromosomes) and the same ending (sporogenesis, with chromosome reduction). Between the beginning and the end is intercalated a vegetative period, short and simple in some forms, and very long and elaborate in others. The history of the development of this vegetative period or the evolution of the sporophyte is a subject far outside of and secondary to the scope of this discussion. We are only concerned with the protoplasmic activities at the beginning (fertilization) and the end (sporogenesis) of the sporophyte generation. We know nothing of the behavior of the chromosomes in types of the thallophytes which illustrate most closely our con- ception of the origin of sex and of the sporophyte generation. OF THE UNIVERSITY No. 463.] STUDIES ON PLANT CELL. VI. 457 I refer to many lower algae such as Ulothrix, forms of the Volvocaceas, CEdogonium, Coleochaete, and many others. How- ever, the homologies of primitive gametes and their origin from types of asexual zoospores is very clear in a number of groups. We can see nothing in the morphology and mode of develop- ment of these reproductive cells to suggest reduction phenomena when gametes are produced. The primitive gamete is generally somewhat smaller than its homologue the zoospore, often because the protoplasm of the gamete mother-cell becomes distributed in a greater number of daughter elements. It is well known that the conditions that lead to conjugation are exceedingly variable, depending upon environmental factors and one often cannot tell at the time whether a swarm spore will show sexual habits or germinate without conjugation. The most satisfactory theory of the origin of sex in plants regards primitive gametes as weaker or lacking in certain potentialities of vegetative growth and the conjugation as a mutually cooperative process resulting in a rejuvenescence of the protoplasm. The fact that many simple types of gametes will germinate without fertilization and produce small and weak sporelings shows that vegetative possi- bilities are not entirely lost. Investigations on the chromosome history among these forms, difficult though they be, are some of the most interesting subjects of botanical research. We know some general principles of the origin and evolution of sex in plants (Davis, :oib, : O3a) but of the chromosome history in the simplest types of gametogenesis nothing is known. With respect to the history of the chromosomes in the sim- plest sporophytes we are also as ignorant as in the simplest types of gametogenesis. We have excellent reasons for believ- ing that the sporophyte generation is represented among the thallophytes in a number of very simple conditions. Numbers of zygospores and oospores (c. g., Ulothrix, CEdogonium, forms of the Conjugales and Volvocaceas, etc.) give rise on germination to several daughter cells. In higher forms this growth period is lengthened to the formation of a reproductive tissue (Coleochaete) and in the great groups of the Rhodophyceae, Ascomycetes, and Basidiomycetes there is present an extensive development from the fertilized female cell (or its equivalent when apogamy obtains) 458 THE AMERICAN NATURALIST. [VOL. XXXIX. involving the development of a vegetative structure before the period of sporogenesis. From the studies of Wolfe (: 04) we know that the sporophyte portion of Nemalion (the cystocarp) contains nuclei with double the number of chromosomes (about 1 6) present in the gametophyte (about 8) and that the period of chromosome reduction is apparently just previous to the devel- opment of the carpospores (sporogenesis). Williams (:O4a and b) has recently determined that the asexual plant of Dictyota is a sporophyte generation with double the number of chromosomes (32) found in the sexual plant (16). The reduction occurs here during a rather long period of preparation on the part of the nucleus in the tetraspore mother-cell and the reduced number appears in the two mitoses that form the tetraspores. These events closely parallel those in the spore mother-cell of higher plants and will be discussed further under " Sporogenesis." William's (: O4b) account of gametogenesis in Dictyota is the most complete that we have for any thallophyte. The oogonia and antheridia are cut off from a stalk cell by a mitosis which presents 16 chromosomes, the number characteristic of the gametophyte. The contents of the oogonium forms a single egg and consequently presents no mitotrc phenomena. The antheridium develops over 1500 sperms thus exhibiting a large number of successive divisions. These all show 16 chromosomes and the mitoses are typical, not differing in any essential from the division in the stalk cell. The entire absence of mitoses in the oogonium and the great number in the antheridium are striking facts which show that no especial significance can be attached to nuclear divisions within sexual organs of this type. There is no place for reduction phenomena within these sexual organs and none precede their development. These studies of Williams and Wolfe justify us in expecting that other thallophytes will support their discoveries that the product of the sexual act will have a fusion nucleus with double the number of chromosomes present in the sexual plant (game- tophyte) and that reduction phenomena may be expected to fol- low the sexual act and not precede it as in animals. In such thallophytes as have no sporophyte generation we may suppose, as Strasburger ('94a) suggested, that the number of chromo- No. 463.] STUDIES ON PLANT CELL. VI. 459 somes is reduced with the germination of the sexually formed cell so that the protoplasm returns at once to the potentialities of the gametophyte. It is quite possible that the_four zoo- spores produced from the oospore of CEdogonium and the four nuclei found in the germinating zygospores of the desmids and Spirogyra may indicate divisions concerned with reduction phenomena similar to those in the tetraspore mother-cells of Dictyota (which may also be expected in the tetraspore mother- cell of the red algae) and in the spore mother-cell of the higher plants. For these reasons we seem to be justified in taking a critical attitude towards the accounts of chromosome reduction at game- togenesis among the thallophytes. The logic of the situation would lead us to expect that every sexual act gives a doubling of the chromosomes and an impulse towards the development of a sporophyte phase in plants which must be worked off before the protoplasm is in condition to reproduce the parent gameto- phyte. Reduction phenomena should follow then every sexual act. If it takes place immediately with the germination of the sexually formed cell there is of course no sporophyte generation. Because the conception of the sporophyte generation with reduc- tion of the chromosomes at sporogenesis is so clearly established in higher groups, those investigators who claim reduction phe- nomena at gametogenesis must expect their views to be severely scrutinized and accept the responsibility of presenting very clear and convincing proof of their conclusions. The author does not think that this evidence is supplied in satisfactory form by any investigation so far. 2. FERTILIZATION. In Section IV of these " Studies " we described the most important phenomena of fertilization under the caption " Sexual Cell Unions and Nuclear Fusions." It will not be necessary to discuss the facts of the phenomena in detail again. This account will take up the more theoretical aspects of the events of ferti- lization and their relation to other critical periods of ontogeny. Plants are in complete agreement with animals in the follow- 460 THE AMERICAN NATURALIST. [VOL. XXXIX. ing chief events and principles of fertilization. Thus Van Beneden's conclusion of 1883 that sexual nuclei are equivalent in their chromatin content at the time of fusion irrespective of differences in size is admirably borne out by Miss Ferguson's (: 04) studies on the pine. In this form as in the gymnosperms generally the male nucleus is much smaller than the female and comes to lie in a depression in the latter before the actual fusion takes place. After the fusion the paternal and maternal chromo- somes are found in two groups side by side preparatory to the first cleavage mitosis and are indistinguishable except for their position ; the chromatin of the two sexes is equal in amount as far as can be seen. Then the observations of the Hertwig brothers, in 1887, and Boveri, in 1889 and 1895, that the sperm nucleus could enter and cause the development of denucleated eggs or their fragments thus taking the part of a female nucleus in parthenogenesis, were established for plants by Winkler's (:oi) experiments on Cystoseira. Winkler was able to divide the egg of this brown alga into a nucleated and a non-nucleated portion and he found that sperms entered the non-nucleated parts and caused them to develop sporelings side by side with the fertilized nucleated portions. The sporelings from the non-nucle- ated fragments, controlled by the sperm nuclei alone, developed about half as rapidly as those from the originally nucleated por- tions which of course were dominated by sexually formed fusion nuclei, but the two sets of sporelings were alike in form as far as they were grown. Only with respect to Boveri's celebrated theory that the sperm brings to the egg in the centrosome the mechanism of cell division, do plants fail to support the conclu- sions of certain zoologists with respect to the most important events of fertilization. This point upon which zoologists are not in full accord will be discussed later. There is general agreement in the view that the male nucleus of plants supplies chromosomes equal in number and equivalent quantitatively to the female, and general accord in the conclusions that the chro- mosomes by their individuality, apparent permanence of struc- ture, and fixed behavior must be bearers of hereditary characters. Evidence from the most recent investigations upon favorable forms of both animals and plants indicates that the chromosomes No. 463.] STUDIES ON PLANT CELL. VI. 461 from both gametes maintain their independence and never fuse at the immediate time of fertilization. We have reason to assume, chiefly from zoological studies, that -the_patern.al and maternal chromosomes of plants remain independent throughout the entire sporophyte generation and that no fusion takes place until the period of chromosome reduction at sporogenesis. If no sporophyte generation is present we should expect the fusion and reduction of the chromosomes to occur after the sexually formed cell had passed through a period of rest (for all reduction phenomena seem to require considerable time) unless there be actually such reduction during gametogenesis in the thallophytes as reported for Fucus and Saprolegnia. The morphology of the chromosomes is probably unchanged by the immediate act of fertilization. The fusion nucleus simply contains double the number of chromosomes present in each gamete nucleus which increases by so much the metabolic possibilities which lie in these structures. Besides chromatin the sperm brings into the egg a certain amount of cytoplasm. Some of this may be the substance of the blepharoplast or other kinoplasm associated with the nucleus but there is often besides considerable granular trophoplasm, sometimes with inclusions of starch and other food substances, and the male gamete of certain thallophytes contains a chroma- tophore. There is no reason to suppose that development especially characteristic of fertilization, the sporophyte genera- tion, has any relation to this trophoplasm with its food inclusions, excepting as it may stimulate growth which is to be expected whenever organic food material is introduced into protoplasm. But we can hardly believe that the formative elements or the rudiments of further development especially those of a sporo- phytic character lie in this region of the protoplasm. They must be sought in the nuclei and in the only stable elements of the nuclei, the chromosomes. It has been held at times by botanists, following the lead of certain zoologists, that the sperm or sperm nucleus introduced a centrosome into the egg which organized the first cleavage- spindle and thereby played a necessary part in starting cell division. Such a centrosome would naturally be sought in the 462 THE AMERICAN NATURALIST. [VOL. XXX IX. blepharoplast which is clearly analogous to the middle piece of the animal spermatozoon. We have no evidence that such events ever take place in the eggs of plants. On the contrary we know that the first cleavage-spindle in the eggs of spermatophytes develops without centrosomes from a mesh of fibrillae. Also the blepharoplasts of the gymnosperms Cycas, Zamia, and Ginkgo remain in the cytoplasm at a distance from the fusion nucleus and Shaw's account of the fern, Onoclea, indicates that similar conditions obtain there. We know less about the history of the blepharoplasts within the egg of thallophytes where the first cleavage-spindle frequently has very handsome centrospheres and asters (e. g., Fucus and Dictyota). Strasburger ('97a) pointed out that one of the asters of the first cleavage-spindle in Fucus arose near the point where the male nucleus united with the female. However, Farmer and Williams ('98) believe that centrospheres of the first cleavage-spindle in Fucus are formed de novo and Williams (: O4b) came to the same conclu- sion for Dictyota. There are some very interesting features in the comparative study that Williams (: <34b) has made on the development of the first segmentation spindle in the fertilized and parthenogenetic eggs of Dictyota. The spindle in the par- thenogenetic egg is multipolar and develops from an intranuclear kinoplasmic mesh and there are no centrospheres. But in the fertilized egg a centrosphere always appears at the side of the nucleus and apparently divides into two which separate until they lie at opposite poles of the mature spindle. Yet Williams after a very careful study concludes that this centrosphere arises de novo and believes that the stimulus of fertilization enables the fusion nucleus to form a centrosphere external to itself, a thing which is not possible for the nucleus of the parthenogen- etic egg. It seems then probable that the only structures of the sperm that preserve their morphological entity in the fertilized eggs of plants are the chromosomes. Whatever may be the relation of the blepharoplast and other cytoplasmic structures as stimuli to the development of the egg they cannot be regarded as fixed factors in the problem of heredity. It is very probable that they introduce valuable food material, perhaps important fer- No. 463.] STUDIES ON PLANT CELL. VI. 463 ments, substances of great service, although possibly not abso- lutely necessary to the successive metabolic processes which characterize growth and development. But the _fact remains that we have in the chromosomes the only new morphological elements. And the progress of research seems ever to strengthen the general view that in the chromosomes are contained the directive rudiments of development and that they are the bearers of hereditary principles. Nuclear studies on apogamous forms will certainly prove of great interest in this connection. We have reason to expect some very important results from thorough cell studies on apogamy and apospory. The best developed theory of fertilization in plants is that of Strasburger and a statement of his views should precede any comments of other authors. Strasburger has written much on the phenomena of fertilization ; important considerations may be found in his papers of '94a, b, '9/c, : ooa, b, :oi, and : O4a. Strasburger points out that the protoplasm of the egg is pre- dominately trophoplasmic in character because of the propor- tionately very large amount of cytoplasm with granular inclusions that are evidently food material or the products of metabolism. On the other hand the cytoplasm of the sperm contains rela- tively little trophoplasm and much kinoplasm, especially when the sperm is a ciliated cell with a large blepharoplast. As Strasburger conceives kinoplasm to be the active substance of spindle formation, -he concludes that the sperm might bring to the well nourished egg, rich in trophoplasm, the substance neces- sary to start the mechanism of mitosis. In its broad aspects this view is very similar to the celebrated theory of Boveri, 1887, that the spermatozoon supplied the animal egg with the centro- some which is conceived as necessary to start mitotic processes and that the egg is powerless to divide before fertilization because it lacks such a structure. Another feature of Strasburger's views (advanced in his paper of : oob) appears to have grown out of the discovery of the so called "double fertilization " in the embryo-sac and other nuclear fusions whose sexual significance is not clear, together with the phenomena of parthenogenesis as produced experimentally in many studies of recent years. Strasburger considers that two 464 THE AMERICAN NATURALIST. [VOL. XXXIX. processes are involved in the sexual act. The first, termed "vegetative fertilization," is simply the stimulus to growth which results from the fusion of two nuclei or other masses of protoplasm. The second, called "generative fertilization," in- volves deeper factors than those of mere growth stimulus. These lie in the union of germ plasm of diverse parentage with the mingling of hereditary racial characters and individual varia- tions and the establishment of a new organism which may have possibilities of development quite different from the parent form. The effects of " vegetative fertilization " may be imparted to protoplasm artificially by chemical and physical stimuli as has been clone in the numerous experiments of Klebs and Loeb on the conditions which induce parthenogenetic development. "Generative fertilization" has a phylogenetic significance and a background which is entirely apart from the mere vegetative- processes of cell growth and division. It is apparent that Strasburger's theory is open to the same line of criticism that has been brought against the universal application of Boveri's hypothesis that the spermatozoon brings to the egg the agent of cell division as a centrosome. The investigations of several zoologists indicate that one or both of the centrosomes in the first cleavage-spindle may be derived from the egg or may be formed de novo (see Wilson, : oo, pp. 196, 208). The kinoplasm of the plant sperm, whether in the form of a blepharoplast or as an ill defined accompaniment of the sperm nucleus has not been shown to take part in the forma- tion of the first cleavage spindle. There is no evidence that the blepharoplast retains its organic entity in the egg to pass over into a centrosome or centrosphere. Of course the kinoplasm which lies immediately without the nuclear membrane of the sperm, and there is sometimes a conspicuous amount of this densely granular protoplasm, must merge with similar kinoplasm associated with the egg nucleus at the time of fusion. For example Miss Robertson (:O4) and Coulter and Land (105) note in Torreya that the sperm nucleus brings to that of the egg a large amount of accompanying kinoplasm which forms an investing layer around the fusion nucleus. It is reasonable to suppose that the mixing of these masses of kinoplasm with the No. 463.] STUDIES ON, PLANT CELL. VI. 465 fusion of the gamete nuclei would give material for a larger and more highly differentiated nuclear figure in the first cleavage of the egg. Williams' ( : 040) observations and conclusions on Dictyota are especially interesting in this connection for he shows that the first cleavage-spindles in the parthenogenetic eggs are intranu- clear and multipolar, showing no dominant kinoplasmic centers while the fertilized eggs form each a well differentiated centre- sphere with radiations, exterior to the nuclear membrane, which clearly guides the whole process of spindle formation. Williams does not hold that this centrosphere comes as an organized struc- ture from either sperm or egg but is developed de novo by the fusion nucleus as the result of the general stimulus of fertiliza- tion. The evidence, then, furnished by studies on fertilization in plants, indicates that the chromosomes alone maintain mor- phological independence throughout the process of fertilization and that the kinoplasmic (archoplasmic) elements play no part in the phenomena as fixed morphological structures but simply con- tribute their substance to the general union of cytoplasm with cytoplasm, and that any specialized kinoplasmic structures of the first cleavage spindle are formed de novo. While it is true that the sperm brings to the egg much kinoplasm it may well be questioned whether such kinoplasm is a necessary factor in the formation of the first cleavage-spindle. It seems more proba- ble that the development of achromatic structures in the first mitosis following fertilization is due rather to the general stimu- lus of cell and nuclear fusion than to particular structures sup- plied by either sperm or egg. The second phase of Strasburger's theory of fertilization con- cerns a separation of the two processes in the sexual act : ( i ) the mere growth stimulus, "vegetative fertilization," that may be expected with the union of any two masses of protoplasm, and (2) the clearly defined sexual phenomena, "generative fer- tilization," which lies in the union of germ plasm of different parentage and diverse potentialities and which leads to the inheritance of these characteristics. It seems clear that the two processes are really present and can be clearly distinguished. But it may be strongly questioned whether the factors charac- 466 THE AMERICAN NATURALIST. [Vou XXXIX. terizing the first (vegetative fertilization) should really be con- sidered a part of the sexual act. Strasburger regards the proc- esses of "generative fertilization" as essential to the sexual act. The growth stimulus "vegetative fertilization" is always to be expected as an accompaniment of fertilization. It may be given to 'cells in other ways than by the sexual act and is found in cell and nuclear fusions which for phylogenetic reasons are plainly not sexual. The experimental work of recent years on the conditions determining artificial parthenogenesis have done much to define the sorts of factors which stimulate growth and division of sexual cells when the process of fertilization is suppressed. Klebs for plants and Loeb for animals have been foremost in these studies and they have shown that what seem to be very minor changes in the environment of the sexual cell may suffice to give a gamete the power of immediate development without fertilization. Thus the egg of the sea urchin will develop parthenogenetically to an advanced stage when placed for a short time in sea water contain- ing magnesium chloride and then brought back to normal sea water. Nathansohn ( : oo) found that a small proportion (about 7 Jfe ) of the eggs of Marsilia vestita would germinate partheno- genetically when the megaspores were cultivated for 24 hours at the rather high temperature of 35 C. arid then left to continue their development at 27 C. There are then a number of fac- tors such as varying osmotic pressure, temperature, and in some cases chemical reagents which may induce gametes to further development without the usual sexual processes. These reac- tions seem to be of a similar character to the processes in that phase of sexual reproduction termed "vegetative fertilization" by Strasburger. They give the stimulus to growth but without that essential feature of sexuality, the mingling of germ plasm of different parentage which distinguishes the processes of " gener- ative fertilization." It seems to the author, for the sake of clearness, that we are trying to include too much under the term fertilization. If the features of "vegetative fertilization," i. e., the growth stimulus, can be introduced experimentally as in artificial parthenogenesis then they cease to be fundamental qualities of the sexual act. No. 463.] STUDIES ON PLANT CELL. VI. 467 They are accompaniments of sexual processes which may always be expected but nevertheless are not the essential characteris- tics. The essence of the sexual act (fertilization) is the union of germ plasm with such possibilities of new developments as come from the inheritance of mixed characters from two lines of ances- try. And the more diverse and complex are the characters of the parents the more conspicuous are the essential features of the sexual act. Among lowly organisms and in simpler types of sexual processes the growth stimulus becomes exaggerated in our attention because the features of heredity are not so promi- nent as in the higher forms. But in the higher groups the varied characters of offspring express clearly the subtle factors concerned with the mingling of diverse germ plasm in the proc- ess of fertilization and the growth stimulus recedes into the background. For these reasons it seems to me that the term fertilization should only be used for the mingling of germ plasm with the possibilities of new combinations in the potentialities of the resulting sexually formed cell and that the growth stimulus should be treated as an accompaniment but quite apart from the essen- tials of the sexual act. And for these reasons I was careful to include in Section IV under the caption " Sexual Cell Unions and Nuclear Fusions " only illustrations in which the sexual nature of the phenomena was clearly shown by applying a mor- phological or phylogenetic test to the elements concerned in the process of cell fusion. The phylogenetic test seems to me the only sure way of determining the sexual nature of the members of a cell fusion and there are very few cases in which there can be any hesitation in deciding whether or not such elements are morphologically gametes. I included under "Asexual Cell Unions and Nuclear Fusions" in Section IV a number of cases in which the sexual nature of the act is under dispute for the reason that none of these satisfy the phylogenetic test. It is perfectly clear that the growth stim- ulus is a conspicuous feature of these cell and nuclear fusions and that in this feature they resemble sexual processes but this does not, to my mind, make them acts of fertilization or the equivalent of sexual processes. The union of sporidia in the 468 THE AMERICAN NATURALIST. [VOL. XXXIX. smuts and of yeast cells, the fusion of nuclei in the teleutospore and basiclium and in the apogamous development of ferns, the double fusion of polar nuclei and multiple nuclear fusions in the embryo-sac (Corydalis) illustrate phenomena which I cannot regard as sexual even though they have in them elements asso- ciated with sexual processes and in certain caseS may be substi- tutes for a former sexual act. In none of these instances can we be positive that the nuclei concerned are morphologically and phylogenetically gamete nuclei. This point was discussed in some detail in Section IV. It seems- to me that Blackman's (:O4a, p. 353) conception of the cell fusions preceding the geci- dium in Phragmidium as "reduced forms of ordinary fertilization" or Farmer's (103) explanation of apogamy in the fern "as a kind of irregular fertilization " leads to a confusion of a substitute process with a true sexual act. The substitute processes have their true place as phenomena of apogamy. They can, however, only have a sexual significance if they represent the origin of a new set of gametes in the organism, a proposition which is not likely to be maintained by anyone. 3. SPOROGENESIS. We are employing the term sporogenesis, as must have been apparent in preceding divisions of this paper, to designate a characteristic and highly specialized type of spore formation that is universal in all plants above the thallophytes. The process always terminates the sporophyte phase in ontogeny of these higher plants, and is especially distinguished as the period of chromosome reduction in the life history. The cell activities of sporogenesis are therefore of particular interest, and, since spore mother-cells are generally large and their nuclear and cytoplasmic structure especially clearly differentiated, we have perhaps ob- tained more knowledge of mitotic phenomena from the study of these elements than of any other tissues of the plant body. The reduction phenomena of sporogenesis have been estab- lished in some forms of the thallophytes, certainly in the tetra- spore mother-cell of Dictyota (Williams, : O4a). There are also reasons for suspecting that the oospore of CEdogonium and the No. 463.] STUDIES ON PLANT CELL. VI. 469 zygospores of Conjugates on germinating present similar events. The teleutospore and basidium are probably also the seat of chromatin reduction (Blackman. : O4b) in the formation of spores either directly or through the promycelium. The ascus holds a position at the end of a sporophyte phase which suggests a similar relation in this group of fungi. Chromosome reduction may also be expected in the tetraspore mother-cell of the Rho- dophyceae, as in Dictyota, but this subject has never been inves- tigated. There are occasional red algae in which the tetraspores are sometimes borne on the same plant with the sexual organs, conditions which may be difficult to explain on the theory that the tetrasporic plant is a sporophyte. Thus Spermotliainnion turneri on the American coast frequently bears both procarps and tetraspores on the same plant, and I have also seen cysto- carpic plants of Ceramium rubrum some of whose branches con- tained tetraspores. Lotsy (: O4a) also reports similar conditions in Chylocladia kaliformis. The other extremely varied methods of spore formation (zoospores, conidia, etc.) in the thallophytes do not concern the present discussion. They seem to have no fixed place in the life history and there is nothing to indicate any relation to reduction phenomena, although we actually know nothing about the chromosome history among these lowly forms. The importance of sporogenesis as a critical period in the life history of higher plants became at once apparent with the dis- covery that fertilization doubled the number of chromosomes in the nuclei of the sporophyte phase and that the double number was reduced during sporogenesis. As stated in our account of gametogenesis, these facts were first established for a number of spermatophytes by the work of Strasburger ('84, '88, and '94), Guignard ('84, '85, and '91), and Overton ('93 a and b). Guig- nard ('91) presented for Lilinm martagon the first complete account of the number of chromosomes in the life history of a plant, and his results were also established independently by Overton ('93 a and b). Then followed confirmatory investiga- tions among the bryophytes in the work of Farmer ('94, '95 a, b, c) and in the pteridophytes by Strasburger ('94, p. 294) for Osmunda. Since 1895 the investigations among the spermato- phytes have so multiplied that we know the number of chromo- 470 THE AMERICAN NATURALIST. [VOL. XXXIX. somes in sporophyte and gametophyte for more than fifty forms. This list may be found in Coulter and Chamberlain's recent text-book, The Morphology of the Angiosperms, 1903, p. 81. Farmer's accounts of the number of chromosomes in the Hepaticae have been confirmed and extended by myself (Davis, '99, :oia) and by Moore (:O3). The more recent literature, especially as it concerns the events of spindle formation in the mitoses char- acteristic of sporogenesis has been treated in our account of the spore mother-cell (Amer. Nat., vol. 38, p. 725, Oct., 1904). There are two chief periods in the processes of sporogenesis as illustrated in all groups above the thallophytes : ( i ) a growth period and (2) a period of cell division. In the growth period the spore mother-cells become differentiated from the general sporogenous tissues through a great increase in the amount of protoplasmic material. At some time in this growth period the nucleus of the spore mother-cell exhibits the phenomenon of synapsis, a very characteristic event, recognized by the very much contracted condition of the chromatin network in the interior of the nucleus. Synapsis is believed to hold fundamental relations to reduction phenomena as the time when chromosomes unite with one another in pairs. The period of cell division fol- lows synapsis and is characterized by two mitoses in the spore mother-cell, the second following immediately upon the first, and a segmentation of the protoplasm, sometimes by two successive divisions, and sometimes by a simultaneous cleavage, into four spores. The two mitoses present certain peculiarities in the structure and behavior of their chromosomes which are unlike the events of typical mitoses. The first is known as the hetero- typic and the second as the homotypic mitosis. These peculiar- ities have been recognized for a long time and have furnished the subject of much investigation and contradictory explanations. They were briefly described in Section III (Amer. Nat., vol. 38, p. 740, Oct., 1904) but recent studies of Farmer and Moore (: 03, : 05) have opened again a discussion which seemed closed at that time. The details of synapsis and the heterotypic and homotypic mitoses will be taken up under the caption, " Reduc- tion of the Chromosomes." Contrary to a statement in Section III of these studies (Amer. No. 463-] STUDIES ON PLANT CELL. VI. 471 Nat., vol. 38, p. 726, Oct., 1904) there is probably a deep sig- nificance in the fact that two mitoses are almost universally present in the spore mother-cell so that four spores are formed. It is probable that these mitoses are always heterotypic and homotypic, although this fact has only been clearly established in comparatively few favorable forms, and that they are indis- pensable to the mechanism of reduction phenomena. The latest accounts describe the first mitosis as the separation of the two portions of a bivalent chromosome, that is of two chromosomes joined either side by side or end to end, giving it a unique posi- tion among the mitoses of the life history. According to these theories the two mitoses of sporogenesis are features of a remarkable mechanism by which the paternal and maternal chromatin after its union in synapsis may become distributed in proportions that can be expressed by mathematical ratios. The peculiarities of the homotypic mitosis depend on a premature fission of the daughter chromosomes of the heterotypic division as will be explained in the next portion of this section. Thus the four spores are the result of these peculiar mitoses and have morphological significance. We are even justified in suspecting that the groups of four spores when found in the thallophytes, as the tetraspores of Dictyota and the red algae, the four spores formed on the basidium and promycelium and the four spores of nuclei present in the germinating oospore and zygospore of QEdogonium and the Conjugates indicate the presence of reduc- tion phenomena simply because the number four is so constant. Williams (: O4a) for Dictyota and Blackman (: O4b) for types of the Uredinales have discovered clear cytological evidence of this reduction phenomenon but we know nothing of the chromosome history in other types. We have already referred to the fact (Section III, Amer. Nat., vol. 38, p. 743, Oct., 1904), that in the spermatophytes the two mitoses characteristic of sporogenesis are very close to the mitoses which differentiate the gamete nuclei. In the male gametophyte of the Angiosperms there are generally only two mitoses between the events of sporogenesis and gametogenesis and in gymnosperms there is a somewhat larger and variable number. The female gametophyte of the angiosperms usually 472 THE A ME RICA N NA TURA US T. [VOL. XXX I X . presents three mitoses after those of sporogenesis before the egg nucleus is formed. But in a number of types in the lily family (e. g., Lilium, Tulipa, Fritillaria, Erythronium, etc.), the mitoses of sporogenesis are actually included in the embryo-sac and the very next mitosis, which is typical, differentiates the egg (see Section III, Amer. Nat., vol. 38, pp. 741-744, Oct., 1904). This is the furthest point attained in the reduction of the gametophyte which in such forms actually includes but a single nuclear division in its history. But however close the mitoses of sporogenesis come to those of gametogenesis it is perfectly clear through the long phylogenetic history in the lower spermatophytes, pteridophytes, and bryophytes that the two are morphologically distinct processes and are always sepa- rate. It is unfortunate that the terms spermatogenesis and oogenesis should be applied to processes of sporogenesis as has been done by several authors, for such usage involves a confusion of two events which phylogeny clearly shows to be different in origin and to have back of them the diverging history of sporo- phyte and gametophyte from the times of thallophyte ancestry, the most remarkable evolutionary history in the plant kingdom. It is conceivable that some plants may finally reach a stage in their evolutionary history when all the gametophytic mitoses in the pollen grain and embryo-sac will be suppressed and the nuclei resulting from sporogenesis become gamete nuclei. But it is clear that in such an event the gametophyte phase would be obliterated and we should have an entirely new type of life history. There would then be only one organism (derived from the sporophyte) whose gametes would be formed immediately with the differentiation of the pollen grain and embryo-sac. Such an organism would present reduction phenomena with the differentiation of the gametes and its type of life history would be identical with that of animals. We should look for such a reduced life history in groups related to forms in which the mitoses of sporogensis are included in the embryo-sac and the gametophyte phase is represented by a single nuclear division (e. g., Lilium, Tulipa, Fritillaria, Erythronium, etc.). Search among some of the most highly specialized Monocotyledon?e may actually reveal examples of the complete suppression of the female gametophyte. No. 463.] STUDIES ON PLANT CELL. VI. 473 The speculative possibilities of a suppression of a sexual gen- eration and the assumption of sexuality by an asexual phase were clearly in the mind of Strasburger when he suggested ('9/j.b, p. 852) the possibility that the two mitoses characteristic of oogenesis and spermatogenesis in animals might signify the remains of a former sexual generation now entirely suppressed in the Metazoa. This suggestion was based on the striking similarity of the events of sporogenesis in plants to those of gametogenesis in animals and on the history of sporogenesis as shown in plant phylogeny. This history is remarkably clear and there can be no question but that the phenomena of sporogenesis have developed as the result of sexual processes and are always associated with an asexual generation (sporophyte). It is also clear that the ancestral primitive sexual generation (gametophyte) h:is steadily degenerated until now it is almost lost in such embryo-sacs as include the two mitoses of sporogenesis within their history. If the sexual generation should become entirely lost the life history of a higher plant would present the same features with respect to the period of chromosome reduction as that of an animal : there would be but one organism, the homo- logue of the sporophyte which would produce gamete nuclei with reduction phenomena previous to gametogenesis just as in- ani- mals. Several authors have expressed views similar to Stras- burger's suggestion ('94-b, p. 852) or carried the speculation even farther than he. Beard ('95a, p. 444) along these lines of argu- ment combined with conclusions from Bower's ('87) studies on apospory, announced a belief that " Metazoan development was really bound up with an antitJietic alternation of generations." Lotsy (:O5, p. 117) expresses unequivocally the view that the animal body represents an asexual phase (2x generation) and that the sexual phase (x generation) is confined to the sexual cells. Chamberlain (:c>5) simultaneously with Lotsy and in much greater detail presents a comparison of the phenomena of sporo- genesis in plants with gametogenesis in animals tracing the resemblance in the events of chromosome reduction step by step and states his belief that "animals exhibit an alternation of gen- eration comparable with the alternation so well known in plants." This is not the place to consider this theory in detail from a 474 THE AMERICAN NATURALIST. [You XXXIX. zoological standpoint since it bears only indirectly upon the material of these papers. Zoologists have, however, discussed critically Strasburger's suggestions (see Wilson, : oo, p. 275, and Hacker, '98, p. 101). The difficulties of accepting this view of a possible antithetic alternation of generations in animals seem insurmountable. In the first place there is not the slightest evidence of antithetic alternation of generations in the Metazoa or for that matter anywhere in the animal kingdom. The examples of alternation of generations which the zoologists present among the Ccelenterates are all illustrations of homolo- gous generations derived from buds. There is no indication of spore formation comparable to the process in the higher plants, so far as I am able to judge, in any group of animals. And also there seems to be accumulating evidence of reduction phenomena previous to the development of sexual cells in the Protozoa which is essentially of the same character as in the Metazoa (see Wilson, : oo, pp. 227, 277, and Calkins, :oi, p. 233). It is very interesting and remarkable that reduction phenomena should show the same order of events in animals and plants and the facts should be clearly recognized. But I cannot follow those botanists who carry over to the animal kingdom the phylogenetic conclusions which are so clear in plants. The remarkable agreement of the events of sporogenesis in plants with gametogenesis in animals appears to me likely to prove only another illustration of similar biological phenomena which have evolved independently of one another, an illustration com- parable with the independent origin of sex, of heterospory, and probably even of the sporophyte generation itself (involving the processes of sporogenesis) in various groups of the plant king- dom. We have considered this comparison of reduction phenomena in plants with animals chiefly to emphasize the clear cut mor- phology of the process as understood by the botanist. It does not matter how close the events of sporogenesis may come to those of gametogenesis in the higher angiosperms, the whole background of plant phylogeny, which is wonderfully clear as a whole, shows that reduction phenomena are the product of the asexual generation. It represents, as Strasburger has so well No. 463-] STUDIES ON PLANT CELL. VI. 475 expressed it ('94a, p. 288), a return on the part of the plant organism in each life history to the condition of an ancestral sex- ual generation (garnet ophyte). Reduction phenomena in them- selves are not the result of a gradual evolution, whatever may be the complicated history of the sporophyte generation, for they consist always in the sudden reappearance of the primitive number of chromosomes, characteristic of the generation in which sex arose (gametophyte). The cause of reduction phe- nomena is phylogenetic. The interval that may separate this phenomenon from the responsible sexual act varies immensely in the plant kingdom according to the evolution of the groups con- cerned. But the suddenness of the appearance of sporogenesis tells in every case the same story of an immediate and total change in the potentialities of the protoplasm in the spore mother-cell, a change which can only be understood as a phylo- genetic process deeply seated in the race. When the events of sporogenesis in plants are considered as processes of spermatogenesis or oogenesis we disregard the most remarkable historic outlines that plant phylogeny can present, to the confusion of clear thought. Botanical science may well be proud of its achievement in outlining with such exactness the relations that the critical periods of gametogenesis, fertilization, and sporogenesis bear to reduction phenomena and too great stress can hardly be laid upon the importance of the results. 4. REDUCTION OF THE CHROMOSOMES. There are perhaps no activities of the cell which have been the subject of more investigation and discussion than those of chromosome reduction in animals and plants. The reasons are clear. The events of gametogenesis in animals and of sporogen- esis in plants have the deepest significance for an understanding of the organization of protoplasm because these are periods when great changes are made evident in the structure of the cells con- cerned and at the same time in their potentialities. We are forced to conclude that some of the structural changes at least are the cause of the new potentialities and the attempt to estab- lish the cause and effect has been one of the most fruitful and 476 THE AMERICAN NATURALIST. [VOL. XXXIX. interesting subjects of cell research. Reduction phenomena also have a deep phylogenetic significance whose history in plants at least can be traced with a remarkable degree of exactness. We are confident that sporogenesis in plants signifies the sud- den return of the organism to the condition of an ancestral sexual generation with the reappearance of a primitive number of chromosomes. The short time consumed in the process and the details and precision of the cell activities show that we are dealing with phenomena whose complicated mechanism can only find explanation in a long phylogenetic history. In the study of reduction phenomena and fertilization we have reached the con- clusion that the chromosomes are intimately concerned with the transfer of hereditary qualities and are probably the chief or even the sole bearers of these characters. And thus we enter upon some of the most far reaching problems of biology, those of heredity, hybridization, and the basis for the remarkable ratios of inherited characters which Mendel first clearly set forth. It seems quite certain for both animals and plants that numer- ical reduction of the chromosomes takes place through an asso- ciation of the paternal and maternal chromosomes in pairs to form the reduced number of bivalent chromosomes (dyads). We have presented in Section IV (" Sexual Cell Unions and Nuclear Fusions ") the evidence which indicates that paternal and maternal chromosomes do not unite at the immediate time of nuclear fusion in fertilization. On the contrary, in all higher animals and plants the paternal and maternal chromosomes are believed to remain separate throughout the long series of cell divisions in the new generation up to the time of sporogenesis in plants and gametogenesis in animals, both events being characterized by reduction phenomena. The fusion of the chromosomes takes place in the growth period which differentiates the spore mother- cell in plants from the archesporium or the primary gametocyte in animals from the preceding gametogenous tissue. The growth period is one of general protoplasmic accumulation and increase in the chromatin content of the nucleus, and is especially char- acterized by that peculiar activity in the nucleus termed synap- sis. Evidence is accumulating that synapsis is the characteristic No. 463.] STUDIES ON PLANT CELL VI. 477 feature of that period when the number of chromosomes is reduced by half. Synapsis is followed very shortly by the two mitoses charac- teristic of sporogenesis. These nuclear divisions ~have given rise to a lengthy literature in which well known investigators have shifted their positions more than once. The discussions have centered on the methods of fission and distribution of the reduced number of bivalent chromosomes which appear in the first mitosis following synapsis. Assuming that the chromatin is organized into smaller units, represented by the chromatin granules (chromomeres, Fol, 1891), which compose the chromo- somes, it is at once apparent that these finer elements may become variously distributed according to the structure of the bivalent chromosomes and the character of the mitoses of sporo- genesis. Each fusion bivalent chromosome is composed of two chromosomes joined (i) end to end or (2) side by side or (3) it is possible that the chromatin is intricately mixed in the struc- ture. With respect to the mitoses a transverse division of the fusion chromosomes might be expected to give a very different proportionate arrangement of the maternal and paternal chroma- tin from longitudinal divisions. Should the chromatin granules differ qualitatively from one another then different parts of a chromosome might be expected to have different characteristics ' which would be distributed by the mitoses of sporogenesis in various proportions or ratios. It has long been known that the mitoses of sporogenesis pre- sent peculiarities in the mode of division and arrangement .of the chromosomes at the nuclear plate which make them unlike the typical mitoses of cell division. These peculiarities have led to the designation of the first mitosis as heterotypic and the second as homotypic, terms which are now applied by both bot- anists and zoologists although we have now a much more extended knowledge of each type than when Flemming first proposed the classification in 1887. We described the charac- ters of the heterotypic and homotypic mitoses in Section III, "The Spore Mother-cell" (Atner. Nat., vol. 38, p. 740, Oct., 1904), and will presently treat them further since some papers of the past year have opened again a discussion which seemed 478 THE AMERICAN NATURALIST. [VOL. XXXIX. closed a few months ago. The chief points of issue in dis- cussions of reduction phenomena have centered around the sig- nificance of the heterotypic and homotypic mitoses. A typical mitosis is believed to present merely a quantitative division of each chromosome into two halves equivalent in their potentiali- ties. The evidence for this view lies in the longitudinal fission by which each chromatin granule on the spirem is supposed to divide and contribute half of its substance to each daughter chromosome. Can there be a qualitative division of a chromo- some by which one of the parts differs in character from the other, and are there such divisions at the time of sporogenesis in plants and gametogenesis in animals when reduction phenom- ena take place ? These have been the chief topics of dispute in studies of this character for two decades. The problem then ultimately concerns the structure of the chromosome and the reason for the constant reappearance of the number characteristic of the species at the beginning of each new gametophyte generation. All the prominent theories of heredity assume that the chromosomes are made up of simpler elements which stand for characteristics of the race. These may form various combinations of higher orders and collectively give the qualities of germ plasm. The simplest members that can be observed in such a series of structures are the chromatin granules (chromomeres) which may be found at almost all times in the nucleus and are especially conspicuous when arranged in a row on the linin thread of the spirem. Weismann has devel- oped the most complex conception founded on the above princi- ples and with the most elaborate terminology. Starting with the chromatin granule, which he named an id, Weismann assumed that this element is composed of still smaller structures called determinants and biophores, the last being the ultimate living units. Groups of ids make up idants or chromosomes. The id was conceived to possess all the essential characters of the specific germ plasm concerned but ids vary somewhat among themselves, determining thus the individual variations of the species. Therefore a chromosome or idant will have a varying structure according to the character and distribution of the ids which compose it. No. 463.] STUDIES ON PLANT CELL. VI. 479 When a chromosome divides longitudinally so that each id splits in half, the daughter chromosomes are exactly equivalent and the division of the chromatin is merely quantitative. But should a chromosome divide transversely then two sets of entire ids would be separated from one another and the two daughter chromosomes would differ in proportion as their component ids varied, i. e., the division of the chromatin would be qualitative. These conceptions of the possible structure and mode of division of chromosomes outline the basis of Weismann's theory of heredity and will serve to illustrate the general attitude of those biologists who approach the subject from the standpoint of pre- formation, although none have cared to formulate such elaborate assumptions as Weismann. However, there is a general agree- ment among biologists of this school that elements are present in the chromatin which do carry hereditary characters and that the chromatin granule and chromosome have a definite architec- ture and organic value because of these elements. Weismann's theory of heredity rests on an interpretation of the complexities of mitosis presented by Roux in 1883. Roux assumed that chromatin was not homogeneous in structure throughout the nucleus, but differed qualitatively in various regions. The elaborate history of mitosis with the formation and division of the chromosomes and their distribution through the mechanism of the spindle seemed inexplicable to Roux except on the theory that portions of the chromatin represented specific characteristics which were sorted and distributed accu- rately according to some system. There could be no need of such a complicated mechanism as mitosis if the distribution of the chromatin was to be merely quantitative for simple direct nuclear division could perform that operation as effectively as mitosis. Mitosis then became a device for the qualitative dis- tribution of chromatin as well as quantitative and the characters of the daughter cells were determined chiefly by the specific ele- ments which were given to one or the other. Weismann siezed upon Roux's suggestion of a possible quali- tative distribution of chromatin in mitosis and this assumption became a very important feature of his theory of heredity. Weismann postulated two methods of mitosis. By the first the 480 THE AMERICAN NATURALIST. [VOL. XXXIX. chromosomes are assumed to split longitudinally into equivalent halves, which are the facts in all vegetative or somatic mitoses so far as is known, and the chromatin is distributed quantita- tively. By the second method chromosomes were conceived to split transversely so that one half is carried to each daughter nucleus, and if the two ends of a chromosome differed in the character of their fundamental elements (ids and determinants) the chromatin would be distributed qualitatively. Weismann prophesied in 1887 that this second type of nuclear division (qualitative mitosis) would be found and ever since investigators have steadily searched for a transverse division of the chromo- somes. They have been reported in connection with the mitoses of chromosome reduction both for animals and plants and the history of these investigations forms an important part of the subject of reduction phenomena. But the present interpretation of these transverse divisions involves the consideration of factors that were unknown to Weismann and are very different from the significance assigned by him. The effect of Weismann's specu- lations, as a stimulus to investigations in these lines can, how- ever, hardly be overestimated. Botanical literature dealing with the two mitoses of sporogene- sis presents a confusion of statements respecting the presence or absence of a transverse division of the chromosomes. Stras- burger has changed his opinion three times. In his early studies Strasburger ('95) believed that the chromosomes divided longi- tudinally in both mitoses of sporogenesis. Then, led by studies of Mottier ('97) he concluded ('97b) that the fission of the chro- mosomes in the second mitosis was transverse. Almost imme- diately, however, Strasburger and Mottier reverted to the former opinion that the chromosomes divided longitudinally, a view which Strasburger maintained in his lengthy considerations of reduction phenomena in igooa. Finally in a recent paper (: O4b) Strasburger gives a very different interpretation of the events of the first mitosis (heterotypic), based on the study of Galtonia, and in general agreement with the most recent conclu- sions of Farmer and Moore (103). Farmer ('95b), Farmer and Moore ('95), Miss Sargant ('96, '97), Guignard ('99a), Gregoire ('99), Lloyd (: 02), and Mottier have also held that the divisions No. 463.] STUDfES ON PLANT CELL. VI. of the chromosomes in the mitoses of sporogenesis were longi- tudinal with somewhat varying views, however, as to the exact time when the two divisions take place. On the other hand Ishikawa ('97), Calkins ('97), Belajeff ('98), and Atkinson ('99, for Trillium) have claimed that the second mitosis presented a transverse division. Dixon ('95, '96, : oo) and Schaffner ('97) held a position apart from all these investigators, believing, that the chromosomes of the first mitosis of Lilium resulted from loops whose free ends became appressed or twisted together finally separating at the angle of the loop and thus constituting a transverse division in this first mitosis. These latter observa- tions accord with the latest conclusions of Farmer and Moore (: 03) and Strasburger (: O4b). Most of this literature is reviewed in detail in Strasburger's paper of igooa. We shall omit an historical discussion of this early work for the entire subject is approached from quite a different standpoint in the series of papers which have appeared in the past three years (1903-05) and which give hope of much clearer information on the mitoses of the spore mother-cell. The remainder of this treatment of " Reduction of the Chro- mosomes " will take up the recent papers and try to show the drift of the present investigations. These papers had not appeared when the author described the behavior of chromo- somes during mitosis in Section II (Amer. Nat., vol. 38, p. 445, June, 1904) and presented the account of the spore mother-cell in Section III (Amer. Nat., vol. 38, pp. 726, 740, Oct., 1904). At that time it seemed probable that Strasburger's conclusions of 1900 held true for all plants, namely, that the chromosomes split longitudinally in both mitoses of sporogenesis as well as in all other mitoses of the life history. Whejther these views may have to be materially changed in the light of the most recent work is now a matter of dispute. Yet the ground has shifted so frequently in these perplexing problems that it is hard to feel sanguine of final conclusions even in the hopeful situation of the present. I shall take up the events of sporogenesis in order, beginning with the growth period and synapsis and ending with the two mitoses of the spore mother-cell. The growth period always extends over a considerable length 482 THE AMERICAN NATURALIST. [VOL. XXXIX. of time and may occupy even weeks or months. During this interval the spore mother-ceils increase to many times the size of the archesporial cells from which they were derived. There is an immense accumulation of protoplasmic material and a cor- responding increase in the size of the nucleus and its chromatin content. The growth may be continued in the spores after the mitoses of sporogenesis, as is characteristically illustrated in the great increase in the size of the megaspores in the pteridophytes and certain embryo-sacs. The most striking nuclear activity of the growth period preceding the mitoses is synapsis. This term is applied to a very characteristic gathering of the chromatin and linin material in a compact tangle or ball at one side of the nucleus and usually near the nucleolus. Nuclei are sometimes in a state of synapsis for several days or perhaps weeks as is shown by the frequency of the stage in sporogenesis. Thus during the entire period of sporogenesis in Anthoceros from the inception of the spore mother-cell to the final differentiation of the spores (which must take many days) the period of synapsis occupies from one eighth to one sixth of the entire time (Davis, '99, p. 104). Synapsis has proved to be a very difficult subject for study and few investigators have made detailed observations upon its events. Some have claimed that synapsis is an artifact due either to poor fixation or to a particularly sensitive condi- tion of the cell nucleus by which the chromatin was especially susceptible to shrinkage but it seems certain now that the phenomenon is entirely normal. Miss Sargant ('97, -p. 195) has observed synapsis in the living pollen mother-cell of Lilinm mart agon. Anthoceros presents a particularly favorable subject for the study of the effects of fixing fluids on spore mother-cells because one may present all stages in the same sporophyte to identical conditions. In a series of experiments on this form (Davis, '99, p. 97) with a number of standard fixing fluids I have always found synapsis at exactly the same period in sporogenesis and at no other time in the process. True synapsis, character- istic of reduction phenomena must be carefully distinguished from other somewhat contracted conditions of the chromatin which are cccasionally found in cells. Thus Miyake {Annals of Bot., vol. 17, p. 358, 1903) noted the resemblance to synapsis No. 463-] STUDIES ON PLANT CELL VI. 483 of an accumulation of granular material in the nucleus of the central cell of Picea and other cases might be cited which super- ficially resemble synapsis but have no fundamental relation to this peculiar nuclear activity. Evidence is steadily accumulating that synapsis is a very important period of sporogenesis. Some authors hold, as will be described presently, that it is the time when paternal and maternal chromosomes, which have remained separate through- out the sporophyte generation, become associated in pairs to give the reduced number of the gametophyte. This conclusion makes synapsis the actual period of chromosome reduction and the two succeeding mitoses become merely distributing divisions of the newly formed chromosomes. Montgomery (: 01) first suggested for animals that synapsis involved a union of maternal and paternal chromosomes in pairs. Other views, however, regard the reduction of the chromosomes as merely the tempo- rary union of paternal and maternal elements, end to end, to form a bivalent chromosome characteristic of the first or hetero- typic mitosis. According to this view the bivalent chromosomes divide transversely so that the halves are distributed as whole chromosomes in the first mitosis. Two very important papers on reduction phenomena have appeared this year (1905) both of which were preceded by pre- liminary publications, that of Farmer and Moore (: 03) and Allen (:O4). These two accounts best represent the attitude of the opposing schools and will be made the chief texts of our treatment. The fundamental points of difference concern the events of synapsis and the heterotypic mitosis while there is complete agreement in the general interpretation of the homo- typic mitosis. All authors have reached essentially the same conclusions as regards the purpose and final results of the reduction divisions but the details of the processes of synapsis and the prophaseof the heterotypic mitosis are described in radically different ways by various investigators. However, as has been stated, the views fall into two groups or schools, one led by Farmer and Moore with whom Strasburger's recent paper, ~the whole length. This is the second longitudinal fission as interpreted by Gregoire ('99), Guignard ('99), Strasburger (: oo), Mottier (: 03), and others, with whom Allen is in full agreement. It is of course a premature division of the chromosomes preliminary to the homotypic mitosis. The second fission is probably com- pleted at this time but the elements of each pair (formerly called granddaughter chromosomes) remain clinging together at one end by a peculiar overlapping of the hooked tips forming thus a V-shaped pair whose apex is drawn to the poles of the heterotypic spindle. The daughter nuclei following the hetero- typic mitosis are not in a true resting condition and the chromo- somes while forming a spirem show abundant evidence of independent structure. They emerge from the spirem at the prophase of the homotypic mitosis as the same morphological entities (i. e., as V-shaped pairs) and are thus brought to the nuclear plate from which they are distributed generally as fairly straight rods to form the nuclei of the pollen grains. Rosenberg's (: O3a, : O4a, : O4b) studies on the hybrids of Drosera furnish further evidence that the chromosomes from different parents fuse in pairs during the prophase of the heterotypic mitosis. The gametophyte number of chromosomes in Drosera rotundifolia is ten and in D. longifolia twenty and those of the former species are larger than those of the latter. The sporophyte number in the hybrid is thirty as would be expected. At the heterotypic mitosis of sporogenesis, however, twenty chromosomes appear in the hybrid, half of which are plainly double structures and consist each of a larger and a 488 THE AMERICAN NATURALIST. [VOL. XXXIX. smaller element. During this mitosis the ten double chromo- somes divide but the single chromosomes remain entire and either pass to one pole or the other or are left out in the forma- tion of the daughter nuclei. The explanation of these conditions must be that ten chromosomes of D. rotundifolia fuse with ten from D. longifolia leaving ten of the latter without mates. Rosenberg's last paper (:O4b) on Drosera describes in consider- able detail the union of chromosomes in pairs in both species of Drosera during sporogenesis. The sporophytic chromosomes which at first are scattered throughout the nucleus in the early prophase of the first mitosis come together in pairs and unite so closely that there is hardly a trace of their dual nature in the resultant larger bivalent chromosomes, which are of course the gametophyte number. Rosenberg is very positive that the pairs of chromosomes are preliminary to a fusion and not the result of a fission of already reduced segments of a spirem thread. Rosenberg believes that the two halves of the bivalent chromo- somes are separated in the first (heterotypic) mitosis and that each splits lengthwise prematurely during the first mitosis in preparation for the second. The fused bivalent chromosomes then appear to divide twice longitudinally but the first division may be only a separation of the two sporophytic chromosomes that entered into the fused pair. We shall consider now the conclusions of Berghs and Gregoire of the Carnoy Institute, Louvain, whose publications have appeared practically simultaneously with some of those which we have just discussed. Berghs has published three papers (:O4a, :O4b, : 05) treating of the early history of sporo- genesis in Allium, Lilium, and Convallaria, and concludes from a study of synapsis that the spirem immediately preceding the heterotypic mitosis arises from the close association, side by side, of two delicate threads. These threads are organized pre- vious to and during synapsis and their coming together brings about that contraction of the chromatic material characteristic of synapsis. The threads contain sporophytic chromosomes of the last mitosis in the archesporium. The apparent longitudinal fission of the spirem which precedes the heterotypic mitosis in the spore mother-cell is interpreted as being these two threads No. 463.] STUDIES ON PLANT CELL. VI. 489 which are believed to have never actually fused during synapsis. The reduced number of segments derived from the spirem pre- ceding the heterotypic mitosis are then bivalent Chromosomes composed of pairs of sporophytic chromosomes lying side by side. The heterotypic mitosis distributes the sporophytic chromosomes in two sets resulting in a numerical reduction of their numbers by one half. It will at once be noted that while Berghs and Allen have independently arrived at similar conclusions respecting the structure of the chromosomes of the heterotypic mitosis there are some important differences in the mode of origin. Allen reports an actual fusion of the two threads (paternal and maternal) during synapsis and a later fission of the spirem previous to the heterotypic mitosis. But the accounts of both authors have much in common in their interpretation of the structure of the spirem and chromosomes of the heterotypic mitosis which is fundamentally different from the accounts of Farmer and Moore, and Strasburger to be described later. Gregoire (: 04) in a general discussion of reduction phenom- ena confirms the observations of Berghs and takes a very posi- tive position against the interpretations of Farmer and Moore and Strasburger. The chief features of his conclusions are in harmony with the results of Allen. The sporophytic (somatic) chromosomes are believed to become associated in pairs by the application of two delicate threads throughout their length during synapsis. These threads are believed to retain their autonomy and never actually to fuse although they may come in close con- tact. Consequently the reduced number of chromosomes are pairs of sporophytic chromosomes which have retained complete independence. Allen, on the contrary, reports a complete union of the two threads involving the fusion of chromomeres in pairs and a later longitudinal division throughout its length of the single (fusion) spirem. Gregoire does not regard the heterotypic mitosis as a true nuclear division but as a special process designed to effect this numerical separation of the sporophytic chromo- somes and intercalated between typical mitoses, while Allen would apparently treat it as a true mitosis and regard the chro- mosome reduction as effected by the fusion of two sporophytic spirems during synapsis. 490 THE AMERICAN NATURALIST. [VOL. XXXIX. Rosenberg (:O5) has recently published a general review of reduction phenomena based on studies upon Listera, Tanecetum, Drosera, and Arum, taking a position in essential agreement with Allen and the investigators of the Carnoy Institute and in opposition to the theory of Farmer and Moore and Strasburger. Rosenberg does not quote Allen's preliminary paper (: 04) which anticipates his conclusions. He finds that the spirem which emerges from synapsis is preceded by a condition when the structure is clearly made up of two threads (spirems) which lie parallel to one another. These two threads are frequently joined together, and in places spirally twisted but here and there they may be seen to be entirely separated from one another. They finally form the single spirem which follows synapsis and which divides into the reduced number of chromatic segments. But the chromatic segments throughout the entire processes are shown to be double in structure (bivalent chromosomes), /'. *?., composed of two chromosomes lying very close together side by side or even united. What appears to be a longitudinal fission of the chromatic segments of the spirem immediately preceding the first mitosis is really then a line of union along which the two independent threads have come together. The phenomenon of synapsis consists of this close association of two threads which are themselves simple spirems into a double spirem which segments into pairs of sporophytic chromosomes each of which may be regarded as a bivalent chromosome. Farmer and Moore published a preliminary communication in 1903 which aroused much interest in their theory of chromo- some reduction. The full account (: 05) has recently appeared. Their studies are upon Lilium, Osmunda, Psilotum, Aneura, and the cockroach, Periplaneta. Lilium and Osmunda among the plants were given chief attention and since the lily was the type studied by Allen it will serve best to contrast the conclusions of these two investigators. The accounts of Allen and Farmer are so fundamentally different as regards the events of synapsis and the prophase of the heterotypic mitosis that it seems scarcely possible that both can be right in their respective material, Lilium candidum, Farmer's type, and L. canadense of Allen's description. Farmer and Moore intro- No. 463-] STUDIES ON PLANT CELL. VI. 491 duce the terms "maiosis" and the "maiotic phase" to cover the whole series of nuclear changes included in the heterotypic and homotypic mitoses. The maiotic phase is regarded _as_ similar in its essential details in both animals and plants but the fact of its appearance at different points in the life histories precludes any probability of relationship in such widely divergent lines. The events of synapsis and the consequent peculiarities of the heterotypic and homotypic mitoses are considered as intercalated between the series of typical mitoses in the life history. Farmer and Moore's conclusions for Lilium c (indicium may be briefly summarized as follows. A definite spirem with the chromatin distributed as granules appears in the young spore mother-cell before its separation from neighboring elements. A " first contraction figure " now appears and the spirem thread becomes densely coiled in the vicinity of the nucleolus, this con- dition persisting for some time. Then the coils of the spirem loosen and become distributed about the periphery of the nuclear cavity, from the point of contraction as a center. A longitudi- nal fission of the spirem thread then appears, the chromatin granules dividing so that they come to lie in two parallel rows on the edge of the split ribbon. The fission is irregular and open loops appear at places. The spirem then shortens and the split gradually closes up and becomes very difficult to recognize. Many of the convolutions of the thread are attached to the nuclear membrane while the remainder form a tangle in the interior around the nucleolus which is believed to give up much of its substance to the chromatic portion of the spirem. Farmer and Moore then fail to find the double thread and its union dur- ing synapsis to form a single (fusion) spirem which is a funda- mental feature of Allen's account. There follows then a stage which has been the subject of much discussion. According to Farmer and Moore the spirem thread becomes pulled out into V- and U-shaped loops, shown with especial clearness where the bend of the loop is attached to the periphery of the nuclear membrane. The arms of the V's then come to lie parallel and so close together as to give the appearance of a fission in a structure which is really the result of an approximation of the two free ends of what was a loop. 492 THE AMERICAN NATURALIST. [VOL. XXXIX. The spirem thread thus breaks up into segments which, how-- ever, lie in pairs represented by the V's in the reduced (gameto- phyte) number. The pairs are bivalent chromosomes, each composed of two sporophytic chromosomes which were arranged serially on a single spirem thread. The pairs are not always organized through the approximation of the arms of V-shaped loops but this is a very characteristic type of structure. The V's have been interpreted by other authors as the approximation of portions of the spirem thread (Dixon, '95, '96, : oo) or the separation of their free ends at the bend of the loop as a trans- verse division of a reduced number of looped chromosomes in the heterotypic mitosis (Schaffner, '97). The two parts of the bivalent chromosomes (which are pairs of somatic chromosomes) now become shorter and thicker and all trace of the original fission of the spirem thread is lost. The essential features of Farmer and Moore's interpretation of the prophase of the heterotypic mitosis are, then : (i) a sin- gle spirem with the sporophytic chromosomes arranged serially, which splits only once longitudinally, the fission afterward becoming obliterated when the chromosomes are organized, and (2) the organization of bivalent chromosomes in the reduced number largely by the approximation of the free ends of loops which entails a separation at the bend of the loops of the two sporophytic chromosomes, giving the appearance of a transverse division. The heterotypic mitosis, then, according to Farmer and Moore involves merely the distribution of the sporophytic chro- mosomes arranged in pairs (bivalent chromosomes) as univalent elements to each daughter nucleus. This is of course the gen- eral conclusion of all recent investigators, the different views being the result of varying accounts of the method of organiza- tion of the bivalent chromosomes. During this distribution in the heterotypic mitosis the split of the original spirem appears in each univalent element (sporophytic chromosome) and the halves open throughout the greater part of their length giving the peculiar V-shaped daughter chromosomes so characteristic of this mitosis in the lily. The arms of these V's become the daughter chromosomes of the homotypic mitosis which are thus No. 463.] STUDIES ON PLANT CELL. VI. 493 formed prematurely during the heterotypic as was first described by Gregoire ('99). However, Gregoire and most botanists have considered the split between the V's as a second longitudinal fission of the original spirem in the spore mother-cell while Farmer and Moore regard it as the reappearance of an original single fission. This view of Gregoire, which has had the sup- port of Guignard ('99), Strasburger (: oo), and Mottier (: 03), is the theory of a double longitudinal splitting of the chromosomes previous to the heterotypic mitosis and is also maintained in Allen's (: 05) recent paper. The homotypic mitosis brings about the final separation of the arms of the V-shaped longitudinally split univalent (sporophytic) chromosomes of the heterotypic division. The fact that the arms of these V's finally break apart at the ends does not con- stitute a transverse division as has been claimed by some earlier writers (Ishikawa, '97 ; Calkins, '97 ; Belajeff, '98 ; Atkinson, '99, for Trillium). The peculiarities of the homotypic mitosis are then due to the premature fission of the univalent chromo- somes during the heterotypic. As a type of nuclear division the homotypic mitosis is not fundamentally different from the typi- cal divisions of other periods of the life history. All recent authors are in agreement on this interpretation of the events of the homotypic mitosis. Gregory (: 04) gives an account of sporogenesis for several leptosporangiate ferns and accepts Farmer and Moore's explana- tion of reduction phenomena. He finds the same sort of U- shaped segments in the reduced number at the heterotypic division and considers them bivalent chromosomes which divide transversely so that the original sporophyte chromosomes are distributed in two sets during this mitosis. The various posi- tions assumed by the limbs of the U-shaped segments give appearances very similar to the tetrads described in the hetero- typic mitosis of animals and which Calkins ('97) reported for Pteris and Adiantum and regarded as resulting from the trans- verse division of the halves of a longitudinally split chromosome. Gregory of course cannot accept the conclusions of Calkins. Williams (1043) applies the theory of Farmer and Moore respecting the bivalent character of the chromosomes in the 494 THE AMERICAN NATURALIST. [VOL. XXXIX. heterotypic mitosis to his studies on the first division in the tetraspore mother-cell of Dictyota. But it can hardly be said that his account offers any material support to the theory. There is a clear synapsis stage preceding the mitosis in this form from which a spirem emerges as a beaded thread. This spirem then becomes split longitudinally and later the chromosomes are organized and show a longitudinal fission. The form of the chromosomes at metaphase of the first mitosis is heterotypic, a ring form being prevalent, and Williams concludes that it is developed by the bending and closing of the free ends of a loop. The events of synapsis are not clearly enough known to make possible a comparison with the accounts of Allen and Berghs. We are now ready to take up the latest conclusions of Stras- burger (:O4b) which are closely associated with views expressed in a recent paper of Lotsy (:O4). Lotsy gives a clear state- ment, illustrated with many diagrams of the various ways in which sporophytic chromosomes may be conceived to unite in pairs previous to the first mitosis in the spore mother-cell and the manner in which the resultant bivalent chromosomes may be divided and distributed by the two mitoses of sporogenesis. Lotsy makes parallel comparisons between sporogenesis in plants and gametogenesis in animals and proposes the term "Gonoto- konten " (" Nachkommenbildner ") for the mother-cells which inaugurate reduction phenomena. The paper presents no new observations but discusses the problems of reduction in their broad aspects. An excellent summary is given by Koenicke 004). Strasburger's (: O4b) most recent paper, " Ueber Reduktionstei- lung," is based chiefly on studies of Galtonia and Tradescantia and presents an entire change of view from his conclusions of 1900. Galtonia seems to be a very favorable form for study since the gametophyte number of chromosomes is only six and the structures are exceptionally clearly differentiated in the spore mother-cells, which Strasburger calls " Gonotokonten " after Lotsy. A single spirem is reported to split longitudinally but the two daughter threads remain close together. The spirem then shortens and thickens and becomes distributed in heavy No. 463.] STUDIES ON PLANT CELL. VI. 495 loops. It finally divides into six segments which are interpreted to be six pairs of chromosomes joined end to end. These six segments are then bivalent chromosomes. The two chromosomes of each pair (segment) finally come to lie side by side in various positions by the bending of the original looped segments and the separation of their two ends in the middle. The halves of the six bivalent chromosomes (segments) are distributed by the first mitosis so that there is the effect of a transverse division of six chromosomes at this time, but really the process is one of the distribution of twelve chromosomes in two sets of six each. The longitudinal fission of the spirem thread becomes more con- spicuous towards the end of the first mitosis so that the twelve chromosomes become partially split and pass as V's to the poles of the first spindle during telophase. This premature division is preparatory for the second mitosis (homotypic) when the sepa- ration is finally effected. There is then only one longitudinal fission of the original spirem in the spore mother-cell and this prepares the chromosomes for the second mitosis, which differs only from the typical mitoses in the premature splitting of its chromosomes. The first mitosis is merely the separation of pairs of chromosomes joined end to end. Strasburger interprets the conditions in Tradescantia and Lilium in a similar way believing that the complications there simply arise from a more involved looping of the spirem thread. Strasburger's account of Gal- tonia then supports in all essentials the theory of Farmer and Moore. Strasburger in the same paper (:O4b) gives an account of synapsis which cannot be brought into harmony with that of Allen. The chromatin granules are reported to gather during synapsis into as many centers, which he names " Gamozentren," as will finally form the reduced number of bivalent chromosomes (six in Galtonia). The " Gamozentren " then become arranged and drawn out into the spirem which emerges from synapsis. The chromatin granules are named " Gamosomen " and the bodies formed in the " Gamozentren " which afterwards become the bivalent chromosomes of the first mitosis are called " Zygo- somen." There are then no organized chromosomes during synapsis and no place in Strasburger's account for the fusion of 496 THE AMERICAN NATURALIST. [VOL. XXXIX. a fully organized paternal and maternal spirem as described by Allen. The identity of the sporophytic chromosomes becomes entirely lost, according to Strasburger's explanation of synapsis, and the chromatin granules (" Gamosomen ") may be variously distributed in the new set of bivalent chromosomes (" Zygoso- men "). These " Zygosomen " are a new creation in the cell. All of the other theories, on the other hand, preserve the mor- phological entity of the sporophyte chromosomes which are of course of maternal and paternal origin but allows their distri- bution in various ratios to one another during the first mitosis of sporogenesis. The chromosome, however, remains a fixed mor- phological structure from one generation to another. These are fundamental differences which have a vital bearing on the discussion of hybridization, which will follow shortly, since one of the most important features of the problems concerns the preservation of the relative purity of the germ plasm. The chief characteristics of the two theories of reduction may be summarized as follows : (i) According to Allen, Rosenberg, Berghs, and Gre"goire, the phenomenon of synapsis presents a close association of two parallel chromatic threads (probably of maternal and paternal origin) which finally unite to form the spirem that precedes the heterotypic mitosis. This single (fusion) spirem is then double in nature and the longitudinal fission which follows, is the sepa- ration of the two threads that entered into its composition. The reduced number of chromatic segments of the heterotypic mitosis are bivalent chromosomes or more precisely pairs of sporophytic chromosomes derived from the two (maternal and paternal) threads of the synapsis stage. The heterotypic mito- sis distributes the sporophytic chromosomes in two sets thus effecting a numerical reduction by one half. The sporophytic chromosomes divide prematurely during the heterotypic mitosis in preparation for the homotypic thus presenting a second longi- tudinal fission of the segments derived from the single (fusion) spirem. A special feature of Allen's studies is the fusion of chromomeres in pairs during the organization of the single (fusion) spirem and a subsequent splitting of each larger chro- momere with the longitudinal fission of this structure. No. 463.] STUDIES ON PLANT CELL. VI. 497 (2) Farmer and Moore, Gregory, Williams, and Strasburger hold that there is primarily only a single chromatic thread in the nucleus of the spore mother-cell which is the spirem of synapsis and the heterotypic mitosis and which most of these authors believe to be composed of the full number of chromosomes (sporophytic) joined end to end. This spirem splits longitu- dinally but the fission is a premature division which prepares the chromosomes for the homotypic mitosis. The chromosomes of the heterotypic mitosis are formed from loops of the spirem which include a pair of sporophytic chromosomes joined end to end. The members of this pair come to lie side by side by an approximation of the arms of the loops and a breaking apart at the head of the structure. This transverse fission of the spirem is not of course a transverse division of a chromosome but merely the separation of a pair of chromosomes joined end to end. The line between the two arms of the loop marks a region of contact due to approximation and not a line of fission. The heterotypic mitosis effects a numerical reduction of the chromo- somes as in the first view but these chromosomes are formed on entirely different principles. A single premature fission of the spirem or its segments prepares the chromosomes- for the homotypic mitosis. Comparing the two schools, it may be noted that they both explain reduction phenomena as a numerical reduction of the double set of sporophytic chromosomes by a distribution in two sets. The fission of the chromosomes is always quantitative and there is no hint in any of the views of a qualitative division in Weismann's sense. Furthermore, most of the investigators are firmly convinced of the individuality of the chromosomes which means that they are convinced as morphological entities persist- ing from one generation to the next. This is an important agreement in relation to theories of heredity and hybridization which we shall discuss at another time (see treatment of "Hybridization"). The differences lie in questions of fact regarding the organization of these chromosomes in the spore mother-cell and their behavior during synapsis' and at other periods of prophase in the heterotypic mitosis. There is entire accord in that the chromosomes of the homotypic mitosis appear 498 THE AMERICAN NATURALIST. [VOL. XXXIX. during the metaphase of the heterotypic but a fundamental dif- ference in the accounts of the manner in which these structures are formed. In conclusion, we may very briefly note the fact that the zoologists are divided into two schools in their accounts of reduction phenomena, apparently along similar lines to those of the botanists. Some recent papers (Winiwarter, : oo ; Schoenfeld, :oi ; and the Schreiners, -.04) have described the union of parallel threads (maternal and paternal) during synapsis to form a single spirem in the rabbit, man, bull, hag-fish, and shark. Winiwarter and the Schreiners regard a later longitudi- nal fission of the spirem as a separation of the two threads which originally entered into the structure. The chromosomes in the hag-fish (Myxine, the Schreiners, : 04) are organized in pairs side by side and a second longitudinal split appears in each. The heterotypic mitosis separates the groups in the plane of the first fission and the two parted chromosomes are divided by the homotypic. This history is essentially similar to Allen's account of the lily. On the other hand there is a large body of observations founded on the investigations of Hacker, vom Rath, Riickert, Montgomery, and others, indicating that bivalent chromosomes are formed consisting of somatic chromosomes joined end to end and that these elements or their derivatives are distributed either with the heterotypic or homotypic mitosis. This of course involves a transverse division which is, however, interpreted as the separation of adjacent chromosomes and not as a qualitative division in Weismann's sense. The attitude of the first group is clearly similar to that of Allen, Rosenberg, Berghs, and Gregoire among the botanists, while that of the second shows many points of similarity to the theory of Farmer and Moore and to Strasburger's last view (: 04). There are a number of accounts of a double longitudinal fission of chromo- somes especially among the vertebrates, which have not been harmonized with the last view but may find explanation along the lines of the more recent investigations. It is of course conceivable that there are two distinct types of arrangement of sporophytic and somatic chromosomes in animals and plants at synapsis during gametogenesis and sporogenesis. No. 463.] STUDIES ON PLANT CELL. VI. 499 It is possible that they may be grouped in pairs (bivalent chro- mosomes) either side by side through two parallel threads (paternal and maternal spirems) or end to end in a siftgle chro- matic thread. But it will certainly be interesting if animals and plants both show variations in these respects and very remark- able if the same genus, as Lilium, should present contrasting types of reduction phenomena. And on these points must . be concentrated the future investigations in this field. While we are making progress in our understanding of the behavior of the chromosomes it must never be forgotten that in them we are dealing only with the most conspicuous form of germ plasm and that there are much finer elements which in their turn will demand attention. We may hold to the view of the individuality of the chromosomes as morphological entities but nevertheless we must recognize the fact that the substance of these bodies which stand for parental characters, the idioplasm of Nageli, may pass through remarkable changes which are far from understood. There is much evidence that the parental idioplasm may mix or combine during synapsis in the organiza- tion of the spirem from which are developed the reduced num- ber of bivalent chromosomes. Allen has described the actual fusion of sets of chromomeres believed to be of maternal and paternal origin and there are many possibilities of the two idio- plasm s reacting upon one another to bring about intimate and fundamental interrelations. These become important principles in discussions of heredity and hybridization and will be con- sidered later. Allen (:O5, pp. 246-252) presents an admirable analysis of these problems. VOL. XXXIX, No. 464 AUGUST, 1905 THE AMERICAN NATURALIST A MONTHLY JOURNAL DEVOTED TO THE NATURAL SCIENCES IN THEIR WIDEST SENSE CONTENTS Page I. A Systematic Study of the Salicaceae . PROFESSOR D. P. PENHALLOW 509 II. Developmental Stages in the Lagenidse . . . . J. A, CUSHMAN 537 III. Studies on the Plant Cell. VII DE. B. M. DAVIS 555 IV. Notes and Literature: Nature Study ; Zoology, Wasps Social and Soli- tary, Trouessart's Catalogus Mammalium, Supplement 601 BOSTON, U. S. A. GINN & COMPANY, PUBLISHERS 29 BEACON STREET New York Chicago London, W. C- 70 Fifth Avenue 378-388 Wabash Avenue 9 St. Martin's Street Entered at the Pott-Office, Boston, Man., as Second Class Mail Mattel'. The American Naturalist. ASSOCIATE EDITORS: J. A. ALLEN, PH.D., American Museum of Natural History, New York. E. A. ANDREWS, PH.D.,/oAns Hopkins University, Baltimore. WILLIAM S. BAYLEY, PH.D., Colby University, Wattrville. DOUGLAS H. CAMPBELL, PH.D., Stanford University. J. H. COMSTOCK, S.B., Cornell University, Ithaca. WILLIAM M. DAVIS, M.E., Harvard University, Cambridge. ALES HRDLICKA, M.D., U.S. National Museum, Washington. D. S. JORDAN, LL.D., Stanford University. CHARLES A. KOFOID, PH.D., University of California, Berkelty. J. G. NEEDHAM, PH.D., Lake Forest University. ARNOLD E. ORTMANN, PH.D., Carmgie Museum, Pittsburg. D. P. PENHALLOW,D.Sc.,F.R.M.S., Me Gill University, Montreal. H. M. RICHARDS, S.D., Columbia University, New York. W. E, RITTER, PH.D., University of California, Berkeley, ISRAEL C. RUSSELL, LL.D., University of Michigan, Ann Arbor. ERWIN F. SMITH, S.D., U.S. Department of Agriculture, Washington LEONHARD STEJNEGER, LL.D., Smithsonian Institution, Washington. W. TRELEASE, S.D., Missouri Botanical Garden, St. Louis. HENRY B. WARD, PH.D., University of Nebraska, Lincoln. WILLIAM M. WHEELER, PH.D., American Museum of Natural History, New York. THE AMERICAN NATURALIST is an illustrated monthly magazine of Natural History, and will aim to present to its readers the leading facts and discoveries in Anthropology, General Biology, Zoology, Botany, Paleontology, Geology and Physical Geography, and Miner- alogy and Petrography. The contents each month will consist of leading original articles containing accounts and discussions of new discoveries, reports of scientific expeditions, biographical notices of distinguished naturalists, or critical summaries of progress in some line ; and in addition to these there will be briefer articles on various points of interest, editorial comments on scientific questions of the day, critical reviews of recent literature, and a quarterly record of gifts, appointments, retirements, and deaths. All naturalists who have anything interesting to say are invited to send in their contributions, but the editors will endeavor to select for publication only that which is of truly scientific value and at the same time written so as -to be intelligible, instructive, and interesting to the general scientific reader. All manuscripts, books for review, exchanges, etc., should be sent to THE AMERICAN NATURALIST, Cambridge, Mass. All business communications should be sent direct to the publishers. Annual subscription, $4.00, net, in advance. Single copies, 35 cents. Foreign subscription, $4.60. GINN & COMPANY, PUBLISHERS. STUDIES ON THE PLANT CELL. VIL BRADLEY MOORE DAVIS. SECTION V. CELL ACTIVITIES AT CRITICAL PERIODS OF ONTOGENY IN PLANTS (Continued}. 5. APOGAMY. APOGAMY is the suppression of the sexual act and the devel- opment of a succeeding generation asexually. The term was first proposed by De Bary in 1878, following Farlow's ('74) discovery of the phenomenon in Pteris cretica. The succeeding generation may arise in one of two ways : ( i ) by the develop- ment of an unfertilized egg or gamete which is termed partheno- genesis, or (2) by some form of vegetative outgrowth from the sexual plant, a process which has been called vegetative apogamy. We shall not attempt to give a detailed account of apogamy in the plant kingdom but will confine ourselves chiefly to the con- sideration of a few detailed studies of recent months which have taken up the cell problems concerned. The cell problems nat- urally treat of the processes which may be substituted for the sexual act in ontogeny and the fundamental problems of the behavior of the chromosomes under these conditions. Parthenogenesis has been known for many years among the thallophytes which furnish illustrations in a variety of groups. In the algae we have the well known examples of Chara crinita, Cutlaria, Dictyota, some species of Spirogyra and Zygnema, and a number of types in the lower Chlorophyceae and Phasophycese whose motile gametes will germinate like zoospores should they fail to conjugate with one another. The recent studies of Wil- liams (:04b) on Dictyota give the only observations which have been made on nuclear activities during the parthenogenetic development of eggs in any algal form and will be considered presently. The fungi furnish beautiful illustrations of partheno- 556 THE AMERICAN NATURALIST. [VOL. XXXIX. genesis in the Saprolegniales. Trow (: 04) believes that some of these forms are sexual but there can be little doubt that the group as a whole is generally apogamous. ^here is probably much apogamy in the Ascomycetes and an almost entire suppres- sion of sexual organs in the Basidiomycetes but no clear instance of parthenogenesis (i. e., a development from a cell whose mor- phology is unquestionably that of an egg) is known in either of these groups. Parthenogenesis is not known in the bryophytes and pterido- phytes excepting for Marsilia (Shaw, '97; Nathansohn, :oo). Although there is much apogamy in the pteridophytes, especially in the leptosporangiate Filicales, the new generation generally develops as a bud-like outgrowth on the prothallus (vegetative apogamy). There have been no nuclear studies on the parthen- ogenetic Marsilia but an interesting preliminary account has appeared announcing nuclear fusions in the apogamous develop- ment of Nephrodium (Farmer, Moore, and Digby, :O3). Parthenogenesis is now known in the spermatophytes for Antennaria alpina (Juel, '98, : oo), several species of Alchemilla (Murbeck, :oia, :oib, : 02 ; Strasburger, : O4c), TJialictrum pnr- purascens (Overton, : 02, : 04), Gnetum (Lotsy, : 03), a number of forms of Taraxacum (Raunkiaer, : 03 ; Murbeck, :O4), sev- eral species of Hieracium (Ostenfeld, : O4a, : O4b ; Murbeck, : 04), Wikstrcemia indica (Winkler, :O5), and is suspected for Ficus Treub, :O2) and Bryonia dioica (Bitter, :O4). A number of cases of polyembryony were formerly considered examples of. apogamy but are now known to be developments from the nucel- lus and consequently vegetative buds of sporophytic origin and entirely independent of gametophytic activities. The best known of these forms are Funkia, Ccelebogyne, Citrus, Opuntia, and Alchemilla pastoralis. Vegetative apogamy is illustrated in the development of embryos from antipodal cells as in A I Hum odormn (Tretjakow, '95 ; Hegelmaier, '97) or from the cells of the endo- sperm as in Belanophora (Treub, '98 ; Lotsy, '99). Synergids have been reported to form embryos in a number of forms but many of these have proved to be cases in which the synergic! is fertilized by a sperm nucleus and not examples of apogamy. However, synergids are known to develop embryos apogamously No. 464.] STUDIES ON PLANT CELL. VII. 557 (or parthenogenetically if the antipodal be considered the homo- logue of an egg) in Alchemilla sericata (Murbeck, : 02). A sum- mary of the various types of vegetative apogamy, parthenogen- esis, and sporophytic (nucellar) budding, supplementing a list of Ernst (:oi) is given by Coulter and Chamberlain (: 03, p. 221). We will now take up the few investigations which consider the cytological details of parthenogenesis. That of Williams (: O4b) on Dictyota is the only one treating of a lower type. It seems probable that parthenogenesis in Dictyota is in no sense normal and would not lead to mature plants, since the germina- tion of unfertilized eggs in the cultures of Williams presented many irregularities. The spindles instead of being formed from asters with centrosomes are intranuclear in origin, multipolar, and very irregular in their form. As a result the 16 chromo- somes become scattered and a cluster of daughter nuclei is formed containing varying numbers of chromosomes, sometimes one and sometimes several. It is clear in Dictyota that the fer- tilization of the egg results in the development of an aster with a centrosome which exerts a directive influence in mitosis pre- venting a scattering of the 32 chromosomes and conducting the mitosis in a normal fashion. Williams does not believe that the centrosome is introduced as an organized structure into the egg by the sperm but that it is formed dc novo as a result of the increased metabolic activities present in the fusion nucleus as compared with that of the unfertilized egg. There have been several important studies on parthenogenesis in the spermatophytes. Some of these papers while establishing the facts of parthenogenesis in various forms, give no details of nuclear history or behavior of the chromosomes. But the studies of Juel (:oo), Overtoil (:O4), and Strasburger (: 04), present some very interesting data on the cytological features of parthenogen- esis in Antennaria alpina, Thalictrum purpurascens, and several species of Alchemilla. Several recent papers indicate that parthenogenesis may prove to be general in certain genera or even characteristic of large groups and therefore a far more widespread phenomenon than has been supposed. Raunkiaer (: 03) (abstract in English in Bot. Centralb., vol. 93, p. 81, 1903) proved by cutting off the 558 THE AMERICAN NATURALIST. [VoL. XXXIX. tops of young flowers that several species of Taraxacum pro- duced normal seeds apogamously and concluded that the embryo must develop parthenogenetically since Schwere, in 1896, traced its origin from the egg. Ostenfeld (: 043., : 040) from failure to find pollen on the stigma of Hieracium and failure to make it germinate in a number of solutions, was led to try similar experi- ments to those of Raunkiaer in cutting off the anthers and stig- mas of flowers. He found that a large number of species of Hieracium were able to set seed apogamously and he believed parthenogenetically but histological investigations were not made to establish the last point. The experiments of Raunkiaer and Ostenfeld are interesting as showing how a form by virtue of its parthenogenetic habits might become segregated and quite re- moved from the probability of hybridization. Murbeck (: 04) in a short paper announced that the embryos in Taraxacum ancj Hieracium, developing from flowers whose stamens were cut out (as in the experiments of Raunkiaer and Ostenfeld) actually do develop from the egg cell and are therefore parthenogenetic. Murbeck also failed to find pollen tubes in the ovules where pollen had been applied to the stigma. Winkler (: 04) reports that Wikstrcemia indica matures very little perfect pollen and produces its seeds apogamously, as proved by experiment. The embryos are stated to develop parthenogenetically from the egg but no details are given in this preliminary paper of the chromo- some history. This group of contributions while very interest- ing, presents no data on the fundamental problems in a cyto- logical explanation of parthenogenesis. Murbeck (:oia) concluded for Alchemilla that true tetrads were formed previous to the differentiation of the embryo-sac but nevertheless found evidence that there were no reduction phenomena so that the nuclei within the embryo-sac contain the sporophytic number of chromosomes. Mur beck's evidence of tetrad formation was not satisfactory and in the light of recent studies of Strasburger (: O4c) cannot be accepted. His view was, however, correct that there is no reduction of the chromo- somes in the formation of such embryo-sacs as produced par- thenogenetic embryos. Juel (: oo) gives a critical comparison of the development of No. 464.] STUDIES ON PLANT CELL. VII. 559 the embryo-sac in the parthenogenetic Antennaria alpina with A. dioica whose ovules are normally fertilized. In A. dioica the embryo-sac is one of a group of four cells (tetrad )_which are formed through two successive mitoses (heterotypic and homo- typic) showing the characteristic features of sporogenesis. A clear stage of synapsis precedes the first mitosis. The type of embryo-sac development in this form is then entirely normal. Not only are tetrads suppressed in the parthenogenetic Anten- naria alpina but there is no trace of the heterotypic and homo- typic mitoses in the embryo-sac. The number of chromosomes is very large (about fifty) and evidently the same as is found in other periods of the life history. There is then no reduction of the chromosomes during the formation of the embryo-sac in the parthenogenetic species and the egg and other nuclei in this structure have consequently the sporophytic number. There is no need of fertilization to bring the egg to a condition when with respect to chromosomes it is prepared to develop a sporophyte embryo. Juel (:O4) notes certain peculiarities in the develop- ment of the embryo-sac of Taraxacum officinale. Tetrad forma- tion is reduced to a single mitosis and this is not heterotypic, since there seems to be no reduction of the chromosomes. Details are not given. Overton (: 04) finds normal reduction phenomena in the pol- len mother-cell of Thalictrum purpurascens which establishes the number of chromosomes to be 24 for the sporophyte and 12 for the gametophyte generations. These mitoses are thoroughly typical of sporogenesis being preceded by a synapsis stage. The development of the embryo-sac is of two types. In some cases a tetrad of four megaspores is formed from a megaspore mother-cell. The nucleus of this cell passes through a synapsis and the first mitosis is heterotypic showing the reduced number of chromosomes. The lower cell of the tetrad becomes the embryo-sac. But many embryo-sacs pass through a different history. There is no heterotypic mitosis and no reduction of the chromosomes which remain 24 in number. Thus in some ovules the mitoses of sporogenesis are omitted and true tetrads are not formed, with the result that the embryo-sac contains nuclei with the sporophyte number of chromosomes (24) in 560 THE AMERICAN NATURALIST. [VoL. XXXIX. place of the gametophyte (12). The details of the nuclear history in these embryo-sacs have not been followed but it is plain that their eggs have the requisite number of chromosomes to develop sporophyte embryos parthenogenetically. The vary- ing proportions of parthenogenetically developed seeds which may be found on plants of Thalictnim purpnrascens indicate that the suppression of normally developed embryo-sacs is not very firmly established in this form. We now come to a recent paper of Strasburger (: O4c) which is the most important contribution to the subject of partheno- genesis that has yet appeared. Strasburger studied a number of species of Alchemilla from the section Eualchemilla, the group which formed the subject of Murbeck's important discov- eries. Most of the forms develop pollen in a normal manner and Strasburger was able to follow reduction phenomena in this process without difficulty. The nucleus of the pollen mother- cell passes through a synapsis followed by a heterotypic mitosis in which the structure of the chromosomes as bivalent elements is apparent. The bivalent chromosomes are in the reduced (gametophytic) number. Similarly Strasburger found that some species (c. g., Alchemilla pentaphylla, gclida, and grossidens} formed embryo-sacs in a normal manner with the presence of a tetrad and a characteristic reduction division (heterotypic). But the development of the embryo-sac in apogamous species (c. g., AlcJicmilla speciosa, splendens, and fallax) cuts out the two mitoses of sporogenesis and no tetrads are formed. The nucleus of the megaspore mother-cell emerges from synapsis with the sporophyte number of chromosomes and the first division which follows is a typical mitosis and not heterotypic. The embryo-sac then comes to contain a group of nuclei with the sporophytic number of chromosomes in place of the gametophytic and a parthenogenetic development of the egg takes place. Stras- burger regards the parthenogenetic tendencies of Eualchemilla as associated with excessive mutations among these forms through which sexual processes are becoming displaced by apogamous methods of reproduction. This clear evidence that the cause of parthenogenesis in Antennaria, Thalictrum, and Alchemilla lies in the suppression No. 464.] STUDIES ON PLANT CELL. VII. 561 of chromosome reduction during the formation of the embryo- sac seems to offer an explanation of other examples of apogamy presented by the embryo-sac. Thus apogamous devejojoments of embryos from synergids as in AlcJiemilla sericata (Murbeck, : 02) or from antipodals as in A Ilium odorum will not seem strange if reduction processes are suppressed in the production of an embryo-sac and its nuclei retain the sporophyte number of chromosomes. Such nuclei have in them the same potentialities of development as do those of the nucellus whose cells form embryos vegetatively and entirely independent of gametophytic activities in a number of forms (e. g., Funkia, Ccelebogyne, Citrus, Opuntia, Alchemilla pastoralis, etc.). This type of apogamy from a gametophyte which retains the sporophyte number of chromosomes may be found to hold a very close relation to apospory for there is the same reduction or omission of the processes of sporogenesis as is found in that phenome- non. However, since we know nothing of the cytological events of apospory it is unwise at present to follow the speculation further. The peculiarities of parthenogenesis in the spermatophytes do not seem so remarkable since the discoveries recorded above. It is not strange that an egg should form an embryo without fer- tilization when its nucleus contains the sporophyte number of chromosomes. The most remarkable feature in this suppression of reduction phenomena in Antennaria, Thalictrum, and Alche- milla is the possibility of developing an embryo-sac with nuclei in the number and arrangement typical of the female gameto- phyte and yet with the sporophyte count of chromosomes. The embryo-sacs with their contents have clearly the morphology of female gametophytes and must be so considered in spite of the fact that their nuclei contain twice as many chromosomes as usual. It is clear that the potentialities of sporophyte and gametophyte involve other factors besides those of the chromo- some count. This is a very important conclusion because we have been accustomed to lay great weight on the number of chromosomes as the cause of sporophytic and gametophytic developments respectively. We must recognize the presence of other factors determining alternation of generations besides the chromosomes. 562 THE AMERICAN NATURALIST. [VOL. XXXIX. There are two types of parthenogenesis in plants : (i) that in the thallophytes where there is no sporophytic generation, and (2) that in higher forms when the life history is complicated by an alternation of generation. We know nothing of the cytologi- cal conditions in the first group including such types as Chara crinita, Cutlaria, some species of Spirogyra and Zygnema and numbers of the lower Chlorophyceae and Phaeophyceae whose motile gametes will germinate like zoospores should they fail to conjugate with one another. But since there is no reason to suppose that there are reduction phenomena at gametogenesis, the unfertilized gamete is fully prepared with respect to the number of chromosomes to continue the parent stock. Dictyota must be excluded from this list since the parthenogenetic devel- opments here are abortive. In the second group parthenogene- sis is likely to prove to be the result of a suppression of reduction processes during sporogenesis by which a gametophyte genera- tion retains the sporophyte number of chromosomes and in consequence is prepared to dispense with sexual processes in the development of a new sporophyte. Parthenogenetic develop- ment in animals seems to be similar in its essential cytological features to parthenogenesis and apogamy in plants. There may be a suppression of reduction processes somewhat comparable to that discussed above, which takes place, however, at the time of gametogenesis, whereby the egg nucleus retains the number of chromosomes characteristic of the parent. Or, through a fusion with the nucleus of the second polar body the egg nucleus is brought back to the normal condition with respect to the num- ber of chromosomes of the parent stock. We cannot, however, consider in detail the forms of parthenogenesis in animals. They have been recently treated by Blackman (: O4b) in comparison with conditions in plants. Apogamous developments which involve wholly or in part other elements than gamete cells and nuclei are likely to be established in a number of groups of the thallophytes. The author has long believed that the cystocarps of some of the Rhodophycese develop apogamously, basing his conclusions on certain general peculiarities of the group and more particularly on a study of Ptilota (Davis, '96). Three species of this genus No. 464.] STUDIES ON PLANT CELL. VII. 563 were investigated and no developments from the carpogonia were found, but the cystocarp in all cases arose from a cell near the base of the group of procarps. These conditions together with the rarity of male plants on the American coasts (none have ever been reported) give strong evidence for apogamy in Ptilota. There are a number of genera of the Rhodophyceae where similar conditions seem to obtain and which lead one to suspect that apogamy may not be very exceptional. However, the subject has been very little studied. As is well known, the Ascomycetes furnish numbers of illus- trations where ascogonia have not been found or appear in what seem to be reduced conditions and even when accompanied by so called antheridial filaments these latter have not been estab- lished as functional. De Bary recognized the possibility of apogamy in the development of the ascocarps of these forms and very little critical study has been given to them since his time. The trend of investigations in this group has been towards the more interesting problems of the establishment of sexuality in a few well known forms (e. g., Gymnoascus, Sphae- rotheca, Pyronema, Monoascus, and among the lichens and Laboul beniaceae . ) It is generally believed that no sexual organs are present in the higher Basidiomycetes (Autobasidiomycetes). But the recent studies of Blackman' (: O4a) in the Uredinales, taken in relation to the well known nuclear fusions in the basidium, pre- ceded by a mycelium containing paired (conjugate) nuclei, make it seem very probable that former sexual processes in the Basi- diomycetes have been replaced by a remarkable type of apog- amous development of a sporophyte generation. Blackman has traced the origin of the paired nuclei in the Uredinales (Phrag- miclium) to a structure preceding the aecidium, a structure. which seems to be the remains of a female sexual organ. We will take up this investigation presently. There is then much reason for believing that a sporophyte generation in the Basi- diomycetes arises apogamously in the creation of the paired nuclei and terminates with their fusion within the teleutospore or basidium. The leptosporangiate ferns have furnished some of the best 564 THE AMERICAN NATURALIST. [VOL. XXXIX. illustrations of apogamy. Since Farlow's discovery in 1874 of an asexual sporophytic growth from the prothallus of Ptcris cretica the list of apogamous pteridophytes has steadily increased until now the phenomenon is known in perhaps 25 forms. Far- low's investigation was followed by an extended study of De Bary ('78) on a large number of forms in the Polypodiaceas and resulted in the establishment of similar sporophytic outgrowths in Aspidinm falcatum and Aspidium filix-mas cristatitm. De Bary proposed the term apogamy ('78, p. 479) for the general phenomenon and distinguished two forms, apandry the suppres- sion of the male sexual organs which results in a parthenoge- netic development of the egg, and apogyny for the suppression of the female. Sadebeck in the following year reported apog- amy in Todea one of the Osmundacese (Schenk's Handbnch dei Botanik, vol. i, p. 231, 1879) thus extending the phenomenon to another family. And later apogamy was found in Tricho- manes alatum one of the Hymenophyllaceae (Bower, '88) and in Selaginella rupestris (Lyon, :O4, p. 287). The most important recent contribution on apogamy in ferns is by Lang ('98, abstract in Annals of Bot., vol. 12, p. 251). This paper presents an able discussion of the phenomenon in its relation to alternation of generations and adds the very interesting discovery of sporangia borne directly on prothalli that were grown from spores. These sporangia were found in clus- ters on a thickened lobe or process from the prothalli of Scolo- pcndrinm vnlgare ramulosissimnm and Neplirodiwn dilatmn cristatum gracile. The sporangia were perfectly normal in structure and they matured spores. It is probable that the process is itself sporophytic in character, i. e., made up of cells with double the number of chromosomes of the true gametophy- tic portion of the prothallus, but cytological details are not known. Lang's study of the apogamous development of sporo- phytic buds on several forms of the Polypodiacese is the most detailed work on apogamy in the pteridophytes yet published. The apogamous growths appeared as the result of cultures which were watered entirely from below and exposed to direct sun- light, important departures from normal conditions surrounding fern prothalli. In all cases the prothalli developed normal No. 464.] STUDIES ON PLANT CELL. VII. 565 embryos when the conditions permitted of fertilization. We shall refer to some general considerations of Lang in our sum- mary and conclusions on apogamy. The spermatophytes present some exceedingly interesting examples of apogamous developments of embryos from nuclei within the embryo-sac other than the egg, as from antipodals (Allium odorwn, Tretjakow, '95 ; Hegelmaier, '97) or synergids (AlcJiemilla sericata, Murbeck, : 02) or nuclei in the endosperm (Belanophora, Treub, '98 ; Lotsy, '99) but in these cases the sporophyte number of chromosomes is apparently present through a suppression of the reduction phenomenon of sporo- genesis in the development of the embryo-sac. We will now consider two studies which describe nuclear fusions preliminary to the appearance of apogamy (Blackman, :O4a; Farmer, Moore, and Digby, :O3). Blackman's (: O4a) observations on Phragmidium have cleared up to a great degree our understanding of the life history of the Uredinales. The chains of secidiospores have been found to arise serially from "fertile cells" which form a group at the spot where an aecidium is to be developed. Each fertile cell has above it a sterile cell which, however, breaks down. The sterile and the " fertile cell " together may represent a female sexual organ, the sterile cell perhaps standing for the remains of a receptive structure similar to a trichogyne. The spermogonium consists of a large mass of antheridial filaments that abjoint sperms which are no longer functional. It is of course uncer- tain whether the " fertile cells " are morphologically the original female gametes since they may readily be other cells drawn into the process of apogamy. The " fertile cell " is stimulated to activity by the entrance of a second nucleus either from an adjacent hypha or from the cell below. The second nucleus does not fuse with the original nucleus in the "fertile cell " but the two come to lie close together as a paired or conjugate nucleus. The two nuclei of the pair divide simultaneously (conjugate mitosis) throughout a long series of nuclear divi- sions, beginning with the formation of aecidiospores and through the vegetative history which follows up to the production of the teleutospores where the members of the last pairs unite to form 566 THE AMERICAN NATURALIST. [VOL. XXXIX. the single fusion nuclei within these reproductive cells. There is much evidence that the period in the life history characterized by the presence of paired nuclei represents a sporophyte phase. Blackman (:O4a, p. 353) regards the process by which the second nucleus enters the "fertile cell," resulting in the conju- gate nuclei, as a reduced form of ordinary fertilization. I have already pointed out in Section IV, " Asexual Cell Unions and Nuclear Fusions," what seem to me to be serious objections to the use of the term fertilization when it is clear that the second nucleus in the pair is morphologically not a gamete nucleus, and the subject was also taken up in the account of fertilization in the present section. Whatever may be the physiological inter- pretation of this remarkable phenomenon it seems to me clearly a substitute process for a former sexual condition and involves other elements than the original gametes and as such is a typi- cal illustration of apogamy. It seems probable that further studies in the Basidiomycetes will determine a similar origin for the paired nuclei preceding the basidium to that of Phragmidium but without any trace of former sexual organs at least in the higher groups. And these conditions must signify the complete disappearance of structures representing sexual organs and the substitution of an apogamous development of the sporophyte generation for the sexual act. In this connection the interesting nuclear fusions in the ascus are of great interest for they may hold relations to degenerate sexual conditions in the Ascomycetes. Farmer, Moore, and Digby (103) have reported some remark- able nuclear fusions preceding the apogamous development of the sporophytes of Nephrodium, which have many points of resemblance to the apogamous phenomena in the Uredinales just described. These authors find that cells of the prothallus from which the sporophytic outgrowths arise, become binucleate through the migration of nuclei from neighboring cells. The two nuclei may remain separate for some time or. they may fuse at once. They regard the whole process " as a kind of irregu- lar fertilization " by which the outgrowth destined to form the sporophyte becomes supplied with nuclei containing the double number of chromosomes. It seems to me unfortunate to asso- No. 464.] STUDIES ON PLANT CELL. VII. 567 ciate the term fertilization with this phenomenon, whatever may be the physiological significance of the nuclear fusions, because we are not dealing with gametes and there cannot be involved in the process anything of the long phylogenetic history of sex- ual differentiation in the group. We considered these matters in some detail in that portion of this section entitled u Fertiliza- tion." With respect to the factors which determine apogamy it must be confessed that we are still in the dark. Lang's ('98) studies on fern prothalli, however, throw some light on the problem. In some twenty forms of the Polypodiaceae apogamy resulted when the prothalli were kept from direct contact with the water (i. e., were watered from below) and exposed to direct sunlight. When watered from above these same forms developed normal embryos from eggs. It is clear that the suppression of condi- tions which make fertilization possible (i. e., water over the sur- face of the prothallus), possibly aided by sunlight which may cause irregularities of growth, induced the development of cylin- drical processes from which the apogamous sporophytes arose and which bore sporangia in two forms. It seems hard to draw more precise conclusions from these experiments other than that the normal life history is checked at a critical period (fertiliza- tion) and the plant is forced into expressions of vegetative activity. The conclusions of Farmer, Moore, and Digby (: 03) offer an explanation of how the developments may take on sporophytic characters through the fusion of nuclei in the tis- sues and the establishment of a sporophyte number of chromo- somes. Strasburger suggests that apogamy in Alchemilla may be the result of a weakening of sexual power associated with excessive mutative tendencies. This would seem to imply that excep- tional vegetative activity, with the appearance of much variation under favoring conditions, may be combined with apogamy. It is of course a well known fact that a high degree of cultivation tends to lessen the fertility of a form unless guarded by careful selection. A weakened sexual fertility due to excessive vegeta- tive activity is likely to be replaced by forms of vegetative reproduction. When the process of sporogenesis becomes so 568 THE AMERICAN NATURALIST. [VOL. XXXIX. reduced or modified that the female gametophyte retains the sporophyte number of chromosomes as in the embryo-sac of Alchemilla and Thalictrum the apogamous development of em- bryos is to be expected. The discovery of apospory in such variable and perhaps mutating genera as Alchemilla, Taraxacum, and Hieracium sug- gests quite a new line of research with possibilities of a clearer understanding of the origin of mutations. It is very interesting that these widespread and successful genera should give evidence of such strong apogamous habits for it seems to indicate an evo- lutionary tendency in the higher plants of great significance. These forms with Thalictrum are representatives of three large, divergent and very successful orders (Ranales, Resales, and Compositales) and it suggests the probability that apogamy will be found to be widespread in the spermatophytes. Its bearing on the establishment of extreme variations and mutations may be of the utmost significance for it is clear that the suppression of sexuality would remove sports and mutants at once from the swamping effects of cross-fertilization. The sudden appearance of mutants in some groups and their ability to hold true may indeed be found to rest on the establishment of apogamy in the form. This is at least a possibility which must be considered in cytological investigations on mutants and has not yet received attention. The subject of apogamy touches another topic of importance, namely, the theory of homologous generations as contrasted with antithetic generations in comparisons of sporophyte with gameto- phyte. We shall not take up this discussion in detail here. It must have been apparent to the reader that the present treat- ment of the critical periods in the life history of plants is based on the conviction of the correctness of the latter view which has had the support of Celakovsky, Strasburger, Bower, Vaisey, and Klebs. The theory of homologous generations as held by Pringsheim and Scott is admirably discussed by Lang ('98) in connection with his studies on apogamy and also in a briefer note (Annals of Bot., vol. 12, p. 583). Lang seemed inclined to the opinion that the facts of apogamy and apospory in ferns lent support to the theory of homologous generations since the No. 464.] STUDIES ON PLANT CELL. VII. 569 prothallus can so readily take on sporophytic potentialities and the sporophyte develop gametophytes vegetatively. But Lang recognized that the importance of this evidence woitkl be mini- mized should it be found to depend on changes of nuclear struc- ture. These nuclear changes have been established at least for apogamy, either in the suppression of the reduction phenomena of sporogenesis or by the substitution of asexual nuclear fusions for the sexual act, and the argument for antithetic alternation of generations seems to the writer stronger to-day than ever before. 6. APOSPORY. Apospory is the suppression of all processes of sporogenesis and the development of a gametophyte generation directly from the sporophyte. The term was first proposed by Vines (Jour, of Bot., 1878, p. 355) in a discussion of the life history of Chara and adopted by Bower ('86, '87) in a general treatment of the subject based on Druery's ('86a, '86b) discoveries of prothalli developed in place of sporangia directly upon the leaves of Athyrium filix-fccmina and its variety clarissima. The phe- nomenon of apospory is best known among the ferns where it has been most extensively studied but so far no cytological inves- tigations have been published. Since apospory results in the development of a gametophyte generation (presumably with the gametophyte number of chromosomes) without the preliminary process of sporogenesis it becomes a very interesting problem to know just how this reduction of the chromosomes is effected. Apospory is probably not uncommon in the mosses and has also been reported for the liverwort Anthoceros. The inde- pendent studies of Pringsheim ('76) and Stahl ('76) established the facts that pieces of the sporophyte stalk (seta) of Hypnum, Amblystegium, Bryum, and Ceratodon when placed on damp soil developed a protonema which in its turn produced leafy moss gametophytes. Stahl also found in Ceratodon that protonemata may arise from the capsule wall and Brizi ('92) discovered a similar development from the atrophied capsule of Funaria hygrometrica. Correns ('99a, p. 421) has confirmed the conclu- 570 THE AMERICAN NATURALIST. [VOL. XXXIX. sions of Pringsheim and Stahl in species of Funaria, Hypnum, and Amblystegium and obtained negative results in a number of other forms, and presents an excellent review of the subject. Lang(:oi) discovered that small pieces of the sporophyte of Anthoceros la-vis when laid on damp sand produced green out- growths which took on the structure of young gametophytes and developed rhizoids. These aposporous gametophytes most com- monly arose from subepidermal cells, but they may come from any layer of the cortex down to the archesporial cylinder. It seems probable that the mosses at least among the bryophytes are able to reproduce themselves apogamously without difficulty, when normal processes of sporogenesis are interfered with and if the sporophytic tissue is in contact with moisture. The leptosporangiate ferns, however, furnish the most con- spicuous illustrations of apospory as they do of apogamy. Indeed, the two phenomena are known to occur in the same form in a number of instances (e. g., AtJiyrium filix-fcemina, NepJirodium filix-mas, Scolopendrium vulgare, Trichomanes ala- tum, etc.). Beginning with the discovery by Druery ('86a, '86b) of apospory in Athyrium filix-fcemina and its variety clans sima the list has steadily grown until now apospory is recorded for about ten forms. In Druery's forms the prothalli developed from arrested sporangia and the spore alone is left out of the life cycle. But Bower ('86) very shortly brought forward in Polystichum angulare pnlcherrimum a form in which prothalli are developed as simple vegetative outgrowths from the tips of the leaves and the life history is thus shortened by the omission of both spores and sporangia. This condition is exactly analo- gous to the development of protonemata from vegetative cells of the sporophytes of mosses and Anthoceros. The following year Bower ('87) presented a very full account of the forms of Athy- rium and Polystichum just described, and a general discussion of the phenomenon of apospory. Bower ('88) then extended the illustrations of apospory to two species of Trichomanes, of the Hymenophyllaceae ; Farlow ('89) reported it for Pteris aquilina, and Druery ('93) in Lastrea pseudo-mas cristata and ('95) for Scolopendrium vulgare crispum. The exceptional amount of fern variation both in nature and under cultivation has not been No. 464.] STUDIES ON PLANT CELL. VII. 571 generally appreciated and the studies on apospory and apogamy indicate that much of it is associated with these fundamental modifications of the life history (Druery, :oi). As to the cause of apospory we are as much in the dark as in the case of apogamy. The phenomenon is clearly associated in some forms with disturbances in the normal vegetative life of the sporophytes. This is particularly true in the cases of mosses and Anthoceros and has been suggested for the ferns. Thus aposporous developments in Pteris aquilina are from leaves which are generally smaller than the normal and whose margins are curled so that the leaf often appears somewhat withered and is easily recognized at a distance. Bower ('87, p. 322) is inclined to regard the phenomenon in the ferns as a sport and does not consider that it has deep morphological significance or that it offers serious difficulty to the acceptance of the theory of an antithetic alternation of generations. As we have stated there have been no cytological studies upon apospory but there seem to be two possible explanations. That which is likely to suggest itself first calls for reduction phenom- ena at the time of the aposporous development by which the nuclei of the sporophytic tissues may come to contain the gametophyte number of chromosomes and are therefore capable of developing the sexual generation. But there is another possibility which has not yet been considered. We know for several of the sper- matophytes (Antennaria, Juel, : oo ; Thalictrum, Overton, 104; Alchemilla, Strasburger, : O4c) that the processes of sporogenesis may be suppressed and yet a structure be developed with the morphology of the gametophyte generation. Thus the embryo- sac will contain the usual number of nuclei grouped in the typ- ical manner but these nuclei still have the sporophyte count of chromosomes. It seems probable then that the development of a gametophyte may result through an interference with the nor- mal life history and under conditions favorable to the game- tophyte even though the nuclei retain the sporophyte number of chromosomes. And it is possible that some of the aposporous developments in bryophytes and pteridophytes may be of this character. It is quite futile at present to carry this speculation further. What is desired is some cytological facts. 572 THE AMERICAN NATURALIST. [VOL. XXXIX. 7. HYBRIDIZATION. This is not to be a detailed discussion of the facts and theories of hybridization, a subject far too extensive for the purposes of our treatment. We shall only consider some of the bearings of the recent studies on fertilization and reduction phenomena upon the problems of hybridization treating it as a critical phase in the life history of the organisms concerned. Until recently the attempts to formulate definite laws for the formation of hybrids and their progeny upon a physical basis have not been satisfac- tory. But the work of a number of breeders all of whom owe their results in large part to a quick appreciation of Mendel's epoch-making contributions have brought much order out of what was a very confused subject. And accompanying the work of this group must be added the equally important con- clusions of a number of cytologists whose investigations on the structure and behavior of nuclei in the critical periods of fertil- ization and chromosome reduction have done much to place Mendelian principles upon a cytological basis. We shall deal with the work of the latter group, for their contributions concern intimately the subject matter of these papers. We shall not review the conclusions of Mendel except to point out the relations of some of his principles to cytological phenom- ena. The two papers of Mendel appeared in the proceedings of a natural history society of Briinn, Austria, under the dates 1865 and 1869. They lay buried until 1900 when De Vries, Correns, and Tschermak independently rediscovered them and called the attention of the scientific world to their worth. Soon after, Bateson published a translation of the two papers (Menders Principles of Heredity, Cambridge, 1902) with an introduction and a defense against the criticisms of Professor Wheldon. There have naturally been many reviews and short discussions of Men- delian theories and among them that of Castle entitled " Mendel's Laws of Heredity" (Science, vol. 18, p. 396, 1903) and Profes- sor Bailey's "Lecture IV" in Plant Breeding, 1904, will per- haps give the reader the clearest and most concise statements. The most striking feature of Mendel's investigations and those No. 464.] STUDIES ON PLANT CELL VII. 573 of others, who have confirmed his conclusions, is the discovery in a number of animals and plants that the germ cells of the hybrid may be pure with respect to certain characters of the parents which are crossed. This principle is not without excep- tions where the conditions are apparently complicated by unusual factors but the phenomenon when present is so striking as to command immediate attention and call for an explanation on a cytological basis. The purity of the germ cells of hybrids means in the words of Castle that " the hybrid, whatever its own char- acter, produces ripe germ cells which bear only the pure char- acters of one parent or the other." Thus if two forms A and B are crossed the hybrid will have embodied in itself the characters AB, one of which however may lie latent, i. e., may not be visi- ble in the hybrid. Such a latent character when present is termed recessive while the prominent character is termed domi- nant. In a simple case some of the offspring of the hybrid AB will be found to have the character of A alone, some of them of B alone, and some of them will again have the mixed characters AB. If experiments are carried out on an extensive scale the proportions of these offspring from the hybrid may exhibit the remarkable fact that there are about twice as many forms of AB as either A or B, i. e., the proportions of A's, AB's, and B's are in the ratio of i : 2 : i . Furthermore the offspring of A when bred among themselves remain absolutely true producing a suc- cession of pure forms all A's and the same results follow when the offspring of B are closely bred. But when forms with the mixed characters AB are bred with one another their offspring break up as before into three types A, AB, and B in numerical proportions expressed by the same ratio 1:2:1. The history is simply told in the following diagram where the number of off- spring is assumed to be 4. f i A 4 A 1 6 A 64 A f 2 A 8 A 12 A Form A, f 4A i6A f 8A' Hybrid AB<| I Form B 4 AB| 8AB-| i6AB 8B L 46 i6B 2B 8B 326 4 B- i6B 646 574 THE AMERICAN NATURALIST. [VoL. XXXIX. This remarkable proportion of forms derived from the hybrids AB, i. e., A, AB, and B in the ratio 1:2:1 can only be explained on the assumption that the germ cells of the hybrid are pure with respect to the characters of either one or the other of the parents. The gametes from the hybrid, with the pure charac- ters of either A or B and approximately equal in number, may unite with one another in three possible combinations AA, AB, or BB forming three types of offspring, one pure A, another mixed AB, and the last pure B. By the law of chance the pro- portions of these combinations ( AA, AB, and BB) in a simple case will be in the ratio i : 2 : i. This assumption of the purity of the germ cells of hybrids has been found to conform with the facts in a number of simple experiments where two characters such as A and B were sharply contrasted. When one of the characters in the hybrid is dominant and the other recessive the ratio can be expressed as D : UR : R as i : 2 : i which is merely a substitution of D and R for the characters A and B. There are of course many factors which tend to modify the ratios as stated above and complicate the results. Thus the normal number of gametes may be of varying vigor and mortal- ity so that there will be proportionately more or less of one type of fusion than is called for by the law of chance. Sometimes the characters of the parents remain evenly balanced in the hybrid and refuse to split up in the succeeding generations, remaining in a stable union in the germ cells produced by the hybrid. Such conditions prove exceptions both to the law of dominance and to that of purity of the germ cells. From these exceptions and particularly the last it is difficult to believe that any large proportion of the germ cells is absolutely pure, i. e., bearing only the pure characters of one parent or the other. H'ovvever, there is much evidence from our knowledge of the distribution of the chromosomes from one generation to the next, that certain relations are possible in the separation of germ plasm which approximate the ratios of Mendel's law and while rarely giving absolutely pure germ cells nevertheless do make possible a large proportion of relatively pure cells. Let us examine now the chromosome history as a possible physical basis for the Mendalian principles. Such considerations No. 464.] STUDIES ON PLANT CELL. VII. 575 must rest on the assumption of what is termed the individuality of the chromosome. This means that the chromosome is believed to be a permanent organ of the cell which-neiLcr loses its organic entity although the form may be frequently obscured, as in the resting nucleus, and which reproduces by fission during mitosis. We have given in other connections the evidence upon which the above view rests, evidence -accumulated from the studies of the critical periods of garnetogenesis, fertilization, and sporogenesis (with its reduction phenomena) in plants and of garnetogenesis and fertilization in animals. All investigations indicate that paternal and maternal chromosomes maintain com- plete independence in the sexually formed cell or fertilized egg and in the mitoses of cleavage so far as these have been fol- lowed. Also, descendants of the chromosomes which became associated with fertilization have been recognized by their form at the end of the life history during the reduction phenomena of garnetogenesis in certain animals (Sutton, : 02, : 03 ; Montgom- ery, : 04) and of sporogenesis in the hybrids of Drosera (Rosen- berg, : O4a, :O4b). Furthermore, the entire history of chromo- some reduction in both animals and plants finds a satisfactory explanation only in the belief that descendants of maternal and paternal chromosomes are distributed as organic entities by the peculiar mitoses of this period. * There is a general agreement that the somatic chromosomes of animals and the sporophytic of plants become grouped in pairs to form bivalent structures before the heterotypic mitosis of the reduction division whether this be present in the primary gametocyte (animals) or the spore mother-cell (plants). The bivalent chromosomes (pairs of chromosomes, dyads) may be- come transformed into tetrads before the heterotypic mitosis by a division of each chromosome in the pair, as is characteristic of animals, or this division may be delayed until a somewhat later period during the heterotypic mitosis, as in plants. We are not concerned now with the dispute as to how the pairs of chromo- somes come to lie side by side to form the bivalent structure or how tetrads are developed, activities which may indeed be vari- ous in different types and which will only be understood by a greater body of observations than we have at present (see dis- 576 THE AMERICAN NATURALIST. [VOL. XXXIX. cussion of " Reduction of Chromosomes "). The important point for us is the belief that the appearance of the bivalent chromosomes during reduction is due to the temporary union of somatic or sporophytic chromosomes in pairs and further that ths reducing divisions distribute the members of the pair, which are believed to be descendants of the maternal and paternal chromosomes of the previous generation, as organic entities to the generation which is to follow. It is difficult to overestimate the importance of this general- ization. If the program prove to be correct as stated above and if the chromosome is established beyond doubt as a self-perpet- uating organ of the cell and a bearer of hereditary characters we have then the possibility of studying the actual manner in which these structures are passed on from one generation to the next and perhaps determine the ratios or combinations through which the distribution is effected. The difficulty of making an exact determination of ratios in any form so far studied lies in our inability to distinguish the chromosomes of maternal and paternal origin. There is much evidence that the pairs of somatic and sporophytic elements, which form the bivalent chromosomes of the reduction mitoses of animals and plants respectively, are of different parentage but we do not know whether, or not there is any rule in the arrangement of the pairs on the spindles of these mitoses although this is hardly to be expected. Cannon (:O2, : O3a) and others have held that the mitoses of reduction brought about the complete separation of the maternal and paternal chromosomes so that two of the resultant four nuclei contain chromosomes from one parent and two from the other, and the germ cells are in consequence abso- lutely pure in character. But this view was soon shown by Sut- ton (103, p. 233; accepted by Cannon, : O3b) to be at variance with the facts of breeding for if germ cells of hybrids are abso- lutely pure there could be no further change by cross-breeding and the first cross would be repeated over and over again with- out any divergence from the type, which is contrary to experi- ence and fact. The pairs of chromosomes are probably arranged in every possible order and the maternal and paternal elements are distributed in every possible combination by the reducing No. 464.] STUDIES ON PLANT CELL. VII. 577 divisions. If this is true then by the law of chance the propor- tions of germ cells of the hybrid which are absolutely pure (con- taining chromosomes entirely from one parent) woukf be small. Likewise there would be a small proportion of germ cells in which the paternal and maternal chromosomes are equally dis-. tributed. And in contrast to these two groups the great major- ity of germ cells would have a marked preponderance of chromo- somes derived from one parent or the other and this condition may be termed one of relative purity. We shall now summarize the cytological evidence for the con- clusions of the paragraph above, first with respect to the actual distribution of the somatic and sporophytic chromosomes as entities during the mitoses of reduction, and second as to the probability of the bivalent chromosomes consisting of a pair of maternal and paternal elements. The evidence on the first point has been treated as regards plants in our own account of " Reduction of the Chromosomes " and need not be repeated. With respect to the possibilities of distinguishing maternal and paternal chromosomes throughout a life history and especially at the period of chromosome reduction we must consider briefly the remarkably favorable studies of Button, Montgomery, Moenk- haus, Baumgartner, and Rosenberg. Sutton (:O2, : 03) discovered in the "lubber grasshopper" (Brachystola magna) a form in which the somatic chromosomes, 23 in number, are markedly different in size, presenting a graded series with respect to pairs in which the two elements are ap- proximately equal. There are then 1 1 types of chromosomes in two groups, a pair of each type, and in addition an accessory chromosome which remains apart from the rest in a special vesicle of its own. These two sets of 1 1 chromosomes appear with regularity throughout the mitoses leading up to the reduc- tion divisions of spermatogenesis. Previous to the reducing divisions the chromosomes of each pair become closely asso- ciated end to end so that 1 1 threads appear which form 1 1 biva- lent chromosomes (dyads) that later become tetrads through the division of each chromosome in the pair. Sutton concludes that the somatic chromosomes which make up each bivalent structure conjugate during synapsis and that the transverse fission which 578 THE AMERICAN NATURALIST. [VOL. XXXIX. appears during the formation of the tetrad simply separates the two somatic chromosomes of the pair, while the longitudinal fission is the usual division of chromosomes, appearing prema- turely at this time. The conclusion is natural that the two series of the 1 1 pairs consist of maternal and paternal chromo- somes which are distributed as organic entities by the reducing divisions. But there are no reasons for supposing that all of the paternal chromosomes will pass into one set of germ cells and all of the maternal into another but rather that the ratios of distribution will be by the law of chance according to which the great majority of germ cells will have a marked preponder- ance of chromosomes from one parent or the other, and will therefore be relatively pure. An exceedingly small proportion of germ cells may, by the law of chance, contain chromosomes entirely of maternal or paternal extraction, and an equally small proportion may contain 6 chromosomes of one parent and 5 of the other. The accessory chromosome divides but once during the mitoses of spermatogenesis so that two of the spermatozoids have ii chromosomes and two 12. No accessory chromosome appears in the mitoses of ob'genesis indicating that the female insect lacks this structure which confirms the belief of McClung (: 02) and others that the accessory chromosome is a determin- ant of the male sex. Montgomery in a series of studies upon insects and Amphi- bians, which are summarized in a recent paper (: 04), reached conclusions in striking support of the theories of the individu- ality of the chromosomes, the association of pairs of chromo- somes during synapsis to form bivalent structures and the prob- ability of the elements of each pair (bivalent chromosomes) being of maternal and paternal origin respectively. His results on the last point are of especial interest in relation to hybridization. In a large number of insects, chiefly Hemiptera, Montgomery has found pairs of chromosomes, which he terms heterochromo- somes, much smaller or much larger than the others and these may be followed through mitosis from one nucleus to another. The heterochromosomes of each pair are known to unite with one another during synapsis to form the bivalent chromosomes of the reduction mitoses and they then become separated, each No. 464.] STUDIES ON PLANT CELL. VII. 579 dividing once, so that every germ cell receives a single hetero- chromosome of whatever sort. Fertilization then brings the heterochromosomes together again in pairs until the Jiexl period of chromosome reduction. This history is then parallel to Button's account of the lubber grasshopper (Brachystola), the difference being that the latter form presents a remarkably graded set of paired chromosomes. Montgomery regards the small chromosomes and the accessory chromosome as structures tending to disappear in a process of evolution from a higher chromosomal number to a lower. Moenkhaus (: 04) crossed reciprocally two, species of fishes (Fnndulus Jieteroclitus and Menidia no tat a} and obtained hybrid embryos which reached an advanced stage of development. The chromosomes of the parents are readily distinguished by size and form. These chromosomes were followed throughout the development of the hybrid embryo and were found to retain their peculiarities so that the two sets may be easily separated in favorable tissues. This investigation furnishes some of the strongest evidence of the individuality of the chromosome and the complete independence throughout the life history of the two sets derived from each parent. Could these hybrid embryos be raised to maturity we should expect to find during spermato- genesis and oogenesis an association of the chromosomes in pairs, those of paternal extraction with those of maternal to form the bivalent chromosomes preliminary to the reducing divi- sions, and a distribution to the sexual cells in varying propor- tions which would, however, give a very large ratio of relatively pure germ cells. Baumgartner (: 04) in studies upon spermatogenesis in the cricket (Gryllus) was able to distinguish the chromosomes by their form, following them through the mitoses of reduction. Most of the chromosomes have the form of straight or bent rods but there are apparently two rings in each set in G. domesticus. The variation in the form of chromosomes in the nucleus is well known but it has not been supposed that a definite form might be characteristic of an element and be maintained throughout the successive mitoses of a life history as seems probable from Baumgartner's results. 580 THE AMERICAN NATURALIST. [VOL. XXXIX. Rosenberg's (: 04.3., : O4b) studies on hybrids of Drosera rotun- difolia (with ten chromosomes in the gametophyte) and D. longi- folia (with twenty chromosomes) offer clear evidence that the chromosomes which unite in pairs to form bivalent structures preliminary to the reduction phenomena of sporogenesis are of different parentage. The sporophyte number of chromosomes in the hybrid is thirty, as would be expected. The reduced number appearing at the first mitosis of sporogenesis is, however, not fifteen but twenty chromosomes, ten of which are plainly double the size of the other ten. The explanation of this inter- esting condition is that the ten chromosomes of D. rotnndifolia unite with one half of the twenty chromosomes of D. longifolia giving ten large bivalent structures accompanied by the ten chromosomes of D. longifolia which are without mates. This explanation finds clear support in the facts that the chromo- somes of D. rotnndifolia are larger than those of D. longifolia and that the bivalent. structure consists of a larger and a smaller element thus giving clear evidence that the pairs of chromosomes which unite in Drosera are of different parentage. The single chromosomes which are without mates may pass to one or the other of the poles of the spindle or may be left behind when the daughter nuclei are formed. This group of investigations illustrates very clearly the charac- ter of the evidence that is leading many biologists to assign to the chromosomes the functions of bearing and distributing hered- itary characters. The question at once comes up as to whether or not the chromosomes may differ among themselves to a greater or less extent even in the same species or individual. Montgomery, Sutton, with others, have established a difference in the size of chromosomes. Baumgartner distinguishes differ- ences inform in the same species and the studies of Moenkhaus and Rosenberg have shown that the chromosomes of different parents may retain their peculiarities of form in hybrids and be really separated. To these investigations should be added the recent conclusions of Boveri ( : 02, : 04), that chromosomes actu- ally differ in function. Boveri found that the chromosomes of eggs of echinoderms that were fertilized by two or more sperms are distributed by multipolar spindles to a varying number of No. 464.] STUDIES ON PLANT CELL. VII. 581 blastomeres which in consequence received a varying number and assortment of chromosomes. Boveri then separated these blastomeres and followed their independent development into larval stages which exhibited marked differences in form that could be correlated with the irregularities in the number of chromosomes contained in each, thus suggesting that specific chromosomes have specific functions. With this sort of evi- dence accumulating from both the morphological and physio- logical side it is not surprising that many biologists believe that specific characters are actually held or are controlled by chromosomes or groups of chromosomes. Such views of course presuppose that the chromosomes retain a high degree of independence of one another and that variation is expressed chiefly through different combinations of chromo- somes and not by modifications of the chromosomes themselves. Yet there is strong evidence of an actual mixing or interchange of the idioplasm among the chromosomes. This possibility which is of course contradictory to the view of the complete independence of the chromosomes finds its chief support in the close association of the pairs of chromosomes with the organiza- tion of the reduced number of bivalent structures during synap- sis. These pairs have been reported so intimately united as to be actually fused. Allen (:O5) has described for Lilium the union of two sets of chromomeres, one believed to be derived from a paternal spirem and the other from a maternal, which come to lie side by side during synapsis and unite to form a spirem with a single series of fusion chromomeres. This single (fusion) spirem later splits longitudinally and the two halves are regarded as again representing maternal and paternal spirems but there are evidently opportunities during the period of fusion for significant reciprocal interaction between the two idioplasms. This conception of the fusion of idioplasm from the two/ parents is an old view which has been held by such well known biologists as Hertwig and Strasburger. De Vries (: 03) has recently discussed the significance of the pairing of chromosomes before the heterotypic mitosis in relation to the theory of pangenesis. He conceives the paternal and maternal chromosomes as coming together during synapsis in 582 THE AMERICAN NATURALIST. [VOL. XXXIX. homologous pairs so that corresponding pangenes or groups of pangenes are brought together and that there may be a mutual interchange or transfer of idioplasm with the result that the chromosomes after separating may contain a mixed set of pan- genes although each is supposed to have a complete assortment. The interchange makes possible all forms of combinations of the pangenes in the two sets, according to the laws of chance, which might be expressed in proportions that would approximate in some cases the ratios of Mendel. If the parents are widely different from one another their idioplasm may not correspond sufficiently to make possible this union and interchange of pangenes so that the process is suppressed and the hybrid is sterile. Allen (:O5, p. 247) points out that the union of two spirems during synapsis with the fusion of two sets of chromomeres, according to his account of the lily, offers a number of possibil- ities with respect to the constitution of idioplasm following the reduction mitosis, (i) There may be such a fusion of elemen- tary units that a single idioplasm is formed different from either parent which would of course be distributed equally to the reproductive cells by the subsequent double longitudinal fission of the single (fusion) spirem. This would be expected to give hybrids of much the same form in every instance and these would remain stable (constant). (2) There may be a greater or less mixing or modification of units but without the actual union and formation of a new idioplasm in the hybrid. Then by the splitting of the single (fusion) spirem there might result a dis- tribution of the mixed idioplasm following ratios or proportions approximating Mendel's law. (3) There may be in part a fusion and in part a mixing of idioplasm which would be expected to result in a .varied combination of parental characters in the off- spring. (4) While the chromosomes may be distributed accord- ing to ratios similar to Mendel's principles their respective characters may be greatly modified by their temporary union during synapsis. (5) Portions of the idioplasm may interact upon one another so that when they are separated by the reduc- tion mitoses their character has become variously modified. (6) Finally, Allen, of course, recognizes the possibility that parental No. 464.] STUDIES ON PLANT CELL. VII. 583 idioplasm may be separated so purely by the longitudinal split- ting of the single (fusion) spirem or through the distribution of unmodified sporophytic or somatic chromosomes as~to-give abso- lutely and relatively pure germ cells through Mendelian laws. Allen's discussion, very briefly summarized above, is impor- tant for the emphasis which is laid upon the significance of a possible mixing of the parental idioplasms in the more or less complete union of chromatic material, which is generally recog- nized as characteristic of synapsis. There is a general tendency to rest content when the chromosomes are accounted for as units while they are merely the grosser form of expression of the idioplasm whose final architecture is intricate far beyond our present powers of analysis. Allen's own studies upon the events of synapsis in the lily with the regular fusion in pairs of chromo- meres of different parentage may well cause one to hesitate in a full acceptance of the chromosome as fixed and unchanged in its organic constitution throughout the life history. The phenome- non of hybridization is far too complex to be explained in terms of simple ratios and while some characters may be paired or correlated in proportions that can be expressed by mathematical formulae there is little probability that the assemblage of charac- ters which make the species can be so definitely grouped as the strongest disciples of Mendel may hope. However, a great forward step has been taken and we may expect important results from the empirical methods so clearly defined by Mendel and by the close investigation that cytologists are making of the history of idioplasmic structures (chromosomes) during ontogeny. 8. XENIA. Xenia is the " immediate or direct effect of pollen on the character of seeds and fruits." The term was first proposed by Focke, in 1881, and is now well established. Xenia has long been known to the plant breeder as one of the most interesting and puzzling problems of hybridization. The botanist has nat- urally looked for the results of hybridization in the development of the embryo from the seed since this structure has received 584 THE AMERICAN NATURALIST. [V<>L. XXXIX. the substance of the sperm nucleus of the male parent. But facts have clearly shown that the pollen may also affect the structure of the endosperm in the seed as well as cause the development of the embryo. Since the endosperm holds no genetic relation to the embryo it has seemed very remarkable that it should take on hybrid qualities. It has also been claimed that other regions of the seed or fruit, such as portions of the pericarp were also affected, but it is doubtful whether this is really so or at least whether such changes are truly a feature of the protoplasmic structure and thus deeply seated in the organ- ism as a feature of hybridization. It is only within recent years that a satisfactory theory has been suggested for the influence of pollen outside of the embryo. And this explanation rests on the discovery of the activities of the second sperm nucleus which enters the embryo-sac and which is known in some cases to unite with the polar nuclei constitut- ing a triple nuclear fusion within the sac that is generally known as "double fertilization." We have briefly referred to the phe- nomenon in the latter part of the account of " Asexual Cell Unions and Nuclear Fusions" in Section IV and shall take it up now in greater detail. The best account of xenia is a very clear treatment by Webber, in 1900. The explanation of xenia upon the facts of "double fertiliza- tion " was proposed almost simultaneously by De Vries ('99, :oo), Correns ('99b), and \Vebber (:oo). Double fertilization was first observed by Nawaschin ('98) in Lilium and Fritillaria and shortly after was described in greater detail by Guignard ('99b) in other species of the same genera and in Endymion. Since these discoveries the phenomenon has been reported by a number of investigators in many other forms representing widely divergent groups in the Monocotyledons and Dicotyledonae and there is every reason to believe that it is widespread in the angio- sperms. A review of the recent literature is given by Coulter and Chamberlain (Morphology of the Angiosperms, 1903, p. 156). There is no fixed order in the events of the triple nuclear fusion of "double fertilization." The polar nuclei may have united at the time when the pollen tube enters the embryo-sac, in which case the second sperm nucleus coalesces with an organized fusion No. 464.] STUDIES ON PLANT CELL. VII. 585 endosperm nucleus. Or, the two polar nuclei and the sperm nucleus may all three fuse together practically simultaneously. And again the sperm nucleus may unite first with o_ne_of the polar nuclei and the second be drawn later into the triple fusion. But no cases seem to have been reported in which but one polar nucleus unites with the sperm leaving the other free although such a combination may be expected. Also, no one has ob- served an independent division of the sperm nucleus within the endosperm, although as we shall see, there are reasons for believ- ing that such a development may sometimes take place. We have already given in Section IV the reason why these triple nuclear fusions may be kept apart from sexual phenomena since we have no knowledge of the phylogenetic history of the processes involved. It seems best at least for the present to regard the phenomenon as a special development associated with the peculiar and highly specialized conditions within the embryo- sac. This detailed and highly difficult problem of phylogeny has no especial bearing on the physiological features of xenia with which we are at present concerned. The best understood examples of xenia are found in the hybrids of maize and are clearly described in the very interest- ing paper of Webber (:oo). As is well known, some of the varieties of corn are distinguished among other characters by the color of the kernels, which are blue, red, yellow, and white, and also by the surface, which is smooth in the starchy corns (flint or dent) and wrinkled in the sugary sweet corns. When well marked pure races are grown out of reach of chance cross- pollination, the offspring remain true to their seed characters but it has long been known that the varieties of corn hybridize very readily so that when grown close together the ears will very frequently present seeds mixed as to color and texture. Thus when exposed to cross-pollination a corn which is characteris- tically yellow or white may bear blue or red kernels or a form with wrinkled and starchy kernels may develop smooth starchy corn if varieties with these characters are in the vicinity. The color character is known to lie in these examples in the outer layer of the endosperm (aleurone layer) and of course the food material whether prevailingly starch or sugar, which gives the 586 THE AMERICAN NATURALIST. [VOL. XXXIX. surface a texture smooth or wrinkled, is stored within the endo- sperm. The clearness of xenia in the maize has led to a number of careful studies on cross-pollination beginning with the work of Vilmorin (1866), Hildebrand (1867), and Friedrich Kornicke (1872). The possible explanation of xenia in maize through "double fertilization" which introduces qualities of the male parent from the pollen into the endosperm was suggested by experiments of De Vries on hybridizing maize in the summers of 1898-99 and Correns and Webber in 1899. De Vries ('99, : oo) pollinated a wrinkled-seeded sugar corn from a variety of smooth starchy corn and obtained smooth starchy kernels which when cultivated in the succeeding summer were found to be true hybrids. He concluded that this furnished experimental proof that the endosperm of the sugar corn was affected by the entrance of a sperm nucleus from the starchy variety according to the theory of "double fertilization " proposed by Nawaschin ('98). Correns ('99b) in the same year expressed similar conclusions in a clear statement of the theoretical aspects of the problem of xenia as found in Zea mays. Correns advanced a number of propositions some of which should be noted for their speculative interest. Thus he states (proposition 7) that the influence of the new pollen (i. e., from the male parent of the hybrid) is expressed as xenia only in the endosperm and (proposition 8) only in the pigment present or the chemical nature of the reserve material whether starchy or sugary. If the two races differ only in the presence of one character, as in the color of the aleurone layer, that character is only found in xenia when brought by the pollen (proposition 10). Xenia is then only expressed in a hybrid (proposition 14) by the formation of a pig- ment which the race of the female parent does not possess or of a more complicated chemical compound (such as starch) in place of a simpler (as dextrin). Correns (: 01) later presented in a lengthy paper, beautifully illustrated, the full results of his studies on xenia in maize with a discussion of the hybrids. Webber (:oo) also simultaneously with De Vries and Correns conducted extensive experiments in hybridizing a number of No. 464.] STUDIES ON PLANT CELL. VII. 587 varieties of corn distinguished by the color of the kernels, which were white, yellow, red, or blue and by the texture whether smooth, hard, and starchy (dent or flint corn) or wrinkled and sugary (sweet corn). The results of his investigation are admir- ably presented with excellent illustrations. He found that the smooth kernel and starchy endosperm of the dent and flint corn were transmitted very conspicuously as xenia when these forms were employed as the male in crossing with the sweet corns whose kernels are wrinkled and sugary. The characters of the sweet corns do not seem to be expressed as xenia when smooth, starchy, dent corn is used as the female member of the hybrid. This experiment would seem to support Correns' proposition number 14 that a more complicated compound is always formed in xenia in place of a less complex. But Webber found that flint corn, which is smooth and starchy, when pollinated with a form of sweet corn developed the wrinkled kernel and sugary type of endosperm of the male member indicating that this rule of Correns is not universal. And McClure ('92) obtained simi- lar results in crossing a white dent race with pollen of Black Mexican which is a sugar corn with black kernels. The product in this case showed xenia clearly in having the wrinkled blue- black kernels of the male sugar corn. Some of Webber's most striking results were obtained in pol- linating yellow and white corns with blue-black and red races. The color was transmitted as xenia in a most striking manner. Webber agrees with other authors that the color is only present in the endosperm of the kernels. Thus the red of certain dent corn, which lies in the pericarp, is not passed on as xenia and McClure observed the same facts in experiments with cranberry corn whose color lies in the seed coat and is not transmitted when employed as the male member in crossing with white corns. Webber's experiments show, as do those of other inves- tigators, that the absence of color in the kernels of the male parent does not seem to affect the tint of the kernels when the female is markedly colored, in agreement with Correns' proposi- tion number 10. But Webber is not convinced that some influence might not be exerted on colored corn when pollinated from races with colorless endosperm, because of certain experi- ments on variegated xenia which will be described presently. 588 THE AMERICAN NATURALIST. [VOL. XXXIX. These experiments of De Vries, Correns, Webber, and others have established experimentally the facts of xenia and Nawas- chin's theory of double fertilization seems to offer the only explanation of the phenomenon. But it was left to Guignard (:oi) to make the concluding observation that a second sperm nucleus does actually enter into the composition of the endo- sperm of maize, and this fact clinched the argument which up to this time had been a speculation. Webber has made a very important addition to the theory of " double fertilization " as an explanation of xenia in some obser- vations and speculations on a mottled condition which is some- times present when white corns are pollinated by colored. He found that the color was sometimes only transmitted in spots as when Hickory King was pollinated by Cuzco, or perhaps only half a kernel may be colored. Webber offers the hypothesis that the second sperm nucleus may enter the embryo-sac but instead of uniting with the two polar nuclei to form a triple fusion may itself divide separately and thus gives rise to a progeny very different from the other endosperm nuclei. There might then be two sets of nuclei in the endosperm one of which is composed of nuclei which would come directly from the male parent. These latter then might become distributed throughout the embryo-sac but would tend to remain in groups as multiplication progressed and would certainly be expected to influence the character of the tissue which is formed later when the walls are developed around the free nuclei. As Webber expresses it, there might be formed islands of tissue in the endosperm whose cells contain nuclei derived directly from the second sperm and such tissue would be expected to show char- acters of the male parent in spots as xenia. Again, if the sperm nucleus should unite with only one of the polar nuclei and the other should give rise to an independent progeny we should expect similar mixed conditions in the endosperm, with xenia only expressed in the areas dominated by nuclei containing material derived from the sperm. There have been reported illustrations of xenia in tissues out- side of the endosperm but we are fully justified in awaiting their confirmation before accepting them, especially since some No. 464.] STUDIES OF PLANT CELL VII. 589 have failed to stand the test of critical investigation, in the light of the present theory. Thus certain investigators have reported xenia in the color of the seed coats of certain peas. But Giltay ('93) in a series of experiments found no instance where color was transmitted to these tissues. The pigments in these plants lie in the cotyledons of the embryo which of course are readily visible through the thin coats of the seed. While the present theory of xenia is very recent and has been critic- ally applied in few forms, it seems thoroughly satisfactory in every particular with no clearly established evidence against it. LITERATURE CITED IN SECTION V, "THE PLANT CELL." ALLEN. : 04. Chromosome Reduction in Lilium canadense. Bot. Gaz., vol. 37, p. 464. ALLEN. :05. Nuclear Division in the Pollen Mother-Cells of Lilium canadense. Annals of Bot., vol. 19, p. 189. ATKINSON. '99. Studies on Reduction in Plants. Bot. Gaz., vol. 29, p. i. BAUMGARTNER. :04. Some New Evidence for the Individuality of the Chromosomes. Biol. Bull., vol. 8, p. i . BELAJEFF. '98. Ueber die Reductionstheilung des Pflanzenkernes. Ber. d. deut. hot. Gesellsch., vol. 16, p. 27. BEARD. '95. On the Phenomena of Reproduction in Animals and Plants. Anti- thetic Alternation of Generations. Annals of Bot., vol. 9, p. 441. BERGHS. : 04a. Depuis le spirem jusqu'aux chromosomes murs dans la micro- sporoge"nese d ^ A Ilium fistulosum et de Lilium lancifolium (speci- osuin). La Cellule, vol. 21, p. 173. BERGHS. : 04b. Depuis la sporogonie jusqu'au spireme de'fmitif dans la micro- sporogdnese de V A Ilium fistulosum. La Cellule, vol. 21, p. 383. 590 THE AMERICAN NATURALIST. [VOL. XXXIX. BERGHS. : 05. La microsporoge'ne-se de Convallaria maialis. La Cellule, vol. 22, p. 43. BITTER. :04. Parthenogenesis und Variabilitat der Bryonia dioica. Abh. nat. Ver. Brem., vol. 18, p. 99. BLACKMAN. : 04a. On the Fertilization, Alternation of Generations, and General Cytology of the Uredineae. Annals of Bot., vol. 18, p. 323. BLACKMAN. : 04b. On the Relation of Fertilization to " Apogamy " and " Partheno- genesis." New Phytologist, vol. 3, p. 149. BOVERI. : 02. Ueber mehrpolige Mitosen als Mittel zur Analyse des Zellkerns. Verh. d. phys.-med. Gesellsch. Wiirzburg., vol. 35. BOVERI. : O4. Ergebnisse iiber die Konstitution der chromatischen Substanz des Zellkerns. Jena, 1904. BOWER. '86. On Apospory in Ferns. Jour. Linn. Soc. London, Bot.,vo\. 21, p. 360. BOWER. '87. On Apospory and Allied Phenomena. Trans. Linn. Soc. London, Bot., II, vol. 2, p. 301 . BOWER. '88. On some Normal and Abnormal Developments of the Oophyte in Trichomanes. Annals of Bot., vol. i, p. 269. BRIZI. '92. Apunti di teratologia briologica. Ann. d. Instit. Bot. d. Roma, vol. 5, p. 53. CALKINS. '97. Chromatin Reduction and Tetrad-formation in Pteridophytes. Bull. Torrey Bot. Club, vol. 24, p. 101. CALKINS. :01. The Protozoa. New York, 1901. CHAMBERLAIN. :05. Alternation of Generations in Animals from a Botanical Stand- point. Bot. Gas., vol. 39, p. 137. CANNON. : 02. A Cytological Basis for the Mendelian Laws. Bull. Torrey Bot. Club, vol. 29, p. 657. CANNON. :03a. The Spermatogenesis of Hybrid Cotton. Bull. Torrey Bot. Club, vol. 30, p. 133. No. 464.] STUDIES ON PLANT CELL. VII. 591 CANNON. : 03b. The Spermatogenesis of Hybrid Peas. Bull. Torrey Bot. Club, vol. 30, p. 519. CORRENS. '99a. Untersuchungen iiber die Vermehrung der Laubmoose durch Brutorgane und Stechlinge. Jena, 1899. CORRENS. '99b. Untersuchungen iiber die Xenien bei Zea Mays. Ber. d. dent. hot. Gesellsch., vol. 17, p. 410. CORRENS. : 01. Bastarde zwischen Maisrassen mit besonderer Beriicksichtigung der Xenia. Bibliotheca botanica, Heft 53. Stuttgart, 1901. COULTER AND CHAMBERLAIN. :03. Morphology of the Angiosperms. New York, 1903. COULTER AND LAND. :05. Gametophytes and Embryo of Torrey a taxifolia. Bot. Gaz., vol. 39, p. 161. DAVIS. '96. Development of the Procarp and Cystocarp in the genus Ptilota. Bot. Gaz., vol. 22, p. 353. DAVIS. '99. The Spore Mother-Cell of Anthoceros. Bot. Gaz., vol. 28, p. 89. DAVIS. : Ola. Nuclear Studies on Pellia. Annals of Bot., vol. 15, p. 147. DAVIS. : Olb. The Origin of Sex in Plants. Pop. Sci. Month., vol. 60, p. 66. DAVIS. : 03a. The Evolution of Sex in Plants. Pop. Sci. Month., vol. 62, p. 300. DAVIS. :04a. Oogenesis in Vaucheria. Bot. Gaz., vol. 38, p. 81. DE BARY. '78. Ueber apogame Fame und die Erscheinung der Apogamie in All- gemeinen. Bot. Zeit., vol. 36, p. 449. DE VRIES. '99. Sur la fdcondation hybride de 1'albumen. Coinpt. Rend. d. Sean. d. I'Acad. d. Sci., vol. 129, p. 973. DE VRIES. :00. Sur la fe'condation hybride de 1'endosperm chez le ma'is. Rev. Gen. d. Bot., vol. 12, p. 129. DE VRIES. :03. Befruchtung und Bastardierung. Leipzig, 1903. DIXON. '95. The Nuclei of Lilium longiflorum. Annals of Bot., vol. 9, p. 663. 592 THE AMERICAN NATURALIST. [VOL. XXXIX. DlXON. '96. On the Chromosomes of Liliuin longiflornm. Proc. Roy. Irish Acad., vol. 3, p- 707. DIXON. :00. On the First Mitosis of the Spore Mother-cells of Lilium. Proc. Roy. Itish Acad., vol. 6, p. i. DRUERY. '86a. Observations on a Singular Mode of Development in the Lady- fern (Athyrium _filix-fa>mina). Jour. Linn. Soc. London, Bot., vol. 21, p. 354. DRUERY. '86b. Further Notes on a Singular Mode of Reproduction in Athyrium filix-fcemina, var. clarissima. Jour. Linn. Soc. London, Bot., vol. 21, p. 358. DRUERY. '93 Notes on an Aposporous Lastrea (Nephrodium). Jour. Linn. Soc. London, Bot., vol. 29, p. 479. DRUERY. '95. Notes upon Apospory in a Form of Scolopendrium vulgare, var. crispum, etc. Jour. Linn. Soc. London, Bot., vol. 30, p. 281. DKUERY. :01. Fern Variation in Great Britain. Bot. Gaz., vol. 31, p. 347. ERNST. :01. Beitrage zur Kenntniss der Entwickelung des Embryosackes und des Embryo (Poly embryonic) von Tulipa gesneriana L. Flora, vol. 88, p. 37. FARLOW. '74. An Asexual Growth from the Prothallus of Pteris cretica. Quart. Jour. Mic. Set., vol. 14, p. 266. FARLOW. '89. Apospory in Pteris aquilina. Annals of Bot., vol. 2. p. 383. FARMER. '94. Studies in Hepaticas : On Pallavicinia decipiens Mitten. Annals of Bot., vol. 8, p. 35. FARMER. '95a. Spore-formation and Karyokinesis in Hepatics. Annals of Bot., vol. 9, p. 363. FARMER. '95b. On Spore-formation and Nuclear Division in the Hepatica;. Annals of Bot., vol. 9, p. 469. FARMER. '95c. Further Investigations on Spore-formation in Fegatella conica. Annals of Bot., vol. 9, p. 666. FARMER. '95d. Ueber Kerntheilung in Lilium-antheren, besonders in Bezug auf die Centrosomenfrage. Flora, vol. 80. p. 56. No. 464.] STUDIES ON PLANT CELL. VII. 593 FARMER AND MOORE. '95. On the Essential Similarities existing between the Heterotype Divisions in Animals and Plants. Anat. Anzeig., vol. n, p. 71. FARMER AND MOORE. : 03. New Investigations into the Reduction Phenomena of Animals and Plants. Proc. Roy. Soc. London, vol. 72, p. 104. FARMER AND MOORE. :05. On the Maiotic Phase (Reduction Division) in Animals and Plants. Quart. Jour. Mic. Set'., vol. 48, p. 489. FARMER, MOORE, AND DIGBY. : 03. On the Cytology of Apogamy and Apospory. Proc. Roy. Soc. London, vol. 71, p. 453. FARMER AND WILLIAMS. '98. Contributions to our Knowledge of the Fucaceae : their Life-history and Cytology. Phil. Trans. Roy. Soc. London, vol. 190, p. 623 FERGUSON. :04. Contributions to the Knowledge of the Life History of Pinus with Special Reference to Sporogenesis, the Development of the Gametophytes and Fertilization. Proc. Washington Acad. Sci., vol. 6, p. I. GILTAY. '93. Ueber den directen Einfluss des Pollen auf Frucht und Samenbil- dung. Jahrb.f. wiss. Bot., vol. 25, p. 489. GREGOIRE. '99. Les cineses polliniques chez les Liliace'es. La Cellule, vol. 16, ? 2 35- GREGOIRE. : 04. La reduction nume'rique des chromosomes et les cineses de matu- ration. La Cellule, vol. 21, p. 297. GREGORY. :04. Spore Formation in Leptosporangiate Ferns. Annals of Bot., vol. 1 8, p. 445. GUIGNARD. '84. Recherches sur la structure et la division du noyau cellulaire. Ann. d. Sci. Nat., Bot., ser. 6, vol. 17, p. 5. GUIGNARD. '85. Nouvelles recherches sur le noyau cellulaire, etc. Ann. d. Sci. Nat., Bot., ser. 6, vol. 20, p. 310. GUIGNARD. '89. Etude sur les phenomenes morphologique de la fecondation. Bull.. Soc. Bot. d. France, vol. 36, p. 106. GUIGNARD. '91. Nouvelles e'tudes sur la fe'condation. Ann. d. Sci. Nat., Bot., svc^ 7, vol. 14, p. 163. 594 THE AMERICAN NATURALIST. [VOL. XXXIX. GUIGNARD. '99a. Le deVeloppement du pollen et la reduction chromatique dans le Naias major. Arch. d'Anat. Mic., vol. 2, p. 455. GUIGNARD. '99b. Sur les anthe"rozoi'des et la double copulation sexuelle chez les ve"ge"taux angiospermes. Compt. Rend., vol. 128, p. 864. GUIGNARD. :01. La double fe'condation dans le mai's. Jour. Bot., vol. 15, p. 37. HAECKER. '98. Ueber vorbreitende Theilungsvorgange bei Thieren und Pflanzen. Verh. d. deut. zool. Gesellsch., vol. 8, p. 94. HEGELMAIER. '97. Zur Kenntniss der Polyembryonie von A Ilium odorum, L. Bot. Zeit., vol. 55, p. 133. ISHIKAWA. '97. Die Entwickelung der Pollenkorner von Alliuin Jistnlosum, L., ein Beitrag zur Chromosomenreduktion im Pflanzenreiche. Jour. Coll. Set. Imp. Univ. Tokyo, vol. 10, p. 193. JUEL. '98. Parthenogenesis bei Antennariaalpina (L.) R. Br. Bot. Centralb., vol. 74, p. 369. JUEL. : 00. Vergleichende Untersuchungen iiber typische und parthenogenet- ische Fortpflanzung bei der Gattung Antennaria. Handl. Sven Vetensk. Akad., vol. 33. JUEL. : 04. Die Tetradentheilung in der Samenanlage von Taraxacum. Ark. f. Bot. K. Svenska Vetens. Akad., vol. 2, nc. 4. KOENICKE. :04. Die neueren Arbeiten iiber die Chromosomenreduction im Pflanzen- reich und daran anschliessende karyokinetische Probleme. Bot. Zeit., vol. 62, p. 305. LANG. '98. On Apogamy and the Development of Sporangia upon Fern Pro- thalli. Phil. Trans. Roy. Soc. London, ser. B, vol. 190, p. 187. LANG. :01. On Apospory in Anthoceros Icevis. Annals of Bot., vol. 15, p. 53- LLOYD. :02. The Comparative Embryology of the Rubiaceae. Mem. Torrey . Bot. Club, vol. 8, p. 27. LOTSY. '99. Belanophora globosa Jurigh. Eine wenigsten ortlich-verwittwete Pflanze. Ann. Jar. Bot. Buiten., ser. 2, vol. i, p. 174. No. 464.] STUDIES ON PLANT CELL. VII. 595 LOTSY. : 03. Parthenogenesis bei Gnetum ula Brogn. Flora, vol. 92, p. 397. LOTSY. : 04. Die Wendung der Dyaden beim Reifung der Tiereier als Stiitze fur die Bivalenz der Chromosomen nach der numerischen Reduk- tion. Flora, vol. 93, p. 65. LOTSY. :05. Die x-Generation und die 2x-Generation. Eine Arbeitshypothese. Biol. Centralb., vol. 25, p. 97. LYON. : 04. The Evolution of Sex Organs in Plants. Bot. Gaz., vol. 37, p. 280. McCLUNG. : 02. The Accessory Chromosome Sex Determinant ? Biol. Bull., vol. 3, p. 43. McCLURE. '92. Corn Crossing. ///. Agric. E.vp. Sta. Bull. 21. MOENKHAUS. : 04. The Development of the Hybrids between Fundulus heteroclitus and Menidia notata with Especial Reference to the Behavior of the Maternal and Paternal Chromosomes. Ainer. Jour. Anat., vol. 3, p. 29. MONTGOMERY. :01. A Study of the Chromosomes of the Germ Cells of Metazoa. Trans. Amer. Phil. Soc., vol. 20. p. 154. MONTGOMERY. : 04. Some Observations and Considerations upon the Maturation Phe- nomena of the Germ Cells. Biol. Bull., vol. 6, p. 137. MOORE, A. C. : 03. The Mitoses in the Spore Mother-cell of Pallavicinia. Bot. Gaz., vol. 36, p. 384. MOTTIER. '97. Beitrage zur Kenntniss der Kerntheilung in den Pollenmutterzellen einiger Dikotylen und Monokotylen. Jahrb.f. wiss. Bot., vol. 30, p. 169. MOTTIER. : 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., vol. 35, p. 250. MURBECK. : Ola. Parthenogenetische Embryobildung in der Gattung Alchemilla. Lunds. Univ. Arsskrift, vol. 36. MURBECK. : Olb. Ueber das Verhalten des Pollenschlauches bei Alchemilla arven- sis (L.) Scop, und das Wesen der Chalozogamie. Lunds. Univ Arsskrift, vol. 36. 596 THE AMERICAN NATURALIST. [VOL. XXXIX. MURBECK. :02. Ueber Anomalien im Baue des Nucellus und des Embryosackes bei parthenogenetischen Arten der Gattung Alchemilla. Lunds. Univ. Arsskrift, vol. 38. MURBECK. : 04. Parthenogenese bei den Gattungen Taraxacum und Hieracium. Botan. Not., 1904, p. 285. NATHANSOHN. : 00. Ueber Parthenogenesis bei Marsilia und ihre Abhangigkeit von der Temperatur. Ber. d. deut. bot. Gesellsch., vol. 18, p. 99. NAWASCHIN. '98. Resultate einer Revision der Befruchtungsvorgange bei Lilium martagon und Fritillaria tenella. Bull. Acad. Imp. Set. St. Petersb., vol. 9, p. 377. OSTENFELD. :04a. Zur Kenntniss der Apogamie in der Gattung Hieracium. Ber. d. deut. bot. Gesellsch., vol. 22, p. 376. OSTENFELD. :04b. Weitere Beitrage zur Kenntniss der Fruchtentwicklung bei der Gattung Hieracium. Ber. d. deut. bot. Gesellsch., vol. 22, p. 537. OVERTON, E. '93a. Ueber die Reduction der Chromosomen in den Kernen der Pflan- zen. Vierteljahrssch. d. naturf. Gesellsch. Zurich, vol. 38. OVERTON, E. '93b. On the Reduction of the Chromosomes in the Nuclei of Plants. Annals of Bot., vol. 7, p. 139. OVERTON, J. B. : 02. Parthenogenesis in Thalictrum purpurascens. Bot. Gaz., vol. 33, P- 363- OVERTON, J. B. : 04. Ueber Parthenogenesis bei Thalictrum purpurascens. Ber. d. deut. bot. Gesellsch., vol. 22, p. 274. PRINGSHEIM. '76. Ueber vegetative Sprossung der Moosfriichte. Monatsber. d. k. Akad. d. Wiss. Berlin, 1876, p. 425. RAUNKIAER. :03. Kimdannelse uden Befrugtning hos Maslkebotte (Taraxacum). BJ(. Tidsskr., vol. 25, p. 109. ROBERTSON. :04. Studies on the Morphology of Torreya calif arnica Torrey. II. The Sexual Organs and Fertilization. New Phytologist, vol. 3, p. 205. ROSENBERG. :03a. Das Verhalten der Chromosomen in einer hybriden Pflanze. Ber. d. deut. bot. Gesellsch., vol. 21, p. no. No. 464.] STUDIES OF PLANT CELL VII. 597 ROSENBERG. : 03b. Ueber die Befruchtung von Plasmopara alpina (Johans.). Bihang til K. Svenska Vet.-Akad. Hand., vol. 28, Afd. 3, No. 10. ROSENBERG. :04a. Ueber die Tetradenteilung eines Drosera-bastardes. Ber. d. deut. hot. Gesellsch., vol. 22, p. 47. ROSENBERG. :04b. Ueber die Reduktionsteilung in Drosera. Meddel. Stock. Hogs. Bot. Inst., 1904. ROSENBERG. :05. Zur Kenntniss der Reduktionsteilung in Pflanzen. Bot. Notts., part i, 1905. SADEBECK. '79. Schenk's Handbuch der Botanik, Breslau 1897. [An Account of Apogamy in Ferns]. Vol. I, p. 231. SARGANT. '96. The Formation of the Sexual Nuclei in Lilium martagon. I. Oogenesis. Annals of Bot., vol. 10, p. 445. SARGANT. '97. The Formation of Sexual Nuclei in Lilium martagon. II. Sper- matogenesis. Annals of Bot., vol. n,p. 187. SCHAFFNER. '97. The Division of the Macrospore Nucleus (of Lilium philadel- phicuni). Bot. Gaz., vol. 23, p. 430. SCHOENFELD. : 01. La spermatoge'nese chez le taureau et chez les mammiferes en general. Arch. d. Bio I., vol. 18, p. I. SCHREINER AND SCHREINER. :04. Die Reifungsteilung bei den Wirbeltieren. Ein Beitrag zur Frage nach der Chromatinreduktion. Anat. Anz., vol. 24, p. 561. SHAW. '97. Parthenogenesis in Marsilia. Bot. Gaz., vol. 24, p. 114. STAHL. '76. Ueber kiinstlich hervorgerufene Protonemabildung an dem Sporo- gonium der Laubmoose. Bot. Zeit., vol. 34, p. 689. STRASBURGER. '84. Neuere Untersuchungen iiber den Befruchtungsvorgang bei den Phanerogamen als Grundlage fiir eine Theorie der Zeugung. Jena, 1884. STRASBURGER. '88. Ueber Kern- und Zelltheilung in Pflanzenreiche nebst einem Anhang iiber Befruchtung. Hist. Beitr., Heft i. Jena, 1888. STRASBURGER. '94a. The Periodic Reduction of the Number of Chromosomes in the Life-history of Living Organisms. Annals of Bot., vol. 8, p. 281. 598 THE AMERICAN NATURALIST. [VOL. XXXIX. STRASBUKGER. '94b. Ueber periodische Reduktion der Chromosomenzahl in Entwick- lungsgang der Organismen. Rial. Centralb., vol. 14, pp. 817 and 849. STRASBUKGER. '95. Karyokinetische Probleme. Jahrb.f. iviss. Bot., vol. 28, p. 151. STRASBURGER. '97a. Kerntheilung und Befruchtung bei Fucus. Jahrb. f. iviss. Bot., vol. 30, p. 351 . STRASBURGER. '97b. Ueber Cytoplasmastrukturen, Kern- und Zelltheilung. Jahrb. f. wiss. Bot., vol. 30, p. 375. STRASBURGER. '97c. Ueber Befruchtung. Jahrb.f. iviss. Bot., vol. 30, p. 406. STRASBURGER. .OOa. Ueber Reduktionstheilung, Spindelbildung, Centrosomen und Cil- ienbildner im Pflanzenreich. Hist. Beitr., Heft 6. Jena, 1900. STRASBURGER. : OOb. Einige Bemerkungen zur Frage nach der doppelten Befruchtung bei den Angiospermen. Bot. Zeit., vol.. 58, p. 293. STRASBURGER. : 01. Ueber Befruchtung. Bot. Zeit., vol. 59, p. 353. STRASBURGER. : 04a. Anlage des Embryosackes und Prothalliumbildung bei der Eibe nebst auschliessenden Erorterungen. Fest. z. yoth Geburtstag Haeckel. STRASBURGER. :04b. Ueber Reduktionsteilung. Sitzber. Akad. Berlin, vol. 18, p. 587. STRASBURGER. : 04c. Die Apogamie der Eualchemillen und allgemeine Gesichtspunkte die sich aus ihr ergeben. Jahrb. f. wiss. Bot., vol. 41, p. 88. STRASBURGER AND MOTTIER. '97. Ueber den zvveiten Theilungsschritt in Pollenmutterzellen. Ber. d. dent. hot. Gesellsch., vol. 15, p. 327. SUTTON. : 02. On the Morphology of the Chromosome Group in Brachystola magna. Biol. Bull., vol. 4, p. 24. SUTTOX. : 03. The Chromosomes in Heredity. Biol. Bull., vol. 4, p. 231. TRETJAKOW. '95. Die Betheiligung der Antipoden in Fallen der Polyembryonie bei A Ilium odoruin L. Ber. d. deul. hot. Gesellsch., vol. 13, p. 13. TREUB. '98. L'organe femelle et 1'apogamie du Belanophora elongata Bl. Ann. Jard. Bot. Buiten., vol. 15, p. i. No. 464.] STUDIES OF PLANT CELL. VII. 599 TREUB. : 02. L'organe femelle et Pembryogdnese dans le Ficus hirta Vahl. Ann. Jard. Bot. Buiten., ser. 2, vol. 3, p. 124. TROW. -.04. On Fertilization in the Saprolegnieae. Annals of Bo/., vol. 18, p. 541- WEBBER. :00. Xenia, or the Immediate Effect of Pollen in Maize. U. S. Dept. Agric., Div. Veg. Path. Phys., Bull. 22. WILLIAMS. : 04a. Studies in the Dictyotaceae. I. The Cytology of the Tetra- sporangium and the Germinating Tetraspore. Annals of Bot., vol. 1 8, p. 141. WILLIAMS. : (Hb. Studies in the Dictyotaceas. II. The Cytology of the Gameto- phyte Generation. Annals of Bot., vol. 18, p. 183. WILSON. : 00. The Cell in Development and Inheritance. New York, 1900. WIMWARTER. : 00. Recherches sur I'ovoge'nese et I'organoge'nese de 1'ovaire des mammiferes (lapin et homme). Arch.d. Biol., vol. 17, p. 33. WINKLER. : 01. Ueber Merogonie und Befruchtung. Jahrb.f. wiss. Bot., vol. 36, P- 753- WINKLER. : O4. Ueber Parthenogenesis bei Wikstrcemia indica (L.) C. A. Mey. Ber. d. dent. hot. Gesellsch., vol. 22, p. 573. WOLFE. :04. Cytological Studies on Nemalion. Annals of Bot.,\o\. 18, p. 607. ar iff 1 STUDIES ON THE PLANT CELL. VTIL 1 BRADLEY MOORE DAVIS. SECTION VI. COMPARATIVE MORPHOLOGY AND PHYSIOLOGY OF THE PLANT CELL. WE shall devote this section to the discussion of a number of topics some of which have received brief mention in the pre- ceding papers of the series but with other subjects will now be considered in some detail. The material will be treated un- der the following five headings : 1. The simplest types of plant cells. 2. Comparisons of the structures of some higher types of plant cell with simpler conditions. 3. Some apparent tendencies in the evolution of mitotic phe- nomena. 4. The essential structures of the plant cell and their be- havior in ontogeny. 5. The balance of nuclear and cytoplasmic activities in the plant cell. i. THE SIMPLEST TYPES OF PLANT CELLS. There are three groups of plants which are conspicuous for the simplicity of their cell structure. They, are : the Cyano- phyceae (blue-green algae), Schizomycetes (bacteria), and the Saccharomycetes (yeasts). All three groups have received much attention and there has accumulated an extensive litera- ture which we shall not attempt to treat in detail, since it has been handled very fully by the specialists in these subjects. We shall, however, present the most important conclusions and 1 This paper concludes the series of studies on the plant cell. The author has a number of complete sets of reprints of this and the earlier sections. Enquiries may be addressed to Professor Bradley M. Davis, University of Chicago. 695 696 THE AMERICAN NATURALIST. [VOL. XXXIX. try to give the present status of investigations in these most difficult subjects. CyanopJiycece (Blue-green Algce). The most recent and com- prehensive papers on the cell structure of the Cyanophycese are by Fischer ('97), Macallum ('99), Hegler (:oi), Biitschli (: 02), Kohl (: 03), Zacharias (:oo, : 03), and Olive (:O4). Olive gives an especially clear analysis of the situation in this field of investigation at the present time and an excellent historical review of earlier literature may be found in Regies (: 01). The discussions center chiefly around (i) the presence or absence of a nuclear structure and its behavior in cell division, (2) the dis- tribution of the blue-green pigment (phycocyan) and the struc- ture of a possible chromatophore, and (3) the nature of certain conspicuous inclusions within the cell, called cyanophycin gran- ules and slime globules. An outline in tabular form of the views of some thirty investigators on these subjects is given by Olive (:O4, p. 10). Writers from the earliest periods of cell studies on the Cyan- ophyceae have recognized the presence of a central body in the interior of the cell more or less sharply differentiated from the peripheral region, which holds the coloring matter and certain inclusions. The central body contains granular material which stains and behaves in other particulars like chromatin. But as a rule this granular material is not confined within a membrane or vacuolar cavity which has proved the most serious difficulty to its acceptance as chromatin and the central body as a nucleus. Then many investigators have not been able to satisfy them- selves that the central body exhibits the phenomena character- istic of nuclear division even in a simple form. Consequently much doubt has been expressed as to its morphology and pos- sible relation to a nucleus. The most recent and detailed investigations have, however, brought forward much evidence to the effect that the granular material in the central body is chromatin which becomes organ- ized into chromosomes that are distributed by a form of mitotic division. In the vegetative cells, which generally divide rapidly, the chromatin is never held within a nuclear membrane but in young heterocysts and spores such inclosing membranes have been found (Olive, : 04). No. 466.] STUDIES ON PLANT CELL VIII. 697 Olive (: 04) has given especial attention to methods of sec- tioning and staining on the slide and presents the most detailed account of the structure and behavior of the chromatin-and the simple apparatus which brings about the division of the central body. The central body is made up chiefly of dense kinoplasm with a fibrillar structure in which lie chromosomes that may be counted and whose number is found to be constant in several species. Thus there are eight chromosomes in a species of Gloeocapsa and Nostoc and sixteen in certain forms of Oscilla- toria, Phormidium, and Calothrix. The chromatin in some cases was observed to be organized into what seemed to be a simple type of spirem (especially clear in Gloeocapsa) within the central body, and the chromosomes are formed by a concentra- tion of material at certain points which are constant in the cells of the same plant. Olive found evidence that the chromosomes split during the process of division of the central body and are gathered in two groups at the ends of the achromatic structure which is gener- ally flattened at the poles and conforms in other particulars to the shape of the cells. The two sets of chromosomes are finally separated by the cell wall which develops from the pe- riphery during cell division and cuts the achromatic structure in the middle region. That portion of the central body which remains between the two sets of daughter chromosomes is regarded by Olive as equivalent to the central spindle so well defined in stages of anaphase and telophase in mitoses of higher plants. The central body during this process of division has certainly very much the appearance of a simple type of spindle although there are not present the large fibers so characteristic of nuclear figures in higher plants. Moreover it can scarcely be held that the division is one of simple fusion when chromo- somes are present in constant numbers and split into two groups with each division of the cell. Olive believes that the achro- matic structure, present during cell division, is a disc-shaped, generally flat-poled spindle, densely fibrous in structure and that the fission of the chromosomes and their separation into two sets constitutes a true mitotic division of the central body, which is a nucleus. 698 THE AMERICAN NATURALIST. [VOL. XXXIX. Other authors as Scott ('88), Hegler (:oi), Butschli (: 02), and Kohl (: 03), also believe that the central body is a nucleus which divides mitotically but none has described the process as so closely similar to nuclear division in higher plants as in the account of Olive. Some of their results are criticized by Olive as based on preparations in which the stain was not properly differentiated or the sections were too thick. Among the recent writers Wager (: 03) stands alone as holding that the nucleus divides directly (amitotically) by a process of simple fission. Both Kohl and Wager conceive the chromatin as in a network or convolute spirem which breaks up into segments which are drawn apart, thread by thread, quite a different process from the splitting of organized chromosomes. Other authors have held that the granules in the central body were chromatin although they were not willing to admit the structure as a nucleus. Thus Macallum ('99) found that the central body contained phosphor- ous and "masked iron" to a conspicuous degree and he, with other investigators, has shown that this structure resists the action of artificial gastric juice, solutions of pepsin, etc. These chemical reactions are considered confirmatory of the theory that the granular material is a proteid of a high order of or- ganization such as would be expected of chromatin. However, such chemical tests are very difficult to apply and do not seem to the writer so important in establishing the nature of the central body as does the careful study of its structure and activity during cell division. The objection that the central body lacks a mem- brane, universally present around resting nuclei of higher plants, is not regarded as vital by Olive. In the first place such a membrane may be found around the resting nuclei in young heterocysts and spores and its absence in vegetative cells is probably explained by the rapidity of the successive cell divi- sions. There are some recent writers, as Massart (: 02) and Zacharias (: oo, : 03) who are still unconvinced that the granules in the central body are chromatin and that the structure is the equivalent of a nucleus. Their papers and figures, however, clearly show that they have failed to find the detailed structures of other investigators. Fischer ('97) has been the most conspicuous opponent of the No. 466.] STUDIES ON PLANT CELL. VIII. 699 view that the cells of the Cyanophyceae and also of the Schizo- mycetes contain nuclei, taking a position in sharp opposition to that of BUtschli ('96). Fischer's conclusions were based on his failure to find that differentiation of the protoplasm within the cell demanded by the conception of the central body and the activities of this structure during cell division as described by other authors. He presented a sharp criticism of the conclu- sions based on the reaction of stains in determining the nature of protoplasmic structures, criticisms largely directed against the investigations of Biitschli. He showed by some ingenious experi- ments upon emulsions of albumen fixed on a slide that stain reactions were a purely physical phenomenon. Thus the same combinations of stains, such as saffranin and gentian violet, may be made to give exactly opposite results in differentiating a mix- ture of large and small globules of albumen when used in reverse order. He attached no importance to the so called affinity of a protoplasmic structure for a particular stain and would not accept such apparent affinity as evidence of its chemical nature. The fact that the central body takes chromatic stains did not seem to him important evidence of its nuclear character and he was very positive in his belief that the cells of the Cyanophyceae do not contain nuclei and that the central bodies should not be consid- ered the phylogenetic forerunners of such structures. This attitude of Fischer towards conclusions based on stain reactions was later presented in more elaborate form in his cri- tique ('99) on methods of fixing and staining protoplasm and has had an important influence on methods of cytological investi- gation and interpretation. The stain reaction is now regarded as probably merely a physical phenomenon but an effective means of differentiating protoplasmic structures. The deter- mination of their morphology rests with an understanding of their history and behavior in the activities of the cell. Although Fischer's general criticism of methods of cell research was timely and in some instances richly deserved, nevertheless his particular conclusions respecting the cell structure of the Cyano- phyceae and the Schizomycetes have not been sustained by investigators who have followed the history of the protoplasmic structures in the cells of these organisms. 700 THE AMERICAN NATURALIST. [VOL. XXXIX. We may pass now to the peripheral region of the cell which holds the blue-green coloring matter of the Cyanophyceae. A number of investigators, as Wager (103), Kohl (103), Hegler (:oi), and Hieronymus ('92), have held that this pigment was contained in minute granules distributed throughout the cyto- plasm under the cell wall. These granules have at times been termed chromatophores or plastids and Hegler has proposed for them the name cyanoplastids. Other authors, especially Fischer ('97), Nadson ('95), Palla ('93), and Zukal ('92) have been unable to find these color-bearing granules and have believed the color- ing matter to be uniformly diffused throughout the peripheral region of the cell. Fischer has made a particularly thorough study of the reactions of the pigmented region to various acids in comparison with the chromatophores of higher algae and con- cludes that no plastids are present but that the color is held in a hollow cylindrical or spherical outer layer of protoplasm which may be termed a chromatophore. Olive supports Fischer, approaching the subject from a very different point of view. If minute plastids are present they should be visible in fixed and stained material and Olive is unable to find any trace of Hegler's cyanoplasts. The granules .of the outer region of the protoplast seem to be colorless inclusions. Perhaps the most confused part of the discussion on the structure of the cell of the blue-green algae is that which deals with certain inclusions. There are apparently two sorts which are very common in the cells: (i) the cyanophycin granules (Borzi) and (2) the slime globules. The cyanophycin granules are very apt to lie along the cross walls in filamentous forms or in other peripheral regions of the cell. They are generally believed to be a form of food material and it has been suggested that they are the first visible product of photosynthetic processes, but their chemical nature is under dispute. The slime globules lie more frequently in the interior region of the cell close to the nucleus and frequently within this structure. They have been termed nucleoli by some authors and also confused with chro- matin. Besides these two bodies, other minute globules have been described as oil or fat and some remarkable crystalloid structures have been figured, especially by Hieronymus ('92)- No. 466.] STUDIES ON PLANT CELL. VIII. 701 Indeed the entire subject is so confused that it does not seem desirable for us to take it up in detail at this time, especially since these inclusions are apparently all secretions or excretions and not morphological features of the cell. The most compre- hensive discussions of the subject will be found in the papers of Hegler (:oi), Kohl (:O3), and Zacharias (: 03). l Schisomycetcs {Bacteria). The history of research upon the cell structure of the Schizomycetes has run in large part parallel with that on the Cyanophycese. The clearest results have come from studies upon the larger forms of the sulphur bacteria, especially certain species of Beggiatoa, and on certain forms of Spirillum. The more minute types and pathogenic forms' in par- ticular have proved very baffling because of their small size and it can scarcely be said that we fully understand their cell struc- ture. As in the Cyanophyceae, investigators of the bacteria fall into two groups : one holding that the Schizomycetes entirely lack a nucleus and the other that there is present a structure, often termed a central body, which is the equivalent of a nucleus. Biitschli ('96, : 02) has been the most conspicuous advocate of the latter view. He described and figured clearly a central body in the cells of Beggiatoa, Chromatium, and Spirillum with the same organization as given in his account of that body in the Cyanophyceae. The central body contains granular material which Biitschli regards as chromatin and the structure is shown in stages of division. Biitschli has no hesitation in giving the central body the value of a nucleus. It lies within a peripheral 1 Since the above was written a lengthy paper by Fischer, " Die Zelle der Cyanophyceen " has appeared (Bot. Zeit., vol. 63, p. 51, 1905), too late to be included in these reviews. Fischer has not changed his conclusions on the chief points as discussed in his earlier papers. The chromatophore is a closed cylin- drical structure ; the cyanophycin granules are proteid in character ; glycogen and another carbohydrate, anabasnin, aie conspicuous substances in the cell; the central body is not a nucleus but the seat of important metabolic processes con- cerned with these carbohydrates, and its contents and behavior in cell division have only a superficial resemblance to nuclear structure and mitosis ; the chro- matin granules of Hutschli, Olive, and others are masses of anabsenin (a car- bohydrate). Fischer's criticisms are fundamental and it is evident that the morphologists must clearly establish the proteid nature of the central body and its contents (especially the" so called" chromatin granules) before they can expect the acceptance of their conclusions as to its nuclear character. 702 THE AMERICAN NATURALIS7\ [VOL. XXXIX. region of protoplasm as in the Cyanophyceae. There is of course no blue-green pigment (phycocyan) in the cells of bacteria and consequently no chromatophore but several sorts of inclusions may be present in the protoplasm. The nature of some of the inclusions is not clear and this subject has not been given as much attention as in the Cyanophyceae. It is significant that this cell structure should be found so clearly in the Beggiatoa since this organism seems very close to Oscillatoria in its mor- phology. Some of the larger species of Beggiatoa may be expected to yield conclusions similar to those of Olive's investi- gation on Oscillatoria if sectioned and critically stained, especi- ally as 1 the cells are very large in some forms and there is probably less extraneous matter to complicate the interpretation of the preparations. As has been stated, investigations upon the smaller species of bacteria and especially upon pathogenic forms have met with great difficulties. These led at one time to the ingenious theory of Biitschli ('90), followed by Zettnow ('97) that possibly the entire protoplast had the value of a nucleus. That is to say, an outer peripheral region of cytoplasm had either never been developed in these organisms or, if present, had become so closely associated with the chromatin that it could not be dis- tinguished as a special region of the cell. A peripheral region of cytoplasm is represented in some of the larger forms by the cilia and by accumulations of protoplasm at the ends of the cells, especially clearly shown in Spirillum (Biitschli, '96 ; Zettnow, '97). Later Zettnow ('99) and Feinberg (: oo) applying the staining method of Romanowski, followed by several later inves- tigators with improved technique (Nakanishi, :oi,and others), succeeded in differentiating a minute body in the cells of smaller bacteria and pathogenic forms, which is regarded now as similar to the central body of the sulphur bacteria and a true nucleus. This structure is very minute since it occupies a portion of these exceedingly small cells. Naturally it will be very difficult to obtain any detailed knowledge of its structure and behavior during cell division. But enough seems to be known to justify the belief that differentiated nuclear structures are probably present even in the smallest types of bacteria. A recent paper No. 466.] STUDIES ON PLANT CELL. VIII. 703 of Vejdovsky (: 04) describes and figures a simple type of spindle in Bacterium gammeri and Bryodrilus eJilersi with a separation of two groups of chromatin granules during mitosis~ The chief critics of the conclusions that the cells of Schizo- mycetes are nucleated have been Migula ('95) and Fischer. The latter author in particular has devoted considerable attention to the group especially in his paper of 1897 which is largely a discussion of Biitschli's ('96) results on studies of the blue-green algae and bacteria. Fischer considers the central body described by Biitschli in the sulphur bacteria as merely a vacuolate region of the cell made conspicuous by the arrangement of the sulphur grains and that the structure does not appear in cells which are free from sulphur. The granular material, considered as chro- matin by others, is regarded by Fischer as reserve material. The central body described by Biitschli in the cells of Spirillum is stated to be a product of contraction. In general the same criticism which Fischer applied to the methods of staining and interpretation of structures in the Cyanophyceae is presented for the Schizomycetes. Fischer cannot justify Biitschli's ('90) view that the smaller bacteria are chiefly composed of nuclear sub- stance, a view which probably has few if any followers to-day and could scarcely claim to be more than a passing suggestion. In short, Fischer finds no evidence of a nuclear structure in the Schizomycetes but curiously ends by declaring that the group has no affinities with the Cyanophyceae but that its forms are closely associated with the Flagellata. SaccJiaromycetes (Yeasts]. The structure of the yeast cell has been perhaps the subject of as long a series of investigations as the cells of the Cyanophyceae and Schizomycetes, and the problems in both cases have many similar features. The chief problem in the yeasts has concerned the presence or absence of an organized nucleus or its equivalent in the form of some sim- pler structure. The accounts range from a complete denial of its presence to descriptions of a nuclear apparatus of considera- ble complexity which passes through some rather involved activi- ties during cell division. It is impossible for us to treat the subject historically. We shall only consider the accounts of the most recent investigators and try to determine the probable 704 THE AMERICAN NATURALIST. [VoL. XXXIX. bearing of these studies. An admirable review of the early lit- erature is presented in Wager's paper of 1898. Wager ('98) himself has made one of the most detailed studies of the yeast cell and his conclusions on the presence of a " nuclear apparatus " will be made the starting point of our dis- cussion. The yeast cell contains a structure, termed by Wager a " nuclear body," generally situated at one side, close to the cell wall. This body resembles the nucleolus of higher plants in its homogeneous structure and reaction to stains. Besides the " nuclear body " Wager finds a vacuole always present which contains granular material and is an important part of the nuclear apparatus. This " nuclear vacuole-" must be carefully distin- guished from other vacuoles of the usual type which merely con- tain glycogen. There are besides some globular bodies in the protoplasm whose nature may be oil in some cases and proteid in others. The "nuclear body" is always in close contact with the " nuclear vacuole " but is never within it. The amount of granu- lar material in the nuclear vacuole is variable but it sometimes contains a dense mass. This content is believed to be chroma- tin from the behavior to stains and insolubility in digestive fluids. Sometimes the nuclear vacuole disappears but in such cases the granular network is found in contact with the nuclear body and sometimes distributed about it in a very regular man- ner. The chromatic granular material appears then to be a per- manent substance in the cell and always closely associated with the nuclear body, sometimes distributed about it and sometimes included within a special vacuole. Wager concludes that the nuclear apparatus consists of (i) a nucleolus (nuclear body) and (2) a store of chromatin in a net- work, either enclosed in a vacuole in close contact with the nucleolus or lying freely about the nucleolus or sometimes disseminated in granules generally throughout the cytoplasm. Wager believes that the nuclear vacuole arises from the fusion of numerous small vacuoles which lie around the chromatin gran- ules which thus come to lie within a common vesicle. This mode of origin seems reasonable from what we know of the history of the nuclear vacuole which arises around the chromo- somes that gather at anaphase of mitosis to form daughter No. 466.] STUDIES ON PLANT CELL. VIII. 705 nuclei in higher plants. The earlier investigators for the most part failed to recognize the chromatic granules and network and considered the nucleolar body (nucleolus) to be ^the- nucleus of the cell. Janssens and Leblanc ('98), however, described a nucleus with a membrane containing caryoplasm and a nucleolus, and other authors noted the vacuole and believed that it held some relation to the nucleus. Both the nuclear vacuole and the nuclear body (nucleolus) take part in the process of bud formation. The bud appears on the opposite side of the cell from the nuclear body and the nu- clear vacuole lies between. The bud contains at first cytoplasm alone ; then the nuclear vacuole begins to pass into it and the nuclear body takes a position in the vicinity, between the mother-cell and the bud. The nuclear body now divides by simple fission and one half enters the bud. The nuclear vacu- ole gradually constricts and is drawn apart in the canal between the two cells. The two daughter nuclear vacuoles and nuclear bodies then pass to opposite ends of the mother- and daughter- cells respectively. If the nuclear vacuole is absent the chroma- tin network is drawn apart so that a division is effected in a similar manner. At the time of spore formation, the chromatin is reported by Wager to become so closely associated with the nuclear body that the two substances cannot be easily separated and behave as one. The resultant structure elongates and divides by con- striction and the subsequent divisions are of the same character. Strands of deeply staining protoplasm between the daughter nuclei are of interest as suggesting the possibility of a simple type of spindle. Wager describes the formation of spore walls around the nuclei enclosing a portion of the protoplasm and thus cutting the spores out from the remaining non-nucleate cell contents. The details of this process are not known and might prove very interesting since the process, from W'ager's account, would seem to be one of free cell formation without, however, the characteristics described by Harper in spore formation within the ascus. It should be more thoroughly studied for it is possible that the division will be found to involve cleavage furrows and really prove to be a type of segmentation by con- striction (Section II, Amer. Nat., vol. 38, p. 453, June, 1904). 706 THE AMERICAN NATURALIST. [VoL. XXXIX. Several papers have appeared on the structure of the yeast cell since Wager's account of 1898. Marpmann (:O2) and Feinberg (: 02) described much simpler conditions than are reported by Wager, and recognize scarcely more than a deeply staining body which they term a nucleus. Hirschbruch (: 02) gives an extraordinary description, accompanied by diagram- matic figures, of a nuclear structure and a body, staining red and blue respectively, which are supposed to fuse previous to the development of a bud, but the account is so unsatisfactory as to merit little attention. Janssens (:O3) reviews the work of these investigators and others in comparison with his earlier results (Janssens and Leblanc, '98). Guilliermond (:O4) has published the most recent paper presenting more completely his conclusions of an earlier investigation in 1902. Guilliermond's conclusions have some points of resemblance to those of Wager. He finds a nuclear vacuole containing a granular network believed to be chromatin and a nucleolar structure. The entire body seems to be a true nucleus, not dif- fering in its essentials from the nuclei of other fungi. Some- times all the material in the nucleus seems to be condensed into a central body, a sort of chromatin nucleolus (chromoblast) somewhat resembling a similar structure in Spirogyra. Guillier- mond figures the nucleus as constricting during the process of budding, one part passing into the daughter cell. His figures show clearly deeply stained material outside of the nuclear membrane in a position similar to that of Wager's nucleolar body (nucleolus). These points of agreement seem to justify at least in part Wager's account, but of course the peculiarities of both lead one to suspect that there are important features in the structure of the nucleus and in the events of nuclear division which have not been determined. It certainly seems probable that chroma- tin is present in definitely organized bodies (chromosomes) some- times within a vacuole and sometimes lying around a nucleolar structure. The latter also holds an intimate relation to the chromatin, which is frequently true in higher plants. There are indications that a simple type of spindle is present at least in the nuclear divisions during spore formation. In view of No. 466.] STUDIES ON PLANT CELL. VIII. 707 Olive's results in studies on the Cyanophycese it does not seem unreasonable to hope that more accurate staining of very thin sections will bring the peculiarities of these accounts into harmony with mitotic phenomena of higher forms. The accounts of conjugation in yeasts (Barker, :oi and .Guilliermond, 103) which were discussed under "Asexual Cell Unions and Nuclear Fusions " in Section IV give no additional information on the essential structure of the yeast cell. 2. COMPARISONS OF THE STRUCTURE OF SOME HIGHER TYPES OF PLANT CELL WITH SIMPLER CONDITIONS. Some of the most fruitful and interesting fields of investiga- tion in cell structure are likely to be in those border groups between the very simplest conditions of the lower algae and fungi and the higher regions where the nucleus and processes of mitosis have clearly the essential features which are generally ascribed to this structure and its activities. At present the gap seems very great between the simple conditions of the Schizo- phyta and the groups of algae and fungi on the next higher general level. But as a matter of fact we know almost noth- ing of the nuclear structure in the lowest groups of the Chloro- phyceae, i. e., among the simplest of the unicellular green algae. It is rather remarkable that this region should have been so neglected. The Nucleus. Comparative studies on the nucleus naturally treat chiefly of the chromosomes and nucleolus. One of the most interesting features of more recent research on the nucleus has been the steady accumulation of evidence indicating that the nucleolus holds a very important relation to the chromatin con- tent. There are types among the lower algae in which the whole or a greater part of the chromatin is gathered into a dense nu- cleolar body in the resting nucleus. Spirogyra is the best- known illustration of this condition and has been studied by several investigators. Similar phenomena have been reported by myself in Corallina (Davis, '98), by Golenkin ('99) for Sphaeroplea, and by Wolfe (: 04) for Nemalion. Some nuclei, however, particularly in the higher plants have nucleoli whose 708 THE AMERICAN NATURALIST. [VOL. XXXIX. substance does not seem to contribute directly to the chromo- somes and these have been regarded as secretions within the nucleus. Strasburger believed that such were masses of reserve material drawn upon by the kinoplasm during the process of spindle formation. The term plastin has been applied to such substance in the nucleolus and also in the linin as cannot be directly connected with chromatin. A nucleolus may consist of plastin alone, or have with this substance varying quantities of chromatin. Nucleoli consisting of chromatin alone may be ex- pected among the lower plants from the studies on Spirogyra, Corallina, Sphaeroplea, and Nemalion. Plastin and chromatin are probably closely related substances. A recent paper of Wager (: 04) indicates that the nucleolus of some higher plants holds a far closer relation to the chromo- somes than has been supposed and rather weakens Strasburger' s theory of the structure as a reserve mass drawn upon during mitotic activities. This study and recent papers by Miss Mer- riman (: 04) and Mano (: 04) have all been upon the cells of root tips while the conceptions of Strasburger and others have been founded largely on the structure and behavior of the nucleolus in the spore mother-cell during the mitoses of sporogenesis. Wager treats of the root tip of Phaseolus, Miss Merriman of Allium, and Mano of Solanum and Phaseolus. They are impor- tant contributions to the subject of the nucleolus and should be considered in any treatment of this structure. The papers appeared too recently to be noted in our brief account of the nucleolus in Section I which is consequently incomplete. Wager's paper especially presents an excellent review of the literature on the nucleolus in the plant cell. Wager concludes that the nucleolus is really a portion of the nuclear network and that the spirem is derived in part at least from this structure. Material from the nucleolus then passes into the chromosomes. Also, in the reconstruction of the daughter nuclei the chromosomes are massed together at a cer- tain stage and from this mass the nucleolus emerges, taking out with it the greater part of the chromatin. Wager then con- siders the nucleolus as a store of chromatin which must be taken into account in theories of heredity based on the morpho- No. 466.] STUDIES ON PLANT CELL. VIII. 709 logical independence of the chromosomes. Miss Merriman reports the origin of the nucleoli as masses among the meshes of chromatin from which they draw their substance. TMafio, in contrast to Wager, holds that the nucleoli appear as globules independent ot the chromatin network and later flow together into a single body. The chromosomes are also believed by Mano to be morphologically independent of the nucleolus and if the latter furnishes material to the former it is not by the emergence of strands as described by Wager. Mano then holds the nucleolus to be an accessory structure without morphologi- cal relation to the chromosomes. The theory of the individuality of the chromosomes is of course vitally concerned with the problem of the morphology of the nucleolus but this topic we have reserved for later treatment under the caption : " The Essential Structures of the Plant Cell and their Behavior in Ontogeny." The chromatin and nucleoli within the nucleus of a higher plant lie in a vacuole whose fluid content is bounded by a plasma membrane similar to that around any vacuole in the cell. Lawson (: 03) and Gregoire and Wygaerts (:O3) have emphasized this structural condition in recent papers but the central idea seems to be an old one run- ning through the writings of Strasburger from an early period. We bring up these striking conceptions of nuclear structure in the higher plants because it seems very probable that a much clearer understanding of the problems will come through inves- tigations upon the simpler conditions in the lower plants. There, we may hope to find evidence of the primitive forms of nucleolar and chromatic associations with perhaps some clues as to the manner of the development of the higher types of struc- ture. Thus the yeast cell, as reported by Wager ('98) with its chromatin sometimes collected within a vacuole and sometimes distributed in the cytoplasm and a nuclear body (nucleolus) in close association with the nuclear vacuole, but not within, is of the greatest interest as presenting intermediate stages in the complexity of nuclear structure and illustrates what may be hoped from further research among the lower forms. The Chromatophore and Plastid. In considering the great variety of chromatophores and plastids exhibited among the 710 THE AMERICAN NATURALIST. [VOL. XXXIX. thallophytes one notices at once certain features of their distri- bution in various groups. The large chromatophores are charac- teristic of the cells of simpler and more primitive groups and the small plastids, numerous in the cells, are generally present in types which are at a fairly high evolutionary level. There are exceptions of course to this general statement but some of these are probably significant of phylogenetic relations. The evidence all indicates that the "primitive type of chroma- tophore was a large structure in the peripheral region of the protoplast and with an ill denned boundary or occupying the entire surface of the cell. This type of structure is at present characteristic of chromatophores of the Cyanophycese and is also present in numbers of the lower groups of green algae. Thus we may find many types in the Pleurococcaceae whose cells con- tain a pigment so diffused that it is impossible to establish definite limits and similar conditions often appear in the cells of some of the higher algae as in Hydrodictyon and certain simple forms of the Ulothncacese. The simple diffused types of chromatophores of the lower algae become replaced in higher groups either by sharply differ- entiated structures of definite form and often showing internal organization in the form of pyrenoids or by numerous plastids. There is considerable evidence that the plastids have arisen by the successive splitting or division of large organized chromato- phores. The most highly differentiated chromatophores are found in the Conjugales and the remarkable size and symmetry of these cells is emphasized by the same peculiarities of the chromatophores. They are generally so placed in the cells as to give an almost perfect balance of protoplasmic structure. This principle is especially clearly illustrated among the desmids and in such forms as Zygnema and Mougeotia while even Spirogyra illustrates the principle strikingly in the distribution of its spi- rally wound chromatophores. Plastids are characteristic of the Siphonales, Charales, most of the Rhodophyceae, the higher Phaeophyceae, and all groups generally above the thallophytes. It seems to be the type of structure best suited to cell activities since with few exceptions it is found in groups in the highest lines of plant evolution in No. 466.] STUDIES ON PLANT CELL. VIII. 7 1 I various directions. The only striking exceptions to this broad principle are Anthoceros, whose cells contain each a single large chromatophore, and Selaginella. Selaginella is especially inter- esting for, while the cells of the meristematic region and young organs contain but a single chromatophore, this structure may divide later in some types to form a chain of discoid plastids in older cells connected with one another by delicate strands of protoplasm. Thus in the life history of certain species of Sela- ginella we have plainly shown the change from a single chroma- tophore to a number of plastids. It seems probable that this history repeats in general outline the evolutionary history of the condition characterized by numerous plastids within a cell from a primitive type of cell structure with but a single chromato- phore. Anthoceros and Selaginella may be regarded as forms whose cells still retain the primitive conditions with respect to the single large chromatophore. There are somewhat similar illustrations in the Rhodophyceae as in Nemalion and Batracho- spermum whose cells hold a single large chromatophore while most of the more highly organized red algae have numerous plastids. A beautiful series of stages illustrating the evolu- tionary principles outlined above might be worked out in the Phaeophyceas. What is the fundamental principle underlying the substitution of numerous plastids in a cell in place of a single chromato- phore ? The author believes that it must have relation to the preservation within large cells of a certain balance of the meta- bolic centers. The fission of a plastid is a process of constric- tion and studies on Anthoceros (Davis, '99, p. 94) indicate that the bounding cytoplasmic membrane exerts pressure upon the elongating structure. It seems probable that the division is due to the mechanical separation of material that is too bulky for the most effective results of photosynthesis which in the case of a single chromatophore are centered in a particular region of the cell. By the division of a chromatophore into numerous plastids the photosynthetic activities are distributed among several centers and a much better balance results within the cell. It is very interesting that the large elaborate chroma- tophores with their peculiar internal differentiations, the pyre- 712 THE AMERICAN NATURALIST. [Vou XXXIX. noids and caryoids, should have been displaced by the much simpler and apparently homogeneous plastids. A comparative study of chromatophores and plastids from the point of view of their evolutionary history is much to be desired and such research would necessitate extensive studies among the lower groups of algae and especially in the Proto- coccales. Such studies would involve far more than the general morphology of the chromatophore and plastid. The structure and activities of the pyrenoid are a very important subject as shown by the investigations of Timberlake on Hydrodictyon and nothing is known of the function of the caryoid. A de- tailed investigation of the chromatophore or plastid throughout ontogeny is yet to be made. The Cytoplasm. There is no region of the plant cell whose structure is more varied and as little understood as that pre- sented by the cytoplasm with its diverse conditions. We have throughout these papers held to the classification of Strasburger that the cytoplasm may be separated into two forms : kinoplasm and trophoplasm, which show certain structural peculiarities and are characterized by very different forms of activity. While it must be acknowledged that kinoplasm and trophoplasm are very similar in certain regions of the cell and at certain periods of the cell history, still the distinctions are in general clearly marked. Kinoplasm is homogeneous in structure, either minutely granular or consisting of delicate fibrillae composed of very small granules placed end to end. The homogeneous condition is characteristically shown in the three forms of plasma mem- branes which cytoplasm places between itself and external or internal surface contacts. The three membranes are: the outer plasma membrane, the nuclear membrane, and the vacuolar membranes. They are certainly closely related and probably identical in structure and appear to be the natural expression of protoplasm to contact with a fluid (water) medium. The fibril- lar condition appears during mitosis and serves important func- tions in the mechanism (spindle) through which the chromosomes are distributed and in most of the higher plants determines the position of the cell wall that is generally formed with each nuclear division. No. 466.] STUDIES ON PLANT CELL. VIII. 713 But the manifestations of kinoplasm during nuclear division and also in relation to cilia-bearing surfaces are exceedingly various and it is among these structures that our ignorance of relationships and modes of origin is deepest. These kinoplasmic structures have been described in various connections through- out this series of papers and especially in Sections I, II, and III, and need not be treated here. But the point which should be emphasized in this connection is the necessity of the close study of their simplest expressions in the lower regions of the thallophytes. The most varied forms of kinoplasm are in the thallophytes where asters, centrospheres, and centrosomes ob- tain and where ciliated cells, presumably with blepharoplasts, may occupy long periods of the life history. It is here that we must search for information that will bring order out of the con- fusion of our present accounts and insufficiency of knowledge. The most vital problems relating to kinoplasm concern the ori- gin and the events of the simplest types of mitotic phenomena and the formation of cilia. We have a fairly clear understand- ing of the general features of mitosis in the groups above the thallophytes and their relation to the lower types and these will be briefly treated in the following portion of this section under the head : " Some Apparent Tendencies in the Evolution of Mitotic Phenomena." But the events of mitosis among the thallophytes are exceedingly various and difficult to understand and nothing is known of their origin or relation to the simpler conditions which must be present in the lowest regions of the Chlorophyceae and in the Cyanophyceae. Trophoplasm comprises all of the cytoplasm included within the plasma membranes. While this region does not give rise to such highly differentiated cell organs as the kinoplasm, never- theless some remarkably interesting structures are developed. Coenocentra and Physodes are specialized structures of exceed- ing interest and our ignorance of the latter is truly remarkable. Nematocysts if trophoplasmic offer another attractive subject for investigation. In a sense, chromatophores and plastids may be considered trophoplasmic but their high grade of specialization and fixity as cell organs gives them a certain independence of other structures in the cell. Respecting the structure of the 714 THE AMERICAN NATURALIST. [VOL. XXXIX. groundwork of trophoplasm, whether fibrillar, granular, or pre- senting the structure of foam, botanical science has as yet fur- nished very little systematic study and this field of research is one of exceptional opportunity for the student of the plant cell. The Cell Wall. The cell wall may be treated from two points of view : either with respect to the strict chemistry of its organization and development or more largely for the biological and morphological features involved. The chemistry of the cell wall is an exceedingly complex subject which has developed a special literature of its own. In the substance termed cellulose we are not dealing with a single body but rather with a large group of closely related bodies. And besides the members of the cellulose group there may be present foreign substances so intimately associated with the carbohydrates as to resist very severe treatment. We cannot even touch this phase of the sub- ject ; a brief review of its complexities and problems is presented by Beer (: 04) and there are further references in Section I of these " Studies." There are, however, some biological features of the process of wall formation, the morphological and physiological aspects of the phenomena as they are related to protoplasm, which offer some exceedingly interesting problems especially among the thallophytes. It has long been a matter of dispute whether the cell wall is a secretion from the surface of a plasma membrane or is formed wholly or in part by the transformation of such a membrane. It seems to be established now that substances of the cellu- lose groups are only formed in contact with plasma membranes, that is, they are not formed actually in the interior of proto- plasm although they may appear to lie in such situations. Thus the material of the capillitium of the Myxomycetes which is of the same character as the chief substance in the exterior cover- ing of the fructification, is laid down within vacuoles in the protoplasm, and is therefore in contact with the surface of vacu- olar plasma membranes precisely as the outer covering lies in contact with the surface of the outer plasma membrane. The morphological relation of capillitium and outer covering to the surface of plasma membranes is therefore precisely the same. No. 466.] STUDIES ON PLANT CELL. VIII. 715 And similarly the cross wall which takes the position of the cell plate at the end of mitosis is not developed from the transfor- mation of a film of protoplasm but is laid down between two surfaces that separate to form a thin vacuole which later spreads to the edge of the cell and the wall is deposited between these two membranes which are almost in contact. There are a number of cases in which large strands or masses of protoplasm have been described as changing directly into cellulose but it is prob- able that these examples upon further study will exhibit the same relation of the cellulose substances to plasma membranes as in the typical cases of wall formation. There are many inter- esting examples of cellulose formation whose precise relation to the protoplasm has not yet been determined. Respecting the exact method by which a cellulose wall is laid down by a plasma membrane there is very little real informa- tion. It is clear now that the cellulose is not a secretion from the plasma membrane comparable to a mineral shell. There is much evidence that protoplasm is actually sacrificed in the de- velopment of cellulose. There are numerous illustrations, as in the tracheids and other cells empty of protoplasm, where the final secondary thickenings are deposited as the protoplast grows smaller and eventually disappears, a large part of its substance evidently contributing to the deposits which are members of the cellulose group. But of course it cannot be supposed that the molecules of the proteids are changed directly into those of the carbohydrates. Nevertheless it does seem clear that the carbo- hydrates appear simultaneously with the disappearance of the proteids and occupy the position formerly held by the latter. It is probable that with the splitting up of the proteid molecule, carbohydrate material is formed which displaces the proteid sub- stances. So in a broad sense the cellulose deposit actually does represent a transformation of a plasma membrane. The evidence in general favors the view that the wall, lamellae, and other deposits of cellulose only increase in amount when in actual contact with a plasma membrane. Some apparent excep- tions to this principle are easily understood. Thus cell walls or portions of such may swell greatly and become much softer in consistency and perhaps even mucilaginous. There are no 716 THE AMERICAN NATURALIST. [VOL. XXXIX. reasons for regarding such transformations as an actual increase in the carbohydrate material for it is clear that the substance is a body with a greater amount of water in its organization than is present in the more usual forms of cellulose compounds. But there are some cases which are not so easily understood and perhaps the most widely known are the jnegaspore walls of cer- tain species of Selaginella. These spores are remarkable for a differentiation of the spore wall in which the outer layer seems to be entirely separated from the inner by a space and yet is able to increase enormously in size and take on marked pecul- iarities of structure, but apparently without any relation to the protoplast. It may, however, be justly questioned whether the apparent space between the inner and outer wall is really a cavity and may not be filled with plastic material which holds the two walls in intimate organic relation to one another and to the protoplast. Miss Lyon has recently given this subject at- tention and announced her belief that the latter condition ob- tains. Her conclusions will be awaited with interest. As regards the way in which a cell wall increases in size we are still limited to the two conceptions termed ( i ) growth by apposition and (2) growth by intussusception. The first method consists in the laying down of successive layers by the plasma membrane and results in a thickening of the cell wall. It is of course a comparatively simple process. Growth by intussuscep- tion is a stretching or expansion of the substance which seems to be greatly increased in quantity although the morphology of the structure remains the same. The current explanation out- lined by Nageli assumes that new molecules of carbohydrates are intercalated among the old. It seems more probable that the increase in bulk is due to some modification or rearrange- ment of existing molecules, involved, perhaps with an increase of material but not through the actual intercalation of new molecules of the same or original carbohydrates. The chem- istry of the carbohydrates is so complex that great changes of form, bulk, or optical properties may be readily assumed which would quite change the appearance of a structure without, how- ever, necessitating the transportation of new carbohydrate sub- stance to it directly. No. 466.] STUDIES ON PLANT CELL. VIII. 717 There are many forms, particularly among the lower plants, where studies on the processes of wall formation are sure to throw much light on the fundamental problems which we have discussed. And a particularly interesting study might be made of the evolutionary history of the cell wall among the thallo- phytes and in the modifications introduced when plants pass from aquatic habits to aerial or terrestrial conditions. Our attention has been chiefly centered on the structure of the pro- toplast and the morphology and behavior of its parts. We are likely soon to give more study to the carbohydrate membranes and walls and this subject is likely to be very fruitful for inves- tigation. 3. SOME APPARENT TENDENCIES IN THE EVOLUTION OF MITOTIC PHENOMENA. Our brief descriptions in Section II (Amer. Nat., vol. 38, p. 431, June, 1904) of the various kinoplasmic structures developed during mitosis in different groups of plants brings up the prob- lem in their relationships to one another, i. e., the evolutionary tendencies in the differentiation of mitotic phenomena. We have seen that the thallophytes present an especially diverse assortment of kinoplasmic structures associated with the spindle and its method of development. The spindle fibers, whether formed within the nuclear membrane (intranuclear) or arising from without (extranuclear), are associated with centrosomes or centrospheres to form asters in a number of well known types as Stypocaulon, Dictyota, Fucus, Corallina, certain diatoms, the ascus, and the basidium. Centrospheres are found in certain phases of the life history of liverworts as in the germinating spore of Pellia. A second type of kinoplasmic structure resem- bling in certain features the aster but with some fundamental differences has been termed the polar cap. The polar cap is an ill denned region of kinoplasm, generally larger than a centre- sphere and without clear boundaries, which forms a region for the insertion of spindle fibers. Polar caps are well illustrated in the mitoses of vegetative tissues and meristematic regions, especially among the higher plants (pteridophytes and sperma- 718 THE AMERICAN NATURALIST. [VOL. XXXIX. tophytes). They sometimes approach the centrosphere very closely in their morphology. The third and highest type of spindle formation in plants is that illustrated in the mitoses within the spore mother-cell which were given special treatment in Section III (Amer. Nat., vol. 38, p. 725, October, 1904). In this remarkable cell the spindles develop from a mesh of independent fibrillae which at prophase more or less completely surround the nucleus. The poles of the spindle arise by the grouping of cones of fibrillae so that a single axis is finally established but without any kinoplasmic cen- ters at the poles. This type of spindle formation which may be termed the free fibrillar type is one of the most interesting cytological peculiarities of plants. It has been found in all types whose sporophytic phase terminates its history with a spore mother-cell, although the accounts in the Hepaticae are not in full accord. Is it possible to connect the various types of spindle formation with one another and to establish any evolutionary tendencies in the processes involved ; and have the different manifestations of kinoplasm such as centrosomes, centrospheres, polar caps, free fibrillar condition, and the mysterious structure called the ble- pharoplast any genetic relation to one another ? The confusion is so great among the thallophytes that the author sees little hope at present of establishing clearly any relationships between the types of centrospheres and centrosomes with their systems of radiations (asters) and we must patiently wait for more infor- mation. And respecting the origin of these structures from the simpler types of mitosis we are absolutely in the dark. But the relation which polar caps and the free fibrillar type of spin- dle formation bear to centrospheres is less perplexing and it seems possible to define certain common features among these structures which hold them together with a degree of unity in their relations to mitosis. That phase of the subject will be considered in this treatment. The Hepaticae as a group occupy an interesting position with respect to the character of mitotic phenomena at various periods of ontogeny, between conditions in the pteridophytes, which are obviously similar to the sperma- tophytes, and conditions in the thallophytes. This was brought No. 466.] STUDIES ON PLANT CELL. VIII. 719 out by the work of Farmer whose accounts of centrosomes and centrospheres in the germinating spores of Pellia and within the spore mother-cell of various liverworts, together- with his account of a " quadripolar spindle ' ' made it evident that the group offered some very interesting cytological problems. They led the author to the study Anthoceros (Davis, '99) and Pellia (Davis, :oi), investigations which have been followed by Van Hook (: oo) on Marchantia and Anthoceros, Moore (: 03) on Pallavicinia, Chamberlain (: 03) and Gregoire and Berghs (:O4) on Pellia, while Ikeno (: 03) has studied the processes of sperma- togenesis in Marchantia. My studies on sporogenesis in Anthoceros and Pellia led me to conclude that the processes of spindle formation did not differ in any essentials from those in the pteridophytes and sper- matophytes. There are present two successive mitoses and the spindles are formed from a surrounding mesh of fibrillae devel- oped from the kinoplasm associated with the nuclear membrane and without achromatic centers (centrospheres or centrosomes). They exhibit clearly the free fibrillar type of spindle formation although in somewhat simpler form than in the pteridophytes and spermatophytes. The poles of the spindles generally end bluntly in areas of granular kinoplasm but these seem to me too indefinite in form to deserve the designation of centrospheres and such granular inclusions as may be present are too variable in number and position to be termed centrosomes. There is clearly present in Pellia during the prophase of the first mitosis a four-rayed achromatic structure which is later replaced by a typical bipolar spindle. This four-rayed kinoplasmic structure is evidently the same as Farmer's " quadripolar spindle " which he described as associated with a simultaneous distribution of the chromatin in Pallavicinia to form at once four daughter nuclei. I was led to doubt this account and to suggest that the "quad- ripolar spindle " might prove to be simply a phenomenon of prophase associated with the peculiar four-lobed structure of the spore mother-cell in the Jungermanniales. I stated my belief that the distribution of the chromosomes during sporo- genesis in all liverworts would be found to take place through two successive mitoses after the usual manner. Moore (: 03) 720 THE AMERICAN NATURALIST. [VOL. XXXIX. has recently studied an American species of Pallavicinia and has failed to confirm Farmer's conclusions. He found the four- rayed figure, which Farmer terms a " quadripolar spindle," a conspicuous feature of the first mitosis here as in Pellia but there was no indication of a simultaneous distribution of quad- rupled chromosomes to form four daughter nuclei as reported by Farmer. The four-rayed figure was merely preliminary to the first mitosis whose spindle at metaphase was bipolar and the first mitosis was followed shortly by a second, so that Pallavi- cinia offers no exception to the essential features of sporogenesis as known in all groups above the thallophytes. Farmer (Bot. Gaz., vol. 37, p. 63, 1904) has taken exception to the restriction of the term spindle by Moore and myself to the structure found at metaphase and holds that the four-rayed structure is a part of the spindle apparatus. In this discussion he appears to avoid the issue, which is not the broader or nar- rower application of the term spindle, a mere matter of usage, but concerns the fundamental character of the mitoses during sporogenesis whether they are two in number and successive in all forms or whether Pallavicinia presents an extraordinary ex- ception in a distribution of the chromatin to form four daughter nuclei simultaneously in the spore mother-cell. Farmer (: 05) has recently reaffirmed his view that the poles of the four-rayed figure in Aneura and presumably in other Jungermanniales are occupied by centrospheres and that sometimes a central body (centrosomes) may be distinguished in each. This statement involves again a matter of usage in which I should differ from Farmer for my studies and those of Moore do not seem to me to justify the application of these terms to regions of kinoplasm whose form is so ill defined and history so transient within the cell. These disputed points which were also discussed in Section III (Amer. Nat.., vol. 38, pp. 727-732, October, 1904) are of importance in relation to the mitotic phenomena in other periods of the life history of liverworts which will now be considered. It may be stated, however, that other investigators who have studied the processes of sporogenesis in the liverworts (Van Hook, : oo ; Chamberlain, : 03 ; Gregoire and Berghs, : 04) sup- No. 466.] STUDIES ON PLANT CELL. VIII. 721 port my general program of sporogenesis with the free fibrillar type of spindle formation. There seems to be little question but that centrospheres are present and conspicuous in the early mitoses within the spore of Pellia. They have been especially studied by Farmer and Reeves ('94), Davis (:oi), Chamberlain (:O3), and Gregoire and Berghs (:O4). All of these authors have agreed that asters are clearly defined in the early mitoses within the spore and most of them have termed the region of kinoplasm in the center of the aster a centrosphere. The struc- tures are less prominent in the third mitosis and are perhaps replaced in later periods of the gametophyte history by kino- plasmic polar caps. Polar caps are characteristic of the mitoses in the seta of Pellia (Davis, :oi). However, Van Hook has described centrospheres with radiations at the poles of the spindles of the archegoniophores of Marchantia, whose centers sometimes contained centrosomes, and it is possible that the centrosphere runs through a considerable period in the life his- tory of liverworts. There is complete agreement that the cen- trospheres when present arise de novo and independently of one another during the prophase of mitosis and that they disappear at telophase. Ikeno has, however, described centrosomes during the mitoses within the antheridium which are said to divide and pass to opposite sides of the nucleus where they become the poles of the spindles. They cannot be found after the mitosis is completed, but are described as formed de novo in the interior of the nucleus and thrust through the nuclear membrane into the cytoplasm previous to each mitosis. After the final divi- sion in the antheridium, the centrosome remains to function as a blepharoplast. Thus we see that the liverworts present during their life history an almost complete range of kinoplasmic structures associated with the nuclear divisions from centrosomes and cen- trospheres to polar caps and that type of spindle formation characterized by free fibrillae gathered into cones but entirely independent of definitely organized centers. There is also pres- ent the blepharoplast. I emphasized this range of kinoplasmic structure in my paper on Pellia and it seemed to me one of the most interesting features of the liverworts. In this paper 722 THE AMERICAN NATURALIST. [VOL. XXXIX. (Davis, :oi, p. 171) are outlined the changes in form which kinoplasm may assume in the mitoses of the liverworts upon which is based a theory of a cycle through which kinoplasm may run in the history of a cell. On this theory, centrosphere, polar cap, and the free fibrillar condition are all secondary devel- opments from a primal finely granular kinoplasm which is the only form of kinoplasm that is in any sense permanent in the cell. This finely granular kinoplasm is always present in char- acteristic form in the plasma membranes of the cell. The sub- stance of centrospheres, polar caps, and fibrillae arises from accumulations of granular kinoplasm during prophase and these structures return to the same undifferentiated granular kino- plasm at the end of mitosis or become lost in the general cytoplasm of the cell. The cycle is from an undifferentiated finely granular kino- plasm through certain specialized conditions either wholly or in part fibrillar in structure back to the granular state. The centrosphere and polar cap are regions from which fibrillae de- velop at least in part and to which they may remain attached as to an anchorage. The polar cap is a less clearly differentiated kinoplasmic center than the centrosphere but does not differ from it in the essentials of its organization. It seems to me that the two structures are very closely related in the liverworts and that in this group we may readily conceive the polar cap as de- rived from the centrosphere. The 'free fibrillar type of spindle formation is a step farther in the direction of such a distribution of the kinoplasm that no very positive centers for the develop- ment of the spindles may be distinguished. The four-rayed structure (quadripolar spindle) so characteristic of the spore mother-cell in the Jungermanniales represents a group of four temporary centers for the formation of fibrillae and there is clearly a gathering of kinoplasm at these points but the regions are so vague in outline as hardly to justify the designation of centrospheres. From the fibrillar state, kinoplasm returns to the finely granular condition by the contraction of the fibers which thus contribute their substance to some common area. The area may lie around the chromosomes of the daughter nuclei where it becomes later in part at least a nuclear mem- No. 466.] STUDIES ON PLANT CELL. VIII. 723 brane. Or the area may be a cell plate whose halves on division finally merge with outer plasma membranes of the cells. The spindle fibers which cut out the spore areas in the-ascus form the basis of a plasma membrane. Thus the fate of all kinoplas- mic fibrillae seems to be a final return to the undifferentiated, finely granular condition so characteristic of plasma membranes which according to this theory is the condition from which they arose. Thus I believe the liverworts present rather striking evidence of a relationship between the centrosphere, polar cap, and the free fibrillar condition of spindle formation and establish an evolutionary tendency from the first two types of kinoplasmic differentiation towards the latter. The free fibrillar type of spindle formation is found in a very simple form in this group, sometimes with temporary centers, as in the four-rayed figure (quadripolar spindle) of prophase, whose poles have accumula- tions of kinoplasm in the position of centrospheres. The polar caps are likely to prove a much simplified type of centrosphere whose kinoplasm is no longer gathered to form conspicuous spherical centers. With respect to the problem of the homol- ogies and nature of the blepharoplast, the liverworts furnish as yet no material assistance and this structure stands at present as one of the most interesting puzzles of plant cytology. As stated in the beginning, the variety of centrosomes and centro- spheres with and without radiations in various types of the thal- lophytes seems to me too confusing to promise an understanding of their relationships at present. Gregoire and Berghs (: 04) have interpreted the structure of the mitotic figure in the germinating spore of Pellia in a very different manner from the accounts of Farmer, Chamberlain, and myself. They consider the asters to arise through a re- arrangement of the cytoplasmic network around the nucleus. They affirm that there are no true centrospheres nor any ac- cumulations of granular kinoplasm to constitute the centers of origin for the spindle fibers or the radiations around the poles of the spindle. The centers of the asters ("vesicules polaires ") are said to have a vesicular structure and neither they nor the nucleus contributes to the building up of the spindle which is 724 THE AMERICAN NATURALIST. [VOL. XXXIX. developed entirely out of the cytoplasmic network. The au- thors are unable to distinguish a kinoplasm distinct from the general network of the cell. These are vital points of differ- ence which are fundamental to the understanding of mitotic phenomena and rest of course on matters of fact. The chief points at issue concern the structure and development of the asters and the nature of the material at their centers. My own studies and those of Farmer and Chamberlain have convinced me that there is an accumulation of substance (kinoplasm) in the centers of the asters and polar caps to such an amount that it must be regarded as a definite structure in the cell and its morphology and relations to the spindle have certainly justified us in considering it as similar to the centrosphere of the thallo- phytes. 4. THE ESSENTIAL STRUCTURES IN THE PLANT CELL AND THEIR BEHAVIOR IN ONTOGENY. The cell is composed of a series of osmotic membranes be- tween which are included a number of protoplasmic structures whose morphology and minute organization is various. They are : the outer plasma, the vacuolar, and the nuclear membranes. Each of these sustains a relation to some fluid which bathes its surface. The fluid nature of the nuclear sap and cell sap is obvious but the outer plasma membrane is also against a moist surface since the cell walls of tissues are normally saturated with water. The structure of the plasma membranes is apparently the same. They consist of the homogeneous finely gran- ular protoplasm that is designated kinoplasm. The protoplas- mic structures included within the plasma membranes may be grouped as cytoplasmic and nuclear. The greater part of the cytoplasm, including that which is termed trophoplasm, has an organization peculiar to itself and very different from that of the plasma membranes. This structure has been described as alveolar or of the nature of foam and sometimes fibrillar and with various large granular inclusions. The cytoplasm also con- tains the characteristic organs termed plastids. The conspicu- ous structures of the nucleus are : the chromatic elements No. 466.] STUDIES ON PLANT CELL. VIII. 725 appearing as chromosomes during mitosis and the nucleoli. These structures are so easily recognized and play such impor- tant parts in the events of nuclear division that they command attention at once as the essential elements in the nucleus. The nucleus may also contain other material such as linin which, however, does not seem to have a fixed form or behavior in the cell. Finally there are certain kinoplasmic structures, as cen- trosomes, centrospheres, and blepharoplasts, whose behavior throughout cell history has been much discussed. We shall now consider the most important of these structures, those which seem essential to the cell in ontogeny. The outer plasma membrane naturally retains its morphologi- cal entity throughout all cell divisions with such slight changes as when new parts are intercalated into its area through the vacuoles that are utilized in the segmentation of protoplasm. Vacuolar membranes are constantly shifting and cannot be fol- lowed during cell division excepting in such cells as have one large central vacuole (the tonoplast of De Vries). Such a cen- tral vacuole is much more characteristic of old cells and tissues than of young or embryonic regions. There is certainly no reason to suppose that it has organic existence through any very extended period of the life history. The nuclear membrane becomes lost during the prophase of mitosis and there is .much evidence that its kinoplasm contributes in some cases to the formation of spindle fibers. Thus the nuclear membrane disap- pears as a structure in the cell during mitosis and new vacuoles are formed around the assemblages of daughter chromosomes during telophase, leading of course to the formation of fresh nuclear membranes at their surface of contact with the sur- rounding cytoplasm. There is perhaps no region of the cell protoplast that presents such different appearances through long periods of the cell -his- tory as the trophoplasm. This is largely due to the varying character of the inclusions which are not in themselves proto- plasmic but which give a mixed structure to the trophoplasmic regions of the cell. The inclusions may be carbohydrate or pro- teid bodies held within spaces in the trophoplasmic groundwork or they may be globules of oil or fatty substances. These 726 THE AMERICAN NATURALIST. [You XXXIX. inclusions occupy small spaces in the trophoplasm which are essentially vacuoles. There is also a class of granular inclusions of a proteid nature which probably represent material in very close organic relation to the substance of protoplasm. Tropho- plasm does not then have so clearly denned a type of structure as do the other regions of the protoplast but it is hardly probable that its essential nature changes very materially throughout the life history. The organization of trophoplasm is itself a matter of dispute but the prevailing views favor an alveolar or foam structure with a fibrous character at times somewhat resembling the texture of sponge. Ever since the classical investigations of Schimper upon the plastid it has generally been held that these structures are per- manent organs of the cell, reproducing by fission, and carried along from one cell generation to the next with as much per- manence as the nucleus. Schimper discovered plastids in the oospheres of certain spermatophytes and in a variety of embry- onic tissues and concluded that the structures passed from parents to offspring as leucoplasts when no trace of color could be found in the reproductive cells or embryonic tissues. There has been, however, no systematic study of the plastid through- out the life history of higher plants and in most of the green thallophytes there are reproductive phases, such as resting spores, where we have no knowledge of the structure or distri- bution of the chromatophores in the cell. It is very important that the plastid be investigated with the same degree of atten- tion which has been given to the nucleus, and that it be fol- lowed through all periods of the life history in forms where the color becomes greatly modified or is absent in the reproductive cells and embryonic (meristematic) regions of the plant. Any- one who has studied the embryonic tissues of plants will realize the difficulties of the investigation which will probably involve the development of methods of technique, especially of staining, somewhat different from those generally employed in cell studies. We may now consider the elements in the nucleus and their behavior during ontogeny. This is one of the most interesting subjects in cell studies, for the importance of the chromosomes No. 466.] STUDIES ON PLANT CELL. VIII. 727 and chromosome history in relation to problems of development, heredity, hybridization, and variation is clearly understood, and these subjects have already been treated in Section V, " Cell Activities at Critical Periods of Ontogeny in Plants." Also some recent papers on the nucleolus of which Wager's (: 04) is the most comprehensive, have brought this structure into very close relation with the chromatin content of the nucleus, and the nucleolus must now be considered in any treatment of the chro- mosomes. The problems hinge on what is termed the individu- ality of the chromosome, which is the question whether or not the chromosome is a structural entity maintaining its independ- ence completely through each and all of the cell divisions in a life history. There is also involved the view that the chromo- somes have come down from a line of ancestral structures, reproducing by fission in every mitosis throughout the history of the race. There are two extremes in the views on this exceedingly interesting conception and also an intermediate position. The one extreme has recently been set forth by Boveri (: 04) in a very clear statement. This view regards the chromosomes as structural entities, possibly elementary organisms, which main- tain an organic individuality and independent existence in the cell. They are further regarded as in their typical form when present as rods or short filaments during mitosis. Their be- havior in the resting nucleus is one of great metabolic activity which affects their morphology for the time being. Those who are inclined to doubt the individuality of the chro- mosomes and to hold off from a full acceptance of the theory, base their attitude on the extreme difficulty or perhaps impossi- bility of following the chromosomes as entities through the rest- ing nucleus from one mitosis to another. These difficulties are well known to those who have studied chromosomes even in nuclei which are most favorable for the investigation of their morphology. The chromosomes which enter the daughter nuclei from a mitosis generally lose their form and the chroma- tin becomes so distributed on a linin network or in a nucleolar structure that the outlines of the original structures become quite lost. Mottier (: 03) in his recent studies on the spore 728 THE AMERICAN NATURALIST. [VOL. XXXIX. mother-cell of certain angiosperms has emphasized these points and Gre^goire and Wygaerts (: 03) have also shown the difficul- ties of following the chromosomes in the resting nuclei of the root tip and spore mother-cell of Trillium, stating that the struc- tures become resolved into an alveolar network. On the other hand Rosenberg (: 04) claims that the chromo- somes may be clearly recognized in the resting nuclei of some forms and cites Capsella bursa-pastoris as a particularly good illustration. In this plant the chromosomes are described as small granular bodies scattered throughout the nucleus in fixed number at various stages of ontogeny. Thus there are 16 in cells of the garnet ophyte and 32 in those of the sporophyte while 48 of these bodies were counted in the nuclei of the endo- sperm as would be expected if these nuclei are descendants of a triple fusion in the embryo-sac. Similar conditions are reported in other forms and there is considerable evidence giving weight to the view that chromosomes may be actually followed through all periods of the nuclear history in some favorable types. Apart from the actual demonstration of the chromosomes in the resting nuclei and their recognition as structural entities through successive cell divisions there is much general evidence in support of the theory of the individuality of the chromosomes. This evidence lies in the nuclear fusions of fertilization and the mitoses of processes of segmentation that follow where the chromosomes are known to remain separate and have been dis- tinguished as maternal and paternal. Also, as we have seen from the discussions of reduction phenomena at sporogenesis and the behavior of the chromosomes in hybridization, there are good reasons for believing that maternal and paternal chromo- somes remain separate all through the sporophyte generation and are distributed to the offspring during sporogenesis. The importance of these events in the minds of all investigators has rested very largely on the behavior of the chromosomes and has led to the very gener'al assumption that they must stand for units of organization and may be counted as constant factors in the problems of heredity. It is not necessary to adopt Boveri's extreme views to hold still the theory of the individuality of the chromosomes. Nor is it necessary to assume that the structures No. 466.] STUDIES ON PLANT CELL. VIII. 729 have a distinct organization which holds throughout the life his- tory. The form of the chromosomes certainly does change with different periods of the cell's history especially within the rest- ing nucleus and yet the centers of chromosome activity may always be present to organize the chromatin into a new set of elements for the next mitosis. It is perhaps difficult to believe that the chromatin granules (chromomeres) find their way back to the same chromosome with the prophase of each mitosis but the existence of chromosome centers may be readily conceived within the resting nucleus which would hold the number of chro- mosomes true to the cell's history. With respect to the nucleolus there is abundant evidence that the structure is not a permanent organ of the cell. When con- taining chromatin, the nucleolus is found in its characteristic globular state only during the resting condition of the nucleus. Its chromatic substance passes into the chromosomes at pro- phase of mitosis and the nucleolus generally disappears before metaphase. Or if any substance is left after the chromosomes are formed the remaining structure either gradually dissolves or is thrust forth bodily into the cytoplasm surrounding the mitotic figure where it disappears sooner or later. The nucleolus in higher types of mitosis never divides to pass on with the chro- mosomes to the daughter nuclei, but such a history is reported in the yeast cell. If the nucleolus has any function in heredity, as has been claimed (Dixon, '99), such function must relate to the chromosomes which contribute to its substance or derive material from it. Besides the nucleoli which are composed wholly or largely of chromatin, there are also those which seem to have little if any relation to the chromosomes. Such are well known in the spore mother-cells of higher plants and no investigator has been able to connect these with the formation of chromosomes as Wager (: 04) has been able to do in the root tip. It was upon nucleoli of this class that Strasburger founded his theory that the structure was a mass of reserve material utilized by the kinoplasm during mitosis in the process of spindle formation. Such nucleoli generally fade away during the pro- phase of mitosis and either entirely disappear or the remaining substance is thrust out into the cytoplasm where it may some- 730 THE AMERICAN NATURALIST. [VOL. XXXIX. times be recognized as deeply staining globules (the so called extranuclear nucleoli). There is left for our consideration that group of kinoplasmic structures termed centrosomes, centrospheres, and blepharo- plasts which, when accompanied by radiations, are called asters. Some authors regard these structures as homologous and believe them to be present in one form or another as permanent organs of the cell in certain types (see discussion of Ikeno, :O4). Against this view stand the well established facts of an increas- ing list of forms, both animals and plants, in which these struc- tures unquestionably arise de novo at certain periods in the cell's history. To the author this evidence seems insurmountable and he cannot believe that the aster is in itself a permanent organ of the cell. We shall not take up the subjects of relationships here for such discussions have proved of little profit except in special cases where the various types of structure are found in closely related forms or in the same life history, and these have scarcely been studied at all. We know so little about the rela- tionships in the thallophytes, where relationships must be sought if present at all, that a satisfactory treatment of the subject is hardly possible at present. One point seems to have escaped attention in the writings of those who have discussed the cen- trosome problem. The active elements of the asters are not the central structures (centrosomes, centrospheres, or blepharoplasts) but the fibrillae which play such important parts as spindle fibers or cilia. This fibrillar condition of kinoplasm has a fixed place in the cycle of cell division appearing with each mitosis and at the time of cilia formation, but the fibrillae are not permanent structures of the cell. There is some evidence that the centro- somes, centrospheres, and blepharoplasts are merely regions for the development and attachment of these fibrillae and as such may stand as the morphological expression of fibrillae-forming dynamic centers rather than as organs which actually induce the development of fibrillae. No. 466.] STUDIES ON PLANT CELL. VIII. 731 5. THE BALANCE OF NUCLEAR AND CYTOPLASMIC ACTIVI- TIES IN THE PLANT CELL. Two regions of the cell are sharply distinguished from one another with respect to both morphology and physiology. They are the nucleus and the cytoplasm. The nucleus soon dies if isolated from cytoplasm and the latter, lacking a nucleus, cannot be kept alive indefinitely unless it be in organic connection with a nucleated mass of protoplasm. The necessary connection may be only through delicate strands, as was established by Townsend ('97), and also seems to be illustrated in the instances of intercellular protoplasm which Michniewicz (: 04) reports are connected by delicate fibrillae (plasmodesmen) with neighboring cells. Some very interesting adjustments of the nucleus and cytoplasm to one another have been reported in a series of investigations of Gerassimow beginning in 1890. His most recent papers of the past year (Gerassimow, : O4a, : O4b) pre- sent a general summary of his studies and constitute a very im- portant contribution to the subject. They will furnish much of the material for this discussion. Gerassimow has found that the cells of Spirogyra and other members of the Conjugales offer admirable material for the study of the relations between the nucleus and cytoplasm, and throw important light on the functions, physiological activities, and interdependence of both structures. By subjecting fila- ments of Spirogyra during cell division to a temperature of o C. or treating them for a short time to the anaesthetic influ- ence of ether, chloroform, or chloral hydrate it is possible to arrest the processes of mitosis at different stages with the result that the protoplasm may become variously distributed in the daughter cells, (i) A daughter cell may be formed lacking a nucleus but containing a portion of the divided chromatophore in a peripheral layer of cytoplasm. (2) A single cell may con- tain the two daughter nuclei either separated from one another or more or less intimately associated and perhaps wholly fused depending upon how far the processes of mitosis have pro- gressed before the cells have been subjected to the shock of the 732 THE AMERICAN NATURALIST. [VOL. XXXIX. experiment. (3) Binucleate cells may continue their growth with subsequent mitoses which when treated as before may give daughter cells with three nuclei and one nucleus respectively or with two each or indeed a cell containing four nuclei. Further- more these nuclei may fuse with one another to give structures with a greatly increased chromatin content. (4) In place of the non-nucleated cells there may be formed chambers containing cytoplasm and chromatophores, but without nuclei, which remain in open communication with the nucleated companion protoplast because the cell wall is not formed entirely across the mother- cell. Gerassimow has made some extended observations on these various types of cells, and presents his results in many elaborate tables and diagrams. We can only give an outline of his con- clusions, (i) Cells which come to contain unusually large nuclei through the suppression of mitosis or by the reuniting of partially divided daughter nuclei increase proportionally in size and their further cell division is postponed. The nuclei of such cells have of course the peculiarity of an increased amount of chromatin content. The large nuclei may later fragment into two or more structures which separate and generally come to lie at a distance from one another in the cytoplasm. The frag- ments finally lose their powers of reproduction and exhibit marked evidence of degeneration. (2) Cells which lack nuclei may form starch in. the usual manner in the presence of light and exhibit for a short time a weaker general growth than nor- mal nucleated cells. The power to develop a gelatinous sheath also becomes markedly weakened. Finally there result a de- crease in the volume of the cell, a fading of the chromatophore, and conditions which lead to eventual death. (3) Chambers which lack nuclei but are in protoplasmic union with nucleated cells may be contrasted sharply with the non-nucleated cells. They exhibit a much stronger growth for a longer time and with a greater power to form starch, although not so marked as in the nucleated cells, and the chromatophores retain their color. There is also a conspicuous development of the gelatinous sheath. Haberlandt, Klebs, Pfeffer, Strasburger, and others have dis- No. 466.] STUDIES ON PLANT CELL. VIII. 733 cussed the relations of the nucleus to the surrounding proto- plasm with respect to both dynamics and morphology. Klebs ('88) indeed anticipated some of the work of Gerassimow, study- ing the non-nucleated cells of Zygnema and Spirogyra and noting the ability of their chromatophores to form starch in considerable quantities but the inability of the protoplast to add to the cell wall. Klebs was able to keep these non-nucleated cells alive in a sugar solution for from four to six weeks. But for the most part the discussions of the balance of nuclear and cytoplasmic activities in the plant cell have been very general in character. Some principles have been, however, widely held for several years and may be summarized. The necessity of the nucleus to the life of the cytoplasm has been clearly understood but the studies of Klebs and Gerassimow indicate that the nucleus is not directly concerned with the process of photosynthesis which apparently may go on in non-nucleated cells as long as the cyto- plasm retains a certain degree of vitality. A non-nucleated cell may enlarge slightly but it is not probable that the amount of protoplasm is increased. An especially interesting feature of non-nucleated cells is the inability of the outer plasma membrane to form cellulose walls or outer membranes. But the very inter- esting studies of Townsend ('97) have shown that this power may be retained provided the non-nucleated mass of protoplasm is connected by delicate cytoplasmic fibrils with a nucleated mass. It thus seems clear that the membrane-forming possi- bilities of the outer plasma membrane are absolutely dependent upon dynamic relations with the nucleus. While the chromato- phore may carry on the processes of photosynthesis independ- ently of the nucleus, nevertheless the general health of the cell requires the activities of the latter so that the nucleus becomes necessary to any extended photosynthetic work. It has frequently been stated that the size of the nucleus is directly proportionate to the amount of cytoplasm in the cell. There are many favorable illustrations of this statement, as the extraordinarily large eggs of the gymnosperms, especially the cycads, whose nuclei are by far the largest in the plant kingdom. And in general an increase in the amount of cytoplasm is accom- 734 THE AMERICAN NATURALIST. [VOL. XXXIX. panied either by a marked enlargement of the nucleus with a corresponding increase in the chromatin content or by mitoses which distribute to the cytoplasm a greater number of nuclei whose sum total of material is very much greater than before. Conversely a sudden increase in nuclear material through nuclear fusions either sexual or asexual is followed almost immediately by general cell growth and increase in the amount of cytoplasm. However, such fixed growth relations between nucleus and cyto- plasm can hardly be an established physiological law for certain highly specialized sperms have an insignificant amount of cyto- plasm proportionately to the chromatin that is contained within the gamete nucleus. It is evident that the interrelations of the nucleus and the cytoplasm are so intimate that the growth activi- ties of the one must benefit the other, but that this principle can be formulated in definite mathematical ratios seems im- probable. The dependence of the nuclei upon favorable situations in the cytoplasm is clearly shown in cells. when a partial or general nuclear degeneration takes place. Thus during the processes of oogenesis in the Peronosporales, Saprolegniales, and in Vau- cheria there is present a period when the most of the numerous nuclei within the oogonia begin to break down and finally become disorganized. The causes of the nuclear degeneration are not entirely clear but apparently the organ is unable to supply all of the nuclei in their respective situations in the cyto- plasm with the conditions necessary for their life. There is con- sequently a sort of struggle for existence among these numerous nuclei and only those that are favorably placed in the cell are able to survive. In all forms the surviving nuclei occupy a situ- ation in the center of the masses of protoplasm which are to become the eggs and those that break down are at or near the periphery of the cell. In several genera (e.g., Albugo, Perono- spora, Pythium, Sclerospora, Saprolegnia, and Achlya) the sur- viving nuclei seem to owe their good fortune to a very close association with the cytoplasmic structure termed the coenocen- trum. The coenocentrum is a clearly differentiated region of the cytoplasm and is probably the morphological expression of a dynamic center in the eggs of these fungi. Stevens' ('99, : 01) No. 466.] STUDIES ON PLANT CELL. VIII. 735 studies on Albugo showed that the coenocentra exert a chemo- tactic influence upon the nuclei in their vicinity, drawing them towards the mass of granular material in this favored region of the cell, and it is clear that they are greatly benefited in this situation since they increase in size while the nuclei at the periphery break done. This subject is discussed in detail in my paper on Saprolegnia (Davis, : 03,. pp. 240-243) a form which also illustrates exceptionally well the same principles of a sur- vival of certain nuclei among many which degenerate, because of their favorable position in the central region of the eggs in close proximity to coenocentra. There are then undoubtedly regions of the cell more favorable for the nutrition of nuclei than others and the positions of these may be marked by morphological characters as illustrated in the coenocentra. That similar dyna- mic centers may also be present when there is little morphologi- cal evidence of their existence is indicated in the processes of oogenesis in Vaucheria (Davis, : 04) which exhibits the same principles "of extensive nuclear degeneration as are found in the Peronosporales and Saprolegniales and the survival of a single nucleus in the oogonium, apparently because it comes to lie in a mass of granular cytoplasm near the center of the oogonium. 736 THE AMERICAN NATURALIST. [VOL. XXXIX. LITERATURE CITED IN SECTION VI, "THE PLANT CELL." BARKER. :01. A Conjugating Yeast. Phil. Trans. Roy. Soc. London, vol. 194, p. 467. BEER. :04. The Present Position of Cell-wall Research. New Phytologist, vol. 3, p. 159. BOVERI. : 04. Ergebnisse iiber die Konstitution der chromatischen Substanz des Zellkerns. Jena, 1904. BIJTSCHLI. '90. Ueber den Bau der Bakterien und verwandter Organismen. Leip- zig, 1890. BUTSCHLI. '96. Weitere Ausfiihrungen iiber den Bau der Cyanophyceen und Bak- terien. Leipzig, 1896. BUTSCHLI. :02. Bemerkungen iiber Cyanophyceen und Bakteriaceen. Arch. f. Protistenkunde, vol. i, p. 41. CHAMBERLAIN. :03. Mitoses in Pellia. Bot. Gaz., vol. 36, p. 29. DAVIS. '98. Kerntheilung in der Tetrasporenmutterzelle bei Corallina officinalis L., var. mediterranea. Ber. d. deut. hot. Gesellsch., vol. 16, p. 266. DAVIS. '99. The Spore Mother-cell of Anthoceros. Bot. Gaz., vol. 28, p. 89. DAVIS. : 01. Nuclear Studies on Pellia. Annals of Bot., vol. 15, p. 147. DAVIS. :03. Oogenesis in Saprolegnia. Bot. Gaz., vol. 35, pp. 233, 320. DAVIS. :04. Oogenesis in Vaucheria. Bot. Gaz., vol. 28, p. 81. DIXON. '99. The Possible Function of the Nucleus in Heredity. Annals of Bot., vol. 8, p. 269. FARMER AND MOORE. : 05. On the Maiotic Phase (Reduction Division) in Animals and Plants. Quart. Jour. After. Set'., vol. 48, p. 489. FARMER AND REEVES. '94. On the Occurrence of Centrospheres in Pellia epiphylla Nees. Annals of Bot., vol. 8, p. 219. No. 466.] STUDIES ON PLANT CELL. VIII. 737 FEINBERG. : 00. Ueber den Bau der Bakterien. Centralb. f. Bak., pt. i, vol. 27, p. 417. FEINBERG. : 02. Ueber den Bau der Hefezellen und iiber ihre Unterscheidung von einzelligen thierischen Organismen. Ber. d. deut. hot. Gesellsch., vol. 24, p. 567. FISCHER. '97. Untersuchungen iiber den Bau der Cyanophyceen und Bakterien. Jena, 1897. FISCHER. '99. Fixirung, Farbung, und Bau des Protoplasmas. Jena, 1899. GERASSIMOW. : 04a. Zur Physiologic der Zelle. Bull. Soc. Imp. Nat. Moscou, no. I. GERASSIMOW. :04b. Ueber die Grosse des Zellkerns. Beih. z. Bot. Centralb., vol. 18, P-45- GOLENKIN. '99. Ueber die Befruchtung bei Sphceroplea annulina und iiber die Structur der Zellkerne bei einigen griinen Algen. Bull. Soc. Imp. Nat. Moscou, p. 343. GREGOIRE AND BERGHS. : O4. La figure achromatique dans le Pdlia epiphylla. La Cellule, vol. 21, p. 193. GREGOIRE AND WYGAERTS. : 03. La reconstitution du noyau et la formation des chromosomes dans les cineses somatiques. i. Racines de Trillium grandiflorum et telophase homoeotypique dans Trillium cernuum. La Cellule, vol. 21, p. 7. GUILLIERMOND. :03. Recherches cytologiques sur les levures. Rev. Ge"n. Bot., vol. 15, pp. 49, 104, 1 66. GUILLIERMOND. :04. Sur le noyau de la levure. Ann. Mycol., vol. 2, p. 184. HEGLER. : 01. Untersuchungen iiber die Organisation der Phycochromaceenzelle. Jahrb.f. wtssensch. Bot., vol. 36, p. 229. HlERONYMUS. '92. Beitrage zur Morphologic und Biologic der Algen. Cohrfs Bettr. z. Biol. d. Pflan., vol. 5, p. 461. HlRSHBRUCH. :02. Die Fortpflanzung der Hefezelle. Centralb. f. Bak., pt. 2, vol. 9, pp. 465, 513, 737. IKENO : 03. Beitrage zur Kenntniss der pflanzlichen Spermatogenese ; die 738 THE AMERICAN NATURALIST. [VOL. XXXIX. Spermatogenese von Marchantia polymorpha. Beth. 2. Bot. Centralb., vol. 15, p. 65. IKENO. :04. Blepharoplasten im Pflanzenreich. Biol. Centralb., vol. 24, p. 211. JANSSENS. :03. A propos du noyau de la levure. La Cellule, vol. 20, p. 337. JANSSENS AND LEBLANC. '98. Recherches cytologiques sur la cellule de levure. La Cellule, vol. 14, p. 203. KOHL. :03. Ueber die Organisation und Physiologic der Cyanophyceenzelle und die mitotische Teilung ihres Kernes. Jena, 1903. KLEBS. '88. Beitrage zur Physiologic der Pflanzenzelle. Unters. a. d. hot. Inst. z. Tub., vol. 2, p. 489. LAWSON. :03. On the Relationship of the Nuclear Membrane to the Protoplast. Bot. Gaz., vol. 35, p. 305. MACALLUM. '99. On the Cytology of Non-nucleated Organisms. Trans. Can. Inst., vol. 6, p. 39. MANO. :04. Nucle'ole et chromosomes dans le me'risteme radiculaire de Solatium tuberosum et Phaseolus vulgaris. La Cellule, vol. 22, p. 57. MARPMANN. : 02. Ueber Hefen und iiber den Zellkern bei Saccharomyceten und Bakterien. Centralb. f. Bak., pt. 2, vol. 9, p. 357- MASSART. :02. Sur le protoplasme des Schizophytes. Recueil d. rinst. Bot. Bru- xelles, vol. 5, p. 251. MERRIMAN. :04. Vegetative Cell Division in Allium. Bot. Gaz., vol. 37, p. 178. MICHNIEWICZ. : 04. Ueber Plasmodesmen in den Kotyledonen von Lupinus-arten und ihre Beziehung zum interzellularen Plasma. Oester. hot. Zeitsch., vol. 54, p. 165. MlGULA. '95. Ueber den Zellinhalt von Bacillus oxalticus Zopf. Arb. a. d. bak. Inst. d. tech. Hochsch. z. Karlsruhe, vol. i, p. 139. MOORE. :03. The Mitoses in the Spore Mother-cell of Pallavicinia. Bot. Gaz., vol. 36, p. 384. MOTTIER. : 03. The Behavior of the Chromosomes in the Spore Mother-cells of Higher Plants and the Homology of the Pollen and Embryo-sac Mother-cell. Bot. Gaz., vol. 35, p. 250. No. 466.] STUDIES' ON PLANT CELL. VIII. 739 NADSON. '95. Ueber den Bau des Cyanophyceen-protoplastes. Scripta Botan, Horti Petropoli, vol. 4. NAKANISHI. : 01. Ueber den Bau der Bakterien. Centralb. f. Bak., pt. i,vol. 30, PP- 97, 145, i'93 225. OLIVE. :04. Mitotic Division of the Nuclei of the Cyanophyceae. Beth. z. Bot. Centralb., vol. 18, p. 9. PALLA. '93. Beitrag zur Kenntniss des Baues des Cyanophyceen-protoplasts. Jahrb.f. wissetisch. Bot., vol. 25, p. 51 1. ROSENBERG. :04. Ueber die Individualitat der Chromosomen in Pflanzenreich. Flora, vol. 93, p. 251. SCOTT. '88. On the Nuclei in Oscillaria and Tolypothrix. Jour. Linn. Soc., Bot., vol. 24, p. 1 88. STEVENS. '99. The Compound Oosphere of Albugo bliti. Bot. Gaz., vol. 28, p. 149. STEVENS. : 01. Gametogenesis and Fertilization in Albugo. Bot. Gaz., vol. 32, p. 77- TOWNSEND. '97. Der Einfluss des Zellkerns auf die Bildung der Zellhaut. Jahrb.f. ivissensch. Bot., vol. 30, p. 484. VAN HOOK. :00. Notes on the Division of the Cell and Nucleus in Liverworts. Bot. Gaz., vol. 30, p. 394. VEJDOVSKY. : 04. Ueber den Kern der Bakterien und seine Teilung. Centralb. f. Bak., pt. 2, vol. 11, p. 481. WAGER. '98. The Nucleus of the Yeast Plant. Annals of Bot., vol. 12, p. 499. WAGER. : 03. The Cell Structure of the Cyanophyceae. Proc. Roy. Soc. London, vol. 72, p. 401. WAGER. : 04. The Nucleus and Nuclear Division in the Root Apex of Phaseolus. Annals of Bot., vol. 18, p. 29. WOLFE. :04. Cytological Studies on Nemalion. Annals of Bot., vol. 18, p. 607. ZACHARIAS. : 00. Ueber die Cyanophyceen. Abhand. a. d. Gebiete d. Naturnvis- sensch, vol. 16. 740 THE AMERICAN NATURALIST. [VOL. XXXIX. ZACHARIAS. :03. Ueber die Cyanophyceen. Jahrb. d. Hamburg, wissensch. Anstal- ten, vol. 21, p. 49. ZETTNOW. '97 Ueber den Bau der grossen Spirillen. Zeitsch. f. Hyg., vol. 24, p. 72. ZETTNOW. '99. Romanowski's Farbung bei Bakterien. Zeitsch. f. Hyg., vol. 30, pi. i. ZUKAL. '92. Ueber den Zellinhalt der Schizophyten. Ber. d. deut, hot. Gesellsch. vol. 10, p. 51. DIADASIA PATTON; A GENUS OF BEES T. D. A. COCKERELL. THE genus Diadasia was first described by Patton in the Bulletin of the United States Geological Survey (vol. 5, p. 475). The type is the Melissodes enavata of Cresson, which, as Patton showed, is nearer to Anthophora than to Melissodes. The genus occurs in our southwestern States, and is, undoubt- edly, of neotropical derivation. Ashmead has recently placed it as a synonym of the South American Ancyloscelis Latreille, but it appears to me to be sufficiently distinct. Our species of Diadasia have not hitherto been tabulated, and as I have now seen* all the species but one, I offer tables for their identification. The species of Cresson are in the collec- tion at the Philadelphia Academy ; I have been permitted to borrow cotypes from that institution, through Mr. Viereck, and this has enabled me to clear up several doubtful points. En- technia toluca (Melissodes toluca Cresson) and Dasiapis ochracea Ckll., are included in the table, as the first has for some years stood in our lists as a Diadasia, while the latter is often mis- taken for a species of that genus. FEMALES. Hair of head and thorax above short and dense, orange fulvous ; abdomen with four clean cut bands of fulvous tomentum on a black ground ; outer side of basal joint of hind tarsi with very long, strongly plumose, dark chocolate-colored hairs ; inner side of this joint with shining dark ferru- ginous hair ; tegulae red ; flagellunrall dark; front rough with very close punctures ....... sumichrasti (Cresson). Hair of thorax not thus colored ; or if fulvous, abdomen not thus banded i . 1. Scopa on outside of hind legs dark gray or blackish (in afflicta paler on basal part of tibiae.) . . . . . . . 2. Scopa on outside of hind legs white, or not gray or blackish . 4. 2. Very small ; less than 8 mm. long ; abdomen with narrow bands of tomentum on apical margins of segments ; mesothorax and scutellum 74i 742 THE AMERICAN NATURALIST. [VOL. XXXIX. minutely, extremely densely punctate all over, therefore rough and not shining ...... Entechnia toluca (Cresson). Larger ; at least over 8 mm. long ; mesothorax well punctured but shin- ing . 3. 3. Large and stout ; about 12 mm. long, or more D. bituberculata (Cresson). Smaller ; about 10 mm. long, or less . . . afflicta (Cresson). 4. Very large species, about 15 mm. long . megamorpha Cockerell. Large stout species, about 13 mm. long ; hair of thorax above pchraceous or fulvous, with the disc bare ... . . . . . 5. Smaller species, less than 12 mm. long . . . . . . 7. 5. Hind spur of hind tibia straight or practically so ; clypeus more closely punctured, the large punctures stronger . . enavata (Cresson). Hind spur of hind tibia strongly bent at end ; clypeus less closely punc- tured, the large punctures weaker .... . . . 6. 6. Legs dark red ; abdominal segments 3 and 4 with a narrow apical fringe, the rest thinly hairy .... australis (Cresson). Legs black ; abdominal segments 3 and 4 with lateral areas where the sur- face is raised and shining black, the hair on it being very sparse and dark . . . . . . australis opuntia (Cockerell). 7. Anterior edge of abdominal bands curved, the basal part of the seg- ments dark ; comparatively large and broad form : hind spur of hind legs curved at end .... australis rinconis (Cockerell). Anterior edge of abdominal bands not curved, the pubescence, except at margin, uniformly distributed ; smaller forms .... 8. 8. Hair on inner side of basal joint of hind tarsi light ferruginous ; abdo- men entirely covered with yellowish tomentum Dasiapis ochracea Ckll. Hair on inner side of basal joint of hind tarsi fuscous or black . 9. 9. Face broad, eyes scarcely converging below ; eyes narrow, especially above ; mesothorax shining, impunctate in middle, at sides with large -scattered punctures ; abdomen broad, with narrow ochreous hair-bands on hind margins of segments 2 to 4 ... laticauda Cockerell. Eyes broader and shorter, distinctly converging below ; mesothorax dul- ler, the sides with very numerous feeble minute punctures diniinuta (Cresson). Larger than the two last (i I mm. long) and at once separated from them by having much fuscous or black hair on the abdomen ; there are ochreous marginal hair-bands . ... . friesei Cockerell. MALES. Hair of face black -nigrifrons (Cresson). Hair of face not black ......... i. I. Apex of abdomen truncate; tongue very long; maxillary palpi not fringed with hair ; size very small . . Entechnia toluca (Cresson). No. 466.] GENUS DJ AD ASIA. 743 Apex of abdomen bidentate . ...'. . . . 2. 2. Abdomen above with much black hair on discs of segments beyond the second . . . . . . . . . -g ^_ 3. Abdomen above without black hair . . " . . . . 6. 3. Large; at least 13 mm. long; apical teeth of abdomen large and divergent ....... bituberculata (Cresson). Smaller; about 10 mm. long; apical teeth of abdomen small and close together . . .''-. . . . . . . 4. 4. Hind tibiae thickened, but shape not remarkable ; basal joint of hind tarsi dark ferruginous, long, slender, and curved, its apex not produced, the hair on its inner side orange ; maxillary palpi not fringed with hair, except a little tuft at the end of second joint ; tegulae light rufous sumichrasti (Cresson). Hind tibiae greatly swollen, narrowing to a very slender base, shaped something like a wine-bottle ; basal joint of hind tarsi dark, not so long, with black or dark fuscous hair on inner side . . . . 5. 5. Tegulae dark but decided red ; second submarginal cell much narrowed above; hair of mesothorax white . . . . afflicta (Cresson). Tegulae piceous : second submarginal cell scarcely narrowed above ; hair of mesothorax and scutellum gray . . afflicta perafflicta Cockerell. 6. Basal joint of hind tarsus ending in a long process ; species covered with gray hair ; maxillary palpi with no fringe of long hairs, but second joint ciliate ........... 7. Basal joint of hind tarsus not ending in a long process . . 8. 7. Larger forms ....... australis (Cresson). Smaller, down to about 10 mm. long . australis rinconis (Cockerell). 8. Very large, about 16 mm. long . . . megamorpha Cockerell. Rather large, length over 10 mm., the pubescence more or less ochrace- ous on thorax, sometimes quite fulvous ; facial quadrangle longer than broad ............ 9. Small, length less than 10 mm. ....... 10. 9. Hair of thorax more or less fulvous . . . enavata (Cresson). Hair of thorax paler .... enavata var. densa (Cresson). 10. Abdomen above shining and sparsely hairy, not banded ; face broad, orbits little converging below (distinctly less than in diminuta) nitidifrons Cockerell . Abdomen hairy, the hind margins of the segments banded diminuta (Cresson). Abdomen covered with appressed white tomentum sphceralcearum Cockerell. D. albovestita Provancher, I have not seen. It was described from the female; length just over 8 mm., flagellum reddish beneath, tegulae brownish, margins of abdominal segments pale yellow and covered with dense whitish pubescence ; apex red- 744 THE AMERICAN NATURALIST. [VOL. XXXIX. dish brown. It must be similar to D. sphceralcearnm, but the antennae are differently colored. The following species are not considered valid : D. tricincta Provancher, from California, is said by Fowler to be a synonym of enavata. This cannot be, from the descrip- tion ; but it is not apparent that it differs from afflicta. D. nerea Fowler, from California, is nigrifrons Cresson ; D. cinerea Fowler, from California, is bituberculata Cresson. Fowler can hardly be blamed for describing these as new, as when he pub- lished his paper Cresson's species were supposed to belong to Melissodes. D. ursina (Cresson) is enavata. D. apacha (Cres- son) is diminuta. The types of apacJia have been in some liquid, presumably alcohol, and this accounts for part of their characters. I formerly separated the specimens of the Middle Sonoran zone as apacha, leaving those of the Upper Sonoran as diminuta ; but the comparison of specimens from various locali- ties appears to show that the characters relied upon are too vari- able to serve for specific distinction. Two forms are new : Diadasia afflicta (Cr.) subsp. perafflicta n. subsp. $ Tegulae piceous ; second submarginal cell scarcely narrowed above ; hair of meso thorax and scutellum gray. ? . This sex does not materially differ from true afflicta. Hab. Clark Co., Kansas, 1962 ft., May (F. H. Snow, 1191) ; Hamilton Co., Kansas, 3350 ft. (F. H. Snow, 460) ; Wallace Co., Kansas, 3000 ft. (F. H. Snow, 852). Three females, from the same three localities, are numbered 851, 1197, and 445. Diadasia sphseralcearum n. sp. $. Length 7^ mm. ; like D. diminuta Cr., but with shorter, perfectly white pubescence, and a narrower, more parallel-sided abdomen ; the pubescence of the abdomen, instead of being loose and suberect as in male diminuta, is appressed (except on first segment) and covers the surface ; aside from the pubescence, the hind margins of the segments are them- selves white ; the apex is bidentate, the teeth being like those of diminuta, but rather larger ; hind legs constructed as in diminuta ; shining hairless triangle of metathorax much smaller than in diminuta ; posterior part of mesothorax almost nude ; tegulae subhyaline, ferruginous, dark at base ; antennas entirely black. No. 466.] GENUS DI AD ASIA. 745 Hab. Between Las Cruces and Mesilla Park, New Mexico, at flowers of Sphceralcea fendleri lobata (Wooton), middle of August (Cockerell). It was accompanied by Macroteropsis latior (Ckll.). The distribution of the species in States, etc., so far as known, is as follows : MEXICO. D. diminuta Cr. ; sumichrasti Cr. ; enavata Cr. (Lower Cali- fornia). CALIFORNIA. D. albovestita Prov. ; afflicta Cr. (tricincta Prov.) ; nigri- frons Cr. ; bituberculata Cr. ; nitidifrons Ckll. ; laticauda Ckll. ; friesei Ckll. ; enavata Cr. ; diminuta Cr. (Palm Spring, Davidson} ; australis rinconis Ckll. ; australis opuntice Ckll. NEVADA. D. bituberculata Cr. ARIZONA. D. diminuta Cr. (Bill Williams' Fork, Snow, Grand Canon, Hopkins) ; australis rinconis Ckll. (Bill Williams' Fork and Oak Creek Canon, Snow) ; enavata Cr. (Oak Creek Canon, Snow). NEW MEXICO. D. diminuta Cr. ; sphceralcearum Ckll. ; australis Cr. ; australis rinconis Ckll. ; enavata Cr. ; megamorpha Ckll. TEXAS. D. australis rinconis Ckll. (part of Cresson's original austra- lis, as shown by a cotype) ; enavata Cr. ; enavata v. densa Cr. (a color variation merely) ; afflicta Cr. KANSAS. D. australis Cr. (Wallace Co., and Morton Co., Snow); enavata Cr. (Wallace Co., Snow) ; diminuta Cr. (Hamilton Co., Snow) ; afflicta perafflicta Ckll. COLORADO. -D. enavata Cr. (Lamar, Snow, Palisade, Gillette, Jules- burg, Ball, Trinidad, Titus) ; enavata v. densa Cr. (Rocky Ford, in beet field, P. K. Blinn) ; diminuta Cr. (Fort Collins, Trinidad, Colo. Agric. Coll) ; australis Cr. D. snmichrasti Cr., is peculiar for the densely punctured mesothorax, but the blade of maxilla is broad at base and nar- row apically, as in true Diadasia. The maxillary palpi are long, 6-jointed. The sexes do not look much alike, but close com- parison confirms their identity. D. australis and its subspecies may be found visiting the flowers of Opuntia. The small species, diminuta and its allies, are addicted to the Malvaceae. D. megamorpha (?) was recorded from the flowers of Sphceralcea angustifolia, but the plant was really 6\ fendleri lobata, which had not then been differentiated. UNIVERSITY OF COLORADO, BOUI.DER, COLORADO.