ANNALS OF BOTANY VOL. XVIII Ojcforb PRINTED AT THE CLARENDON PRESS BY HORACE HART, M.A. PRINTER TO THE UNIVERSITY Annals of Botany ft* EDITED BY ISAAC BAYLEY BALFOUR, M.A., M.D., F.R.S. KING’S BOTANIST IN SCOTLAND, PROFESSOR OF BOTANY IN THE UNIVERSITY AND KEEPER OF THE ROYAL BOTANIC GARDEN, EDINBURGH D. H. SCOTT, M.A., Ph.D., F.R.S. HONORARY KEEPER OF THE JODRELL LABORATORY, ROYAL BOTANIC GARDENS, KEW AND WILLIAM GILSON FARLOW, M.D. PROFESSOR OF CRYPTOGAMIC BOTANY IN HARVARD UNIVERSITY, CAMBRIDGE, MASS., U.S.A. ASSISTED BY OTHER BOTANISTS % VOLUME XVIII With Forty-one Plates and Sixty-one Figures in the Text Bon&on HENRY FROWDE, M.A., AMEN CORNER, E.C. OXFORD: CLARENDON PRESS DEPOSITORY, 116 HIGH STREET TQ04 '\f ; j. i' J ri J,. / rT r\ 'i n D P I ; J "’V Oy v j y < r M.'ri'ra: ► : . < .. , i : . ; . . ' Y l V'l'i *{■'•* --:r :nr .vw/Yrcv: >{< > <"A'\ .r/- wt fr - j >" : riqurjc./ia;? ;.s s • /.'? f: : :m'T mo j;T oY- ,i\M /ITOOY .11 XI vvMVf :r tom ,v . v ■" . t -j. •:<: :<:r a / .< .. ’ A r/; uir mm / a. / r: oi : • 1' OYOO /.O' ■ Y. ■ ’ v. A TY5; i : : J: : <10 ZO.SHZ .to ! CONTENTS. No. LXIX, January, 1904. PAGE Lawson, A. A. — The Gametophytes, Archegonia, Fertilization, and Embryo of Sequoia sempervirens. With Plates I-IV . . . . . . . . . 1 Wager, H. — The Nucleolus and Nuclear Division in the Root-apex of Phaseolus. With Plate V 29 Worsdell, W. C. — The Structure and Morphology of the e Ovule/ An Historical Sketch. With twenty-seven Figures in the Text . . . . . . .57 Cavers, F. — On the Structure and Biology of Fegatella conica. With Plates VI and VII and five Figures in the Text 87 Potter, M. C. — On the Occurrence of Cellulose in the Xylem of Woody Stems. With Plate VIII I2i Williams, J. Lloyd. —Studies in the Dictyotaceae. I. The Cytology of the Tetrasporangium and the Germinating Tetraspore. With Plates IX and X 141 Benson, Miss M. — Telangium Scotti, a new Species of Telangium (Calymmatotheca) showing Structure. With Plate XI and a Figure in the Text 161 NOTES. Hem§ley, W. Botting. — On the Genus Corynocarpus, Forst. Supplementary Note . . 1 79 Weiss, F. E. — The Vascular Supply of Stigmarian Rootlets. With a Figure in the Text . 180 Ewart, A. J. — Root-pressure in Trees 181 No. LXX, April, 1904. Williams, J. Lloyd. — Studies in the Dictyotaceae. II. The Cytology of the Gametophyte Generation. With Plates XII, XIII, and XIV . 183 Bower, F. O. — Ophioglossum simplex, Ridley. With Plate XV ...... 205 Parkin, J. — The Extra-floral Nectaries of Hevea brasiliensis, Miill.-Arg. (the Para Rubber Tree), an Example of Bud-Scales serving as Nectaries. With Plate XVI . . .217 Church, A. H. — The Principles of Phyllotaxis. With seven Figures in the Text . . . 227 Mottier, D. M. — The Development of the Spermatozoid in Chara. With Plate XVII . 245 Weiss, F. E. — A Mycorhiza from the Lower Coal-Measures. With Plates XVIII and XIX and a Figure in the Text ........... 255 Reed, H. S. — A Study of the Enzyme-secreting Cells in the Seedlings of Zea Mais and Phoenix dactylifera. With Plate XX 267 Vines, S. H. — The Proteases of Plants 289 NOTES. Massee, G. — On the Origin of Parasitism in Fungi . . . . . . . .319 Salmon, E. S. — Cultural Experiments with ‘ Biologic Forms ’ of the Erysiphaceae . . . 320 Oliver, F. W., and Scott, D. H. — On the Structure of the Palaeozoic Seed Lagenostoma Lomaxi, with a Statement of the Evidence upon which it is referred to Lyginodendron . 321 VI Contents . No. LXXI, July, 1904. PAGE Blackman, V. H. — On the Fertilization, Alternation of Generations, and general Cytology of the Uredineae. With Plates XXI-XXIV - 323 Darbishire, O. V. — Observations on Mamillaria elongata. With Plates XXV and XXVI . 375 Lawson, A. A. — The Gametophytes, Fertilization, and Embryo of Cryptomeria japonica. With Plates XXVII-XXX 417 Gregory, R. P. — Spore-Formation in Leptosporangiate Ferns. With Plate XXXI and a Figure in the Text 445 Massee, G. — A Monograph of the genus Inocybe, Karsten. With Plate XXXII . . . 459 Boodle, L. A. — On the Occurrence of Secondary Xylem in Psilotum. With Plate XXXIII and seven Figures in the Text 505 NOTE. Scott, D. H. — On the Occurrence of Sigillariopsis in the Lower Coal-Measures of Britain . 519 No. LXXII, October, 1904. Engler, A. — Plants of the Northern Temperate Zone in their Transition to the High Mountains of Tropical Africa 523 Trow, A. H. — On Fertilization in the Saprolegnieae. With Plates XXXIV-XXXVI . . 541 Lang, W. H. — On a Prothallus provisionally referred to Psilotum. With Plate XXXVII . 571 Burns, G. P. — Heterophylly in Proserpinaca palustris, L. With Plate XXXVIII . . 579 Ford, Miss S. O. — The Anatomy of Psilotum triquetrum. With Plate XXXIX . . *589 Wolfe, J. J. — Cytological Studies on Nemalion. With Plates XL and XLI and a Figure in the Text 607 Ganong, W. F. — An undescribed Thermometric Movement of the Branches in Shrubs and Trees. W7ith six Figures in the Text 631 NOTES. Wigglesworth, Miss G. — The Papillae in the epidermoidal Layer of the Calamitean Root. With three Figures in the Text 645 Fritsch, F. E. — Algological Notes. No. 5 : Some points in the Structure of a young Oedogonium. With a Figure in the Text 648 Pertz, Miss D. F. M. — On the Distribution of Statoliths in Cucurbitaceae .... 653 Hill, T. G. — On the Presence of a Parichnos in Recent Plants 654 . * INDEX. A. ORIGINAL PAPERS AND NOTES. Benson, Miss M. — Telangium Scotti, a new Species of Telangium (Calymmatotheca) showing Structure. With Plate XI and a Figure in the Text Blackman, V. H. — On the Fertilization, Alternation of Generations, and general Cytology of the Uredineae. With Plates XXI-XXIV Boodle, L. A. — On the Occurrence of Secondary Xylem in Psilotum. With Plate XXXIII and seven Figures in the Text Bower, F. O. — Ophioglossum simplex, Ridley. With Plate XV Burns, G. P. — Heterophylly in Proserpinaca palustris, L. With Plate XXXVIII . Cavers, F. — On the Structure and Biology of Fegatella conica. With Plates VI and VII and five Figures in the Text * Church, A. H. — The Principles of Phyllotaxis. With seven Figures in the Text . Darbishire, O. V. — Observations on Mamillaria elongata. With Plates XXV and XXVI . Engler, A. — Plants of the Northern Temperate Zone in their Transition to the High Mountains of Tropical Africa ........... Ewart, A. J. — Root-pressure in Trees Ford, Miss S. O. — The Anatomy of Psilotum triquetrum. With Plate XXXIX . Fritsch, F. E. — Algological Notes. No. 5 : Some points in the Structure of a young Oedogonium. With a Figure in the Text ......... Ganong, W. F. — An undescribed Thermometric Movement of the Branches in Shrubs and Tre6s. With six Figures in the Text . Gregory, R. P. — Spore-Formation in Leptosporangiate Ferns. With Plate XXXI and a Figure in the Text Hemsley, W. Botting. — On the Genus Corynocarpus, Forst. Supplementary Note Hill, T. G. — On the Presence of a Parichnos in Recent Plants Lang, W. H. — On a Prothallus provisionally referred to Psilotum. With Plate XXXVII Lawson, A. A. — The Gametophytes, Archegonia, Fertilization, and Embryo of Sequoia sempervirens. With Plates I-1V The Gametophytes, Fertilization, and Embryo of Cryptomeria japonica. With Plates XXVII-XXX Massee, G. — A Monograph of the genus Inocybe, Karsten. With Plate XXXII . On the Origin of Parisitism in Fungi . . Mottier, D. M. — The Development of the Spermatozoid in Chara. With Plate XVII Oliver, F. W., and Scott, D. H. — On the Structure of the Palaeozoic Seed Lagenostoma Lomaxi, with a Statement of the Evidence upon which it is referred to Lyginodendron . Parkin, J. — The Extra-floral Nectaries of Hevea brasiliensis, Miill.-Arg. (the Para Rubber Tree), an Example of Bud-Scales serving as Nectaries. With Plate XVI . Pertz, Miss D. F. M. — On the Distribution of Statoliths in Cucurbitaceae .... Potter, M. C. — On the Occurrence of Cellulose in the Xylem of Woody Stems. With Plate VIII Reed, H. S. — A Study of the Enzyme- secreting Cells in the Seedlings of Zea Mais and Phoenix dactylifera. With Plate XX Salmon, E. S. — Cultural Experiments with * Biologic Forms’ of the Erysiphaceae . Scott, D. H. — On the Occurrence of Sigillariopsis in the Lower Coal-Measures of Britain See Oliver, F. W. Trow, A. H. — On Fertilization in the Saprolegnieae. With Plates XXXIV-XXXVI . Vines, S. H. — The Proteases of Plants . .......... Wager, H. — The Nucleolus and Nuclear Division in the Root-apex of Phaseolus. With Plate V PAGE l6l 323 505 205 579 87 227 375 523 181 589 648 631 445 179 654 57i 4i7 459 319 245 321 217 653 121 267 320 519 54i 289 29 Vlll Index . PAGE Weiss, F. E. — A Mycorhiza from the Lower Coal-Measures. With Plates XVIII and XIX and a Figure in the Text 255 The Vascular Supply of Stigmarian Rootlets. With a Figure in the Text . 180 Wigglesworth, Miss G. — The Papillae in the epidermoidal Layer of the Calamitean Root. With three Figures in the Text 645 Williams, J. Lloyd. — Studies in the Dictyotaceae. I. The Cytology of the Tetrasporangium and the Germinating Tetraspore. With Plates IX and X . . . . .141 Studies in the Dictyotaceae. II. The Cytology of the Gametophyte Generation. With Plates XII-XIV 183 Wolfe, J. J. — Cytological Studies on Nemalion. With Plates XL and XLI and a Figure in the Text 607 Worsdell, W. C. — The Structure and Morphology of the ‘ Ovule.’ An Historical Sketch. With twenty-seven Figures in the Text -57 a. Plates. I-IV. V. VI, VII. VIII. IX, X. XI. XII-XIV. XV. XVI. XVII. XVIII, XIX. XX. XXI-XXIV. XXV, XXVI. XXVII-XXX. XXXI. XXXII. XXXIII. XXXIV-XXXVI. XXXVII. XXXVIII. XXXIX. XL, XLI. B. LIST OF ILLUSTRATIONS. Sequoia sempervirens (Lawson). Root-apex of Phaseolus (Wager). Fegatella (Cavers). Cellulose in Woody Stems (POTTER). Dictyotaceae (Lloyd Williams). Telangium Scotti (Benson). Dictyotaceae (Lloyd Williams). Ophioglossum simplex (Bower). Hevea brasiliensis (Parkin). Spermatozoid of Chara (MotTier). Mycorhiza from Coal-Measures (Weiss). Enzyme-secreting cells (Reed). Uredineae (Blackman). Mamillaria (Darbishire). Cryptomeria japonica (Lawson). The Reduction-division in Ferns (Gregory). Characters of Hymenium in Inocybe (MASsee). Psilotum (Boodle). Fertilization in Saprolegnieae (Trow). Prothallus referred to Psilotum (Lang). Proserpinaca palustris (Burns). Psilotum (Ford). Nemalion (Wolfe). 1. Figures. 1-27. 28-32. 33* 34* 35-41* 42. 43* 44-50. 5i* 52-57* 58-60. 61. The Structure and Morphology of the Ovule (Worsdell) 59, 62, 72, 73, 76-78 The Structure and Biology of Fegatella conica (Cavers) 99, 102, 106, 109 Telangium affine (Benson) 164 The Vascular Supply of Stigmarian Rootlets (Weiss) . . . .181 The Principles of Phyllotaxis (Church) . 229, 230, 232-234, 237, 239 A Mycorhiza from the Lower Coal-Measures (Weiss) .... 258 Diagrams of the Reduction-division in Cyclops (Gregory) . . . 454 On the Occurrence of Secondary Xylem in Psilotum (Boodle) . . 509 Cytological Studies in Nemalion (Wolfe) . . . . . .619 Thermometric Movement of Branches in Shrubs and Trees (Ganong) 632, 634, 635, 638, 640 Papillae in the epidermoidal Layer of the Calamitean Root (Wiggles- worth) 645-647 Some points in the Structure of a young Oedogonium (Fritsch) . .651 Vol. XVIII. No. LXIX. January, 1904. Price 14s Annals of Botany EDITED BY ISAAC BAYLEY BALFOUR, M.A., M.D, F.R.S. KING’S BOTANIST IN SCOTLAND, PROFESSOR OF BOTANY IN THE UNIVERSITY AND KEEPER OF THE ROYAL BOTANIC GARDEN, EDINBURGH D. H. SCOTT, M.A., Ph.D., F.R.S. HONORARY KEEPER OF THE JODRELL LABORATORY, ROYAL BOTANIC GARDENS, KEW AND WILLIAM GILSON FARLOW, M.D. PROFESSOR OK CRYPTOGAMIC BOTANY IN HARVARD UNIVERSITY, CAMBRIDGE, MASS., U.S.A. 1 ASSISTED BY OTHER BOTANISTS s ' V ' * ■■ v r II r.friT* . t'V* > .. A , : ty&j x I /f ■/' .. , - 1\ £ ondon HENRY FROWDE, AMEN CORNER, E.C. ©jcfo** CLARENDON PRESS DEPOSITORY, 116 HIGH STREET 1904 Printed, by Horace Hart, at the* Clarendon Press, Oxford. PAGE CONTENTS. Lawson, A. A.— The Gametophytes, Archegonia, Fertilization, and Embryo of- Sequoia sempervivens. With Plates I-IV . . i Wager, H. — The Nucleolus and Nuclear Division in the Root-apex of Phaseolus. With Plate V 29 Worsdell, W. C.— The Structure and Morphology of the ‘Ovule.’ An Historical 'Sketch. With twenty-seven Figures in the Text 57 Cavers, F. — On the Structure and Biology of Fegatella conica. With Plates VI and VII and five Figures in the Text . . 87 Potter, M. C. — On the Occurrence of Cellulose in the Xylem of Woody Stems. With Plate VIII ...... 121 Williams, J. Lloyd. — Studies in the Dictyotaceae. I. The Cytology of the Tetrasporangium and the Germinating Tetraspore. With Plates IX and X 1 4 1 Benson, Miss M. — Telangium Scotti, a new Species of Telangium (Calymmatotheca) showing Structure. With Plate XI and a Figure in the Text . . . 161 NOTES. Hemsley, W. Botting.— -On the Genus Corynocarpus, Forst. Supplementary Note . . . 179 WEISS, F. E. — The Vascular Supply of Stigmarian Rootlets. With a Figure in the Text . .180 Ewart, A. J.— Root-pressure in Trees 181 NOTICE TO SUBSCRIBERS. The subscription-price of each- volume is thirty shillings, payable in advance : the Parts, four in number, are supplied as they appear, post free to subscribers in the United Kingdom, and with a charge of is. 6d. per annum for postage to subscribers residing abroad. The price of individual Parts is fixed at a higher rate. Intending subscribers should send their names, with subscription, to Henry Frowde, Oxford University Press Ware- house, Amen Corner, London, E.C. As the earlier volumes of the Annals of Botany are becoming scarce, Vol. I will only be sold as part of a complete set; and Parts will not as a rule be sold separately, after the publication of the volume to which they belong. A few extra copies of particular Parts at present remain on hand, for which special application must be made to the Editors, Clarendon Press, Oxford. NOTICE TO CONTRIBUTORS. Contributors in America should address their communications to Professor Farlow, Harvard University; and all other contributors, to the Editors, at the Clarendon Press, Oxford. Papers sent in with a view to publication must be type-written ; and the illustrative figures should be planned so as to properly fill a 4to or an 8vo plate. Attention to this will conduce to the rapid publication of the papers if accepted. Each contributor to the Annals of Botany is entitled to receive gratis twenty-five separate copies of his paper, and may purchase additional copies if he informs the Editors of his wishes in this respect when he returns corrected proofs. The price of these additional copies will depend upon the amount of text and the number of plates in the paper. The Gametophytes, Archegonia, Fertilization, and Embryo of Sequoia sempervirens. BY ANSTRUTHER A. LAWSON, Ph.D. Instructor in Botany, Stanford University , California, US. A. With Plates I-IV. Introduction. ..THOUGH investigations among the Gymnosperms, especially among the Cycadales and Ginkgoales, have, in the last few years resulted in most important and startling discoveries, this field of research has not received the attention it deserves. The literature on the Coniferales is gradually accumulating and revealing much that is of interest and of importance. These contributions are, however, too fragmentary to deserve the appreciation which they might otherwise receive. Representatives of every family of the Coniferales have been investi- gated, especially in regard to the gametophyte generation. Of the numerous types selected by the various investigators many phases in the life-history have been revealed which are of great morphological interest and importance. In spite of the large number of forms that have been worked upon, however, Pinus is the only Conifer in which a connected account of the important events completing the life-cycle has been compiled. The works on the other Conifers are nevertheless of great importance, and although they are at present but disconnected chapters in the life-history, their true value will no doubt be fully appreciated as soon as the missing chapters have been written. The interesting genus Sequoia is represented by two living species, Sequoia gig ante a and Sequoia sempervirens. The former species is confined to very narrow limits in California, while the latter extends along the coast ranges of middle and northern California and for about twelve miles into the State of Oregon. Of the latter species there are at present over one hundred trees growing on the campus of Stanford University. As the majority of these trees are healthy and vigorous, and although young, produce cones every year, and especially as many of them grow in the immediate vicinity of the Botanical Laboratory, excellent opportunity for the daily collections of material was afforded. Taking advantage of these exceptional facilities, I have thought it worth while to work out the morphology of the gametophytes, with the hope of filling in the gaps left [Annals of Botany, Vol. XVIII. No. LXIX. January, 1904.] B 2 Lawson . — The Gametophytes, Archegonia, Fertilization , by Arnoldi, Shaw, and Strasburger, the only writers who have contributed to our knowledge of the gametophytes of this interesting Conifer. It is therefore the object of the present work to give a connected account of the events leading to the development of the gametophytes, sexual organs, fertilization, and embryo, thus completing as far as possible the life-history of Sequoia sempervirens. While the observations recorded by Shaw and Arnoldi are of great interest, they are by no means complete, and as we shall see later, some are even inaccurate. Shaw (1896) has given a description of how the male and female flowers develop in Sequoia sempervirens. When very young he finds that the macrosporangium is about as long as it is broad. The integument at this time consists of an epidermis and two layers of hypodermal cells. The integument develops rapidly and soon comes to be about twice as long as the nucellus. Soon after pollination the upper and inner layers of epidermal cells enlarge and by their elongation finally close the micropyle. By this means the pollen-grains are completely enclosed in a subconical cavity at the apex of the nucellus. Within the nucellus several sporogenous cells now make their appearance. Shaw reports that these cells divide twice, each one giving rise to four macrospores. Upon germination the macrospores develop a number of female prothallia. As the embryo-sacs increase in size they contain several nuclei. The further development of the prothallia was not observed. The archegonia were found to be very numerous and distributed over the upper portion of the prothallium and each one has but a single neck-cell. At the time of pollination each microspore consists of two cells, a large central cell with a large nucleus and a much smaller cell. As the pollen-tube develops the nucleus of the larger cell moves forward and enters the tube. The tube extends down between the nucellus and the integument and as often as not it branches. The further course of the tubes was not followed. Later a number of long suspensors bearing the young embryos on their tips were formed in the endosperm. Arnoldi (1900) has given a description of the manner of endosperm- formation in Sequoia sempervirens. According to this description the form of the embryo-sac may vary considerably. When young it consists of a very large central vacuole surrounded by a parietal layer of cytoplasm in which numerous nuclei are found. As the sac develops the parietal layer increases and the protoplasm accumulates in great abundance in the lower end and at the same time the free nuclei divide repeatedly. In the central region a portion of the vacuole remains, and here the cellular endosperm is formed by means of ‘ Alveolen 5 as Sokolowa describes for other Conifers, while the remaining endosperm is formed by ordinary free cell-formation. The development of the archegonia is confined to the tissue derived from the ‘Alveolen/ so that the prothallium has a distinct generative region. Embryo of Sequoia sempervirens. 3 In the development of the endosperm in Sequoia , Arnoldi sees a striking similarity to that which occurs in Gnetuni. In the same year (1900) Arnoldi published some observations on the archegonia and pollen-tubes in Sequoia sempervirens . He finds that the archegonia arise from peripheral endosperm-cells and are present in large numbers and may appear singly or in groups. Each archegonium has two neck-cells but none contain a ventral canal cell. The position taken by the pollen-tubes is between the nucellus and the endosperm and they eventually lie opposite the archegonia. In his more recent work, Arnoldi (1901) touches upon fertilization and the development of the embryo. In addition to Sequoia this short paper also discusses these phases in the life-history of Taxodium , Cryptomeria , Cunninghamiay Glyptostrobus, and Sciadopitys . In Sequoia sempervirens , which concerns us more particularly, he finds that the pollen-tube eventually contains two male cells and two free nuclei, of which one is the tube nucleus. At the time of fertilization the male cell becomes elongated or even spirally twisted. The male and female nuclei fuse in the middle of the egg and then move to the base of the archegonium, where the first segmentation-spindle is developed. Following this division two cells are organized, one behind the other. The lower of these divides again so that the embryo now consists of a row of three cells. The lower cell of the first division functions no further and soon becomes disorganized. On the sporophyte of Sequoia , Peirce (1901) has contributed some interesting and important observations on fasciation, albinism and vegeta- tive reproduction. Methods. There are few groups of plants that offer more difficulties in the way of cytological research than the Coniferales. The structures that are of greatest cytological interest are usually buried deep in the other tissues, thus requiring very careful dissection before being placed in the killing fluids. Then, again, if resin is present, as is usually the case, a rapid penetration of the fluid is impossible. These and many other difficulties probably account for the frag- mentary nature of the work that has been done. A brief statement of the methods adopted in the following work on Sequoia may be useful to others working in this field. The fixing fluids experimented with were as follows : — ■ 1. Flemming’s weak solution — 25 c.c. of 1 °/o chromic acid 10 c.c. of 1 °/o acetic acid 10 c.c. of 1 °/o osmic acid 55 c.c. distilled water. 4 Lawson . — The Gametophytes , Archegonia , Fertilization , <2/^ Flemming’s strong solution. 3. Chrom-Acetic mixture. 4. Chromic Acid — 1 °/o sol. 5. Alcohol Acetic. Of these Flemming’s weak solution probably gave the best results although equally satisfactory fixation was generally obtained by the Chrom-Acetic and one per cent Chromic. The Alcohol Acetic solution proved to be a failure. The fixing fluids were always taken into the field and the material deposited in them immediately. In the very early stages of the ovules and also of the pollen no dissection was necessary. On account of the air present in them these structures had a tendency to float. This diffi- culty was, however, overcome by sinking the material in the fluid by means of cotton plugs. In the very early stages the entire ovules were removed and im- mediately killed without further dissection, but in all the later stages it was found necessary to remove the integument. This, however, was not resorted to until after the pollen-tube had penetrated the nucellus. To insure rapid fixation most of the dissections were made with the material immersed in the fluid. The ovules were removed one by one, placed in a watch-glass containing the fixing reagent and while in the fluid the in- tegument was immediately removed by means of a sharply pointed scalpel and forceps. With a little experience this may be accomplished very rapidly. The material was allowed to remain in the fixing fluid from ten to twenty-four hours and then washed in running water from four to six hours. Care was now taken in transferring the material to alcohol. For this purpose Schleicher and Schull’s diffusion shells were used. The shells were cut to the height of small beakers and the material placed in the bottom, and 95% alcohol placed in the beakers. Water was now poured in the shell in sufficient quantity to make the combined solutions about 7o°/o alcohol. By placing the shell, containing the material and water, in the beaker containing the 95 °/o alcohol, a gradual diffusion took place which was not sufficiently rapid to cause shrinkage. In two or three hours the material was transferred directly to 95 °/o alcohol. I found the shells much more convenient than ordinary parchment paper. In preparation for imbedding the material was thoroughly dehydrated in absolute alcohol. Bergamot oil was used to precede the infiltration of paraffin. After dehydration the material was placed in a mixture of 1 part absolute alcohol and 1 part bergamot oil ; then into pure bergamot oil ; then into a mixture of 1 part Bergamot oil and 1 part melted paraffin ; and finally into pure paraffin. Minot’s wheel microtome was employed for cutting and the sections varied from 2 //, to 8 /x in thickness according to the detail desired. 5 Embryo of Seqttoia sempervirens. It was found very desirable to use albumen instead of alcohol as a fixative. When the staining was not satisfactory, the sections fixed on the slide with albumen were bleached and restained without trouble. With the alcohol method, however, restaining was impossible, as the sections were invariably washed off the slide. By restaining, many valuable demonstrations were restored. The triple stain safranin, gentian, and orange G., was found to be the most satisfactory in differentiating the various cell-structures. The Male Gametophyte. The reduction-division of the microspore mother-cell leading to the formation of the tetrads takes place during the first week in December. The first division is rapidly followed by the second, and within a week or ten days after the first division the tetrads have separated and the pollen-grains formed. The microspores remain within the sporangium at least three weeks before pollination takes place. During this period they become larger, spherical in form, and surround themselves with a hard thick wall. The cytoplasm is very granular and contains a small amount of starch. The nucleus is comparatively small and is always centrally situated (PL I, Fig. i). While yet in the sporangium and about a week before pollination the nucleus of the microspore enlarges and divides ; so that at the time the pollen is shed there are two nuclei in each grain. Sections made before and after pollination showed a considerable difference in the size of the nuclei, the one being about twice the size of the other. The larger one was centrally situated, while the smaller one was invariably found near the spore wall. The smaller nucleus was surrounded by a sharply differen- tiated zone of very granular cytoplasm, which suggested the presence of a membrane between the two nuclei as shown in Fig. 2. The chromatin of the smaller nucleus was in the form of small granules closely packed together ; it consequently stained more deeply than the larger nucleus, where the meshes of the chromatin network appeared to be much more loosely arranged. A study of the further history of these nuclei has convinced me that the larger nucleus is the so-called tube-nucleus and the smaller one represents the generative cell. A very careful search was made with the hope of finding a vestige of the vegetative tissue of the gametophyte. One or more vegetative cells have been reported for the Cycads Ginkgo and Pinus> but a most searching examination failed to reveal even a vestige of such a cell or nucleus in Sequoia. I am strongly inclined to believe that the develop- ment of these evanescent structures has been entirely suppressed. Observations of two years indicated that pollination takes place during the first week in January, just about the time the female flowers 6 Lawson . — The Gametophytes , Archegonia , Fertilization , make their appearance. During this time the trees are constantly en- veloped in a cloud of pollen, so that it would be almost impossible for any of the exposed ovules to escape the reception of at least a few of the grains. At this time the integument of the ovule is about on a level or a little above the apex of the nucellus, and from four to six pollen-grains are here deposited. The grains remain in this position for three or four weeks without further germination, when the integument grows over them and closes the micropyle in the manner described by Shaw (1896). The first indication of the further germination of the pollen-grains was the splitting off of the hard thick wall. (If some ripe grains are examined in water under the microscope, it will be seen that the casting off of the outer wall takes place suddenly and with considerable force, leaving a thin delicate membrane underneath.) The pollen-tubes now push out over the tops of the nucellus, and one or two of them may grow down between the nucellus and the integument, as shown in Fig. 3. Material collected during the first week in March frequently showed the pollen-tubes extending more than halfway down the side of the nucellus. In such cases the tube-nucleus was invariably near the tip of the tube, with the generative nucleus considerably in the rear. Both nuclei were found in the central axis of the tube, suspended in a broad strand of cytoplasm which contained an abundance of starch grains. While one or two of the tubes may follow the course between the nucellus and the integument, others may penetrate the nucellus immediately at the top, as shown in Fig. 4. The penetration of the tube is accompanied by a breaking down and a probable absorption of the cells of the nucellar tissue through which the tube forces its way. From a study of longi- tudinal and cross-sections in series it became quite evident that the course taken by the pollen-tubes may vary considerably. At a later stage cross- sections of the endosperm showed the tubes in various positions. Usually three or four tubes develop and become functional. In the majority of cases they are found situated between the female prothallium and the remaining tissue of the nucellus. Many were found partially surrounded by endosperm, while others were completely surrounded. In no case was I able to find any evidence of branching of the tube, although Shaw (1896) reports that 4 quite as often as not it branches.’ Just about the time the tube penetrates the wall of the nucellus, the generative nucleus, having increased to quite the size of the tube-nucleus, divides. As shown in Fig. 4, there are now three nuclei in the tube, one large one and the two smaller ones. The largest of these, situated nearer the tip of the tube, is no doubt the tube-nucleus, while the other two are the stalk- and body-nuclei. Of these latter two there is a slight difference in size. As the larger one appears to be preparing for further activity, I regard it as the body-nucleus and the smaller one as the stalk- 7 Embryo of Sequoia sempervirens. nucleus. During the further development of the pollen-tube, the three nuclei remain in close proximity to each other. During these changes the size of the tube-nucleus remained the same, while that of the stalk-nucleus increased slightly. The body-nucleus, however, increased to at least three or four times the size of the stalk-nucleus, as shown in Fig. 5. During its development the body-nucleus surrounds itself with a dense zone of granular cytoplasm. This zone increases until it is about half the diameter of the nucleus in thickness, when it becomes shut off from the rest of the cytoplasm in the tube by a distinct membrane. The tube now contains one large cell and two free nuclei (Fig. 6). Previous to the formation of the body-cell, the free nuclei lie close together suspended in the same strand of cytoplasm, which contains an abundance of starch. The most of the starch was present in the vicinity of the body-nucleus, and it later becomes confined within the cytoplasm of the body-cell. The stalk-nucleus remains close to the body-cell (Fig. 5), even up to the time the male cells are formed. The changes resulting in the organization of the body-cell showed considerable variation as to the time of their occurrence. In some cases the mature body-cell was found in material collected early in May, while others were found as late as the middle of June. This irregularity as to the time of the changes was also noticeable in all later changes in the development of the male prothallium, even in the matter of fertilization. Soon after the body-cell has been fully organized, its nucleus enlarges and prepares for division. By extreme good fortune the spindle of this division was found. As this is the division which results in the formation of the two male cells, it demands a careful examination. It was during the division of the body-cells in Cycas , Zamia , and Ginkgo (Hirase, 1898 ; Ikeno, 1896-8; Webber, 1897-1901) that the centrosome-like bodies known as blepharoplasts were discovered. It was thought probable that a vestige of such an organ might be found in Sequoia, , but an examination of the cytoplasm surrounding the spindle failed to reveal a trace of any body that might be interpreted as a blepharoplast. It must be remembered, however, that the blepharoplasts are only concerned with the development of the cilia, and as these latter structures have never been found in con- nexion with the male cells in Conifers, it is not surprising that the organs responsible for their formation should also be missing. Fig. 7 shows the spindle dividing the body-nucleus with the daughter-nuclei at the poles. The formation of the cell-plate that separates the nuclei and divides the body-cell into two was not actually observed. But an examination of Fig. 7 where the connective fibrils curve out laterally, and of Fig. 8 where the two daughter-cells are lying side by side, makes it obvious that the cell-plate is developed in the usual way. After the wall separating the male cells has been formed, the latter remain close together for some 8 Lawson . — The Gametophytes , Archegonia , Fertilization , time, and a section of the two together has the outline of an ellipse that has been cut in half. They are rounded on one side and flat on the other. They are of equal size, and, as we shall see later, are both functional. Just before the male cells separate from each other, the nucleus in each has increased to about twice its original size. The chromatin is in the form of a network which contains a large irregularly shaped nucleolus. At the time of fertilization the male cells become almost spherical, and are perfectly similar in regard to their size and structure. Arnoldi (1901) has reported that the male cells may become elongated, and he figures one that has a spirally twisted form. I was unable to find such conditions, and feel confident that they are abnormal or due to shrinkage by poor fixation. In all the cases I have examined the male cell was spherical, and, as we shall see later, its spherical form may persist for some time after its nucleus has been injected into the egg. As we shall point out later, under the head of fertilization, only the nucleus of the male cell enters the archegonium, the rest of the male cell remaining outside in the pollen-tube. The nucleus is first liberated and, with but a small film of protoplasm surrounding it, passes between the neck-cells of the archegonium and immediately fuses with the egg-nucleus. In all other Gymnosperms in which observations have been recorded, at least one male cell enters the archegonium, so that in this respect the spermatogenesis of Sequoia is unique. According to the following diagram, Coulter and Chamberlain have compared the spermatogenesis of the Cycads with that of Isoetes. Isoetes. Cycads . Sequoia sempervirens. 0 Generative cell @ Generative cell . . a It will be observed that the male gametophyte is complete with the organization of the male cells, and this is true for all other gymnosperms where spermatogenesis has been worked out. If, however, we construct a similar diagram for Sequoia , the additional step of the discharge of the male cell-nucleus suggests more strongly the spermatogenesis of the Pteridophytes than even that of the Cycads, although the male cells of the later are ciliated. Embryo of Sequoia semper virens. 9 The Female Gametophyte. Shaw (1896) has given an accurate account of the development of the macrosporangium and the integument, but has, however, given a very meagre description of the sporogenous cells and the macrospores. There may be as many as five or six macrospore mother-cells organized from the hypodermal cells of the sporangium. Many preparations were made of this stage in the development of the sporangium, and a special study was made of the mother-cells as soon as they became differentiated from the surrounding sterile cells. They first become recognized as mother-cells by their large deeply staining nuclei. In the beginning they are not much larger than the other hypodermal cells, but they very soon enlarge, and their cytoplasm becomes very dense and granular and stains very readily with orange G. They are further characterized by the absence of large vacuoles. They are situated just about the centre of the sporangium ; about five or six layers of cells lying between the uppermost of them and the epidermis at the apex. A careful study was made of the number of mother-cells formed, and there seems to be a slight variation in this respect. Six was the largest number found. In all the sporangia studied five or six was the prevailing number, but in no case were fewer than four found. Shaw (1895) has reported that these sporogenous cells divide twice, each cell giving rise to four spores. As there are nearly always five or six of these sporogenous cells developed, and if each one gave rise to four spores, this would result in the formation of twenty or twenty-four macrospores. We shall show later that no such large number of macrospores were formed, but that, on the contrary, ten or twelve were the prevailing numbers met with. Owing to Shaw’s statement, it was thought probable that some of the sporogenous cells had failed to divide and that others had divided twice. Accordingly a very vigorous search was made for the stages showing a second division. Although the spindles of the first division were found frequently, some with the chromosomes at the equator of the spindle (Fig. 10), and others with the daughter-nuclei formed, in no case was there any evidence of four spores having been formed from a single mother-cell. From these observations I feel tolerably certain that the five or six sporo- genous cells (Fig. 9) differentiated from the hypodermal cells of the sporangium are the macrospore mother-cells, and after dividing twice give rise to the ten or twelve macrospores ; one cell of the second division fails to develop. Soon after their organization, about the first of March, the mother-cells divide, each producing two macrospores. As this division of the mother- cell is the reduction-division, which marks the beginning of the gametophyte, the character and number of the chromosomes demanded considerable attention. As far as the writer is aware, the reduction of the chromosomes io Lawson. — The Gametophytes , Archegonia , Fertilization , in the division of the macrospore mother-cells has never been actually observed in any of the Conifers. It was hoped that, as the spindles were found, some light might be thrown upon this most important step in the life-history of Sequoia , but after the examination of the chromatin, all such hope was lost. The spindle of the reduction-division was found in several preparations, but the chromosomes were too many to allow of an accurate estimate of their number. To add to the difficulty, the chromosomes were in the form of large V-shaped structures. The arms of the V were very long, and when at the equator they extended almost to the poles of the spindle. It also invariably happened that several of the chromosomes overlapped their neighbours, making it almost impossible to observe the method of splitting and separation. It was interesting to note, however, that as the daughter-chromosomes approached the poles of the spindle they were very much smaller than the mother-chromosomes at the equator ; this condition, however, was also observed in the divisions during the for- mation of the prothallium and also in the young sporophyte. Although the actual reduction could not be observed in the division of the macrospore mother-cell, we shall point out later that the number of chromosomes in the prothallium and in the development of the archegonium is obviously half that in the embryo. Reduction-spindles are shown in Fig. io. Soon after the macrospores are formed, they surround themselves with distinct walls and almost immediately begin to germinate. The germination is first noticeable by a slight increase in the size of all of the spores. This increase in size is always accompanied by a division of the nucleus, so that each of the ten or twelve germinating spores or female prothallia has two free nuclei in its cytoplasm. Their further growth is apparently at the expense of the sterile tissue in which they lie imbedded. The cells in this tissue, in contact with the young growing prothallia, show every sign of disorganization ; the nuclei appearing as homogeneous deeply staining masses lying in the distorted and more or less fragmented cells. At this stage (Fig. n) there are about eight or ten layers of sterile cells between the uppermost macrospore and the apex of the sporangium. The majority of the young prothallia show no further development after the first nuclear division, but two or three of them continue to elongate, and their increase in length is always directed toward the chalaza. Such a condition is shown in Fig. ii. Here eight young prothallia are represented, five of which have grown but very little, while the remaining three are many times the size of the original spores. The smaller prothallia function no farther, and are probably absorbed by the growth of the larger ones. Of the two or three larger prothallia represented at this stage, and which continue their growth down through the nucellar tissue, one grows more rapidly than the others, and for convenience we will speak of it as the primary prothallium. The one or two remaining prothallia persist in their further development, Embryo of Sequoia sempervirens. 1 1 and since (as we shall point out later) they have a considerable influence on the ultimate form of the primary prothallium, we will designate them as the secondary prothallia. The growth of the primary and secondary prothallia is always accom- panied by a rapid division of their free nuclei. Stages were found showing two, four, eight, sixteen, and thirty-two free nuclei. It was too difficult to count the numbers greater than this in sections, but this was sufficient to indicate that there is a regular and successive division of all the free nuclei, at least in these early stages. By the time the prothallia have reached the two or four nuclei stage the sporangium has increased to over twice its length. As Shaw (1895) has pointed out, the growth resulting from the elongation of the nucellus is very much greater in the region between the spores and the chalaza. At the time when there are eight or sixteen nuclei in the prothallium, as shown in Fig. 12, there is a layer of cytoplasm lining the walls, and there may be several vacuoles separated by strands of cytoplasm which contain many starch-granules. The nuclei at this time are suspended in a broad strand of cytoplasm which runs the length of the young prothallium and are usually arranged in a row, as shown in Fig. 12. The further increase in the size of the primary prothallium is always accompanied by an increase in the size of the vacuoles as well as an increase in the number of free nuclei. Although the primary prothallium is at this time somewhat larger than the secondary, the two present much the same condition. From now on, however, the rate of their development is very different. The development of the primary prothallium proceeds very rapidly, and is followed very slowly by that of the secondary, which, as we shall see later, never produces true cellular prothallial tissue. The initial stages in the development of the prothallium agree in general with the conditions found in Taxus by Jager (1899) and Campbell (1902), and in Pinus by Coulter and Chamberlain (1901), with this difference, that more than one prothallium in Sequoia reaches an advanced stage of development The following events, however, which lead to the formation of the cellular prothallium differ in many interesting respects from any of the Conifers in which the endosperm-formation has been worked out. As Arnoldi (1899-1900) has pointed out, the events are strikingly similar to those found by Lotsy (1899) in the formation of the endosperm in Gnetum. As the vacuoles in the primary prothallium grow, they eventually fuse together to form an enormous single vacuole which forces the cytoplasm and the nuclei to the walls. As represented in Fig. 13, the prothallium now consists of a large central vacuole, surrounded by a comparatively thin layer of cytoplasm in which a large number of free nuclei are distributed along the wall which surrounds the whole structure. The amount of cytoplasm 12 Lawson. — The Gametophytes, Archegonia , Fertilization , now increases considerably, and free nuclear division proceeds at a rapid rate. The cytoplasm and nuclei, as they increase, accumulate in greater abundance at the lower end,, and as this accumulation goes on, the large vacuole becomes correspondingly smaller and is confined to the upper end of the prothallium. Up to this time the shape of the prothallium has been more or less irregular in outline, the upper region remaining more or less constricted as compared with the broad middle and lower portions, and there has been no trace whatever of any cell-plate formation between the free nuclei. A portion of the prothallium in this condition is shown in Fig. 14. In his recent work on the formation of the endosperm in Sequoia sempervirens , Arnoldi (1900) finds that as the parietal layer of cytoplasm thickens it accumulates in great abundance at the lower end of the prothal- lium. Free cell-formation now proceeds in the lower and upper regions, but in the region surrounding the vacuole, structures (‘ Alveolen ’) similar to those described by Mile. Sokolowa (1891) in the formation of the endosperm of other Conifers are organized. The ‘Alveolen’ ultimately give rise to ordinary cellular tissue. He also reports that the archegonia are developed only in the tissue derived from the c Alveolen.’ My own observations on the thickening of the parietal layer and the accumulation of the cytoplasm and nuclei in the lower and upper regions agree very closely with those of Arnoldi. I was, however, unable to observe the formation of ‘ Alveolen ’ as he describes. As the long narrow vacuole continues to decrease in size, the cells surrounding it become larger. And these are no doubt the structures Arnoldi interprets as ‘ Alveolen.’ As we shall point out later, the archegonial initials are not confined to this region. A careful study of this stage in the development of the prothallium has convinced me that the final division of all the free nuclei which immediately precedes cell-plate formation is nearly simultaneous. In a single section over two hundred and fifty mitotic figures were counted. Many of these spindles showed the chromosomes at the equator. Some of them, even at the lower and the upper regions, showed the daughter-nuclei formed with the connective fibrils between them (Fig. 15). The nuclei in the vicinity of the vacuole divided in the same way and at the same time as the other nuclei, and I therefore could not confirm Arnoldi’s observations on the formation of ‘ Alveolen ’ in this region. That this was the final division of the free nuclei was shown quite conclusively by the fact that cell-plate formation had already begun between some of the daughter- nuclei in various regions of the prothallium. With so many nuclei undergoing division just about the same time they presented every possible phase of mitosis. Some of the nuclei were enlarging and just preparing to divide, many showed the chromosomes at the equator, others showed the chromosomes on the way to the poles, others again 13 Embryo of Sequoia sempervirens. showed the daughter-nuclei well organized, with the kinoplasmic fibrils stretching between them. As near as could be estimated, the number of chromosomes appeared to be sixteen. These various stages of the dividing nuclei had an influence upon the rate at which the cell-plates were formed, and consequently the various regions of the prothallium showed all stages in the development of the plates. While the daughter-nuclei resulting from this division are being organized, the continuous fibrils of the spindle persist and increase in number ; the result is that each daughter-nucleus is surrounded by a system of kinoplasmic radiations. These radiating fibrils not only join the sister- nuclei but they connect with the fibrils radiating from the other neighbouring nuclei. The result is that certain regions of the prothallium show large numbers of nuclei all joined together by systems of radiating fibrils, as shown in Fig. 15. Each nucleus at this time is enveloped in a dense granular zone of cytoplasm, from which the system of kinoplasmic fibrils radiates. The radiations show all the characters of ordinary spindle- fibrils, and they apparently have the same origin, that is, they are differ- entiated out of the cytoplasm. The first indication of the plate appears in the form of small granules or thickenings on the fibrils. These thickenings occur about midway between the nuclei, and as they increase in size the fibrils become less numerous in the vicinity of the nuclei. This would suggest that the fibrils were transformed into plate-forming substance, in much the same way as Timberlake (1899) has indicated. As illustrated in Fig. 1 6, the thickenings occur on all the fibrils stretching between the nuclei, so that each nucleus becomes completely boxed in by the developing plates. Different regions of the prothallium showed various stages in the formation of the plates. In a single section we may have conditions represented in Fig. 15 and Fig. 16, or where the daughter-nuclei are just being organized. In the upper portion of the prothallium some of the cells may be very large and elongated. These are no doubt the structures which Arnoldi (1901) has described as ‘Alveolen.’ I find, however, that the archegonial initials may develop much below this region, and I therefore cannot endorse Arnoldi’s view that there is a distinct generative tissue in the prothallium of Sequoia sempervirens . It is interesting to note that Arnoldi has reported a very different method of endosperm-formation in Sequoia gigantea . Here the entire endosperm is formed in much the same way as in other Conifers described by Sokolowa (1891), that is, by means of ‘Alveolen.’ When such great difference exists between two species of the same genus, it tends to eliminate the endosperm-formation as a means of establishing relationships among the Coniferales. During the development of the primary prothallium as described 14 Lawson . — The Gametophytes , Arckegonia , Fertilization , and above, one or two of the secondary prothallia have continued their growth, but at a much slower rate. Previous to the formation of the cell-plates, it is difficult to distinguish between the primary and the secondary prothallia, but after free cell-formation has ceased in the primary prothallium, the form of the secondary prothallium becomes well defined. Its course is long and tortuous, and in many cases winding in and out through the tissue of the primary prothallium, as illustrated in Fig. 18. Its shape is greatly modified, as its growth is limited to the spaces left by the more rapidly growing primary prothallium. Instead of having a single large vacuole and a parietal layer of cytoplasm, several long narrow winding vacuoles are present, separated by strands of cytoplasm. In several cases the spindles were found, showing that nuclear-division is carried on in the same manner as in the primary prothallium. In no case, however, was there any trace of cell-plate formation observed, so that I am inclined to believe that the secondary prothallia never develop sufficiently to produce cellular pro- thallial tissue. A very interesting relationship between the primary and secondary prothallia was noticed, which no doubt explains the sluggish development of the latter. After the cell-plates were formed in the primary prothallium, the parietal protoplasm of the secondary prothallium clings very closely to the cells of the former. These cells grow out into the free protoplasm and evidently act as absorbing organs. This intimate relationship is shown in Fig. 39. Here the newly formed cells are seen projecting into the proto- plasm of the secondary prothallium in a dovetail fashion. The nuclei of these cells are very much larger than the others, and have the appearance of being engaged in very active metabolism. There is no doubt that these projecting cells absorb the protoplasmic substance of the secondary pro- thallium, and thus retard its development. The primary prothallium of Sequoia takes just about three months to mature. The macrospores are formed during the first week in March, and the first archegonial initials were observed in material collected June 8. At this latter date the nucellar tissue had been almost completely absorbed, little more than the epidermal layer of cells remaining. Within the integument we now have a most confusing complex of structures, for in addition to the primary cellular prothallium there are usually present one or two secondary prothallia in an advanced stage of development and three or four pollen-tubes. The Archegonia. Very soon after the nuclei of the endosperm have been shut off from each other by the cell-plates, and after true cellular prothallial tissue has been organized, certain cells in the upper half of the prothallium become differentiated into the archegonial initials. These cells are quite numerous, and occupy a position not at the periphery, but near the central axis of the *5 Embryo of Sequoia semper virens. prothallium. In cross-sections of this region, as shown in Fig. 26, they may be easily distinguished from the surrounding vegetative cells. When first differentiated, their distinguishing characters are their large size, highly granular cytoplasm, which stains readily with orange G., and each one has a large, centrally situated, deeply staining nucleus. When first distinguish- able, the archegonial initials are not much larger than the ordinary vegeta- tive cells by which they are surrounded. They rapidly increase in size, however, and are soon several times their original dimensions. As they do not all enlarge simultaneously, they present various shapes and sizes, and sections very frequently showed the matured archegonia and the young initials in the same plane. As the initial grows in size, the cytoplasm appears more conspicuously granular, and the nucleus becomes much larger as it prepares for the division which cuts off the primary neck-cell. This first mitosis of the archegonial initial takes place during the latter part of June. Sections of material collected June 25 showed an abundance of the various stages of the spindle. Fig. 20 shows one of them with the chromo- somes on the way to the poles. These various stages of mitosis afforded an excellent opportunity of confirming the observation made during the endo- sperm-formation, that the number of chromosomes in the gametophyte is half that of the sporophyte. A comparison of Fig. 20, which represents the last spindle but one in the life-history of the gametophyte, with Fig. 32, which represents the first mitosis in the history of the sporophyte, shows without much question the difference in the number of the chromosomes in the two generations. In addition to the archegonial initials numerous jacket-cells are also differentiated. These are distributed very irregularly among the arche- gonia. Some archegonia may be almost surrounded by the jacket-cells, while others may be completely devoid of them. In the early stages of differentiation it is impossible to distinguish between the jacket-cells and the archegonial initials. They are structurally identical in regard to their cytoplasm and nuclei. This fact, accompanied with their irregular distribu- tion over the prothallium, suggests very strongly that the jacket-cells are archegonial initials which have become sterile. As soon as the first mitosis is completed, the archegonial initial is divided into two by a distinct cell-wall. The two cells thus formed are the central cell and the primary neck-cell. The central cell now grows very rapidly, and, as shown in Fig. 21, comes to be many times the size of the primary neck-cell. It was noticeable that the elongation of the central cell always took place in the direction of the neck-cell, so that the latter was forced forward toward the periphery of the prothallium. As it is pushed forward by the elongation of the central cell, the neck-cell divides, so that before it reaches the periphery of the prothallium the archegonium has two distinct neck-cells. It may be mentioned that Shaw (1895) reported the 1 6 Lawson.— The Gametophytes , Archegonia , Fertilization , presence of but a single neck-cell in Sequoia sempervirens , but this error has recently been corrected by Arnoldi (1900-1), who has found two. It was noticeable in examining my preparation that the two neck-cells were not always seen in longitudinal sections, but in cross-sections they were very distinctly made out. The neck-cells are not only to be distinguished by their position and shape, but their contents differ from those of other prothallial cells. The cytoplasm is devoid of large vacuoles, and, being highly dense and granular, stains more readily with the orange G. than do the other cells. My observations agree with those of Arnoldi (1901), both in regard to distribution of the archegonia over the upper half of the prothallium and that two is the typical number of neck-cells. I have, however, found four distinct neck-cells in a considerable number of archegonia (Fig. 24). This larger number is, however, exceptional. A somewhat similar variation in the number of neck-cells has been reported by Murrill (1900) for Tsuga , and according to Coker (1901) the number of cells in the neck of the arche- gonium in Podocarpus may vary from two to twenty-five. It was interesting to observe the position that many of the archegonia take in relation to the pollen-tubes. Long before the archegonia are formed, the pollen-tubes have grown down and their courses are well established, so that their growth is not directed toward the archegonia. On the other hand, the growth of the archegonia is invariably directed towards one or the other of the pollen-tubes. It has been pointed out (p. 6) that the course of the pollen-tube varies considerably. Some may grow down alongside the prothallium, others are partially surrounded by prothallial tissue, and others completely surrounded. In every case there were found numbers of arche- gonia pointing towards one or the other of the tubes, and their necks in contact with the tube-wall (Fig. 27). For some time after the neck-cells have been organized, the central cell is completely filled with a very dense granular cytoplasm, in the centre of which is suspended a very large nucleus. The next step in the develop- ment of the egg-cell is important and interesting, because it bears directly on the question of the general occurrence of the ventral canal-cell in the Conifers. Arnoldi (1900), in his careful investigations, failed to find any vestige of such a cell in Sequoia. He also denies its existence in Crypto - meria> Cunninghamia , and Taxodium . On the other hand, the ventral canal-cell or nucleus has been found by Strasburger (1879) and Belajeff (1893) in Juniperus ; by Blackman (1898), Chamberlain (1899), and Fer- guson (1901) in Pinus ; by Coker (1900-2) in Taxodium and Podocarpus ; by Murrill (1900) in Tsuga ; and by Land (1902) in Thuja. In Juniperus , Pinus , and Tsuga the spindle dividing the central nucleus into the egg and ventral nuclei has been described and figured, so that there can be little, doubt of its presence in these forms at least. Embryo of Sequoia sempervirens. 17 After a very careful search I was unable to find the spindle that gives rise to the ventral canal-cell nucleus in Sequoia , but enough evidence was found to convince me that such a nucleus is cut off from the central nucleus. At a stage soon after the neck-cells have been organized, several archegonia showed two distinct nuclei in the cytoplasm. As shown in Fig. 22, the nuclei are of the same shape and size, and are situated at opposite ends of the archegonium. The nucleus at the base of the archegonium becomes the egg-nucleus, and there is little doubt that the one nearer the neck-cells represents a vestige of the ventral canal-cell. In Pinus and Tsuga a distinct cell-plate is formed, which separates the ventral canal-cell from the egg-cell. In the majority of the Conifers investigated, however, the ventral canal-cell is only represented by a nucleus, no cell-plate being formed. This is evi- dently the case in Sequoia , for the ventral canal-cell nucleus functions no farther, and very soon becomes disorganized. In one or two archegonia it was found more or less flattened against the neck-cells, and in others it was more or less fragmented. That this is not the nucleus of the male cell was shown quite conclusively by the fact that the neck-cells were in no way disturbed. It apparently breaks down soon after it is cut off from the central nucleus. By the time the archegonium is ready for fertilization the ventral canal-cell nucleus has entirely disappeared. This very short period of its existence probably accounts for its having been overlooked by Shaw and Arnoldi. Soon after the disappearance of the ventral canal-cell nucleus, the egg- nucleus moves forward and occupies a position about halfway between the centre of the archegonium and the neck-cells. Meanwhile a large vacuole is developed in the lower part of the archegonium, as shown in Fig. 23. This figure shows a typical mature archegonium ready for fertilization. Fertilization. It has already been pointed out that the courses taken by the pollen- tubes have been well established long before the archegonia have been organized. It is therefore obvious that the direction taken by the tubes is not at all influenced by the female organ. On the other hand, however, it would seem that the pollen-tube had an influence upon the direction taken by the developing archegonia, for the growth of these latter structures is almost invariably directed towards the nearest tube. Fig. 27 shows how the archegonia appear in a cross-section of the prothallium. In this section there are four archegonia arranged in a semicircle, and with their neck-cells in contact with the wall of the pollen-tube. A longitudinal section would show ten to fifteen archegonia in a row, with the neck-cells directed toward the tube. The archegonia do not all point toward one tube, but each tube — and there are usually three or four present— has a number of archegonia directed toward it. Considering the diverse positions occupied by the c 1 8 Lawson . — The Gametophytes , Archegonia , Fertilization , dwa? tubes, this position is peculiar, as it suggests a case of the female organ seeking the male. Owing to the large number of archegonia present, and their peculiar arrangement around the pollen-tubes, fertilization is easily accomplished. When the archegonia are mature, that portion of the wall of the tube opposite the opening between the two neck-cells is the only structure inter- vening between the egg and the male cells. At this time the two male cells have become spherical and lie one behind the other in the tube, and take up a position near the wall which touches the neck of the archegonia. The nucleus of the male cell is very large, containing one or two large nucleoli, and a very regular network of chromatin. The latter is in the form of small spherical granules suspended on the threads of linin. The egg-nucleus is slightly different in this respect. The chromatin here is much more finely granular, and consequently does not stain as deeply as the male nucleus. The wall of the tube opposite the archegonium about to be fertilized apparently dissolves, for a small portion of the male cell now penetrates the archegonium by forcing the two neck-cells asunder, as shown in Fig. 29. Through the narrow communication thus established the male nucleus finds its way into the archegonium. As it squeezes through the narrow passage between the neck-cells, it becomes constricted and much elongated, as shown in Fig. 31, and immediately fuses with the egg-nucleus. It is a remarkable fact, as shown in Fig. 29, that only a very small amount of the cytoplasm of the male cell enters the archegonium. After the discharge of the nucleus into the archegonium, the greater part of the male cell remains outside and is discarded. As far as I am aware, such a condition has never before been reported for any of the Conifers, and it goes to establish the generally accepted view that the essential thing in fertilization is the fusion of the nuclei. As shown in Fig. 29, the denucleated male cell retains its spherical form, but has a vacuolated appearance in the central region once occupied by the nucleus. The section immediately following the one from which this figure was drawn showed the nucleus of this male cell in the act of fusing with the egg-cell in the archegonium. An examination of a large number of preparations has convinced me that this is the typical method of fertiliza- tion in Sequoia sempervirens. For some time after fertilization has been affected, the denucleated male cell is invariably found in the tube outside of the archegonia. Even as late as the early stages of development of the embryo, shrunken fragments of it were found. According to Coulter and Chamberlain (1901), nearly the whole of the contents of the tube in Pinus is injected into the cytoplasm of the egg. Goroschankin (1883) reports that both male cells pass into the archegonium, and Strasburger (1884) finds a similar occurrence in Picea vulgaris . In Pinus silvestris , Dixon (1894) and Blackman (1898) find that all the structures present in the pollen-tube pass into the egg. This has also been observed by Ferguson (1901) in 19 Embryo of Sequoia sempervirens . Pinus Str obits, in Taxodium by Coker (1900), and in Cephalotaxus by Arnoldi (1900). It therefore happens that at the time of fusion of the male and female nuclei in many Conifers the archegonium contains both male cells as well as the stalk- and tube-nuclei. In his recent work on Thuja , Land (1902) finds that the tube- and stalk-nuclei may sometimes enter the egg, but in the majority of cases they do not enter at all, but become dis- organized in the space above the archegonium complex. Compared with these other Conifers, Sequoia sempervirens stands unique, in that only the nucleus and a very small amount of cytoplasm of the male cell enters the egg. Only the two sex-nuclei are present in the egg at the time of fusion. As the male cell passes through the narrow canal of the neck, it immediately advances toward the egg-nucleus. As shown in Fig. 31, it has a long-drawn-out appearance. As it advances, the forward portion becomes rounded and very much wider than the long tapering hinder portion. While in this condition it first flattens against and then pushes in the membrane of the egg-nucleus. The long tapering end now draws in, and the male nucleus is more or less spherical. At the time of their fusion the male and female nuclei are of equal size, and in this respect differ from those in Tsuga and Pinus (Murrill, 1900; Blackman, 1898), where the male is much smaller than the female. Each cell has a large nucleolus, and the chromatin is in the form of small granules suspended in a regular network of linin threads. The chromatin in the female nucleus appears to be more finely granular than in the male, and consequently the male stained more deeply with safranin. In Pinus the conjugation of the sexual nuclei has been worked out with considerable detail by Blackman (1898), Chamberlain (1899), and Ferguson (1901). According to these writers, the membranes of the sexual nuclei remain intact for some time after the penetration of the male into the female. Blackman further finds that the first segmentation- spindle begins its formation before the male and female nuclei lose their identity. This has been partially confirmed by Chamberlain (1899), and Ferguson (1901), in Pinus , and by Woycicki (1899) in Larix , who were able to distinguish the male and the female chromatin as distinct groups inside the walls of the female nucleus. In Sequoia sempervirens the behaviour of the conjugating nuclei differ slightly from that described above for Pinus and Larix. As the male nucleus pushes into the female it becomes only partially surrounded by the membrane of the latter. This is no doubt due to the fact that the conjugating nuclei in Sequoia are of equal size. The membrane separating the two nuclei apparently breaks down much earlier than in Pinus. The chromatic contents of the two nuclei flow together, forming a common net- work in which the male and female elements can no longer be distinguished. The fusion-nucleus thus formed is very large and occupies a central 20 Lawson.— -The Gametophytes , Archegonia , Fertilization , and position, and in the few cases observed the chromatin was in the spireme condition. From the fact that the first cleavage-spindle was frequently met with, and the fusion-nucleus in but a few cases, I conclude that a very short time intervenes between the conjugation of the sexual nuclei and the formation of the first cleavage-spindle. Before passing on to the events that follow fertilization, we have yet to explain what becomes of the second male cell in Sequoia. Being func- tional, but unlike the other Conifers mentioned above, it does not fertilize the same archegonium. In Pinus (Blackman, 1898) only one male nucleus is functional. This is also true for Cephalotaxus (Arnoldi, 1900), and probably for other Conifers. In these forms both male cells enter the same female organ, so that one pollen-tube can accomplish the fertilization of but one archegonium. In Sequoia sempervirens this is clearly not the case, for the two male cells of a single pollen-tube were found in the process of fertilizing two separate archegonia. The behaviour of the second male cell is a duplicate of that described for the first male cell. They discharge their nuclei into two neighbouring archegonia just about the same time. As each male cell functions in separate archegonia, we have in Sequoia a case where each pollen-tube may bring about the fer- tilization of two archegonia. As there are usually three or four pollen- tubes surrounded by a larger number of archegonia, and since each male cell fertilizes a separate archegonium, there may be as many as six or eight embryos developed. We shall see below that this generally happens. The Embryo. As stated above, very little time elapses between the fusion of the male and female nuclei and the development of the first cleavage-spindle. This spindle was frequently met with, and always found with its poles in the long axis of the archegonium. In some cases observed the chromosomes were at the equator, or on the way to the poles ; others again showed the daughter-nuclei already organized with the continuous fibrils con- necting them. One of the most noticeable features of the spindle is the large number of chromosomes as compared with the number present in the gametophyte. As in the gametophyte, the chromosomes are very long V-shaped structures, and after they have split and moved toward the poles they appear to be about half the size of the mother-chromosomes when at the equator. As near as could be estimated, there are about thirty-two chromosomes in the sporophyte and sixteen in the gametophyte. Fig. 32 shows the first cleavage-spindle of the sporophyte with the large V-shaped chromosomes at the equator. The events following the first cleavage-spindle proved to be very interesting, because they differ so widely from those of other Conifers in which the early stages of the embryo have been investigated. In most Embryo of Sequoia sempervirens. 21 Conifers the fusion-nucleus gives rise to a number of free nuclei, which take up a position in the plane at the base of the oospore. These free nuclei, by dividing, give rise to three tiers of cells with complete walls. The uppermost of the three tiers remains in the oospore, the middle tier develops the suspensors, and the lowest tier the embryo. This is no doubt true for Pinns and many other Conifers, but in Sequoia the embryo de- velops in quite another way. Soon after the daughter-nuclei of the first cleavage have been organ- ized, a cell-plate is formed between them, and this results in the develop- ment of two distinct cells, each surrounded by its own cell-wall. The first cleavage, therefore, does not result in the formation of free nuclei. Fig. 33 shows the two daughter-nuclei of the first cleavage. It also shows the cell-plate developing on the kinoplasmic fibrils which extend between the nuclei. The two cells thus organized from the first division occupy almost the entire cavity of the oospore, and lie one behind the other. These two cells now divide in the same plane as the first division, and the pro-embryo now consists of a single row of four large cells, as shown in Fig. 34. An examination of this figure makes it apparent that the division of the first two cells of the pro-embryo is nearly or quite simul- taneous. In this connexion it is interesting to note that Arnoldi reports the presence of but three cells in the embryo at this time. He states that the uppermost cell of the first cleavage functions no further, but that the lower one divides and organizes two cells which become the suspensor and the embryo proper respectively. As shown in Fig. 34, there can be no doubt as to the division of both daughter-nuclei of the first cleavage. The next stage observed showed five cells in the oospore, and, judging from the position of them, the fifth one arises by a division of the lowest cell in the row of four. The five cells take up a position as shown in Fig- 35- The fifth or lowest cell in the row now enlarges and divides. Of the last two cells thus organized, the end one becomes the embryo proper and the other one develops into the suspensor. As soon as the embryo-cell and the suspensor-cell are organized the latter enlarges very rapidly and becomes very much elongated. As it increases in length it bends downward and carries the embryo-cell, at its apex, down through the endosperm. As the embryo forces its way through the prothallial tissue the cells of the latter show every sign of disorganization, and are no doubt absorbed by the developing embryo. Fig. 37 shows a suspensor with its large single nucleus and a two-celled embryo at its apex. Summary. For some time after the separation of the tetrads, the microspore contains a single nucleus. Before leaving the sporangium this nucleus divides, so that at the time of pollination there are two nuclei present. 22 Lawson. — The Gametophytes , Archegonia , Fertilization , and The larger of these nuclei is the tube-nucleus, and the smaller, which is situated near the wall of the spore, is the generative nucleus. No trace of nuclei or cells representing the vegetative tissue of the male gameto- phyte were found. No further division of the nuclei in the pollen-grain takes place until the pollen-tubes are partially developed. There are usually three or four pollen-tubes present, and they pursue different courses. One or two of them may grow down between the nucellus and the integument for a considerable distance, while others penetrate the nucellus immediately at the apex. Just about the time the tube penetrates the nucellus the generative nucleus divides, giving rise to the stalk- and body-nuclei. The body-nucleus becomes very large, and surrounds itself with a zone of dense cytoplasm, which in turn is surrounded by a membrane. The tube now contains one large cell and two free nuclei. During the latter part of June the body-cell divides and gives rise to two male cells. At first the male cells are flat on one side but soon become spherical. They are of equal size and both are functional. There are from four to six macrospore mother-cells organized from centrally situated hypodermal cells of the macrosporangium. Each mother-cell divides twice, but one cell fails to develop into a spore, so that there are from eight to twelve macrospores formed. The first is the reduction-division which marks the beginning of the gametophyte. It takes place about the first of March. Each macrospore begins to germinate. Some of the spores enlarge more rapidly than others, but the enlargement is always accompanied by a nuclear division. Although of various shapes and sizes, each macrospore now contains two free nuclei. The majority of the sacs or young prothallia show no further development, but two or three of them grow very rapidly and extend down through the tissue of the nucellus in the direction of the chalaza. Of the two or three sacs that continue their growth, one develops much more rapidly than the other one or two. In the early stages of the prothallium there is a progressive and simultaneous division of the free nuclei. These are at first distributed throughout the length of a long central strand of cytoplasm. By the development of large vacuoles which eventually fuse together, the cytoplasm and the free nuclei are forced to the wall. The prothallium at this time consists of a very large central vacuole, surrounded by a parietal layer of cytoplasm in which the numerous free nuclei are distributed. As the prothallium increases, the nuclei divide rapidly and frequently. The cytoplasm and nuclei now accumulate in great abundance in the lower, and less so in the upper portions, and the vacuole is reduced Embryo of Sequoia semper virens . 23 to a comparatively small narrow area in the upper half of the prothallium. The final division of all the free nuclei which immediately precedes the formation of the cellular endosperm is almost simultaneous. The spindle- fibrils which connect the daughter-nuclei in this division persist and increase in numbers, so that each nucleus is surrounded by a system of delicate radiating kinoplasmic fibrils. These fibrils not only connect with the sister-nuclei but reach out and join the fibrils of the neighbouring nuclei. The result is that large numbers of nuclei become joined together by radiating systems of kinoplasmic fibrils. The cell-plates are formed in the usual way between the nuclei. The formation of the endosperm in Sequoia sempervirens therefore does not follow the method described by Sokolowa (1891) for other Coniferales. As the vacuole in the upper region disappears, the cells in this region are very large and elongated, and these structures are no doubt the ‘ Alveolen ’ described by Arnoldi. The archegonia are, however, not con- fined to this tissue in the prothallium. During the development of the primary prothallium one or two secondary prothallia advance much more slowly. Their growth is confined to the narrow limits left by the primary prothallium. Their form is there- fore very irregular, and they never develop cellular tissue. The cells of the primary prothallium which are in contact with the protoplasm of the secondary prothallia act as absorbing organs. This absorption of the protoplasmic substance of the secondary prothallium by the cells of the pri- mary prothallium no doubt retards the development of the former and prevents it from forming tissue. Soon after the endosperm is formed, certain cells deep within the upper portion of the prothallium become differentiated into archegonial initials. When quite small the primary neck-cell is cut off, and the central cell enlarges rapidly and carries the primary neck-cell forward towards a pollen-tube. During the enlargement of the central cell the primary neck- cell divides once. Occasionally four neck-cells were formed in the arche- gonium, but two seemed to be the typical number. Just before the archegonium has reached its full size the central nucleus evidently divides, for we now have two large nuclei present. These nuclei are of equal size, and are situated at opposite ends of the archegonium. The one near the neck represents the ventral canal-cell, and the lower one is the egg-nucleus. The ventral canal- nucleus very soon becomes disorganized and disappears entirely. The lower portion of the archegonium develops a large vacuole, and at the time of fertilization the egg-nucleus is centrally situated. As the archegonia develop, their elongation is always directed towards a pollen-tube, so that each tube becomes partially surrounded by the necks of several archegonia. 24 Lawson. — The Gametophytes, Archegonia, Fertilization , and As soon as the archegonia are mature and ready for fertilization, the two male cells move toward the wall of the pollen-tube and take up positions immediately opposite the necks of two neighbouring archegonia. The wall of the male cell and of the tube in the region opposite the neck- cells evidently become dissolved, for the nucleus of the male cell, with a very small amount of cytoplasm surrounding it, squeezes through the narrow canal between the neck-cells and immediately advances toward the egg-nucleus. During its passage through the canal the male nucleus becomes very much constricted and elongated, but as it approaches the egg-nucleus it soon resumes its spherical form. The denucleated male cell remains outside of the archegonium and retains its spherical form for some time after fertilization has been effected, when it becomes disorgan- ized. The nuclei of both male cells are functional, but they fertilize two neighbouring archegonia. At the time of fusion the sex-nuclei are of equal size, and as the male pushes against the female it becomes partially surrounded by the mem- brane of the latter. The chromatin of both nuclei are in the spireme stage, and when the membrane between the two breaks down, a common chro- matin network is formed in which the male and female elements can no longer be distinguished. Very soon after the complete fusion of the sex-nuclei the first cleavage- spindle is developed. There are no free nuclei formed in the pro-embryo. The first cleavage results in the formation of two distinct cells, each surrounded by a complete cell-wall. Both of these cells divide, so that the embryo now consists of a row of four cells. The lowest of these en- larges and divides again. The two cells resulting from this last division give rise to the suspensor and embryo proper. Each fertilized archegonium gives rise to but a single embryo. As near as could be estimated, there are sixteen chromosomes in the gametophyte and thirty-two in the sporophyte. In conclusion, I wish to express my sincere thanks to my friend Capt. C. B. Hudson, who assisted me in collecting material. Note. Since the above went to press Dr. Coker’s valuable work (1903) on Taxodium has appeared. I regret that it is too late to make a detailed comparison of Sequoia and Taxodium as worked out by Dr. Coker, as the two forms differ so widely in every essential detail. They should certainly be placed in different families. Embryo of Sequoia sempervirens. 25 Literature Cited. Arnoldi, W. (’00) : Beitrage zur Morphologic und Entwicklungsgeschichte einiger Gymnospermen. I. Die Entwicklung des Endosperms bei Sequoia sempervirens. Bull. Soc. Imp. Nat. Moscou, 1900. Beitrage zur Morphologie und Entwicklungsgeschichte einiger Gymnospermen. II. Ueber die Corpuscula und Pollenschlauche bei Sequoia sempervirens. Bull. Soc. Imp. Nat. Moscou, 1900. (’01) : Beitrage zur Morphologie und Entwicklungsgeschichte einiger Gymnospermen. V. Weitere Untersuchungen iiber die Embryologie in der Familie der Sequoiaceen. Bull. Soc. Imp. Nat. Moscou, 1901. Belajeff, W. (’93) : Zur Lehre von dem Pollenschlauche der Gymnospermen. Ber. der Deutsch. Bot. Gesell. II, p. 196. Blackman, V. H. (’98) : On the Cytological Features of Fertilization and Related Phenomena in Pinus silvesh is (L.). Phil. Trans. Roy. Soc. 190, p. 395. Campbell, D. H. (’02) : A University Textbook of Botany. New York, 1902. Chamberlain, C. J. (’99) : Oogenesis in Pinus Laricio. Bot. Gaz. xxvii, p. 268. Coker, W. C. (’00): Observations on the Gametophyte and Embryo of Taxodium distichum. Johns Hopkins Univ. (’02) : Notes on the Gametophytes and Embryo of Podocarpus. Bot. Gaz. xxxiii, p. 89. (’03) : On the Gametophytes and Embryo of Taxodium. Bot. Gaz. xxxvi, pp. 1 and 1 14, 1903. Coulter, J. M., and Chamberlain, J. C. (’01): Morphology of Spermatophytes. Part I. New York, 1901. Dixon, H. N. (’94) : Fertilization in Pinus silvestris. Ann. Bot. viii, p. 21. Ferguson, M. C. (’01) : The Development of the Pollen-tube and the Division of the Generative Nucleus in certain Species of Pinus. Ann. Bot. xv, p. 193. — — The Development of the Egg and Fertilization in Pinus Strohus. Ann. Bot. xv, p. 435. Goroschankin, J. (’83) : Zur Kenntniss der Corpuscula bei den Gymnospermen. Bot. Zeit. xli, p. 825. -# HiRAsi, S. (’95) : Etudes sur la tecondation et l’embryogenie du Ginkgo biloba. Jour. Coll. Sci. Imp. Univ. Tokyo, viii, p. 307. Ikeno, S. (’96) : Das Spermatozoid von Cycas revoluta. Bot. Mag. Tokyo, x, p. 367. Jager, L. (’99) : Beitrage zur Kenntniss der Endospermbildung und zur Embryologie von Taxus baccata. Flora, lxxxvi, p. 241. Land, W. J. G. (’02) : A morphological Study of Thuja. Bot. Gaz. xxxvi, p. 249. Lotsy, J. P. (’99) : Contribution to the Life-history of the Genus Gnetum. Ann. Jard. Bot. Buitenzorg, II. i, p. 46. Murrill, W. A. (’00) : The Development of the Archegonium and Fertilization in the Hemlock Spruce ( Tsuga Canadensis') . Ann. Bot. xiv, p. 583. Peirce, G. J. (’01) : Studies on the Coast Redwood, Sequoia sempervirens . Cal. Acad. Sci., 3rd ser. Bot. ii. 3. Shaw, W. R. (’96) : Contribution to the Life-history of Sequoia sempervirens. Bot. Gaz. xxi, P- 332* Sokolowa, Mile C. (’90) : Naissance de l’endosperme dans le sac embryonnaire de quelques gymnospermes. Bull. Soc. Imp. Nat. Moscou, 1890 (1891), p. 446. Strasburger, E. (’72) : Die Coniferen und Gnetaceen, 1872. (’97) : Die Angiospermen und die Gymnospermen, 1879. — (’84) : Neue Untersuchungen iiber die Befruchtungsvorgange bei den Phanero- gamen als Grundlage fur eine Theorie der Zeugung. Jena, 1884. Timberlake, H. G. (’00) : The Development and Function of the Cell-plate in Plants. Bot. Gaz. xxx, p. 73. 26 Lawson. — The Gametophytes , Archegonia , Fertilization , Webber, H. J. (’97) : Peculiar Structure in the Pollen Tube of Zamia. Bot. Gaz. xxiii, p. 453. The Development of the Antherozoids of Zamia. Bot. Gaz. xxiv, p. 16. Notes on the Fecundation of Zamia and the Pollen Tube Apparatus of Ginkgo. Bot. Gaz. xxiv, p. 225. (’01) : Spermatogenesis and Fecundation of Zamia. U. S. Dept. Agri. Bur. of Plant Industry. Bull, ii, 1901. Woycicki, Z. (’99): Fertilization in the Coniferae. Warschau, 1899. Reviewed in Jour. Roy. Micro. Soc. p. 482, 1900. EXPLANATION OF FIGURES IN PLATES I -IV. Illustrating Dr. Lawson’s Paper on Sequoia sempervirens. All the figures refer to Sequoia sempervirens , and were drawn with the aid of the Camera lucida. For the finer cytological details, Zeiss’s homogeneous immersion obj. ^ apert. 1-25 with compensating ocular No. 6, and for the lower magnifications obj. and £ and ocular No. 4, were used. Fig. 1. A microspore as it appeared about two weeks before pollination, showing the single centrally situated nucleus. Material collected Dec. 15, 1901. x 900. Fig. 2. A microspore at the time of pollination, showing the large centrally situated tube-nucleus {t. n.) and the smaller laterally situated generative nucleus ( g . n.). Jan. 1, 1902. x 900. Fig. 3. A longitudinal section of the ovule, showing the relative height of the macrosporangium and the integument. The micropyle is nearly closed, and the pollen-tubes are shown growing down between the sporangium and the integument. Within the sporangium six macrospore mother-cells are represented. Material collected March 12, 1902. x 125. Fig. 4. A pollen-tube penetrating the nucellus near the top. The generative nucleus has divided, so that there are now three free nuclei in the tube. The large tube-nucleus ( t . n.) is in advance of the stalk- ( S.n .), and the body- ( B . n.) nuclei, x 750. Fig. 5. The lower part of the pollen-tube, showing the body-cell (p. c.) fully organized and the stalk-nucleus lying close beside it. Material collected June 12, 1902. x 750- Fig. 6. The tip of the pollen-tube containing the tube-nucleus ( i . n.), the stalk-nucleus ( S . n.)} and the large body-cell (b. c.) in the rear. Material collected June 15, 1902. x 750. Fig. 7. The body-cell undergoing division. The daughter-nuclei are organized and are connected by a series of kinoplasmic fibrils. The fibrils curve out toward the cell- wall in prepara- tion for cell-plate formation. Material collected June 29, 1902. x 750- Fig. 8. Two male cells as they appear immediately after the division of the body-cell. At this time they are flat on one side, and each contains a large centrally situated nucleus. Material collected June 29, 1902. x 750. Fig. 9. A longitudinal section through the central portion of the macrosporangium, showing six macrospore mother-cells. The cytoplasm of these cells is very dense and granular. The nuclei are comparatively large, and judging from the condition of the chromatin they are preparing for mitosis. Material collected March 2, 1902. x 800. Fig. 10. Section same as Fig. 9. The macrospore mother-cells are undergoing division. Two spindles of the reduction-division are shown. March 12, 1902. x 800. Fig. 11. A longitudinal section through the macrosporangium, showing eight germinating macrospores or young prothallia. Three of the prothallia are larger than the others, and each contains two nuclei. Their growth is directed towards the chalaza. April 25, 1902. x 175. Fig. 12. A longitudinal section, showing two young prothallia after repeated nuclear division. The free nuclei are distributed along a central strand of cytoplasm which extends from end to end of the prothallium. April 25, 1902. x 175. Fig. 13. A longitudinal section of a primary prothallium at a later stage, showing the very large central vacuole and parietal layer of cytoplasm in which the free nuclei are distributed. June 8, 1902. x 175. 27 Embryo of Sequoia semper virens. Fig. 14. A longitudinal section of the middle and lower portion of the prothallium, showing the accumulation of cytoplasm and free nuclei, and the corresponding decrease in the size of the vacuole. The free nuclei have increased enormously in numbers, and are distributed uniformly throughout the cytoplasm. June 8, 1902. x 175- Fig. 15. From a longitudinal section of the prothallium immediately after the final division of the free nuclei. Each nucleus is surrounded by a system of kinoplasmic radiations in such a way that a large number of nuclei are connected together by fibrils. This is the first step in preparation for cell-plate formation. June 12, 1902. x 1000. Fig. 16. A section of endosperm, showing a stage immediately following that shown in Fig. 1 5. The cell-plate has developed from the kinoplasmic fibrils stretching between the nuclei. Between some of the nuclei the cell-plate is already fully formed, while in others the granules are seen on the fibrils which eventually become the cell-plate. June 12, 1902. x 1000. Fig. 17. A longitudinal section of the upper portion of the prothallium, showing the distribution of the archegonia. The archegonia are shown in various stages of development. Archegonial initials, jacket-cells, and mature archegonia are shown in the same region. The growth of the older archegonia is directed toward the periphery of the prothallium. July 8, 1902. x about 150. Fig. 18. A longitudinal section, showing a secondary prothallium in an advanced stage of development. The primary prothallium has already developed cellular tissue, and the course of the secondary appears to wind in and out through this. June 29, 1902. x 150. Fig. 19. A cross-section, showing the close relationship between the primary and secondary prothallia. The cells of the primary prothallium in contact with the secondary extend outward and act as absorbing organs. These cells extend into the protoplasm of the secondary prothallium in a dovetail fashion, and their nuclei are very large and have the appearance of being engaged in very active metabolism. June 27, 1902. x 100. Fig. 20. An archegonial initial in process of division. The spindle shows the large V-shaped chromosomes on the way to the poles. The cells resulting from this division are the central cell and the primary neck-cell. In order to bring out more clearly the character of the chromosomes, the figure was drawn at a somewhat higher magnification than the other figure of the archegonium. June 25, 1902. x 850. Fig. 21. A young archegonium, showing the large central cell and the small primary neck-cell. June 25, 1902. x 750. Fig. 22. A later stage of the archegonium, showing two neck-cells. The archegonium contains two large nuclei of equal size. The nucleus nearer the neck is the ventral-canal nucleus, and the one at the opposite end is the egg-nucleus. June 25, 1902. X 750. Fig. 23. A typical mature archegonium ready for fertilization. There are two neck-cells present. The ventral-canal nucleus has disappeared, and the egg-nucleus is centrally situated. A large vacuole occupies the lower portion of the archegonium. June 29, 1902. x 750. Fig. 24. A mature archegonium, showing four distinct neck-cells. June 29, 1902. x 750. Fig. 25. A cross-section of a pollen-tube completely surrounded by female prothallial tissue. The tube contains the body-cell and the tube- and stalk-nuclei. June 25, 1902. x 150. Fig. 26. A cross-section of the upper portion of the female prothallium, showing the position of the archegonial initials and jacket-cells when they first become differentiated. June 25, 1902. x 100. Fig. 27. A cross-section of the upper portion of the female prothallium. The prothallial tissue partially surrounds a pollen- tube. Four mature archegonia are shown with their necks in contact with the wall of the tube. This position of the mature archegonia in relation to the pollen-tubes is very characteristic. June 27, 1902. x 100. Fig. 28. A cross-section of the female prothallium, showing sections of their pollen-tubes. As shown in this figure, the outline of the prothallium is always modified to the shape and position of the pollen-tubes. Two of the tubes show the male cell. June 25, 1902. x 100. Fig. 29. A denucleated male cell as it appears after the nucleus has been injected into the egg. It retains its spherical form, but the central region once occupied by the nucleus is distinctly vacuo- lated. The figure shows quite clearly that only a very small amount of cytoplasm from the male cell accompanies the nucleus through the narrow canal between the neck-cells into the egg. June 25, 1902. x 750. Fig. 30. An archegonium immediately after the entrance of the male nucleus. The male nucleus is shown pushing in the wall of the female. June 25, 1902. x 750. 28 Lawson . — Gametophytes of Sequoia sempervirens. Fig. 31. Another stage in the fusion of the male and female nuclei. The rear portion of the male nucleus has a long-drawn-out appearance, due to its passage through the narrow canal between the neck-cells. June 27, 1902. x 750. Fig. 32. The first cleavage-spindle with the large V-shaped chromosomes at the equator. June 25, 1902. x 750. Fig. 33. A later stage of the first cleavage-spindle with the daughter-nuclei organized at the poles, and the cell-plate developing from the continuous fibrils. June 25, 1902. x 750. Fig. 34. Each daughter-nucleus of the first cleavage has divided, so that the pro-embryo now consists of four cells placed one behind the other, and each surrounded by a distinct cell- wall. June 25, 1902. x 750. Fig. 35. A later stage of the embryo, showing five distinct cells. June 25, 1902. x 750. Fig. 36. An early stage in the development of the suspensor with the first cell of the embryo proper at its apex. June 25, 1902. x 750. Fig. 3 7. A later stage of the suspensor, showing its single large nucleus and two cells of the embryo proper as it is carried forward down through the endosperm. July 8, 1902. x 750. t yfnnods of Botany G.N. LAWSON, GAMETOPHYTES OF ...r. N. SEQUOIA. A.A.Iawson , clel. Voim^pu. University Press , Oxf ord LAWSON, GAMETOPHYTES OF SEQUOIA ^ylnnals of Botany G. N. rn.mil, pi. i University Press. Oxford zjlnndZs of Botany A. A. Lawson, del. LAWSON GAMETOPHYTES OF SEQUOIA. Vol. IVIII; PI II tjUnnaZs of Botany Vol. XVIII, PI II A. A. Lawson, del. LAWSON, GAMETOPHYTES OF SEQUOIA. University Press, Oxford c //TznaZs of J3otam/ LAWSON, G AM ETO PHYTES OF SEQUOIA Voi.mii pun MMm, ^••♦--•Vfc ‘,C2'VS- KV'* University Press, Oxford c yfruiaZs' of Botany Vol.XVJII, PI. Ill ■f'f m University Press, Oxford LAWSON GAMETOPHYTES OF SEQUOIA ^Annals of Botany Vol.XVIIL PI.IK The Nucleolus and Nuclear Division in the Root- Apex of Phaseolus. BY HAROLD WAGER. With Plate V THE great prominence of the nucleolus in nearly all cell-nuclei, its definite form, its avidity for aniline dyes, and its behaviour during nuclear division led, more than twenty years ago, to the conclusion of Strasburger, Flemming, Guignard, and others that it is in some way connected with the growth and increased stainability of the chromosomes during the division of the nucleus. This view is still maintained by many cytologists, although the micro- chemical researches of Zacharias, and the later investiga- tions of Strasburger are opposed to it. From a series of observations which I have made upon the changes which take place in the nucleolus during the process of nuclear division in the root-cells of Phaseolus , it appears to me that not only is the nucleolus concerned in the formation of the chromosomes, but that there is a definite morphological connexion between them. Briefly stated, it is found that the nucleolus is intimately connected with the nuclear reticulum ; that it contains nearly all the chromatin of the nucleus ; that this is transferred, previous to division, into the nuclear thread, which is then segmented into chromosomes ; and that, in the reconstitution of the daughter-nuclei, the chromosomes become fused into a number of more or less spherical or irregular masses which unite to form the daughter-nucleoli 1. Although there are numerous observations which indicate a close relationship between nucleoli and chromosomes, the existence of such a definite morphological connexion between them had only been previously observed in a few cases, of which Spirogyra among plants and Actino- sphaerium among animals are the most prominent examples. More recently a somewhat similar conclusion has been arrived at in the case of the root-cells of Vicia. 1 These results were stated in a Paper read before Section K at the Bradford meeting of the British Association in 1900. See Brit. Ass. Reports, p. 944. [Annals of Botany, Vol. XVIII. No. LXIX. January, 1904.] 30 Wager . — The Nucleolus and Nuclear Division Literature. In the following brief account of the literature, reference is only made to the more important papers in which the relations of the nucleolus to the chromatin and chromosomes are dealt with, and more especially to those papers published since 1897. An excellent summary of the whole of the literature on nucleoli, down to 1897, is given by Montgomery in his valuable memoir on the Morphology of the Nucleolus1. According to Flemming2 the nucleolus is a special organ of the nucleus or cell for the collection or elaboration of chromatin. It does not actually consist of chromatin (or nuclein), but of a substance in which the chromatin is elaborated, or, it may be, of a homogeneous substance which may be a chemical modification of chromatin or a preliminary stage in its formation. It possesses a definite form, but with no membrane around it, and it may be vacuolate with fluid contents in the vacuoles. Strasburger 3 regarded the nucleolus as an inert mass consisting of a reserve substance or substances, allied to chromatin, for the formation of the chromosomes. During division the nucleoli become dissolved in the nuclear sap and are then taken up into the nuclear thread, to reappear again in the chromatin-network of the daughter-nuclei. Pfitzner 4, Guignard5, Went6, and others give expression to similar views. Strasburger, in his more recent memoirs 7, adopts a different view from that just stated. He now believes that the chromatin is contained in the nuclear sap, and that one must look for its origin in the cytoplasm and not in the nucleolus. The nucleolar substance serves for the building up of the spindle. The evidence for this appears to be the complete or partial disappearance of the nucleolus immediately preceding the formation of the spindle ; the active and quantitative condition of the kinoplasm rises or sinks as the nucleolus dissolves or reappears, and the solution of the nucleolus is followed by the highest point of spindle-development. He brings forward the observations of Nemec, who states that the kinoplasm is directly transformed into nucleoli, and that granules which show the peculiarities of nucleoli arise at the poles of the division-figure through transformation of the spindle-threads. Their appearance is especially due 1 Jour, of Morph, xv, 1899. 2 Zellsubstanz, Kern- nnd Zelltheilung. Leipzig, 1882. 3 Ueber den Theilungsvorgang der Zellkeme und das Verhaltniss der Kerntheilung zur Zell- theilung, Arch. f. mikr. Anat., xxi, 1882. Die Controversen der indirekten Kerntheilung, Arch, f. mikr. Anat., xxiii, 1884. 4 Beitrage zur Lehre vom Bau des Zellkerns und seinen Theilungserscheinungen, Arch. f. mikr. Anat., xxii, 1883. *5 Nouvelles recherches, &c., Ann. Sci. Nat., Bot., Ser. 6, xx, i88"5. 6 Beobachtungen iiber Kern- und Zelltheilung, Ber. d. deutsch. Bot. Gesell., v, 1887. 7 Karyokinetische Probleme, Jahr. f. wiss. Bot., xxviii, 1895. Ueber Reduktionstheilung, Spindelbildung, Centrosomen und Cilienbildner im Pflanzenreich. Jena, 1900. 3i in the Root- Apex of Phaseolus . to a lower temperature, and Strasburger remarks that Hottes 1 has shown that a low temperature promotes the appearance of extranuclear nucleoli whilst a higher temperature promotes the formation of the spindle-figure. Zacharias 2, in a series of important papers in which the structure and micro-chemical reactions of the nucleolus are dealt with, contends that by the action of digestive fluid nucleoli are clearly differentiated chemically from chromosomes, and that this difference is also brought out by staining in methyl green in which the chromosomes stain more deeply than the nucleoli. Nucleoli may be vesicular and may exhibit a differentiation of structure into a homogeneous peripheral portion and a more refringent, granular or vesicular central portion. They appear to consist of plastin with albuminoids, but contain no chromatin. By observations on the living cell, especially the rhizoids of Chara , he finds that the nucleoli disappear just as the nucleus is about to divide, and that, just before this happens, they undergo amoeboid changes of form. Several nucleoli appear in the daughter-nuclei, and these afterwards fuse together into one. No definite conclusions as to what becomes of the nucleolus during division, or its function can be stated. In the nucleolus of Spirogyra , which many observers maintain is morphologically connected with the formation of the chromosomes, he finds no chromatin ; but he remarks in his most recent paper that could such a morphological connexion between nucleolus and chromosomes be established it would follow from the combination of morphological and chemical data that during the formation of the elements of the nuclear plate definite chemical changes would take place. Carnoy’s observations 3 led him to distinguish four kinds of nucleoli — (i) ‘nucleoles nucleiniens,’ (2) ‘ nucleoles-noyaux,’ (3) ‘nucleoles plasmatiques/ and (4) ‘ nucleoles mixtes,’ a combination of (1) and (3). He thinks that the plasmatic nucleoli may be concerned in the formation of the spindle. The ‘ nucleoles nucleiniens } are part of the chromatin-network and may consist of an amorphous mass of chromatin or of a chromatin-thread. Macfarlane, in his earlier contributions 4 and in his latest 5 paper, con- tends that the nucleolus is the most specialized portion of the cell and contains the main mass of the chromatin-substance. 1 Reduktionstheilung, Spindelbildung, Centrosomen und Cilienbildner im Pflanzenreich, p. 127. 2 Zacharias, E., Ueber den Zellkern, Bot. Zeit., xl, 1882. Ueber den Nucleolus, Bot. Zeit., xliii, 1885. Ueber das Verhalten des Zellkerns in wachsenden Zellen, Flora, lxxxi, 1895. Ueber einige mikrochemische Untersuchungsmethoden, Ber. d. deut. Bot. Gesell., xiv, 1896. Ueber Nachweis und Vorkommen von Nuclein, Ber. d. deut. Bot. Gesell., xvi, 1898. Ueber die achromatischen Bestandtheile des Zellkerns, Ber. d. deut. Bot. Gesell., xx, 1902. 3 La biologie cellulaire. Etude comparee de la cellule dans les deux regnes. Lierre, 1884. La cytodi^rese chez les Arthropodes. La Cellule, i, 1885. 4 Trans. Bot. Soc. Edinburgh, xiv, 1881. Trans. Royal Soc., Edinburgh, xxx, 1885, and xxxvii, 1892. 5 Current Problems in Plant Cytology. Contributions from the Botanical Laboratory, University of Pennsylvania, ii, 1901. 32 Wager.— The Nucleolus and Nuclear Division Mann 1 considers it probable that the nucleus and nucleolus are con- cerned in the assimilation of food-material, and that the nuclear chromatin may be less highly elaborated material than the nucleolar chromatin. He suggests that the nucleolus may be either ‘ an organ for the further trans- formation of substances already elaborated by the nucleus, or simply a storehouse for food-material, which has been already transformed by the nucleus into substances directly available for the nourishment of the achromatic elements of the cell.’ In Spirogyra , according to Tangl 2, Meunier 3, Moll 4, Decagny 5, Henneguy 6, Mitzkewitsch 7, and Van Wisselingh8, the nucleolus contains chromatin and the chromosomes are derived entirely or in part from it. Moll points out that only the nucleoli retain gentian violet with obstinacy ; from all other parts it is extracted without difficulty. The nucleoli are found in three different forms — (i) homogeneous, (2) vesicular, (3) exhibiting a skein structure. The last is the more common. Chromatic substance does not exist to an appreciable amount outside the nucleolus. The nuclear segments are formed by the transference of the chromatic substance from the nucleolus into the nuclear plasm. ‘ It seems as if the chromatic substance were squeezed from the nucleolus by an aperture 5 into the nuclear plasm in which it appears ‘ as small fragments, ranged in an intermediate, achromatic thread, like the beads of a necklace ; and thus a skein, containing chromatic substance, is formed.’ Mitzkewitsch states that in the process of division the nucleolus increases in size, its membrane disappears and it becomes irregular in shape. It then becomes differentiated into a number of deeply stained granular chromosomes, which form the nuclear plate, and a less deeply stained substance in contact with them. In the daughter-nuclei the chromosomes can be distinguished still surrounded by this less deeply stained substance, with which they eventually become incorporated to form the daughter-nucleoli. A somewhat different account is given by Van Wisselingh, who believes that two only of the chromosomes are derived from the nucleolus, the others being derived from the nuclear thread. In the reconstitution 1 The Embryo-sac of Myosurus minimus , L. A Cell Study, Trans, and Proc. Bot. Soc. Edinburgh, 1892. See also Brit. Ass. Reports, 1892, p. 753. 2 Ueber die Theilung der Kerne in Spirogyra-ZellQu. Sitz. d. k. Akad. d. Wiss. in Wien, lxxxv, 1882, p. 268. 3 Le nucl^ole des Spirogyra. La Cellule, 1888. 4 Observations on Karyokinesis in Spirogyra. Verhand. der k. Akad. van Wetenschappen te Amsterdam, 1893. 5 Bull. Soc. Bot. France, xlii, 1895, p. 319. 6 Lemons sur la Cellule, 1896. 7 Ueber die Kerntheilung bei Spirogyra. Flora, lxxxv, 1898, p. 81. 8 Ueber den Nucleolus von Spirogyra : ein Beitrag zur Kenntniss der Karyokinese. Bot. Zeit., lv, 1898, p. 195, and Flora, lxxxvii, 1900. in the Root- Apex of Phaseolus, 33 of the daughter-nuclei the halves of these two chromosomes again give rise to the new nucleoli. Whatever may be the exact method of division which takes place, it seems clear, as Strasburger says 1, that we have here a special behaviour of the nucleolus in giving rise to the chromosomes which, although opposed by Zacharias on micro-chemical grounds, appears to be morphologically decisive. According to Golenkin 2 we have a somewhat similar phenomenon in Sphaeroplea. The nucleolus breaks up into a number of fragments, which arrange themselves in a nuclear disk, and then appear to split up and move to the two poles where they fuse into daughter-nucleoli. All the chromosomes appear to originate from the nucleolus. He says that similar nuclei occur in other green Algae, including all Volvocineae, and in Musci. Rosen 3 states that in some plant-cells two kinds of nucleoli may be present, ‘ Eunucleoli ’ and ‘ Pseudonucleoli.’ The Pseudonucleoli form part of the chromatin-network, and are used up in the formation of the chromo- somes. The Eunucleoli are like the ordinary nucleoli, and do not disappear until a later stage in nuclear division. In the root-apex of Phaseolus there is only one kind of nucleolus present, which becomes lobular in the pro- phases of division, and does not entirely disappear in some cases until the separation of the daughter-groups of chromosomes is completed. Macallum 4 has shown that some nucleoli give an intense reaction for iron. This indicates that they contain chromatin. He finds in Erythronium at least three kinds of nucleoli: — (1) nucleoli which give a weak reaction for iron, and which therefore contain little or no chromatin ; ( 2 ) nucleoli rich in iron, and which give a reaction for iron in every respect like the nuclear thread ; (3) nucleoli found in the embryo-sac. The nucleoli of the embryo-sac appear in the filaments during the retrogressive stage as spherical elements containing little iron. As the chromatin in the fila- ment becomes reduced they give a stronger reaction for iron, and eventually are found to consist chiefly of chromatin, ‘and in stained preparations appear to contain nearly all the chromatin of the nucleus.’ ‘ When mitosis again commences the filament forms at their expense, the increase in size of the filament keeping pace, apparently, with the decrease in the quantity of chromatin which the nucleoli contain.’ V. Hacker considers 5 that nucleoli are not to be regarded as reserve 1 Loc. cit., 1900. 2 Bull. Soc. Imp. Nat. Moscou, 1899 C1!?00)* P« 343* See J. R. M. S., 1901, p. 65. 3 Beitrage zur Kenntniss der Pflanzenzellen, Cohn’s Beitr. z. Biol. d. Pflanzen, v, 1892, and vii, 1895. 4 On the Distribution of Assimilated Iron Compounds, other than Haemoglobin and Haema- tins, in Animal and Vegetable Cells, Q, J. M. S., N. S., xxxviii, 1895. 5 Praxis und Theorie der Zellen- und Befruchtungslehre. Jena, 1899. Die Vorstadien der Eireifung : zusammenfassende Untersuchungen iiber die Bildung der Vierergruppen und das Verhalten der Keimblaschen-Nucleolen. Arch. f. mikr. Anat., xlv, 1895, p. 200. D 34 Wager . — The Nucleolus and Nuclear Division substance, but as products of excretion resulting from the metabolic change taking place between the cytoplasm and nucleus, and especially a secretion of the chromatin, which is thrown out of the nucleus during division. Farmer1 states that in the Liverworts he examined 'the nucleolus was associated with the chromosomes in an unmistakable and remarkable manner.’ Threads of delicate texture run from the nucleolus to the linin. These are especially well seen in Fegatella. In a considerable number of cases the decrease of nucleolar substance is contemporaneous with the growth of the chromosomes. In Fossombronia , for example, 4 the nucleolus becomes angular, and, in extreme cases, almost star-shaped, whilst at the same time the linin begins to exhibit a very striking increase in the amount of chromatin which it contains.’ In a later stage the very much distorted nucleolus is often connected with several of the chromosomes, and soon after disappears. In the daughter-nuclei the nucleoli appear early, first as two or three small bodies which finally fuse to one large one, and the linin concomitantly loses its chromatin constituent. This does not necessarily imply that the chromatin passes, as such, into the nucleolus. But some constituents of the chromatin may find their way into this body, since both chromatin and the nucleolus readily yield albumen on appro- priate treatment. Miss Ethel Sargant’s observations on the formation of the sexual nuclei in Lilium Martagon 2 support the view that the nucleoli are con- cerned in chromosome-formation. The linin-thread during its growth appears to be fed from the partially dissolved nucleolus. Drops of nucleolar matter are found attached to the linin-thread in certain stages of its development. During the later stages of chromosome-formation the seg- ments of the thread become shorter and thicker, and ultimately ‘ the colouring (staining) of the chromosome segments becomes uniform. Each is apparently homogeneous. There is no contrast between cyanophilous granules and erythrophilous ribbon, but the whole chromosome stains uniformly like chromatin.’ Miss Sargant has very kindly allowed me to examine her preparations of the embryo-sac nuclei, and I find that, as she and Professor Farmer 3 point out, there is a definite relation between the linin and the nucleoli. Both in the resting stage and in the synapsis the nucleolus is connected to the linin- network by delicate threads, and in the later stages the chromosomes in process of formation are also often connected to it by threads. Carnoy and Lebrun4 describe a series of complicated changes in the nucleus of the egg of the Batrachia from which it appears that the 1 On Spore-Formation and Nuclear Division in the Hepaticae, Ann. Bot., ix, 1895, p. 469. 2 The Formation of the Sexual Nuclei in Lilium Martagon , Ann. Bot., x, 1896 and xi, 1897. 3 Loc. cit. (Hepaticae), p. 491. 4 La cytodi^rese de 1’oeuf. La vesicule germinative et les globules polaires chez les Batraciens. La Cellule, xii, 1897, and xiv, 1898. 35 in the Root- Apex of Phaseolus. chromosomes are definitely derived from the nucleoli. The young nucleus contains a much convoluted, apparently continuous thread. A portion of this becomes transformed into nucleoli (primary nucleoli). The remainder becomes resolved into granules, some of which dissolve, and the others are converted into secondary nucleoli. The nucleus now contains only nucleo- plasm and primary and secondary nucleoli. Both primary and secondary nucleoli become resolved into filaments which present very complicated figures. They are ephemeral, however, and again break up into granules, some of which contribute to the formation of new secondary nucleoli. These again produce new filamentous figures which are also ephemeral and from which comes a third generation of secondary nucleoli, and so on for a number of generations. All the filamentous figures found in the nucleus up to the time of the first polar kinesis have thus a nucleolar origin, and finally a portion of the products of this nucleolar resolution is dedicated to the formation of the nuclein elements (chromosomes) of the first maturation- spindle. Montgomery 1 comes to the following conclusions concerning the structure and function of the nucleolus. The ground-substance of the nucleolus is more or less dense, homo- geneous, or granular, and either fluid or viscid in consistency. In Spirogyra it has a true membrane. Vacuoles are normal structures. The alveolar structure of nucleoli described by various observers (e.g. Cavara) is probably referable to the regular distribution of equal-sized vacuoles in the nucleolus. Nucleolini, granules within the nucleolus, have frequently been observed, but no particular morphological significance can be attached to them. They appear to be only detached portions of the nucleolar substance, and may in some cases be only small vacuoles. Two kinds of nucleolus may be seen in some animal egg-cells, the nucleolus proper and the paranucleolus or Nebennucleolus. The para- nucleolus usually stains less deeply with nucleolar stains. It is probably not present in plant-cells. Some observers consider the paranucleoli to be derivatives of the nucleolus. Hacker regards them as secretions of the chromatin. Montgomery considers that they may represent those portions of the nucleolar substance which are deposited last, after the nucleus has undergone important physiological and chemical changes, the first portion produced being the nucleolus proper. In some cases a double nucleolus is found, the component parts of which may each represent a true nucleolus ; or such a double nucleolus may consist of a true nucleolus in apposition to a chromatin-nucleolus. The chromatin-nucleolus may be a metamorphosed chromosome (in Pentatomoi)) or, as in the larva of Carpo- capsa , it may originate from one of the granules of the nuclear reticulum 2. 1 Comparative cytological Studies, with especial regard to the Morphology of the Nucleolus,, Jour, of Morph., xv, 1899, p. 265. 2 Loc. cit., p. 519. D % 36 Wager. — The Nucleolus a?id Nuclear Division This seems to show that the chromatin-nucleolus always stands in genetic relation to the chromatin, whilst the nucleolus proper does not. The mode of transportation of the nucleolar substance to the daughter- nucleoli is probably different in different objects. A discharge of nucleolar substance into the cytoplasm appears to take place in some cells and may disappear, hence ‘the substance of the parent-nucleolus may be in many cases not transferred to the daughter-nuclei, but the latter (perhaps as a rule) may produce their own nucleoli de novol There is no substantial basis for Zimmermann’s conclusion * omnis nucleolus e nucleolo,’ and there is no evidence that the nucleoli are genetically related to the chromatin. The nucleolus ‘is derived in the first place from the cytoplasm,’ and consists ‘ of a substance, or different substances, taken into the nucleus from the cell-body.’ ‘ It seems probable that these substances stand in some relation to the nutritive processes of the nucleus, and in a relation to the growth of the latter.5 Thus growing nuclei are ‘ characterized by an especially large amount of nucleolar substance.5 From this one might conclude that the nucleolar ‘ substance stands in some connexion with the processes of nutrition5 and is either: (i) ‘nutritive in function,5 or (2) ‘represents that portion of substances assimilated by the nucleus from which all nourishment has been extracted, and in this case it would be a waste product,5 or (3) ‘ may represent accumulations of nutritive substance retained in the nucleus as a reserve supply ; but this does not seem to be very probable, for by this assumption it would be difficult to explain the uniformity in the size of the nucleoli in a given species of cell.’ Cavara1 has made an important contribution to this subject by his researches on the nucleoli of various plant-cells. I abstract briefly his chief conclusions. Nucleoli in a state of rest in. cells still capable of division or growth are composed of two substances, one, internal, which forms the major part, a homogeneous and specially refractive substance, only slightly stainable, which corresponds to the plastin of Zacharias or pyrenin of Schwarz ; the other, peripheral, of variable density, much more stainable than the plastin, with characteristics that connect it with chromatin or a modification of chromatin. These two substances are associated in various ways : sometimes the chromatin-like substance is uniformly distributed around the plastin, or more frequently is not so uniform, but presents breaks in its continuity which give to the nucleoli an alveolate structure or sometimes even a real reticulate structure. During the prophases of division the nucleoli decrease somewhat in volume and break up. At the same time the linin-thread contracts and breaks 1 Intorno ad alcune strutture nucleari. Estratto dagli Atti del R. Istituto Botanico dell’ Universita di Pavia, nuova serie, v, 1898. Breve contribtizione alia conoscenza del nucleole. Bollettino della Societa Botanica Italiana, 1902, p. 108. Bot.Cent., xxxix, 1902. 37 in the Root-Apex of Phaseolus . up into chromosomes. The nucleolar remnants which may remain in many cells, after the formation of the chromosomes, have no longer the characteristics of the nucleolus as presented before karyokinesis. They have become reduced in volume and in capacity for stains. This means that the chromatin-like substance has been subtracted from the nucleolus to build up the chromosomes, leaving the plastin which is employed in another manner, as in the formation of spindle-fibres and the new cell-wall. As the nucleoli are reconstructed during the anaphase, they become centres of attraction, not only for plastin, which is taken up from the nuclear sap, but also in part for chromatin, which is taken up from the nuclear thread. The differences which have been observed in the capacity of nucleoli for stains can be explained by the fact that they contain varying quantities of the two substances (see also Hertwig, loc. cit.). Without impugning the constancy of the characteristics of the nucleoli and chromosomes, ‘ non vi ha dubbio che gli uni e gli altri siano organi di una certa mutualita e transitorieta, e che come avviene dissoluzione di nucleoli o pirenolisi in seno al nucleo, altrettanto si possa dire di dissoluzione della cromatina, o di cromatolisi'. Chromatolysis ought not then to be regarded merely as an abnormal or pathological phenomenon, but it represents a normal condition, sine qua non , of the evolution of the cell. Cavaras conclusions are combated by Longo1, who says that from his observations the nucleolus is not formed of two distinct substances, and that the so-called plastin portion of the nucleolus represents only a vacuolization. Buscalioni % also concludes that there is no connexion between the formation of the chromosomes and the disappearance of the nucleoli. Bradley M. Davis 3 shows that in the tetraspore mother-cells of Corallina officinale the chromatin is scattered in a finely divided state in the nucleus, and that a linin-network is wanting. Each nucleus contains a single nucleolus which stains differently from the chromatin-granules. After division the chromosomes fuse together into a large deeply stained body on the surface of the prominent centrospheres. A nuclear membrane is formed and then the nucleolus reappears, at first smaller than the chromatin body, but afterwards becoming larger. The chromatin body then begins to fragment into a number of minute granules surrounding the single con- spicuous nucleolus, 1 Existe cromatolisi nei nuclei normali vegetali ? Rendiconti della R. Acc. dei Lincei, 7a ser., v, 1898. Contribuzione alia cromatolisi (picnosi) nei nuclei vegetali. Estratto dal vol. 90 degl’ Ann. del R. Istit. Bot. di Roma. 2 Osservazioni e ricerche sulla cellula vegetale. Ann. del R. Istit. Bot. di Roma, vii, 1898. See Bot. Cent., lxxix, 1899. 3 Kerntheilung in der Tetrasporenmutterzelle bei Corallina officinalis, L., var. mediterranea. Ber. d. deut. Bot. Gesell., xvi, 1898. 38 Wager. — The Nucleolus and Nuclear Division Macallum 1 shows that the reaction for phosphorus obtained in nuclei by means of his nitric-molybdate reagent indicates the presence of chromatin in the nucleoli. ‘ The eosinophilous nucleoli in animal and vegetable nuclei give a strong reaction for phosphorus, but less marked than in the case of chromatin. On the other hand, the nucleolar elements in the nucleus of the ovary of Erythronium which are rich in “ masked ” iron, give a deep reaction for phosphorus. A similar result was obtained in the nucleoli of the nuclei of the embryo-sac of the same form, in the peripheral nucleoli of the maturing ovarian ova of Menobranchus , in the nucleoli of Corallorhiza multiflora and of Spirogyra, all rich also in “masked’' iron.’ A very interesting case, on the animal side, in which the nucleolus takes part in the chromosome-formation is that of Actinosphaerium. Here, according to Hertwig 2, the nucleolus, at one stage in the life-history of the organism, consists of plastin only (plastin-nucleolus) ; at another stage both plastin and chromatin are collected into a single homogeneous body, the plastin-chromatin-nucleolus. There is, however, no sharp distinction be- tween them, the one passing gradually into the other so that the more chromatin the nucleolus contains the more it reacts like the chromatin- nucleolus, and vice versa. In mitosis the plastin-chromatin-nucleoli produce fine ramifying threads, which exhibit a granular appearance. The granules react like chromatin, whilst the ground-substance reacts as plastin. The equatorial plate is formed out of these threads. In the division of the primary cysts, which result in the formation of secondary cysts (gametes), the daughter-nuclei are reconstituted in such a manner that the chromatin remains distributed on the nuclear network, whilst the nucleolus is com- posed of plastin. In mitosis the chromosomes are formed out of the nuclear thread. As they become grouped to form the equatorial plate, the plastin-nucleoli become drawn out into fine granular filaments, which sur- round the chromosomes and become more or less incorporated with them. The spindle is formed out of the nuclear network and cytoplasm, but the connecting fibres between the two groups of daughter-chromosomes appear to be formed out of the plastin-substance of the nucleolus. The rudiments of the plastin-nucleoli appear as blisters in the chromatin masses, from which Hertwig suggests the possibility that both plastin and chromatin are modifications of, or represent the same element in, the nucleus. According to Nemec3, the nucleoli consist of a substance like plastin. The development of the spindle coincides with the disappearance of the 1 On the Detection and Localization of Phosphorus in Animal and Vegetable Tissues, Proc. R. Soc., lxiii, 1898. 2 Ueber Kerntheilung, Richtungskorperbildung und Befruchtung von Actinosphaerium Eichhorni . Abh. bayer. Akad. d. Wiss., 1898. Ueber die Bedeutung der Nucleolen, Sitzungsber. Bot. Ges. Munchen, xiv, 1898. See L’Ann^e Biologique, iv, 1898. 3 Cytologicka pozorovani na vegetacnich vrcholech rostlin : Vestnik krai. Cesk^ spolecnosti nauk Prag, xxiii, 1897 (see L’Annee Biologique, iv, 1898) ; Zur Physiologie der Kern- und Zelltheilung, Separatabdruck aus Bot. Cent., lxxvii, 1899. 39 in the Root- Apex of Phaseolus . nucleolus. The mantle-fibres of the spindle condense during the anaphase into a granular mass, which becomes the nucleolus and passes into the daughter-nucleus. The extranuclear nucleoli which are sometimes seen at the anaphase along the new cell-wall also originate from the connecting fibres. Nemec, however, in a later paper1, points out that, although his obser- vations indicate the connexion between nucleoli and spindle-fibres, other explanations of the phenomena are possible, and remarks that these obser- vations only prove certainly that there is some definite relation between the nucleolus and the division process, and that there is a definite connexion between the size of the nucleolus and the power of the nucleus to divide ; in this he agrees with Schwarz 2. The solution of the nucleolus and the formation of the spindle-figure might be two quite different processes, having this only in common that they go on side by side. Chamberlain3 finds a distinct connexion between nucleoli and chro- matin in the oosphere-nucleus of Pinus Laricio. The chromatin takes the form of nucleoli, which collect from all parts of the nucleus to a definite area near the centre, and there develop into a typical spireme. Duggar, in his studies on the pollen-grain and embryo-sac in Symplo- carpus , Peltandra, and Bignonia 4, has some very interesting observations on the relation of the nucleoli and the chromatin-elements. The nucleolus takes the chromatin-stain mainly, the nuclear reticulum very little. During the formation of the chromosomes they remain attached to the nucleolus by minute linin-threads, and the nucleolus is often drawn out by these into a fusiform shape. In Bignonia , ‘ in the dispirem of the daughter-nuclei the chromosomes become very irregular in outline, and gradually diminish in size, while there is gradually formed a large nucleolar- like body with irregular outlines. This body takes the chromatic dyes as did the nucleolus generally before/ Speaking of the second division, the author says : ‘ Everything indicates that the nucleolus of the microspore- nucleus has thus resulted from the direct or indirect fusion of chromatic material used in division/ Mottier 5 concludes that in Dictyota c the behaviour of the nucleolus, during both the development of the nuclear figure and the construction of the daughter-nuclei, indicates that this body represents a substance which is utilized by the chromatin and not by the spindle-fibres/ 1 Neue cytologische Untersuchungen : Sonderabdruck aus Fimfstiick’s Beitragen zur wissenschaftl. Bot, iv, 1900. 2 Die morphologische und chemische Znsammensetzung des pflanzlichen Protoplasmas. Cohn’s Beitr. z. Biol. d. Pflanzen, v. 3 Oogenesis in Pinus Laricio , Bot. Gaz., xxvii, 1899. 4 On the Development of the Pollen-grain and the Embryo-sac in Bignonia mnusta, Bull Torrey Bot. Club, xxvi, 1899. Studies in the Development of the Pollen-grain in Symplocarpu foetidus and Pdtandra undulata , Bot. Gaz., xxix, 1900. 5 Nuclear and Cell Division in Dictyota dichotoma , Ann. Bot., xiv, 1900. 40 Wager . — The Nucleolus and Nuclear Division Wiegand 1 points out that the nuclei in Potamogeton are all very pecu- liar, differing from the ordinary type in having the chromatin mostly aggregated in a ball at the centre of the cavity, instead of being distributed on the linin-network. Andrews 2 shows that in the nuclei of Magnolia and Liriodendron the nucleolus is large, stains deeply, and has a conspicuous vacuole. The linin- network is chiefly connected with the nucleolus. The threads of the net- work are at first smooth and uniform in diameter. At a later stage granular masses of chromatin appear on it, which gradually increase in size and become the chromosomes. The nucleolus at about the same time disappears entirely. c It is probably utilized as food in the growth of the chromatin masses, for they stain much more readily at this time than at an earlier stage.’ The chromosomes therefore arise from the resting nucleus as irregular masses without a previous formation of the usual spire m, and their identity from the first to the second mitosis is not maintained. Blanche Gardner 3, on the root-cells of Vicia Faba, comes to the con- clusion that the chromosomes derive ‘ at least a large part of their material from the nucleolus.’ ‘ The nucleolus is related to the nuclear reticulum in such a way that the fibres penetrate its substance.’ Previous to the nuclear division the nucleolus divides into two. Then the thin, long, almost con- tinuous nuclear thread can be seen ‘ at one or at several points ’ to ‘ dip into the nucleolus.’ ‘ The nucleolus now begins to transfer its contents into the nuclear thread.’ The thread becomes thicker and stains just like the nucleolus, which gradually disappears. This spirem thread then divides transversely into the chromosomes. In the daughter-nuclei the chromo- somes ‘ aggregate to form a small, dense, blue-black coil.’ Out of this mass the nucleolus is formed. The chromosome-coil gives up its chromatin and gradually loses its dark colour. ‘ At first these chromosomes are full of the dark chromatin-granules ; these become fewer and fewer as the nucleolus becomes larger and more distinct V 3 Methods. The root-apices of varieties of the common French Bean, Phaseolus vulgaris , L., were used. Sections were cut both by hand and by the micro- tome. Various fixing-fluids were used, but Perenyi’s fixing-fluid was the most useful. Among the many staining methods tried, Heidenhain’s iron 1 Wiegand, Karl M., The Development of the Embryo-sac in some Monocotyledonous Plants, Bot. Gaz., xxx, 1900. 2 Karyokinesis in Magnolia and Liriodendron with special reference to the Behaviour of the Chromosomes, Bot. Cent., xi, 1901, Beih., p. 134. 3 Studies on Growth and Cell-division in the Root of Vicia Faba . Contributions from the Botanical Laboratory, University of Pennsylvania, ii, 1901. ' 4 Chamberlain, in a recent paper (Bot. Gaz., xxxvi, 1903, p. 28) concludes that in Pellia the nucleolus contributes material to the chromosomes and spindle. 41 in the Root- Apex of Phaseolus . haematoxylin gave the best results. Other stains, however, were used for special purposes. Very careful staining and washing-out is necessary in order to exhibit the finer details of nuclear structure, and a good illumination with high powers is required for the microscopic examination of the preparations. The apochromatic object-glass of Zeiss (2 mm. N. A. 1-40) and an oil im- mersion condenser, or Powell and Lealand’s dry apochromatic condenser (N. A. -98) were almost always employed ; but a Leitz ^ oil immersion lens was also found valuable. Daylight was often used for general work, but the light from an incandescent mantle passed through a bull’s-eye con- denser was used to determine the finer details of structure. I am much indebted to my friend Mr. Norman Walker, Assistant- Lecturer in the Yorkshire College, who was kind enough to make some of the preparations for me. The Resting Nucleus. The resting nucleus does not appear to differ materially in structure from the ordinary nuclei of plant-cells. It is limited towards the cytoplasm by a thin deeply stained layer — the nuclear membrane — which exhibits in thin sections a fine granular structure. Inside this is a finely meshed nuclear network, which forms a thin layer at the periphery of the nucleus in close contact with the nuclear membrane. On it are distributed a num- ber of small granules, which stain deeply in nuclear stains. Each nucleus contains one or more nucleoli, which are usually much more conspicuous than the nuclear reticulum, and stain more deeply. The number present varies according to the age of the cell. It is only in the young cells, however, that two or more occur ; as the cells come to maturity the nucleoli fuse together so that in all the older cells one nucleolus only is present. Each nucleolus is lodged in a cavity in the nuclear network, and is surrounded by a clear space. As in many other vegetable nuclei, it is suspended in this cavity by a number of delicate threads, radiating from it on all sides to the nuclear network (PI. V, Figs. 3-7). These threads are only visible in stained specimens. They appear to be continuous with the nuclear network, and to form a part of it. The relative size of the nucleolus and amount of nuclear network varies in different cells. In the actively growing and dividing cells of the meriste- matic region the nucleolus is the most prominent feature in the nucleus, the nuclear network forming only a very thin layer in close contact with, and scarcely distinguishable from, the nuclear membrane. In the cells of the root-cap the nucleolus is much smaller, but the nuclear network is relatively more abundant, and in the outermost older cells of the root-cap layer is in its turn the most prominent feature in the nucleus, the nucleoli being so small in some cases as to be scarcely visible. The same differences are 42 Wager. — The Nucleolus and Nuclear Division observable in other parts of the young root. It thus appears that the amount of nucleolar substance present in a nucleus is to some extent a measure of the nuclear activities. Structure of the Nucleolus. The nucleolus is spherical in shape or nearly so, and is usually placed in or near the centre of the nucleus. It does not lie free in the nucleus, but, as already pointed out, is in close connexion with the nuclear network, in which it is suspended by fibres which not merely surround it, as Mont- gomery states (loc. cit., p. 506), but actually penetrate its substance, and in many cases appear as if drawn out of it. The nucleolus in fact appears to form a part of the nuclear network. I have observed similar connecting fibres in the embryo-sac nuclei of Lilinm in Miss Sargant’s preparations, both in nuclei in the resting stage and in synapsis and later stages, in the apical cells of the stem of Elodea , in the root-cells of the Oak, and in many other cases. They have been observed by Farmer in the nuclei of Liver- worts and Lilium ; by Duggar (loc. cit.) in various plants ; by Rosenberg in the nuclei of the suspensor of Zostera marina , L., and by other observers in various nuclei — both of animal and plant cells. Zimmermann’s statement 1 that the nucleolus in no case appears to be in direct connexion with the chromatin-network is therefore not correct. In some cases the nucleolus appears to be homogeneous throughout, but in the majority of nuclei it possesses one or more vacuolar spaces filled with a substance which appears to differ from the rest of the nucleolus (Figs. 1-7, 37). This vacuolization always appears to take place as the nucleus comes to maturity. It is absent only in young nucleoli. I have not seen the alveolar structure described by Cavara (loc. cit.), but it is obvious that the presence of a large number of vacuoles in a nucleolus would produce some such appearance as he describes. In some of the larger nucleoli I have observed darker coloured granules (nucleolini) as described by Macfarlane (loc. cit.) and others, but I agree with Montgomery (loc. cit.) that no particular morphological significance can be attached to them at present. In the cases observed by me they seemed to form only the central portion of the vacuolar substance. The vacuolate structure of nucleoli appears to be very common, and was first of all noticed in vegetable-cells by Schleiden, who speaks of the nucleolus as ‘ a small, sharply defined body, which, judging from the shadow that it casts, appears to represent a thick ring or hollow globule V Nageli also observes that nucleoli may be homogeneous or ‘ hollow in 1 Die Morphologie und Physiologic des pflanzlichen Zellkernes. Jena, 1896, p. 40. 2 Contributions to Phytogenesis. Translated by Henry Smith in 1847 for the Sydenham Society, p. 234. See also Schleiden’s Principles of Botany, Eng. Ed., translated by E. Lankester, 1849, P- 32. in the Root- Apex of Phctseohts . 43 the centre,.’ and that ‘there may be one, two, three, or many cavities of various sizes and arrangement V Macfarlane 2 regarded the vacuoles as endonucleoli and ascribed to them an important part in the nuclear economy, and he still insists on their probable importance in living cells, and suggests that they play a ‘ special part in furnishing to the nucleolar substance some ferment or compound which may be utilized during the division period 3.’ Mann4 regards the endonucleolus as ‘the trophic centre for all the organs concerned in assimilation and dissimilation,’ and says that it plays an important part in the conjugation of cells. Cavara5 contends that the vacuoles indicate the separation of the nucleolar substance into two distinct parts, chromatin and plastin, and Zacharias, who is opposed to the chromatin-plastin theory of the nucleolus, nevertheless concludes that it appears to consist of two distinct substances, a more refringent vesicular substance surrounded by a homogeneous, less refractive substance 6. Chamberlain has described a case in which the central, less deeply stained portion appeared to be separable from the outer, more deeply stained peripheral part, for ‘ there seemed to be a crack in the nucleolus, and upon applying a gentle pressure the central portion came out of the shell7.’ In haematoxylin-stained specimens of Phaseolus it is very clearly seen that the outer layer becomes more deeply stained than the vacuolar sub- stance ; in some cases there is a very marked distinction between them, and in some nucleoli the vacuolar substance is in contact with the exterior through an opening in the outer layer on one side of the nucleolus (Fig. 5). In methyl green and fuchsin the nucleolus stains bluish-red, the nuclear network red, and in specimens from which the stain has been well washed out the outer layer of the nucleolus is coloured light blue, the vacuolar portion remaining colourless or nearly so. Similar results are obtained with methyl green and eosin. In gentian violet the outer layer is deeply stained violet and the vacuolar portion light blue. But in many of the larger nucleoli the gentian violet shows up a more complex structure. In such cases we find a deeply stained thin outer layer surrounding a less deeply stained inner layer, and in the centre one or more deeply stained masses which are irregular in shape, with coarse radiations into the lighter stained part, or even exhibit a structure akin to a coarse network, which recalls the description given by Carnoy of his nucle ole s-noyaux. The existence of a chromatin-like substance in the nucleolus is indicated by the fact that in all these stains the chromosomes stain like the nucleolus, except that there may be a difference in the intensity of the 1 Memoir on the Nuclei, Formation, and Growth of Vegetable Cells, translated from Schleiden und Nageli’s Zeit* f. wiss. Bot., 1844, by W. Henfrey, p. 239. Ray Society, 1845. 2 Loc. cit, 1882-5. 3 Loc. cit., 1901, p. 194. 4 Loc. cit., 1892, p. 396. 5 Loc. cit. 6 Loc. cit., 1885. 7 Loc. cit., Bot. Gaz., 1899. 44 Wager.— The Nucleolus and Nuclear Division colouration. Thus in gentian violet and safranin the nucleolus stains red, the chromosomes deep red, whilst the cytoplasm and nuclear network are stained violet. In safranin and picro-nigrosin, well washed out, the larger part of the nucleolus stains deep red, but a thin peripheral layer stains blue, sometimes giving to the nucleolus the appearance of being surrounded by a membrane. The chromosomes stain light red or reddish-blue, the spindle deep blue, and the cytoplasm and nuclear network blue. In gentian violet the chromosomes are usually more deeply stained than the nucleoli of the resting nucleus ; but this appears to be due partly to the fact that the stain is more easily washed out of the nucleoli. The chromatin-like substance thus appears to occur mainly in the outer more deeply stained portion of the nucleolus. This is further sub- stantiated by the following observations. In sections cut by hand from fresh root-apices and stained in dilute acetic acid solution of methyl green the protoplasm and nucleus both stain bluish-green, but the nucleolus is the most deeply stained, especially in the actively growing cells, and the outer layers of the vacuolate nucleoli are the most deeply stained. If fresh sections are placed in pepsin solu- tion for from three to six hours, at a temperature of 36° C., a contraction and distinct diminution of the cell contents are shown. The nucleolus becomes much smaller and presents a bright glistening appearance. On staining in the methyl green solution, the protoplasmic remnant stains green, while the nucleolar remnant stains bright green, and in some cases a few granules, probably chromatin-granules in the nuclear network, take a similar bright green colour. The nuclear network stains like the cytoplasm. Again, if sections, either fresh or in spirit, be treated according to the method of Macallum, we get a strong reaction for phosphorus in the nucleoli and chromosomes, but very little in the nuclear reticulum. In vacuolar nucleoli this reaction is confined mainly to the outer layers ; we get very little in the vacuolar portion. And further, we get an intense reaction for phosphorus in the nucleolar remnant which is left after treat- ment with digestive fluid. In gentian violet-stained specimens the nucleolar remnant left in contact with the chromosomes at the time of the formation of the nuclear plate (Fig. 21) is often much less deeply stained than the chromosomes, and in some cases probably represents that inner portion of the resting nucleolus which does not become stained deeply in nuclear stains. It is impossible to definitely decide this point, however, by microscopic examination. The observations just described all point to the conclusion that there are at least two different substances in the nucleolus, one of which at any rate possesses the reaction of chromatin. It seems to me probable that, as Cavara suggests, the explanation of the different accounts which have been 1 Loc. cit., 1898. 45 in the Root- Apex of Phaseolus . given of the structure and staining and chemical reactions, may be due to the fact that nucleoli at different stages of their development, or from different plants, may vary as to the relative amounts of plastin and chro- matin substance which they contain. It seems almost certain that there are nucleoli which do not contain any chromatin or very little (the plas- masomes, true nucleoli, & c., of various observers), and others which contain a large quantity of chromatin (the so-called chromatin-nucleoli, nucleoles- noyaux , &c.), and between these two extremes we may have nucleoli with varying relative amounts of the two substances 1. Miss Ferguson mentions, in her paper on Pinus Strobus (Ann. Bot., 1901, p. 433), that ‘ the occurrence of unstained nucleoli in the same nucleus in which others were deeply coloured was common, especially at about the time of synapsis.’ Again, ‘ the attitude of this (egg) nucleolus towards dyes varies much at different periods of its history. It may or may not take the safranin of Flemming’s triple combination ; it may stain intensely with gentian violet or iron haematoxylin ; it may show a weak reaction or may be absolutely unaffected by them.’ It is not clear from the evidence available that the vacuolization of the nucleolus indicates a definite separation of the contents of the nucleolus into chromatin and plastin. It is more probable that the ground-substance of the nucleolus is plastin, and that the outer layer of it may become impregnated with chromatin or a modification of it. Changes in the Nucleolus during the Prophase. In the resting condition the nucleolus is suspended to the peripheral network by delicate threads, which are only visible in carefully stained specimens. As the nucleolus increases in size the suspending threads become more prominent, and it is then seen that they are intimately connected with the nucleolus, appearing as if drawn out of its substance, and on the other side are continuous with the peripheral network (Figs. 6 a and 7). In Fig. 6 a is seen a nucleolus at this stage which has become dis- placed from a nuclear cavity. The connexion of the threads with the nucleolus is clearly seen. In the root-apex of Allium Cepa Nemec2 points out that in younger cells the nucleoli are surrounded by a clear space, and are connected by achromatic fibres to the nuclear network. The hyaline appearance of this space may be simply due to the fact that chromatin- granules are absent from these threads. In somewhat older nuclei he points out that the radiating threads become more prominent, and chro- matin-granules are now seen lying thickly on the surface of the nucleolus, and appearing as if fused with its outer layer. His figure illustrative of this stage corresponds almost entirely with my Fig. 6, for which I offer 1 Cf. Montgomery, Woods Holl. Bio. Lectures, 1898. 2 Ueber die karyokinetische Kerntheilung in der Wurzelspitze von Allium Cepa , Pringsh. Jahrb. f. wiss. Bot., xxxiii. 46 Wager —The Nucleolus and Nuclear Division a different explanation, viz., that it indicates the beginning of the transfer- ence of nucleolar material to the nuclear thread 1. The peripheral network of the resting nucleus as seen from the surface is shown in Fig. 2. It consists of a lightly stained network with numerous small chromatin granules. It stains less deeply in nuclear stains than the nucleolus, and is not very conspicuous in the resting stage. At the time that the suspending threads become more prominent, certain portions of the peripheral network become also more prominent and more deeply stained (Fig. 9), a number of thicker threads being visible, connected to one another by finer filaments of the original net- work. The impression conveyed to an observer on looking through a large number of such stages in the nuclear development is that the substance of the nucleolus is being conveyed into the surrounding threads, and this is borne out by observation of later stages, where the threads radiating from the nucleolus become larger and more definite, and stain exactly like the nucleolus, which at the same time is becoming smaller and more irregular or amoeboid in shape (Fig. 8). It is interesting to note that Zacharias, in his observations on the division of the nucleus in living rhizoids of Char a , observed that in the process the nucleolus becomes irregular in shape, and undergoes amoeboid changes of form, and then disappears just before the formation of the chromosomes. The stage of this amoeboid condition of the nucleolus in Chara corresponds exactly with what is observed in stained specimens of Phaseolus. In Phaseolus the amoeboid condition of the nucleolus coincides exactly with the thickening of the nuclear threads connected with it, and an inspection of Figs. 14 to 16 indicates pretty clearly that as the nucleolus decreases in size the nuclear thread becomes more and more prominent, while still in definite continuity with the nucleolus, and staining reactions show that it becomes stained in a similar way. While these changes are taking place, kinoplasmic fibres appear at the poles of the nucleus, elongated in the direction of the future spindle. The nucleus contracts, and the nuclear thread tends to be drawn more closely around the nucleolus (Fig. 10). The nuclear wall breaks down gradually, and the fibres penetrate the cavity of the nucleus (Figs. 11 and 12). This may take place before the formation of the chromosomes, and the spindle-fibres may come into contact with the unsegmented nuclear thread (Figs. 15 and 16). The thread then becomes bent sharply at different points, the nucleolus still being in connexion with it (Fig. 16), and then breaks up into short rod- like chromosomes. Fig. 17 indicates a portion of a nucleus at this stage, in which can be seen the remnant of the nucleolus and some of the chromo- 1 The reference to the Figure in Nemec’s paper, p. 316, appears not to be correct, it seems to be Fig. 5, not Fig. 42, to which his description refers. 47 in the Root-Apex of Phaseohis . somes just on point of separation ; Figs. 13, 18, and 19 show various appearances of the nuclei at this stage, with an irregularly shaped indefinite nucleolar mass surrounded by chromosomes. The chromosomes become shorter and thicker, and arrange themselves to form the nuclear plate (Fig. 20). The remains of the nucleolus can still be seen, and in Fig. 20 there is still visible an indication of its connexion with the chromosomes in the faintly stained band between the drawn-out portion of the nucleolus-remnant and the chromosomes (cf. Fig. 16). It appears that the numerous connecting threads around the nucleolus gradually disappear, until only one is left, the nucleolus being drawn out at this point into a kind of tail (Figs. 16 and 20). It is extremely difficult, however, to be certain of the exact sequence of events, as the observations have to be made entirely on stained specimens. In many cases the nucleolus appears as if it was becoming directly trans- formed into chromosomes (see Figs. 12 and 13), but this I think is due to a contraction of the nuclear network around the nucleolus just at this time. Finally all connexion of the nucleolus with the chromosomes ceases, It is now much smaller in bulk, stains less intensely than before, and often exhibits a somewhat spongy texture, and at the same time begins to divide into two generally unequal portions (Fig. 21) which separate to opposite poles of the spindle-figure (Fig. 23). Rosen has already observed this phenomenon of nucleolar division in Phaseolus , and similar phenomena have been observed in some other plants. These nucleolar remnants are also surrounded by a clear space as in the resting nuclei, and in Fig. 22 is seen a case in which the single nucleolar remnant left at one of the poles of the spindle is surrounded by a clear space across which suspending fibres are seen. At a later stage these nucleolar remnants entirely disappear, and coincident with this the spindle becomes more prominent : the fibres increase in number and stain more deeply, so that one might easily con- clude that there was some connexion between the two, and that a portion of the nucleolus is concerned in the formation of the spindle. Reconstitution of the Daughter-nuclei. We have now to consider the changes which take place in the chromo- somes during the reconstitution of the daughter-nuclei. These have an important bearing upon the question of the relation of nucleoli to chromo- somes, for it is clear I think, from the observations about to be described, that the nucleoli in the daughter-nuclei definitely originate by the fusion of the chromosomes, first of all into a number of small nucleolar masses, connected together by a deeply stained network, and then by a further fusion into the large nucleoli found in the mature cells. 48 Wager, — The Nucleolus and Nuclear Division The equatorial plate first of all splits into two groups of daughter- chromosomes which separate to opposite poles of the spindle. As this takes place the chromosomes are distinctly seen as very short rods or granules (Figs. 25, 26). Whether they are actually separated from one another or remain more or less connected together by fine anastomosing threads I could not definitely determine. On arrival at the poles of the spindle they become aggregated together into a more or less homogeneous mass, in which the chromosomes can with difficulty be recognized (Fig. 27). At this stage the cell-plate appears across the equatorial region of the connecting fibres which are now very numerous. A nuclear membrane then appears, and the nucleus begins to open out or expand, exhibiting the chromosomes connected together by a deeply stained network (Fig. 28). The nucleus appears at this stage as if lodged in a cavity in the cytoplasm, the limiting layer of the nucleus or nuclear membrane apparently consisting of the peripheral layer of the nuclear network in close contact with the cyto- plasmic layer1. The cell-plate now gives place to the new cell-wall in the middle region of the figure, from which the connecting threads are fast disappearing, but at the periphery of the spindle the cell-plate forma- tion is still going on, as indicated by the prominent and deeply stained con- necting fibres. There is no indication whatever of any Concentration of the spindle-threads to form nucleoli, as Nemec states, although it is quite possible that a portion of them have been absorbed into the daughter- nuclei, and may enter into the constitution of the prominent network with which the chromosomes are connected. The chromosomes now begin to fuse together into somewhat irregular masses, or in some cases into a thick, irregular band or thread, still con- nected by a well-defined and deeply stained network of threads (Fig. 29). The connecting fibres in the middle region of the cell have disappeared, but those at the periphery of the nucleus are still very prominent. Already there are indications that the nucleolar masses are the centres of radiation for suspending fibres. This becomes very clear at a later stage, as shown in Fig. 31, in which the nuclei contain two and four nucleolar masses respectively. Around each large mass and one of the smaller ones clear spaces are to be seen. Fig. 30 shows a still further stage of chromosome fusion. In the upper nucleus there are now only two large nucleolar masses, and the way in which these have arisen by the fusion of smaller ones is shown in the lower of the two nuclei, in which four smaller nucleolar masses are to be seen fusing together in pairs. The nuclear network is still very prominent, and numerous granules of chromatin-like substance are visible on it. The 1 Gr6goire and Wygaerts, Beihefte, Bot. Cent., 1903, p. 13, and Lawson, Bot. Gaz., xxx, 1903, p. 305, also conclude, from their observations on various plants, that the nucleus is lodged in a cavity similar to a cell-vacuole. 49 in the Root- Apex of Phaseohis. cell-wall now extends nearly all across the cell. The formative fibres are completely separated from the nuclei, and appear free in the cytoplasm at the periphery of the cell, as shown in Figs. 30 and 31. Fig. 31 shows a later stage of chromosome fusion. There are in each nucleus one large nucleolar mass, which has a form indicative of its production by the fusion of two smaller ones, and one or more smaller ones. In the upper nucleus one of the smaller nucleolar masses appears to be on the point of fusing with the larger. In both nuclei the large mass and one of the smaller ones is surrounded by a clear space. In all the cases of division figured, it is interesting to note that in each the two daughter-nuclei exhibit much the same conditions as regards fusion of the chromosomes, and are just in the same stage of development (Figs. 28-37). Fig. 32 figures a case in which the transverse cell-wall is completely formed. In the daughter-nuclei are to be seen two large nucleolar masses in the same stage of fusion, and two smaller masses. The nuclear network is also clear, but with very few chromatin-granules. In the next stage the fusion of the nucleolar masses is carried a step further (Fig. 33). In the lower nucleus there is one large nucleolus and near it are two smaller ones, of which one is just beginning to fuse with the large mass. The upper nucleus contains a single irregular nucleolus which has obviously just arisen by fusion with it of at least two smaller ones. I think we may consider, that the two prominent projections on it indicate the result of fusion of smaller masses with it. It is not always the case that the result of the nucleolar fusion leaves the nucleus with one large and one or more smaller nucleolar masses. In Fig. 34 we have a case in which there are three about equal-sized masses of nucleolar substance which are just beginning to fuse together. In all these cases a delicate nuclear network is visible in contact with the fusing nucleolar masses, and in Fig. 35 is shown an interesting case of an irregular mass of nucleolar substance connected at one point in a very prominent manner with the linin-network (cf. Fig. 5). When the nucleolar masses have become completely fused, the result- ing nucleolus is at first a homogeneous irregular body of uneven outline (Figs. 33 and 35), but it gradually becomes more or less spherical (Fig. 36), and finally its homogeneity disappears, and it becomes vesicular (Fig. 37). As this proceeds the nuclear network becomes restricted to the peripheral part of the nucleus (Figs. 36 and 37), but the nucleolus remains connected to it, as already mentioned, by delicate connecting threads (Figs. 1-5). During the time these changes are taking place, each nucleus gradually increases in size with the increase in size of the cell ; the nucleolus also adds to its bulk, and goes on growing until, having reached a certain stage of development, the nucleus again begins to divide. E 50 Wager . — The Nucleolus and Nuclear Division General Considerations. It is not necessary here to enter into any detailed discussion as to the importance of the question of the function of the nucleolus. It will be sufficient to point out that the part ascribed to the nucleus, and especially the chromatin, as the bearer of the hereditary qualities in fertilization, renders necessary as precise a knowledge as possible of the chemical nature and function of each part of the nucleus before we can come to a definite conclusion that any one part or parts of it is more concerned in the process than another. The prominence of the chromosomes at certain stages during the division of the nucleus led to the enunciation of the doctrine that the hereditary qualities are transmitted by them ; that they retain their individuality, more or less, in the resting nucleus ; and that they must be regarded as individual units having an independent existence in the nucleus or cell. The observations described in this paper show that we have in Phaseolus a phenomenon which, if it is found to be a widely spread one, must modify our conception of the significance of the chromosomes and nucleolus in heredity. The nucleolus is intimately bound up with the formation of the chromosomes, and Strasburger’s contention that it is only concerned in spindle or kinoplasmic formation does not hold good, although it is not impossible that a portion of it — the plastin or pyrenin of Zacharias and Schwarz — may be used up in this way. There is no evidence either that the nucleolus originates from the spindle-fibres as stated by Nemec; and Hacker’s view that it is a product of excretion finds no support. A portion of it persists for a long time, even up to the stages of metaphase and anaphase, but it seems to disappear entirely within the region of the nuclear activity, that is in connexion with the chromosomes and spindle-fibres. It seems clear also that the nucleolus does not originate de novo either from nuclear substance or from the cytoplasm. There is a definite continuity of nucleolar substance from mother-nucleus to daughter- nucleus through the chromosomes. How far this supports Zimmermann’s conclusion, 4 omnis nucleolus e nucleolo/ is perhaps difficult to determine. But if the nucleolus is simply a part of the nucleus in which nutritive substances are stored and perhaps partly elaborated, and not an indepen- dent nuclear organ, then, while there may be a definite nucleolar continuity, it seems to me that Zimmermann’s conclusion simply becomes absorbed in the larger and more important — omnis nucleus e nucleo. It is almost impossible to avoid coming to the conclusion that the nucleolus must be regarded simply as a part of the nuclear reticulum, in which chromatin-substance is stored for the use of the chromosomes during 5i in the Root- Apex of Phaseolus . division or other active condition of the nucleus. But it is not easy to arrive at any definite conclusions as to their chemical relations, or as to the exact role of the nucleolus in the metabolism of the nucleus. The increase in size of the nucleolus in the resting stage probably takes place at the expense of materials brought into the nucleus from the cytoplasm. In the transformation of these materials into chromatin there are, it seems to me, three alternatives as to the way in which they may be dealt with : — (i) They may be taken up by the nucleolus directly from the cell-sap, and elaborated into chromatin, or (2) they may be first of all elaborated in the nuclear thread, and then passed on to be simply stored up in the nucleolus, or (3) they may be partly elaborated in the nuclear thread, and then passed on to be more completely elaborated in the nucleolus. Farmer’s suggestion (loc. cit, p. 512), ‘ that the nucleolus, though not in itself containing chromatin, is able to furnish at least one, and that prob- ably the albuminous constituent of this substance/ would not explain the presence of the phosphorus-containing substance, which both Macallum and I find in such abundance in many vegetable-nucleoli. Miss Huie 1 has made some interesting observations which bear upon this point and which seem to support the second, or possibly the third, alternative. In the gland-cells of Drosera during food-assimilation the nucleolus becomes smaller, whilst the chromosomes (nuclear network) become larger. The cytoplasm becomes impoverished and scanty in amount. Some time after feeding, the nucleus again becomes normal and the cytoplasm returns to its former condition. This indicates that the nucleus is the seat of metabolic activity, and that it is in the nuclear net- work, and not in the nucleolus, that the changes take place. But the diminution in size of the nucleolus, which is coincident with the increase in size of the nuclear network, seems to show that nucleolar substance is required before this metabolism can take place, or, in other words, that the activity of the nuclear thread is set up by the passage of nucleolar substance — chromatin or nuclein — into it, and that when the nucleus re- sumes its normal condition the nucleolus becomes restored to its original size by taking up the chromatin-material again from the nuclear thread 2. This is in harmony with Kossel’s conclusion that the formation of new organic matter is dependent on the nucleus, and that nuclein (chromatin) plays a leading role in this process 3. It appears to be always the case, that when cells are in an active metabolic condition, the nuclear thread becomes prominent, whilst the 1 Changes in the Cell-organs of Drosera rotundifolia , produced by feeding with Egg-albumen, Q. J. Mic. Soc., New Ser., xxxix, 1897, p. 387. Further Study of Cytological Changes produced in Drosera, Part II, Q. J. Mic. Soc., xlii, 1899, p. 203. 2 Cf. Farmer, Ann. Bot., ix, 1895, p. 495. 3 See Wilson, The Cell, &c., p. 340. 52 Wage r. — The Nucleolus and Nuclear Division nucleolus becomes reduced in size or disappears, as in the guard-cells of stomata, in the gland-cells of Chironomus , and in the male sexual cells of plants and animals. In all such cases the nucleolar substance probably migrates into the nuclear thread, and remains there so long as the nucleus is in an active state, to be stored up in the nucleolus again when the activity ceases. Farmer and Williams 1 and Strasburger 2 have shown, for example, that the sperm-nucleus of Fucus is very rich in chromatin, and at the time of fusion with the ovum-nucleus exhibits a network-structure but no nucleolus. After fusion, a second large nucleolus appears in the ovum-nucleus, apparently derived from the chromatin of the sperm-nucleus. Here we might suppose that the male nucleus, on its entry into the quiescent female nucleus, loses its intense activity, and that the chromatin-substance therefore becomes accumulated in the form of a nucleolus. So also Ikeda 3 has shown, in his observations on the nutritive function of the antipodals in Tricyrtis hirta , that during their stage of metabolic activity the nucleolus decreases in size, whilst the original scanty chromatin- network shows an extraordinary increase of chromatin, which becomes variously aggregated within the nucleus. It may be objected that chromatin-nucleoli, such as we have in Phaseolus , are only masses of chromatin-substance, and ought not to be confounded with those nucleoli which do not give a chromatin-reaction. There certainly appear to be two distinct types of nucleolus in some animal-cells, and it is possible that the same may exist in plant-cells also ; but the evidence before us as to their chemical constitution, and staining, and other reactions, is not at present sufficient to differentiate them into two distinct categories 4. We must rely mainly upon their morphological behaviour, and so far this has not produced much evidence of such a differentiation. As Montgomery suggests, it may be that all these bodies, which some observers are inclined to regard as fundamentally different structures, may be regarded as ‘ true nucleoli of a different chemical nature5.’ In conclusion, it appears to me, from a careful consideration of the facts as presented to us by various observers, that the following statements are probably justified as a summary of our present knowledge of plant- nucleoli : — (i) That the rounded bodies present in nearly all plant-nuclei, which 1 Fertilization of Fucus, Phil. Trans., 1898. 2 Kerntheilung und Befruchtung bei Fucus, Pringsh., Jahrb. f. wiss. Bot., xxx, 1897, p. 364. 3 Studies in the Physiological Functions of Antipodals and Related Phenomena of Fertilization in Liliaceae : (1) Tricyrtis hirta, Reprint from Bulletin, Coll, of Agriculture, Tokyo Imperial Univ., v, 1902, p. 41. 4 Cf. Fischer, Fixirung, Farbung und Bau des Protoplasmas, Jena, 1899, pp. 98-102, and Mann, Physiological Histology, Oxford, 1902. 5 Wood’s Holl. Bio. Lectures, 1898. in the Root- Apex of Phaseolus . 53 differ among themselves as regards various stains and reagents, are all to be regarded as ‘ nucleoli.’ (2) That these nucleoli may be composed of plastin (or plastin-like substance) only, or of plastin combined with chromatin in varying quantities, and that this variation in composition partly accounts for the varied accounts which are given of their staining reactions, &c., in various plant-cells (cf. Montgomery). (3) That in those cases where the chromatin-network is prominent and gives a strong reaction for chromatin, the nucleolus may either be absent or, if present, may give only a slight reaction for chromatin or none at all, and that where the chromatin-thread is not prominent the nucleolus is large and gives a strong reaction for chromatin. (4) That the nucleolus simply forms a part of the nuclear network, in which chromatin or chromatin-substance may be stored, and possibly to some extent elaborated, and that it is not therefore an independent organ of the nucleus. (5) That the nucleolus is concerned in the formation of the chromo- somes, and possibly also in the production of the spindle, and that a portion of it may in some cases be extruded into the cytoplasm, and there disappear. (6) That in the reconstruction of the daughter-nuclei the chromosomes unite together in a more or less irregular mass or thick thread, out of which is evolved the nucleolus and nuclear network, the major part of the chromatin passing ultimately into the nucleolus, except in cases where division again immediately takes place. (7) That the vacuolar structure of nucleoli is general, and may in- dicate, either the separation or partial separation of the nucleolar substance into plastin and chromatin, or a greater accumulation of chromatin in its peripheral layer. (8) That Zimmermann’s conclusion, ‘ omnis nucleolus e nucleolo,’ is not justified by the evidence before us, and that the nucleolus cannot, in the majority of cases at any rate, be regarded as an independent organ of the nucleus. (9) If these conclusions are correct, it is obvious that the conception of the part played by the chromosomes in heredity will have to be modified, and, as Dixon has already suggested 1, the nucleolus, as well as the chromo- somes, will have to be taken into account in any new hypothesis that may be put forward. 1 Ann. Bot. xiii, 1899. 54 Wager . — The Nucleolus and Nuclear Division EXPLANATION OF FIGURES IN PLATE V. Illustrating Mr. Harold Wager’s paper on Phaseolus. Fig. i. Resting nucleus, showing the large vacuolar nucleolus and delicate threads suspending it to the peripheral network. Fig. 2. Peripheral network seen from the surface of a resting nucleus. The linin-network with minute chromatin-granules is seen. Figs. 3 and 4. Resting nuclei, with nucleoli containing a single vacuole of irregular outline, less deeply stained than the peripheral portion. Fig. 5. A nucleus, showing the nucleolus with an interruption in the peripheral deeply stained portion at one point, giving it the appearance of a flask-like structure enclosing a less deeply stained substance. The suspending threads are more prominent and more numerous around this opening. Fig. 6. Nucleolus of irregular shape with prominent suspending threads. The substance of the nucleolus appears as if drawn out into these threads. Fig. 6 a. Nucleolus, which had fallen out of a nucleus, showing threads drawn out of the substance of the nucleolus. Fig. 7. Irregularly shaped nucleolus, showing its substance drawn out into the suspending threads. Fig. 8. Later stage than Fig. 7. The nuclear membrane not clearly visible. The nucleolar threads are prominent. Fig. 9. Surface-view of a nucleus at about the stage between Figs. 6 and 8. A portion of the peripheral network has become thicker and more prominent, with a delicate linin-network in the meshes of it. Compare with Fig. 2. Fig. 10. A nucleus, showing the kinoplasmic filaments at the poles. The nuclear membrane is still visible. Fig. 11. Nucleus, showing the nuclear membrane disappearing on one side, and the penetration of the kinoplasmic filaments into the nucleus. Fig. 12. Slightly later stage than Fig. 11. Fig. 13. The nuclear membrane has disappeared, and the bipolar spindle is now visible around the irregular mass of chromosomes and nucleolus. Figs. 14-20. Show the formation of the chromosomes. Fig. 14. A slightly later stage than Fig. 8. The connexion of the nucleolus with the nuclear network is well seen. Fig. 15. Later stage. The filaments of the spindle are now in close contact with the nuclear network, which has not yet begun to break up into chromosomes. Fig. 16. About the same stage as Fig. 15, but from a different point of view. Fig. 17. A section taken through a nucleus at a later stage than Fig. 16, showing the segmenta- tion of the nuclear thread into rod-like chromosomes. Fig. 18. Shows the irregularly lobed nucleolus surrounded by chromosomes. About the same or slightly later stage than Fig. 19, but a larger nucleus. The connexion of the nucleolus with the chromosomes not visible. Fig. 19. Section through a nucleus at about the same stage as Fig. 18, showing nucleolus and a few of the chromosomes. Fig. 20. The grouping of the chromosomes in the equatorial plane. Equatorial plate. The spongy remnant of the nucleolus in the midst of them, less deeply stained, can be seen, with apparently some connexion still to the chromosomes. Fig. 21. The less deeply stained remnant of the nucleolus shown dividing into two unequal parts, one of which will pass to each pole of the spindle. Compare Fig. 23. Fig. 22. A nucleolar remnant at one pole only of the spindle. This is surrounded as in a resting nucleus by a clearer space across which suspending fibres are seen. Fig. 23. Two unequal spherical nucleolar remnants approaching opposite poles of the spindle. Each is surrounded by a clear space, and stains much less deeply than the chromosomes. 55 in the Root- A pex of Phaseolus. Fig. 24. The nucleolar remnants have entirely disappeared ; as they disappear it may be noted that a deeper staining and increase in number of the spindle-fibres takes place. An indication of this is shown in the Figures. Compare Figs. 21 and 24. Fig. 25. Separation of the chromosomes. Anaphase. Fig. 26. Later stage of anaphase. Fig. 27. Fusion of chromosomes at ends of spindle to form the daughter-nuclei. Formation of cell-plate. Fig. 28. Daughter-nuclei with chromosomes connected together by linin-network. The new cell-wall is beginning to form in the centre of the connecting fibres, which are now more numerous at the periphery where the formation of the cell-wall is still in progress, and will continue until it reaches the lateral walls of the cell. Fig. 29. The chromosomes fuse together into larger masses, still connected by a linin-network. The central connecting fibre here entirely disappeared, leaving peripheral fibres only. Fig. 30. Later stage of the chromosome-fusion. One daughter-nucleus contains two large masses only (nucleoli), the other four nucleolar masses, which are on the point of fusing together to form two. The cell-wall now extends nearly across the cell. Fig. 31. Shows each daughter-nucleus at a later stage, with only one large nucleolar mass and one or more smaller ones. Each larger mass is surrounded by a clear space across which suspending fibres are visible as in a resting nucleus. In the nucleus in the upper part of the figure, two of the smaller nucleolar masses are in this clearer space, and one of them is on the point of fusing with the larger mass. Fig. 32. Shows cell-division completed. The two daughter-nuclei contain fusing chromatin- masses and a linin-network. Fig. 33. Later stage than Fig. 32. The fusion of chromatin-masses is nearly complete. In the upper daughter-nucleus there is a single irregular lobed mass. In the lower one a large spherical mass and two smaller ones, one of the latter is on the point of fusing with the large mass. Fig. 34* Nucleus, showing the last stage in the fusion of chromatin-masses to form the nucleolus. The three masses have just begun to fuse together. Fig- 35. The fusion is now complete, but the nucleolus has not yet rounded itself off. Fig. 36. A still later stage, showing the nucleoli as more or less spherical homogeneous bodies. Fig. 37. A still later stage, showing the vacuolization of the nucleoli. The daughter- nuclei are now in the resting stage. 1 ! 0 : . . : : ! '.J ; '{ . . :;J. " 1 .■ i,;I : 5 t • >■ i ■ ; ill • :■ . v • .Lf ■' , . ; 1 • ... i : ;VI : ■ > :• : . ' ' r : 7 . . ■> ! .'1 :f.;i i ■ : ; • : i , • r • : 11 Yt .•til ; ' ■> rr '!KI ;■■■■ o.--i l : yj v\ ;iv. ' Ai \. ■ "i) ,1! .r^rvil, .0.1 J - :o r i o : v;r -[' in r.'l ••• ,lo ..wl T vkco 'W+noX ) : ; • ■ ft ■ hl'-n eafiffr -(J:f :1 .7:1 ; ' ; !f;[ a flu visit1) '.Tj \:t j->}sd r ' : ■; f ' . 'y.rjtafl’) . \ .via ■'M : 1, •' 1 "v. ; yj, . ' _;1 : ;■[ : " ' ' I i ■; li ,;r:i iv'.O.V' 10 3 HO 1 - I 1 1 ! : i : i 1 ' :v , > r; ‘ • 11 ’• d iv-". ■■ ■ lf)\. 1: fn l .'1 T . , a •••••"/.<•{ Oiofhi ' •. : r;r) ov/t v:':’ ' i ■. uv: V.\-./ '|y , ^ ,-n|'*S. . ■ ■ ■ n flT. .o'vUirno^ V).73'- 7i f*:> Dn'i'f.'U? * V) {;f ■ ii ’ : ;■ f i ‘ : : ' ! . : > ' -;'V i ■' •-< ’ ."■! i i : ’ j;,.m ; ■ -ft ' ■'* I V.> 'ii- i - .'Vr; • ;r ).• ! ■.y.) ■y--[-y.yl-, ' ;:M .t.;; ,y"'i | : '•( i : : i: r i,; • 1 ;r .fo;;;. 1 ;t . ]... r; i ■> vz-t. •; . 1 iU i! i [i iTn .;{ *1;^ ; :'i/' A , >> -3r. 7‘7-r..f i .■ C : ; : ; >■> n i; 5 2.1; ,’i ! if;.* -i _ A fnnxiLs of Botany: WAGER - ROOT- APEX OF PHASEOLUS Voi.xvm,Pi.v: University Press, Oxf or d. rfnruxZs of Botany: Voi.xvm,pi.v: Wmjr v< University PrBss,OxfoTd. WAGER ROOT-APEX OF PHASEOLUS The Structure and Morphology of the ‘ Ovule/ An Historical Sketch. BY W . C. WORSDELL. With twenty-seven Figures in the Text. Introduction. HE great aspect of botanical science known to us as Morphology A comprises that department of study which is concerned with the form and the differential characters of the various structures or organs composing the individual plant. As this study advanced in the past it became clear that all the various parts or organs of the higher plants are capable of being grouped or classified into a few main categories, the raison d'etre for the existence of these latter being that the organs con- stituting each, although exhibiting an extensive range of variation in accordance with the equally varied environmental conditions to which they inevitably become subjected, yet possess certain well-defined, exclusive characters of form, structure, and position which have rendered them during the course of ages of progressive differentiation so stereotyped and fixed as to preclude the possibility of the existence of any intermediate or tran- sitional forms between any two of these categories. It is true that in some cases it is excessively difficult, perhaps even impossible, to determine the morphological category to which a given organ or structure pertains, and this owing to the extreme modification of the latter, resulting from certain special adaptive requirements, e.g. the submerged bladder-bearing portion of the Utricularia- plant and the seminiferous scale of the Abietineae. But this, it seems to the writer, is due solely to our ignorance and in- capability of tracing all the stages of adaptive modification which, during its long phylogenetic history, the organ in question has undergone, the original and all the intermediate forms having long since become extinct ; the ontogeny, or individual development, may yield us no clue ; for in the majority of these cases the highly modified organ arises congenitally as such. If we are unable to discover, say, in the submerged organ of U tricularia a prevalence either of the distinctive characters of the pbyllome [Annals of Botany, Vol. XVIII, No. LXIX. January, 1904.] 58 Worsdell. — The Structure and Morphology of the * Ovule! or of those of the caulome, we surely dare not conclude that this organ exhibits within itself a fusion of those two categories ! for if, in this particular case, such a fusion actually exists, we ought occasionally to find here and there in other plants, normally or abnormally, true transitional forms between, say, stem and leaf, or leaf and root, &c. If the existence of these could be demonstrated it would, in the writer’s opinion, prove our morphological categories to be mere figments of the imagination ; but he has no hesitation in saying that he believes such transitions never will be demonstrated. Until within comparatively recent years it was usual to regard the categories as four in number, and as consisting of the caulome or stem, the phyllome or leaf, the trichome or hair, and the root. A few writers, however, have advocated the introduction of a fifth category, viz., that of the sporangium ; a discussion of the validity of this view will be afforded at the proper time and place in this thesis. The ovule , which forms the subject of the present paper, does so on the ground of its being one of those highly complex (as may be assumed from the evolutionary point of view), much modified structures whose mor- phological nature has on that account remained for so long a doubtful quantity and the cause of deep debate and argument on the part of many able investigators. It will be the writer’s present object to endeavour to afford a presentation of the various views on the subject held by many of the leading botanists of the century, leaving his readers to judge for themselves which of those views contains the fullest measure of the truth. The facts thus collected and presented in a concise and accessible form will also, he hopes, prove useful to the student and the teacher of morphology. At the outset let us consider the few, simple facts connected with the structure and position of the ovule as they are familiar to us to-day. The ovule is well known to appear in very various positions on the plant ; in the majority of cases it appears as an outgrowth from the margin of the carpel ; in other cases as a development from the inner surface of that organ, as in Butomus and Nelumbium ; in older types of plant, the writer would submit, as terminal to a carpel of radial symmetry of structure, as in the fossil Bennettites ; in some instances, as in Caryophyllaceae and Primulaceae, it is apparently a product of an upgrowing axile placenta; in yet others there are indications of its actually terminating the floral axis, as in Compositae 1, Piperaceae, Najas , Polygonaceae, Taxus. The facts connected with the position of the ovule have had no small share in in- fluencing the decision of botanists as to its morphological nature. There have, indeed, been not a few who have placed their whole reliance on this set of data, constituting it the central pivot of their deductions. The ovule, whatever its position on the plant, consists, as a general rule, of 1 But the ovule here is really lateral. An Historical Sketch. 59 three distinct parts : a central structure, the nucellus , which assumes a terminal position on the ovular rudiment, and two integuments , which onto- genetically arise as annular outgrowths from the basal portion of that structure ; the order of development of these integuments is basipetal, the inner, as regards the great majority of cases, preceding the outer (Fig. i). To these facts of the normal structure a few exceptions occur : the integu- ment may be entirely absent, as in Crinum ; there may be only a single integument present, as in Ranunculaceae, Piperaceae, and apparently also in the group of the Gamopetalae 1 ; on the other hand, the number of integuments may be increased, for C. Schimper is said to have found three in Reseda lutea , the third integument arising in the acropetal direction, i. e. within the two normal ones ; such a case as this latter, however, requires confirmation. In the present thesis the writer is purposely omitting any reference to the nature of the ovule in parasitic plants, as he hopes to treat this subject separately on some future occasion. The facts of the normally constructed ovule being given, viz., a central papilla, the nucellus, enveloped by two basipetally-developed integuments, the problem which both past and present botanists have set themselves to solve is this: to which of the morphological categories does this im- portant structure belong? The views which have been held on this subject may be classified i Fig. i. Fig. 2. Fig. 3. Fig. 1. Diagrammatic representation of the ovule to illustrate the actual structure apart from its morphological interpretation. Fig. 2. Diagrammatic representation of the ovule to illustrate its interpretation according to the axial-theory (after £elakovsky). Fig. 3. Diagrammatic representation of the ovule to illustrate its interpretation according to the foliolar theory (after telakovsky). n, nucellus ; i. int ., inner integument ; 0 . int., outer integument. under three main headings, according as the ovule has been held to possess the morphological value of: — 1 . a shoot ; 2. a leaflet ; 3. a new structure or sporangium. 1 In this group of plants, at any rate, the * single ’ integument is probably due to congenital fusion of the two integuments. 60 WorsdelL— The Structure and Morphology of the 'Ovule! These views may be termed (i) the axial, (2) the foliolar , and (3) the sui generis theory respectively. The most striking of the arguments which, during a period of some seventy years, have been advanced in support of each of the above theories will now be presented. Historical Sketch. As regards the first of these views, the shoot-, bud-, or axial-theory — whichever term most appeals to the reader — appears to have been the one prevalently held about the middle of the last century (Fig. 2). Its most weighty supporter is undoubtedly the great German botanist Alexander Braun (15), who, nearly twenty years after Schleiden, in a paper dealing more particularly with certain phenomena in the life-history of Coelebogyne , embodies ideas which, at first appearing to support the foliolar-theory of the ovule, are eventually seen to practically dispose of this in favour of the axial view. ‘ Leaves/ he says, ‘ suppose an axis ; and as the idea that placentae are axile in nature as regards the majority of plants has been decisively refuted, the notion that ovules are entire leaves also falls to the ground.’ So that c the first explanation of ovules must be accepted, viz., that they are parts of a leaf, marginal structures, either peculiarly modified teeth, lobes , or pinnae of the carpel/ as Roeper and Brongniart stated, but with the recog- nition that the ovule must be regarded as something beyond a mere con- tinuation of the carpel. But this idea of the ovule as a marginal lobe, &c., of the carpel cannot be of universal application, as seen in the case of free central placentation and in cases where the ovules are distributed, multi- seriately or irregularly scattered on very thick or extended placentae. These cases show that ovules are not mere marginal structures, but out- growths from the surface, comparable to the normal or abnormal emergences of many leaves. He says that numerous observations show that the flowers, leaves, or leaf-segments arising on expanded structures in cases of antholysis do not represent an entire ovule, but only part of such. There is to be dis- tinguished a stalk (Trager), which must be regarded as a portion of the carpel, or in rare cases as an independent leaf, and sprouting from this a new structure — a bud. In metamorphosed ovules of Adonis autumnalis and Nigella Damascena he observed the outer integument expanded into a leaf- like structure. He remarks that in these cases there is no reason to regard it as part of the carpel, nor for aught else but an independent leaf belonging to the ovule (Eiknospe1). c If this is correct, the bud-nature of the entire ovule is assured.’ He is not decided as to which, the carpel or the ovule, the funicle belongs. The bud-nature of the ovule, the latter being 1 The terms ‘ Eiknospe,’ * Samenknospe ’ (‘ Egg-bud/ ‘ Seed-bud ’), clearly indicate the prevailing view as to the nature of the ovule held in Germany about the middle of the last century. The term ‘ Eichen ’ (‘ ovule’), as now generally adopted, is preferable. An Historical Sketch. 6r a structure capable of further development, is proved by the following phenomena : — (1) Multiplication of integuments, the latter arising in acropetal suc- cession (he thinks that it is not strange that the abnormal integuments should arise in the opposite direction to that of the normal ones, seeing that the regions from which the annular protuberances arise are already formed). (2) Unilateral expansion of one or both integuments. (3) Proliferation of the ovule into an elongated shoot. (4) Production of a leafy bud in place of the nucellus within the normal integuments. On one leaf of this bud he found an ovule, which he regards as a case of ‘ one ovule arising out of another.5 Having introduced Al. Braun first, as the most important of the sup- porters of this theory, let us now take up the views of the remaining authors in the historical order of their publication. We find that in 1840 Aug. de St. Hilaire (6), in his noteworthy ‘ Legons de Botanique 5 (p. 543), in support of the axial-theory of the ovule remarks: ‘ We can only regard the ovule as a miniature branch composed of an axis and appendicular organs. The placenta, as we already know, is a continuation of, and represents the stem, whose branches are the ovules.5 4 The primine and secundine are the appen- dicular organs of the young branch, and are comparable to the sheathing leaf- bases found in great numbers of Monocotyledons . . . ; it is therefore not sur- prising that the ovular axis, the least vigorous of all axial structures, should produce nothing but such sheaths.5 On p. 490 he further says : ‘ We know that every part of a plant bearing ovules — or, rather, every placenta — springs from the axial system, and is nothing more than a prolongation of this latter ; so that in cases of axile or parietal placenta it is clear that the axis of the flower, after producing the carpellary leaves, has, in order to produce ovules, to divide by means of a partition, for supplying branches, double or equal in number to that of the carpels, but which, in the latter case, are susceptible of forming double placentas.5 This same view, as to the universally axial nature of the placenta, is put forward by the author in his earlier memoir on the Resedaceae. As afford- ing a slight variant on the axial-theory, the coupled names of Endlicher and Unger (8) may be introduced, who held the ovule to be of the nature of a disc (‘ Nebenaxe ’) produced on the axial placenta. Schleiden (14), who, in 1839, as also in his subsequent work the ‘ Grund- ziige der wissenschaftlichen Botanik,5 published in 1843, promulgated, like St. Hilaire, the view that the placenta is always an axial structure, and that the ovules borne on it are of the nature of buds. Doubtless, this view of the matter, owing to the great weight of the epoch-making work containing it, would have had much influence both among the author's contemporaries and among many of those who came after him. 62 WorsdelL— The Structure and Morphotogy of the 'Ovule! Payer (23), in his great work on the organogeny of the flower, evi- dently— at least, in certain cases — regards the ovule, although he nowhere explicitly states it, as the morphological equivalent of a portion of the axis, this being the view definitely entertained by him with regard to all placental structures. In describing the development of the ovary in the Polygonaceae he says : ‘ At the bottom of this cup [the depression arising between the developing carpels] the apex of the axis is observed, which becomes suc- cessively clothed with two envelopes, thus constituting an erect and ortho- tropous ovule.’ He applies the same interpre- tation to the nature of the ovule in Urticaceae, Chenopodiaceae, Paronychieae, and Amaran- taceae. Magnus (44) offers a precisely similar ex- planation of the ovule, terminal to the axis, in Najas. These two latter authors may be regarded as typical examples of those who, in their endeavours to determine the morphological value of any given organ are primarily influenced by the position in which it arises on the plant. Hence we may speak of such as topical morphologists. The well-known observations made by Peyritsch on abnormal structures in the ovaries of Cruciferae and other plants will be referred to more fully hereafter. Penzig (57) examined proliferated ovules of Scrophularia vernalis L.f which he found, in the extreme metamorphosed condition, were toothed leaflets bearing a nucellus at different levels on their upper surfaces or apex. He also found nucelli borne at the apex of elongated structures growing on the placenta which were often fused with the leaflet on the dorsal side ; the appearance of the bundle in the nucellus showed the latter to be of ovular nature. He came to the conclusion that such phenomena are in greatest harmony with the bud-theory of the ovule. Many other authors have put forward this theory of the ovule, reference to whose works will be found in the subjoined bibliography. A discussion of the validity of the axial-theory will be deferred until Celakovsky’s views on the whole subject are brought forward. We have next to notice that which we have termed the sui generis theory of the ovule ; in this connexion there are the writings of four or five botanists to be taken into account. Schmitz (43), in his study of floral development in the Piperaceae, and especially of Peperomia repens , makes this general remark : that ‘ as the ovule arises sometimes as an emergence on a leaf, and sometimes by meta- morphosis of the vegetative apex, its relation to the shoot as a whole Fig. 4. Portion of flower of Polygonum , showing both ovary and ovule terminal to the axis (after Payer). An Histoi'ical Sketch. 63 cannot have the same morphological meaning. The ovule does not always possess the same morphological value.’ The idea of an ovule, he thinks, only includes a tissue which encloses the embryo-sac and nothing more ; and there need necessarily be no differentiation into integument and nucellus. ‘ The origin of the embryo-sac is not confined to the members of a definite morphological category.’ He considers that the ovule in Piperaceae represents a continuation of the axis of the flower 1, the integu- ments being its leaves, and equivalent to the carpels and stamens. In order to form the ovule the apical portion of the axis undergoes a meta- morphosis through the disappearance of all internal differentiation ; the integuments arise basipetally ; and on these grounds he regards the ovule as a new structure , for it cannot be made to fit into the morphological categories of caulome, phyllome, or trichome. From this author’s views on the nature of the ovule generally, it appears justifiable to include his name under our present heading. Sachs (51), in the second edition of his Textbook, enunciates a very similar view ; on p. 573 (Eng. ed.) he says : c I am induced to ascribe different morphological significations to the ovules, according to their mode of origin and their position.’ On p. 575, after citing the various theories held up to that time on the ovule, he says : ‘ Of these views, the one which appears to be most true to nature is the one which allows the greatest latitude ; but it is not always possible to refer an ovule to one of the categories, caulome and phyllome, for its position does not necessarily indicate its morphological significance. Thus a lateral ovule, as in Com- positae and Primulaceae, might be either a leaf or a bud ; its probable leaf-nature in these cases depends entirely on teratological evidence, which is of very doubtful value, for an organ in a monstrous condition does not necessarily assume its primitive archetypal form.’ The following sentence affords us the key to his whole position and characteristic view of the matter : f The difficulties met with in endeavouring to regard the ovule as a caulome or phyllome may be transcended by regarding it as an “ emergence” borne sometimes on an axial, sometimes on a foliar member.’ (This is practically equivalent to regarding it as an organ sui generis.) Strasburger (47), in his time, has held two distinct theories on the subject ; the earliest of these, embodied in his classical work, ‘ Die Coniferen und Gnetaceen,’ supported the position of Braun and Payer. This view arose naturally from his researches on the Gymnosperms, the ovule of which he regarded as a metamorphosed bud whose stem was the nucellus and the integuments the leaves. But in course of time his views changed, until in ‘Die Angiospermen und die Gymnospermen ’ we find him fully qualified for being classed among those who support the sui generis theory of the ovule, in regarding the latter as homologous with an independent 1 In this order the single ovule terminates the floral axis. 64 WorsdelL — The Structure and Morphology of the 'Ovule! structure like the sporangium of the Vascular Cryptogams, and with an emergence , which may arise promiscuously on either a foliar or an axial organ. It is to be noted that he founds his morphological views entirely on the phenomena of the individual development or ontogeny of the ovule. After stating that the funicle and sporangial-stalk, the nucellus and spore- capsule, are respectively parallel structures, he goes on to say that ‘the integuments of Angiosperms cannot be directly identified with the indusia of Ferns, for they arise not from the structure which bears the ovule, but from the ovule, thus the sporangium itself, and from the upper margin of the funicle.’ ‘ The nucellus is terminal to the funicle ; the integuments, on the contrary, are of lateral origin.’ After a careful investigation of certain ovular ‘ monstrosities ’ in Rumex scutatus and Helenium Hoopesii , A. Gray, he arrived at the following conclusions with regard to metamorphosed ovules generally : that metamorphoses (sports) are not retrogressive phenomena 1, ‘ but rather the expression of a competitive supplanting of one structure by another. In place of generative, vegetative rudiments arise, in accordance with the place of origin ; so that, e. g., pinnae are formed as divisions of the carpels and buds as prolongations of the floral axis.’ During the process of struggle between these two opposing tendencies, transitional forms occur, which will vary according as the one or the other tendency gains the upper hand. If the phenomena of oolysis were really retrogressive in character one would expect to see a cryptogamic sporangium appearing now and again, which, however, is never the case. In cases where an emergence appears on the surface of the leaflet, he regards the former as representing an entire ovule and not merely the nucellus ; nor is the leaflet to be considered as the extreme form of the integuments, for he does not believe in the existence, v as put forward by Celakovsky and others, of a series of transitional forms consisting at one extreme of almost normal integuments, and at the other of a vegetative leaflet. ‘ Each case must be considered in and for itself alone, and represents a compromise which has been arrived at between the struggle to form an ovule on the one hand and that to produce a leaflet on the other.’ And again he says : ‘ If only a mere emergence is present on the leaflet, I regard this as the consequence of the tendency to leaflet- formation having early gained the upper hand, and not as the result of a reversion of already formed integuments into the leaflet. The hybrid-cases observed cannot be regarded as constituting a series of developmental stages which must be traversed in order to arrive at the extreme forms.’ He agrees with Celakovsky that where buds occur on fully formed parts of the carpel these are usually to be regarded as adventitious. But from his own standpoint, he must adhere to the view that the ovule may be directly supplanted by a vegetative bud. 1 This has reference to Celakovsky’s earlier view on the subject, which will be introduced in its proper place. An Historical Sketch . 65 Finally, with reference to ‘ sports ’ generally, I may quote a sentence from his ‘ Lehrbuch der Botanik,’ p. 132, where he says: ‘ Sports (Miss- bildungen) can only in rare cases be made use of in drawing morphological conclusions.’ Goebel (56) is a staunch upholder of the sui generis theory of the ovule and a severe critic of the method of settling morphological problems by the aid of ‘ sports.’ In his ‘ Organographie ’ he recapitulates his views on the subject, put forward much earlier in Schenk’s ‘ Handbuch der Botanik’ and elsewhere. Speaking of the replacement of sporangia by vegetative organs, as in cases where stamens become foliaceous or petaloid, he says that ‘ a metamorphosis of the former into the latter does not occur, and the transitional stages between the normal and abnormal con- dition do not establish such a metamorphosis, but only show that there may be a varying degree of disturbance of the normal state of affairs.’ The morphological character of stamens is determined by means of the development and a comparison with vascular Cryptogams. In treating of proliferated ovules, he considers it unjustifiable to regard these as reversions, and absurd to represent the simple leaf into which the ovule has become metamorphosed as the most primitive phylogenetic stage of development. ‘ The only conclusion it would be possible to draw from the proliferations would be that the integuments are formed from carpellary substance or represent outgrowths of the carpel which are the better adapted to vegeta- tive development in proportion as the reproductive organs (the nucellus) are hindered in their growth.’ We are to regard these proliferated ovules as crippled structures which have suffered a pathological modification of form ; they may not be used for the determination of homologies. ‘ We are surprised to hear the assertion made that a leaflet bearing the abortive nucellus is homologous with the sorus- or sporangium-bearing leaflet of a fern. As if an abortive rudiment, showing no sign even of an embryo-sac, could in the remotest degree have anything to do with a sporangium.’ He urges that ‘it is a more profitable task to consider how such abnormalities have arisen and to discover the causes which condition the deviation from the normal development,’ and this is the burden of his article on ‘ The Teratology of Plants’ which appeared a few years ago in ‘ Science Progress.’ Here too we may place Eichler (49), who during his career held in turn all three views on the subject, but at length appears to have found repose for his ideas among those who support the theory we are now considering. In 1875, in the first volume of his ‘ BlUthendiagramme,’ he upholds the axial-theory of the ovule, influenced largely by the facts resulting from Schmitz’s recent work on the Piperaceae, and by those cases where the ovule is apparently replaced by a shoot ; further, from the teachings of the theory of descent he felt bound to regard the ovule F 66 Worsdell. — The Struchire and Morphology of the 4 Ovule! as everywhere of similar morphological dignity ; the placenta, on the other hand, as possessing a varying morphological value. But in 1878, in the second volume of the same work, he becomes an adherent of the v foliolar theory, and says : ‘ I must side with Celakovsky and accept both his placental and ovular theory from beginning to end.’ It is in the Botanisches Centralblatt for 1882, in his well-known thesis on the female flower of the Coniferae, that we find him finally throwing in his lot with the professors of the sui generis hypothesis ; his observations on the coni- ferous ovule have evidently altogether influenced him in taking this step, for, after referring to the varying positions of this organ in the different genera of Coniferae, he says : ‘ We must, therefore, give up the idea that the ovule everywhere corresponds either to a leaf-segment or everywhere to a bud arising as a rule from the metamorphosis of those structures ; it is rather the more or less modified macrosporangium of the higher plants [Vascular Cryptogams] which has been inherited by the Phanerogams, representing, like that organ, a structure sui generis. It may be compared to an emergence which it cannot be the exclusive privilege of leaves to produce, nor may it be regarded as invariably occurring on shoots ; on the contrary, we must recognize that the ovule, like other emergences, may arise as well on the one as on the other organ, or on the margin of both, i. e. in the axil : this is not only clearly the case in the Coniferae, but undoubtedly also in the Angiosperms.’ Bayley Balfour’s view (72) resembles that of Strasburger, as clearly shown by the following statement : 4 1 do not share a view which sees in integuments or other parts of the ovule anything of an axile or of a foliar nature. To me the funicle is a sporangiophore or a sporangial stalk, and the integumentary system is an outgrowth of the sporangial primordium of somewhat variable origin and development,’ &c. As regards the schools of botanists who interpret the ovule from the standpoint of teratology and vascular anatomy respectively, he says : 4 1 do not accept the starting-point of either the one or the other.’ Speaking of the sporangium his view is as follows: 4 All recent investigations . . . tend to confirm the view that it is, and always has been, an organ sui generis . To that category the nucellus of the ovule is now pretty generally admitted. It is the body of a sporangium.’ His general views on the nature of the ovule appear (as in the case of Strasburger) to be chiefly influenced by the facts of the individual development, or ontogeny of that organ ; and he seems to hold the idea that the integuments, for instance, are structures arising de novo in the life-history of the nucellus or sporangium to subserve the special physiological function of water-carriers and food- reservoirs. We have now to take up the consideration of the third, the foliolar * theory of the nature of the ovule, that theory which explains this mysterious An Historical Sketch. 67 organ as in fact representing the homologue of a segment or leaflet of the carpel which bears it (Fig. 3). The notable founder of this view of the matter is Brongniart (9), who in 1844 made in the first place observations on abnormal carpels of Del- phinium e latum, and noted the occurrence of transitions between the triden- tate lobes of the foliaceous carpel and the ovules themselves. In his con- clusions formulated from these facts he says that the veins of the placenta are really the lateral veins of the carpellary leaf ; each ovule corresponds to a lobe or a large tooth of this leaf, while the funicle, as also the raphe, is formed by the median vein of this lateral lobe ; the outer integument is the terminal portion of this leafy lobe folded on itself and forming a hood, the nucellus being a new production — a cellular protuberance arising on the upper surface of this lobe. It is thus impossible to regard the ovules together with the placenta as a production distinct from the carpellary leaf, and as part of the main or lateral axis. In Brassica Napus again, he observed transitions between the ovary and two foliage-leaves, and between the ovules and lobes of the leafy carpel. In abnormal ovaries of Anagallis arvensis the ovules are represented on the axial placenta by tiny leaves. In this case he evidently regards the ovules as the homologues of entire leaves, and also states his belief in a double origin for placentas : from the margins of carpels and from a prolongation of the floral axis. Finally, his estimation of the value of abnormalities, in the opening words of the paper, are worth quoting : ‘ There is hardly a botanist at present who does not recognize how much light the study of these aberrations from the normal structure, to which the name of monstrosities is given, sheds on the essential and fundamental structure of certain parts of the plants, or on that which is peculiar to certain groups of plants.’ Robert Brown (35), in his paper on Rajflesia (p.211), says: ‘ The principal point in which the antherae and ovaries agree, consists in their essential parts, viz. the pollen and ovula, being produced on the margins of the modified leaf.’ On p. 21 1 he says further: ‘The marginal production of ovula not infrequently becomes apparent where its formation is in some degree imperfect, and is most evident in those deviations from the regular structure where stamina are changed more or less completely into pistilla,’ as in Sempervivum tectorum , &c. On pages 379 and 556 of the first volume of his Miscellaneous Botanical Works he speaks of the margin of the carpels as the proper place for ovules, and he gives instances of exceptions to this rule, as in Nymphaeaceae, Mesemhryanthemum. On p. 563 we find the remark that ‘ovules belong to the transformed leaf or carpel, and are not derived from processes of the axis united with it, as several eminent botanists have lately supposed. That the placenta and ovula really belonged to the carpel alone is at least manifest in all cases where stamina are changed into pistilla.’ 68 Worsdell. — The Struchtre and Morphology of the ‘ Ovule ! Caspary (25), in a paper, accompanied by beautiful illustrations, on abnormal ovules of Trifolium repens, states that the funicle in this plant is morphologically equivalent to the lower part of the leafy lobe or pinna of the carpel ; at no stage is any limit between the funicle and the outer integument to be found. The inner integument, being so similar, must possess the same meaning and origin. It may disappear in the tissue of the leaflet, just as the outer integument has done in the funicle. ‘The funicle, along with the integuments, is in Trifolium repens the morphological equivalent of a leaflet (pinna), whose stalk or midrib is in the lower part of the funicle and whose campanulate or conical outgrowths of the upper part are the integuments. The nucellus is the new shoot seated on this leaflet.’ He thinks the nucellus also probably forms an integral part of the carpel. Through that classical work of his, the ‘ Bildungsabweichungen,’ Cramer (30) established his reputation as one of the most distinguished exponents of the morphological nature of the ovule. He observed meta- morphosed ovules of Primulaceae ; these were in the form of small foliar organs, developed on the free central placenta, which he regarded as entire leaves. He regards the nucellus as the product of a leaf. As for the integuments, if, he says, they were two distinct leaves and not part of one leaf they should both proliferate. That they are not such is shown by their becoming concave towards the upper side. The metamorphosed ovules are to be regarded as ‘ the outcome of the working of ideal successive combinations of formative forces fighting against each other, and not as developmental stages which have succeeded each other from time to time.’ He is strongly in favour of the use of abnormalities for the solution of morphological problems. To him, the ovule of Compositae is, like that of Primulaceae, an entire leaf. He also observed proliferous flowers of Delphinium elatum in which the green carpels bore either sterile or vegeta- tively developed ovules in the form of lobes ; the latter he held to be integumental in nature and the protuberances borne by them nucelli. In his paper ‘ On the Morphological Significance of the Ovule,’ occurring in the same volume, he sets forth that the nucellus is a new structure and not of the nature of a shoot, as he does not believe that the latter, as asserted, could ever arise from an ovule. The nucellus is merely an emergence , like pollen-sacs and fern-sporangia. His researches in the development of the ovule in Centaur ea jacea , Lysimachia punctata , and Anther icum liliago show that the primary papilla is not the nucellus but later gives rise to both this and the integuments. It is the funicle which first appears, and on this the nucellus is developed laterally. From his observations on Trifolium repens he concludes, with Caspary, that it is the outer, and not the inner integument which proliferates and which bears the nucellus in the middle of its upper surface. Prantl (53) stands conspicuous as one of the bold, yet not necessarily An Historical Sketch . 69 rash, thinkers who feel themselves at full liberty to compare certain organs of the Phanerogams with what they regard as related organs in the Vascular Cryptogams. It will suffice to quote a remark from page 11 of his work of 1875, where he says : ‘ Perhaps the integument (of Cycas) may be regarded as the equivalent of the indusium ’ (of Ferns). This view will be further elaborated when the tenets of the last author on our list are considered. Again, on page 13 he says : ‘ From my point of view I am unable to believe in actually axile ovules or anthers ; these organs are derivatives of sori,’ which are parts of the leaf. Further, he remarks : ‘ Ovules I must regard as under all circumstances originally parts of the carpel.’ At first (as seen from his paper of 1872, in which he says : € Ovules are certainly most frequently metamorphosed axes ’) a defender of the axile nature of the ovule, Warming (46), in his most valuable thesis : ‘ De l’ovule,’ published in 1878, and containing a series of researches into the development of the ovule-rudiment of the nucellus, unequivocally upholds the Brong- niartian theory of the ovule. ‘ In every ovule we have considered two parts essentially different : the funicle and the integuments, which are of foliolar nature, and the nucellus, which is a new creation, a sporangium, a “ sorus ” composed of a single sporangium, as Prantl would say.’ The origin and mode of development of the ovule-rudiment is similar to that of leaves, leaf-lobes, metablastema, emergences, and buds. Histogeny tells us nothing as to the morphological nature of the rudiment ; it only informs us as to how the latter arises on the placenta as a new formation. ‘ There is but one method which can lead us to the goal : the gradual comparative study of allied forms, relying on all the means at the disposal of the morphologist.’ On page 195 he writes: * The comparative study of the carpels and placen- tation in the entire vegetable kingdom, the scrupulous examination of antholyses and the course of the vascular bundles, lead us, as Celakovsky and Van Tieghem have recently proved, to the conclusion that carpels and placentas are everywhere phyllomes, and that the ovule-rudiment is a leaf-lobe ; I do not know if, in certain cases, they should not be regarded as metablastema, but the difference between the latter and a leaf-lobe is not essential and cannot be everywhere sustained.’ On page 200 he proceeds : £ The so frequent foliar transformation of the ovule-rudiment which will later become the funicle, and its fusion with the carpel, is indomitable proof that this organ is really a leaf-lobe, a conclusion which had already been rendered very probable by the position and order of appearance of the ovules. Ovules are not buds ; I know of no complete and well-studied teratological transformation which absolutely confirms the contrary, and I dare add that it will not be discovered. The ovular rudiment is therefore a leaf-lobe.’ The developmental history and teratology show us : — > 1. That the nucellus is a new formation on the ovule-rudiment, which is itself merely a lobe of the carpel. 70 Worsdell. — The Structure and Morphology of the ‘ Ovule l 2. That the pollen-sac and nucellus are identical as to their mode of development, which is here a proof of true homology, confirmed besides by a comparative study of these organs in the whole vegetable kingdom and by a series of teratological cases. 3. That the pollen-sac, like the sporangium (the common fundamental form of pollen-sac and nucellus), is everywhere attached to a leaf; con- formably with this truth we are assured, in a totally different manner, that the ovule-rudiment is, in fact, of foliar nature. This last view is supported by the study of Cycads. And he further says that we have seen that these metablastemata are everywhere attached to phyllomes and not to caulomes. As regards the true position of the nucellus, he says that in a general way it is to be regarded as terminal to the ovular rudiment. On the other hand, the teratological cases show it to be lateral and appear to indicate the true, or at least the primitive, relations existing between the two organs. At the end of Chapter III he says : ‘ It seems proved that the integument of Angiosperms is a special structure pertaining to the ovular leaflet and of foliar nature.’ This author’s treatise must be regarded, from its thoroughness of detailed investigation and breadth of treatment, as one of the most weighty contributions to this great and difficult subject, and of great interest in demonstrating how far developmental data serve in fixing the morphological dignity of the ovule. Celakovsky (38) stands out as a very brilliant exponent of the foliolar theory of the ovule, his contributions to the subject being, in the writer’s opinion, remarkable for their deep insight into, and ingenious handling of, one of the most difficult problems in botanical science. As the writer is also inclined to believe that in the sum of the numerous treatises of this author on the subject is contained a probable solution of the great problem as to the real morphological nature of the ovule, more space will be allotted to the consideration of his views than has been the case with the other authors cited above. Firstly, with regard to the various methods employed in determining the morphological nature of any doubtful structure, these may be said to be four in number, and may be termed : (a) the developmental, ( h ) the comparative, {c) the anatomical, (d) the teratological. Applying these methods in turn to the solution of the nature of the ovule, it will be seen, in the first place, that the study of the development or ontogeny of this organ tells us simply that it is a protuberance borne sometimes on an axis, sometimes on a foliar organ, producing in basipetal order two lateral, cup- shaped appendages. Relying entirely on the ontogenetic facts and assuming that the ovule is a bud would surely be dangerous ; for we cannot be sure that it is not an organ of a totally different nature, arising from the earliest stage of development, as a congenitally-modified structure. Of this latter An Historical Sketch. 7i the ovary of the Primulaceae is another example, affording us excellent proof of the unreliability of developmental evidence in such cases. The comparative method is a useful one, but taken alone is not finally reliable, for the standard of comparison is likely to be ever a shifting quantity varying with each botanist who employs it ; indeed, it seems impossible to apply this method by itself to the solution of the nature of an obscure structure like the ovule. The Bohemian Professor further regards anatomi- cal data as powerless when directed to the same purpose ; Van TieghenTs method of research along these lines must be regarded as stilted and artificial ; for vascular tissue is developed where it is needed and is quite independent of the morphological character of the organ which it supplies, and no absolute reliance is to be placed on the character of the course or structure of the bundles. Yet all three of these methods may be usefully employed as collateral aids and adjuncts in solving the morphology of a doubtful structure. For this latter purpose, however, the direct and principal method is that of the study of teratological phenomena ; it is on the ‘ metamorphogenesis ’ that Celakovsky rests the whole weight of his argument ; for he maintains that these abnormalities, or deviations from the normal structure, are not of the nature of ‘sports’ or ‘ monstrosi- ties,’ as is usually supposed, but, on the contrary, are phenomena controlled by definite laws of development, producing thereby structures which are not new or monstrous (except, indeed, as regards the limited life-cycle of the particular organisms or group of organisms concerned), but whose homologues will inevitably be found occurring as normal structures in organisms belonging to other sections of the vegetable kingdom. For all vegetable life is one, and the four types of organs : caulome, root, phyllome, and trichome, are the common heritage of all members of the vegetable world. But the question arises: how is the morphological value of any given doubtful organ to be determined by means of the abnormal structure which it assumes under the influence of these laws of ‘metamorphogenesis’? It would not be justifiable, on the ground of a mere supplanting or replacement of the abnormal structure, to regard the two as homologous in nature. But if (and this the writer regards as Celakovsky’s unassailable position) there can be traced in the metamorpho- genesis a series of gradual transitions (whether occurring in a single in- dividual or scattered over a number of individuals, matters not) between the normal form and the extreme stage of the metamorphosed structure replacing it, then this must prove the absolute homology of the two, i. e. that they possess an identical morphological value. It seems to the writer that a certain shallowness of thought and the influence of preconceived ideas on the subject are the cause of the rigid adherence by so many eminent botanists to the oft-reiterated statement that ‘ abnormalities can be made to prove anything ’ ! It betrays an utter neglect of the possibility 72 WorsdelL — The Structure and Morphology of the ‘ Ovule ! that, as stated above, these same ‘abnormalities’ may be the outcome of the working of rigid laws of development whose activity is reproduced in other departments of the vegetable kingdom. It is a view which can be easily refuted along the lines laid down above. Celakovsky regards the ovule as the homologue of a trilobed leaflet or segment of the carpel, of which the terminal lobe, involuted towards the upper surface to form a cup-shaped structure enclosing the nucellus (this latter being an organ of the nature of an emergence or sporangium borne on the upper surface of the lobe) is the equivalent of the inner integument; while the two lateral lobes, fused by their inner margins across the upper surface of the leaflet, represent the outer integument (Figs. 5, 6). It is to be noted that, in accordance with what the author terms the ‘law of laminar inversion,’ which ordains that two foliar laminae fg Figs. 5 and 6. Diagrams of a trilobed leaflet showing how, by fusion of the two lateral lobes across the upper surface of the terminal lobe through union of the lines fa , ac and gb , be , a structure homologous with the virescent ovule is obtained. The nucellus is seen seated as an emergence on the terminal lobe of the leaflet (after Celakovsky). funi funicle. are invariably in contact with each other by means of the similar surfaces of each, the lower surface of the outer integument contacts the lower surface of the inner integument. The above-outlined general position of the author derives its entire support from the facts revealed by the so- called ‘ monstrosities ’ of ovules where gradual transitional forms between the normal ovule and the three-lobed or simple leaflet have been observed. In the case of abnormal ovules of the Crucifer, Alliaria , it was the inner integument which exhibited the greatest amount of proliferation or virescence ; and another remarkable feature consisted in the preponderating tendency to proliferation of the funicle rather than of the outer integument. So that the ovular leaflet (the virescent ovule) sometimes assumes the form An Historical Sketch. 73 of a leafy structure (the funicle) bearing the inner integument, subtended by the rudimentary sheath of the outer integument, on its lower surface (Fig. 8) ; or, in some cases, the outer integument may be completely ab- sorbed in the funicular lamina. (For all details, both in this and other cases of abnormal ovules described by our author, the reader is referred to the original papers, which amply illustrate the various stages of the metamorphogenesis.) In the case of Trifolium repens it was also the funicle which chiefly proliferated, assuming the form of a bilobed structure at the Fig. 7. Fig. 8. Fig. 9. Fig. 11. Fig. 12. vrvt O vn£ • - ' Fig. 13. Fig. 14. Fig. 7. Alliaria officinalis', normal ovule. Fig. 8. Alliaria officinalis', ovule in which funicle has proliferated, bearing at its base the outer integument ensheathing the inner integument. Fig. 9. Alliaria officinalis : ovule showing funicle subtending the much proliferated bifid inner integument, bearing the nucellus on its upper surface and an adventitious shoot ( ad) at its base. Fig. 10. Trifolium repens : ovule in which funicle has proliferated, bearing inner integument in sinus between its two lobes (cf. Fig. 16). Fig. ii. Trifolium repens : virescent ovule proliferated as three-lobed leaflet. Fig. 12. Hesperis matronalis : virescent ovule with proliferated outer integument bearing inner integument on its lower surface. Fig. 13. Hesperis matronalis : proliferated outer integument bearing several inner integuments on its lower surface (Figs. 7-13 after Celakovsky). Fig. 14. Cupressus : seminiferous scale (= proliferated outer integument) bearing several ovules (= inner integuments and nucelli) on its lower surface (diagrammatic). base of whose sinus sometimes occurred the small sheath of the outer integument (from the upper surface of which the funicular lamina is an out- growth), enclosing the weakly-developed inner integument either in the form of a cup-shaped organ or as a simple leaflet bearing the nucellus on its upper surface (Figs. 10, 11). This is a very important stage, and will be referred to hereafter, 74 WorsdelL — -The Structure and Morphology of the ‘ Ovule! Hesperis matronalis differed from both of the last two plants, inasmuch as the leafy structure bearing the inner integument consisted solely of the outer integument , as is shown by its sheathing base and by the fact that the margins of the lamina passed gradually over into this sheath (Figs. 12 and 20). Hence our author terms it the ‘ basal lamina ’ (Grundspreite). The funicle takes therefore no part in its formation. In this plant the ovule frequently appeared as a simple leaflet bearing the nucellus on its upper surface ; this leaflet cannot be the outer integument, because for- mation of the inner integument being always the primary process this latter organ could not arise as an emergence from the surface of the outer integument ; it must also be situated invariably on the lower surface of the latter. The leaflet in question must therefore contain within itself both the inner and the outer integuments. This case resulted from the proliferating tendency setting in at the period when the ovule was nothing but a mere undifferentiated rudiment, this latter containing within itself the two integuments in potentia. The integuments, whether the outer or the inner, once laid down as completely sheathing structures , never pro- liferate as laminae. Hesperis is particularly interesting as having exhibited a case of a proliferated outer integument bearing two or more inner in- teguments, the extra ones occurring on the lateral lobes of the leaflet (Fig. 13) ; this case, as our author has elsewhere pointed out, is of con- siderable value for the interpretation of the female parts in Cupressus (Fig. 14). With Aquilegia and its various stages of metamorphosed ovules our author winds up his investigations. The virescent ovules of this plant are probably the most difficult to understand and to unravel. The first stage, in which proliferation sets in rather late, shows the inner integument seated on the upper surface of the * basal lamina ’ ; the two lobes of the latter are bent back and fused together behind instead of, as in all other cases, in front of the inner integument. As this lamina is to constitute the outer integu- ment, there here occurs an apparent contradiction to the usual law of ‘ laminar inversion ’ ; but our author finds it to be only apparent , for differ- entiation into an upper and a lower surface has not yet taken place in the inner integument. In the second stage there is an anatropous cup-shaped structure which, from the mode of development and the various modifica- tions occurring during the metamorphoses, is shown to be the inner integu- ment, with which the outer integument is intimately fused along its whole length ; the whole constituting a single undivided structure. This plant differs from Trifolium and Alliaria in the fact that the lower portion of the leaflet never grows out to form a separate individualized lamina. In the second stage just mentioned, where ■ proliferation * sets in early, the outer integument remains stationary, while the inner integument alone proliferates as the apical portion of the entire leaflet. The usual relation- An Historical Sketch . 75 ship between the outer integument and the ovular leaflet is described as follows : the former is an upgrowth out of the lower surface of the lower portion of the leaflet, after this latter has become inverted and folded in towards the upper surface . There exists, therefore, no essential difference between the case of Aquilegia and that of the other plants mentioned above. As demonstration of the fact that these variable structures, known as the ‘ ovular leaflet ’ or virescent ovule, are all governed in their development by definite laws of growth and sequential differentiation, and are not the fortuitous and haphazard result of ‘ sportive 5 tendencies on the part of the plant producing them, it may be mentioned that precisely identical phenomena have been observed as curious modifications of the entire foliage-leaves of the common Lilac ( Syringa vulgaris ). Again, exact homologues both of the metamorphosed ovule and of the unaltered ovule or ‘ ovular leaflet 5 may be found as normal structures in other departments of the vegetable kingdom. Celakovsk^ finds such in the apparatus of the female ‘flower’ of the Coniferae. The Taxaceae present the case of the normal ovule along with its two integuments ; the remaining groups that of the semi-proliferated ovule, of which the semi- niferous scale (or rather one-half thereof, seeing that each scale possesses two ovules) is the vegetatively developed outer integument bearing the involuted nucellus-producing inner integument on its lower (dorsal) surface. The case of Cupressus , in which a single seminiferous scale bears several such inner integuments on its lower surface (Fig. 14), finds its counterpart, as we have seen, in Hesperis ; such a structure as this might conceivably arise out of a compound ovular leaflet , the terminal segment of each lobe becoming, as in the simple three-lobed leaflet, the inner integument borne on that lobe’s lower surface. Descending lower in the scale, precisely the same set of structures (although naturally modified in accordance with the idiosyncrasies of the special group of plants in which they occur) are exhibited as normal stages of development in the sporophylls of the Ferns. In Thyrsopteris and Hymenophyllaceae (Figs. 15, 16) we see the case of the normal ovule of Angiosperms and of Taxaceae, in which the receptacle bearing its numerous sporangia (homologue of the nucellus), terminal in position, is ensheathed by the integuments, of which the indusium is the morphological equivalent of the inner, while the outer integument is represented by the laminar extension (when present) of the pinnule- segment on either side of the indusium. If now this structure be compared with the virescent ovule in Trifolium repens , its similarity to the bilobed funicular lamina enclosing the inner integument is obvious : the two structures are to be regarded as homologous although naturally presenting differing degrees of development of the respective parts. In Dicksonia the sorus is also terminal to the leaf-segment, but the indusium is here two- 76 WorsdelL — The Structure and Morphology of the ‘ Ovule! lipped instead of cup-shaped. In Davallia and Microlepia the first stage in the projection of the sorus on to the lower surface is seen, this being caused by the elongation of the upper side of the indusium, which becomes green and is (in part) an extension of the pinnule-segment (the outer integument (Fig. 17). The final stage is seen in Cystopteris , where the sorus, with the lower lip of the indusium, is projected completely on to the lower surface of the pinnule-segment (Fig. 18). In Cibotium and Cyathea there is a distinct cup-shaped indusium situated on the lower surface of the segment (Fig. 19); in the former this is at first terminal and marginal, becoming sub- Fig. 15. Thyrsopteris : sorus with indusium {ind) and terminal receptacle bearing sporangia {spg). Fig. 16. Trichomanes : receptacle and cup-shaped indusium (= inner integument) seated in sinus between two lobes of leaflet (= outer integument) (cf. Fig. io). Fig. 17. Davallia : early phylogenetic stage in projection of indusium on to lower surface of leaflet ; indusium one-sidedly developed. Fig. 18. Cystopteris : later stage in same process. Fig. 19. Cyathea : same stage, but indusium here completely formed ; spg, receptacle bearing sporangia. Fig. 20. Hesperis : proliferated outer integument bearing inner integument on its lower surface; this is precisely the same structure as that in Cyathea. (All after Celakovsky.) sequently displaced into an inferior position ; in the latter it is inferior from the first. In these two latter cases we have the exact counterpart of an inner integument situated on the lower surface of a ‘basal lamina’ or proliferated outer integument or funicle, such as occurs in Alliaria or Reseda (Fig. 20). As regards the Rhizocarps : in the Salviniaceae the fruit is equivalent to an ovule with one integument ; the indusium represents the inner integument and the leaf-lobe bearing the sorus is probably homologous with the outer integument. There is a striking resemblance, admitted by most botanists, between the monangic sorus of Azolla and an ovule (Fig. 21). Prantl regarded the two as homologous structures, and Campbell suggests the same thing. In the Marsiliaceae the fruit has the value of a compound fruit of Salviniaceae. In Pilularia it is homologous with a pinnately quadri-foliolate leaflet of the entire leaf of Marsilia ; in Marsilia with a pinnately multifoliolate leaflet ; in these cases the leaflet of the sporangiferous leaf possesses the compound structure of the vegetative leaf, An Historical Sketch . 77 while the leaflet of the latter is simple. The outer wall of the sporocarp is homologous with the upper surface of the outer integument of the ovule ; here, as in Hesperis and Cupressus , it is polysorous, bearing a number of indusia or inner integuments (each with its enclosed sporangia) on the lower surface (Figs. 22-25). Fig. 24. Fig. 25. Fig. 21. Azolla : megasporangium enclosed by indusium (after Campbell). Fig. 22. Pilularia\ diagram of morphological structure of sporocarp from two pairs of leaflets, each bearing a sorus on its lower surface. Fig. 23. Pilularia : transverse section (taken through dotted line in preceding figure) of leaflet, showing sorus and indusium; ma = mega-, mi = microsporangia. Fig. 24. Marsilia : diagram of morphological structure of sporocarp, from several pairs of leaflets, each bearing sorus on its lower surface. Fig. 2 5. Marsilia : transverse section (taken through dotted line in preceding figure) of leaflets here supposed to be folded together from opposite sides of midrib to form the sporocarp ; sorus shown. For a further elaboration of this subject of the homologies in the different groups the reader is referred to our authors exhaustive paper in ‘ Pringsheim’s Jahrbucher.’ In the Lycopodiaceae are again found equivalents for our ovule and its various parts. The ligule of the Selaginelleae our author, in the above- mentioned paper, regards as the same leaf-lobe which later in the Schi- zaeaceae grows out over the sporangia, but here, owing to the ventral position of the sporangium, the ligule is also ventral, and as such is so weakly developed that it is partly fused with the carpel, so that it de facto springs from the carpel ; the sporangium, as in the Schizaeaceae, springing 78 Worsdell. — The Structure and Morphology of the 4 Ovule ! Fig. 26. Lepidocarpon : dia- grammatic. lateral view of sporo- phyll, showing ‘ integument ’ and its mode of attachment to sporo- phyll ; position of ligule and sporangium are seen through the supposedly transparent ‘ integu- ment.’ from the lower surface of this modified leaf-lobe. He regards the velum as 4 unquestionably equivalent to the indusium of the Ferns/ £ The ligule along with the velum and sporangium is homologous with a leaf-segment of Cyathea with its integumented sorus, and in both cases the development of the leaf-segment precedes the formation of the integumented sorus.’ ‘As now a leaf-segment of a Fern with inferior indusium is homologous with a doubly-integumented ovule ; thereby is also the homology of Isoetes set forth : the velum corre- sponds to the inner integument, the ligule to the basal lamina of the half-proliferated ovule, the leafy equivalent of the outer integument.’ The present writer regards the 4 integument ’ of Lepidocarpon as homologous with the velum of Isoetes ; being more especially comparable with this organ in I. echinospora , where it com- pletely envelopes the sporangium. In Lepido- carpon he regards the ligule as really (i.e. morphologically or ideally) situated outside the ‘integument,’ which is here open at its distal end (Figs. 26, 27). In other genera either the velum or both this and the ligule have quite aborted or never developed : the result, probably, of an efficient protection of the sporangia being afforded by the peltate ends of the sporophylls, as in Lepido - dendron and Spencerites. The same may be said with regard to the Equisetaceae. Before proceeding further it is necessary to direct at- tention to one very important consideration. Celakovsky, until comparatively recently, regarded the virescent con- ditions of the Angiospermous ovule as retrogressive phenomena, as reversions to the primitive ancestral structures of the organ. After a deeper study, how- ever, of the phylogenetic relationships of the various organs of plants, from which he gathered that the primitive position of the sporangium on the sporophyll was a terminal one, he was eventually led to regard the normal structure of the ovule as representing a more ancestral state of affairs, inasmuch as the nucellus or sporangium occupies a position terminal to the ovular leaflet. The more vegetatively developed is any structure, the more modified will it be in the direction of an advance away from the primitive reproductive condition. The virescent condition of the ovule merely reveals to us the homologies of the latter ; it tells us nothing as to its phylogenetic origin ; this can alone be determined by the comparative method of research. Fig. 27. Lepido- carpon : diagrammatic transverse section taken through dotted line in preceding figure, int = integument ; sph = sporophyll. An Historical Sketch. 79 We have thus been able to trace the homologue of the ovule in three great phyla of the vegetable kingdom ; and this is only what one would a priori expect, for if there is a unity underlying vegetable life, and if all forms of plant-life own a common origin in the far past, the same organs, under a more or less modified or disguised form, will be produced by plants belonging to most of the great phyla of evolution. Finally, some of the objections urged by Celakovsky against the other two theories of the ovule may be brought forward. Against the bud-theory may be placed the fact that in almost all cases the integuments arise basipetally and not acropetally, a strange and ex- ceptional state of affairs if they really represent foliar appendages of the nucellus. Eichler, Schmitz, and at one time Warming, held that the terminal organs belong to the caulome-category on account of their position ; but this reliance on topical morphology is useless, as terminal organs other than axes are known. Hanstein showed that where the apex of the stem ceases to grow the apical periblem continues to do so and forms a terminal organ, which is a leaf, as the latter is always formed from periblem, whereas it is by growth of the plerome in length that an axis is continued in growth. The terminal cotyledon of Monocotyledons is another case in point. Thus, terminal ovules are not necessarily axial in nature. As regards the occurrence of adventitious buds in connexion with the ovules, our author determines, as far as his own investigations on Alliaria are concerned (Fig. 9), that the bud always represents a patho- logical new formation, and not a transformation of the ovule ; he found never a trace of an axis on which integuments are borne as lateral leaves. He enters into a lengthy criticism of Peyritsch’s views and observa- tions, which is in itself of considerable value towards the further elucidation of the nature of the ovule. The conclusions of this author, that the adventitious shoots are derived from and homologous with ovules, is shown to be groundless. Celakovsky says : £ It may be stated with certainty that shoots replacing ovules are never met with in virescent structures, and much less is this the case with transitions between ovules and such shoots.’ Peyritsch’s study of ovular metamorphosis is throughout influenced by the idea that the nature of the ovule can be determined from developmental data. He regards the ovular rudiment and the nucellus as identical (shown by Warming to be incorrect) ; that, therefore, the nucellus arises directly on the placenta and bears the integuments : but this never happens. He thinks, further, that the terminal ovule of Rumex scutatus is a shoot owing to its position ; but it is no more so than is a terminal nucellus. Other criticisms of Peyritsch’s observations and theories could be 8o Worsdell. — The Structure and Morphology of the ‘ Ovule l given, all going to show that that writer entirely misconstrued the nature of the various structures which came under his notice. Referring to the case described by Penzig of adventitious shoots on ovular leaflets of Scrophidaria vernalis which bore nucelli at their apex , our author holds that this phenomenon arises from a coincidence of the two constructive forces of the two structures respectively. The shoot arises below the nucellus after the latter has been formed and carries it upon its apex ; but here no intermediate forms can exist between the two ; ‘ in all truly intermediate forms an homology exists between all the parts, so that a transformation of the same basic structure obtains.’ An analogy for such an enclosed apex is found in a case of Helianthus annuus , mentioned by Sachs, in which the normal apex was damaged. From all this we gather that inasmuch as no true transitional struc- tures between an ovule and a shoot have ever been seen, but that, on the contrary, all so-called ovular shoots are either axillary productions of the carpels or else of the nature of adventitious buds on the ovular integuments or the placenta, therefore the ovule can no longer be held to possess the morphological nature of a shoot. Finally, as regards the sui generis theory of the ovule, Celakovsky s criticism of the views of Strasburger, the chief champion of this view, are very instructive and illuminating. This author’s main theory, as also in the case of the abnormal female parts of Coniferae, with regard to the various virescent conditions of the ovule, is this, viz., that they represent the varying results of a simultaneous strife between two forces : the generative , tending to produce the ovule which is of the nature of a sporangium or emergence, and the vegetative , tending to form a leafy structure. Now, Celakovsky maintains that, according to this view of Strasburger’s, the transitional forms which occur must be those between two quite heterogeneous plant-organs : a leaflet or segment of a carpel on the one hand, and a sporangial emergence on the other ; this is, on the face of it, an absurdity, and represents the same fallacy as that underlying his position with regard to the female flower in Coniferae. For the transitional forms clearly betray their derivation from the self-same organ, their origin from the same morphological substratum. Moreover, the virescent phenomena clearly show that the generative and vegetative forces assume possession of the ovular leaflet successively rather than simultaneously. Further, ‘if the ovule is not a metamorphosis of the carpellary segment, but a macrosporangium, an emergence, and this purely in place of the leaf- segment, the ovule with its integuments should contract more and more and gradually vanish, the leaflet or segment, on the other hand, become more fully developed.’ There can be no other compromise. And yet, as has been repeatedly shown, this is precisely what does not happen, for the An Historical Sketch . 8r integuments increase in size and become fused with and an integral part of the carpel. Celakovsky does not suppose, as was thought by Strasburger, that the various transitional forms represent the different stages of development which are to be run through in order to reach the extreme form ; nor does he regard it as ever possible that the integuments could, once they are laid low, revert back into a carpellary segment. A series composed of metamorphosed structures is of greater morphological importance than a developmental series. Strasburger asserts that each of these forms ought to be treated and studied separately, in and for itself alone. Our author, on the contrary, affirms that £ each single case taken by itself is of little value ; it is only the comparatively arranged complete series which can guarantee us real insight into the essential being, metamorphosis, and origin of the ovule.’ Our author further urges that a comparison of an orthotropous ovule with the macrosporangium and sporocarp of Azolla tends to destroy Strasburger’s views founded on the ontogeny ; for the integument of Azolla arises not from the sporangium, but from the leaf-segment beneath it. And, moreover, as the present writer would add, Warming has clearly shown that both nucellus and integuments develop independently from the ovular rudiment. The terminal position of the nucellus does not prevent it from being altogether distinct from the ovular rudiment producing it. As regards the view held, for instance, by Balfour, that the nucellus is an organ sui generis , Celakovsky points out that inasmuch as, in the virescent ovule, the nucellus, as an emergence from the surface of the ovular leaflet, may become completely absorbed into, and form an integral part of, the latter, it cannot possibly constitute one of the distinct morpho- logical categories ; for the respective organs belonging to these latter have never been known to merge or be absorbed into each other in such a manner. The same will be true of the eusporangium of Vascular Crypto- gams, which is the homologue of the nucellus. General Summary. In conclusion is appended a brief resume of the various theories on the morphology of the ovule. Axial Theory. St. Hilaire (1830, 1840), Schleiden (1839, 1843), Payer (1859), Braun (i860), Peyritsch1 (1872-76): the nucellus is of the nature of a bud bearing the two integuments as lateral foliar appendages. 1 It must be borne in mind that this author nowhere makes any definite statement as to the morphology of the ovule, but his writings betray the trend of his ideas on the subject. G 82 Worsdell. — The Structure and Morphology of the ‘ Ovule l Sui Generis Theory. Schmitz (1870); Sachs (1874); Goebel (1882-1901); Strasburger (1879) 5 Balfour (1901) : the ovule does not (necessarily) belong to any of the morphological categories, but is an independent structure, borne either on stem or foliar organs. Foliolar Theory. Brongniart (1844); R. Brown (1845-66); Caspary (1861); Cramer (1864); Prantl (1875); Warming (1878); Celakovsky (1874-1900): the ovule belongs morphologically to the category of the phyllome ; it is the homologue of a (usually) three-lobed leaflet or segment of the carpel (or female sporophyll), the outer integument and funicles representing the lower portion of the leaflet whose lateral lobes, fused across its upper or ventral surface, constitute a lamina having the lower surface directed towards that of the involute cup-shaped terminal lobe ; the representative of the inner integument. The nucellus is, like the eusporangium of Ferns, of the nature of an emergence , borne on the upper surface of the leaflet’s terminal lobe. The morphological value of the various parts of the ovule here set forth is ascertained and demonstrated by the occurrence, in teratological conditions of the ovule, of gradual transitions between the structure of the normal ovule and that of the extreme virescent organ in the form of a carpellary leaflet. This latter must not be regarded as a reversion to the primitive condition ; on the contrary, the normal ovule possesses a structure essen- tially primitive and archetypal. Bibliography. 1. Mirbel: Observations sur la famille des vegetaux coniferes, Ann. du Mus. d’Hist. Nat., vol. xv> P- 473, iS1©- Recherches sur la structure et le developpement de l’ovule vegetale, 1828. and Spach: Notes sur l’embryogenie des Pinus Lari cio, &c., Ann. d. Sci. Nat.,s£r. 2, vols. xix and xx, 1843. 2. Jaeger : Missbildungen der Gewachse (Aguilegia), 1814. 3. Roeper : Enumeratio Euphorb. (. Delphinium ), 1824. 4. Richard, L. C. : Commentatio botanica de Coniferiset Cycadeis, 1826. 5. Schimper, K. Fr. : On Umbelliferae, Cruciferae, Resedaceae : Salix babylonica, Papaver somniferum , Symphytum Zeyheri , Flora, vol. ii, 1829. — Beschreibung des Symphytum Zeyheri , &c., Geiger’s Magazin fur Pharmacie, vol. xxviii, 1835. 6. St. Hilaire, A. de : Premier memoire sur la structure et les anomalies de la fleur des Resedac^es, Ann. de la Soc. d’Orleans, 1830. — Deuxieme memoire sur les Res^dacees, Montpellier, 1837. Le?ons de Botanique comp., princip. la morphologie vegetale, 1840. An Historical Sketch. 83 7. Engelmann: De Antholysi ( Sisymbrium , Torilis , Gilia), 1832. 8. Endlicher : Linnaea, vol. vii, p. 21 (axial placenta), 1832. and Unger : Grundziige der Botanik (disk- theory), 1843. 9. Brongniart : Beobachtungen an Primula chinensis, Annales des Sciences Naturelles, sdr. 2, vol. i, p. 308 (leaf-theory), 1834. — — Examen de quelques cas de monstruosites vegetales propres a eclairer la structure du pistil et l’origine des ovules, Archives du Museum d’Hist. Nat, Paris, vol. iv, p. 43, 1844. 10. Fritsche: Ueber die Entwickelung des Pflanzeneies in seinen friihesten Zustanden, Wieg. Archiv. ii, p. 229, 1835. 11. Mohl, von : Sur la metamorphose des antheres en carpelles, Ann. d. Sci. Nat., s£r. 2, vol. viii, 1837. 12. Treviranus : Physiologie der Gewachse, vol. ii, p. 333, 1838. 13. Valentin : Antholysen von Lysimachia ephemerum , Act. nat. cur.,. vol. xix, i, p. 225, 1839. 14. Schleiden : Sur la signification morphologique du placentaire, Ann. des. Sciences Nat., s^r. 2, vols. ix and xii, 1839. Grundziige der wissenschaftlichen Botanik, p. 389, 1843. 15. Braun, Al. : Vortrag auf dem Congres scientifique de France, 10. Sitzung, Strassburg, pp. 172-2 (seminif. scale of Coniferae), 1842. Ueber Polyembryonie und Keimung von Coelebogyne (axial theory), i860. p)je Frage nach der Gymnospermie der Cycadeen, &c., Monatsber. d. Konigl. Akad. Wiss. Berlin, 1875. 16. Reisseck : Ueber das Wesen der Keimknospe, Linnaea, vols. xvii and xx ( Alliaria , leaf-theory), 1843. 17. Schlechtendahl : Pflanzen-Monstrositaten, Bot. Zeit., 1847. 18. Unger : Antholysen von Primula chinensis , Act. nat. cur., vol. xxii, ii, p. 543, 1850. 19. Wigand : Grundlegung der Pflanzenteratologie, p. 39 {Reseda, axial theory), 1850. — Beitrage zur Pflanzenteratologie, Botanische Untersuchungen, p. 23 ( Rosa , Turritis, Crepis ), 1853. 20. Germain de St. Pierre : LTnstitut, no. 1021, p. 255, 1853. 21. Rossmann : Die Entwickelung der Eiknospen aus dem Fruchtblatte und Deutung des Samentragers, Flora, pp. 657, 705 (. Aquilegia , axial theory), 1855. 22. Norman, J. M. : Observations sur les Chloranthies, Ann. d. Sci. Nat., Bot., ser. 4, vol. ix, 1858. 23. Schacht : Grundriss der Anatomie und Physiologie der Gewachse (disk-theory), 1859. 24. Payer : Grganogenie de la fleur, 1859. 25. Wydler : Beschreibung einiger Antholysen von Alliaria , Denkschrift der Regensburger Bot. Gesellsch., vol. iv, Abth. i, p. 77, 1859. 26. Caspary : De Abietinearum floris feminei structura morphologica, Ann. d. Sci. Nat., s^r. 4, vol. xiv, i860. Vergriinungen der Bliithe des weissen Klees, Konigsberger Schriften der physik.-okon. Gesellschaft, p. 50, 1861. 27. Baillon : Recherches organogeniques sur la fleur femelle des Coniferes, Adansonia, vol. i, 1 860-1. Nouvelles recherches sur la fleur femelle des Coniferes, Adansonia, vols. v-vi, 1864-5. Description d’une Primulacee a fleurs monstrueuses, Adansonia, vol. iii {Lysimachia), 1862-3. 28. Hofmeister : The Higher Cryptogamia, Ray Society, 1862. 29. Pringsheim : Zur Morphologie der Salvinia natans, Pringsh. Jahrb. f. wiss. Bot., vol. iii, 1863. 30. Marchand : Monstruosites vegetales, ier fasc., Adansonia, vol. iv, 1863-4. 31. Cramer : Bildungsabweichungen bei einigen wichtigeren Pflanzenfamilien und die morpho- logische Bedeutung des Pflanzeneies, Zurich, 1864. 32. Parlatore: Studi organografici sui fiori e sui frutti delle Conifere, Opuscula botanica, 1864. 33. 0rsted : Bidrag til Naaletrseernes Morphologi, Videnskab. Meddelelser Nat. Foren., Kjobenhavn, 1864. G 2 84 Worsdell. — The Structure and Morphology of the 'Ovule! 34. Hallier: Phytopathologie ( Verbascum ), 1864. 35. Salter : On a Sexual Monstrosity, consisting in the Development of Polliniferous Ovules in two species of Passiflora, Trans. Linn. Soc., vol. xxiv, p. 143, 1864. 36. Brown, R. : Miscellaneous Botanical Works, vol. i, 1866. 37. Meyen : Beitrage zur Kenntniss der Azollen, Act. Acad. Leop. Carol., vol. xxxii, 1867. 38. Masters : On Polliniferous Ovules in a Rose {Rosa arvensis ), Seemann’s Journal of Botany, vol. v, p. 318, 1867. Pflanzenteratologie, trans. by U. Dammer, Leipzig, 1886. 39. Celakovsky : Ueber die allgemeine Entwickelungsgeschichte des Pflanzenreiches, Konigl. bohmische Gesellschaft der Wissenschaften, Naturwiss.-mathem. Section, 1868. 7 Ueber den Zusammenhang der verschiedenen Methoden morphologischer Forsch- ung, Lotos, Oct., 1874. Ueber die morphologische Bedeutung der Samenknospen ( Anagallis , Dictamnus), Flora, 1874. Die Vergriinungsgesckichte der Eichen von Alliaria officinalis , Botanische Zeitung, 1875. Zur Discussion fiber das Eichen, Bot. Zeitung, 1875. Vergleichende Darstellung der Placenten in den Fruchtknoten der Phanerogamen (, Scrophularia , Anagallis ), Abhandl. der bohm. Ges. d. Wissensch., 1876. Die Vergriinungsgeschichte der Eichen von Trifolium repens, Botan. Zeitung, 1877. Ueber Chloranthie der Reseda lutea , Bot. Zeit., 1878. — Ueber vergriinte Eichen der Hesperis matronalis, Flora, 1879. Die Vergriinungsgeschichte der Eichen von Aquilegia als neuer Beleg zur Foliolar- theorie, Bot. Centralblatt, vol. x, 1882. Untersuchungen fiber die Homologien der generativen Producte der Frucht- blatter bei den Phanerogamen und Gefasskryptogamen , Pringsh. Jahrb. f. wiss. Botanik, vol. xiv, 1883. • Neue Beitrage zur Foliolartheorie des Ovulums, Abhandl. d. bohm. Ges. d. Wissensch. ( Raphanus ), 1884. Die Gymnospermen : eine morphologisch-phylogenetische Studie, Abhandl. d. konigl. bohm. Ges. d. Wiss., 7. Folge, vol. iv, 1890. Nachtrag zu meiner Schrift iiber die Gymnospermen, Engler’s Botanische Jahrbiicher fiir Systematik, &c., vol. xxiv, 1897. Epilog zu meiner Schrift iiber die Placenten, Sitzungsb. d. Konigl. bohm. Ges. d. Wiss., 1899. ■ — Ueber den phylogenetischen Entwickelungsgang der Bliithe, &c., part ii, p. 192. Sitzungsb. d. Konigl. bohm. Ges. d. Wiss, 1900. — — Neue Beitrage zum Verstandniss der Fruchtschuppe der Coniferen, Pringsh. Jahrb. f. wiss. Bot., vol. xxxv, part iii, 1900. 40. Van Tieghem : Anatomie comparee de la fleur femelle et du fruit des Cycad6es, des Coniferes et des Gnetac6es, Ann. d. Sci. Nat., ser. 5, vol. x, 1869. Traitd de botanique, vol. i, p. 405, 1884. Elements de botanique, vol. i, p. 398, vol. ii, p. 266, 1898. 41. Koehne : Ueber Bluthenentwickelung bei den Compositen, p. 65, Inaugural-Dissertation, Berlin, 1869. 42. Sperk : Die Lehre von der Gymnospermie, Mem. de l’Acad. Imp^r. des Sciences Nat., ser. 5, Bot., 1868-9. 42A. Rohrbach : Die Samenknospe der Typhaceen ; Bot. Zeit., 1869. 43. Hanstein and Schmitz : Ueber die Entwickelungsgeschichte der Bliithen einiger Piperaceen, Botanische Zeitung, 1870. 44. Schmitz : Die Bluthenentwickelung der Piperaceen, Hanstein’s Botanische Abhandlungen, vol. ii, 1875. 45. Magnus : Beitrage zur Kenntniss der Gattung Najas , 1870. Ovula der vergriinten Bliithen von Reseda lutea , Sitzungsb. d. Ges. naturf. Freunde zu Berlin, Bot. Centralbl., vol. ii, p. 430, 1882. 46. Peyritsch : Ueber Bildungsabweichungen bei Cruciferen, Pringsh. Jahrb. f. wiss. Bot., vol. viii, 1872. An Historical Sketch. 85 46. Peyritsch : Teratologie der Ovula, Festschrift d. Wien. Zoolog. -hot. Ges., pp. 115-144, 1876. — Zur Aetiologie der Chloranthien einiger Arabis- Arten, Pringsh. Jahrb. f. wiss. Bot., vol. xiii, 1882. 47. Warming : Recherches sur la ramification des Phanerogames, Schriften d. danischen Gesell. d. Wiss., Ser. 5, vol. x, 1872. Bemerkungen fiber das Eichen, Bot. Zeit., 1875. De P ovule, Ann. d. Sci. Nat., Bot., 1878. Bidrag til Cycadeernes Naturhistorie, K. Dansk. Vidensk. Selsk. Forhandl., 1879. 48. Strasburger : Die Coniferen und die Gnetaceen, 1872. Einige Bemerkungen fiber Lycopodiaceen, Bot. Zeit., 1873. — Ueber Azolla, Jena, 1873. Die Angiospermen und die Gymnospermen, 1879. 49. Jur\nyi : Ueber die Entwiekelung der Sporangien und Sporen der Salvinia natans , 1873. Ueber die Gestaltung der Frucht bei Pilularia globulifera, Botanisches Centralblatt, p. 201, 1880. 50. Eichler : Sind die Coniferen gymnosperm oder nicht ? Flora, 1873. — : i Blfithendiagramme, vol. i (axial theory), 1875. — 1 Ifithendiagramme, vol, ii (foliolar theory), 1878. Ueber die weiblichen Blfithen der Coniferen, Monatsber. d. Konigl. Akad. d. Wiss., Berlin, 1881. Ueber Bildungsabweichungen bei Fichtenzapfen, Sitzungsb. d. Konigl. Akad. d. Wiss., Berlin, 1882. Ueber die weiblichen Blfithen der Coniferen, Bot. Centralbl., vol. ix, p. 49 ( sui generis theory), 1882. 51. Chatin : Etudes sur le ddveloppement de 1’ovule et de la graine dans les Scrofularinees, &c., Ann. d. Sci. Nat., ser. 5, vol. xix, 1874. 52. Sachs: Lehrbuch der Botanik, 4th edition, 1874. 53. Burck : Indusium der Varens, Haarlem, 1874. 54. Prantl : Bemerkungen fiber die Verwandtschaften der Gefasskryptogamen, Verhandl. d. Phys. Med. Ges. in Wfirzburg, 1875. — Morphologische Studien, Flora, 1875. Untersuchungen zur Morphologie der Gefasskryptogamen, vol. i, Hymenophyllaceen, 1875 ; vol. ii, Schizaeaceen, 1881. 55. Beketoff : On Proliferated Flowers of Cichorium Intybus, M^m. de la Soc. Nat. d. Sc. Nat. de Cherbourg, vol. xxi, 1877. 56. Dyer : The Morphology of Selaginella , Nature, 1877. 57. Goebel : Ueber Sprossbildung auf /soeieshlattern, Bot. Zeit., 1879. Beitrage zur Entwickelungsgeschichte der Sporangien, Bot. Zeit., 1880-2. Die vergleichende Entwickelungsgeschichte der Pflanzenorgane, Schenk’s Handbuch der Botanik, vol. iii, part, i, pp. 114, 120. Archegoniatenstudien, Flora, 1892-3. — Organographie der Pflanzen, 1901. 58. Magnus, P. : On virescent Ovules of Aquilegia atrata , Verh. bot. Ver. Brandenburg, vol. xxi, 1880. 59. Pax, F. : Metamorphogenese des Ovulums von Aquilegia, Flora, vol. lxv, 1882. 60. Penzig : Ueber vergrfinte Eichen von Scrophularia •vernalis, L., Flora, vol. lxv, 1882. — Pflanzenteratologie, vol. i, 1890. 61. Nageli : Mechanisch-physiologische Theorie der Abstammungslehre, p. 496, 1884. 62. Meunier : La Pilulaire: Etude anat.-genetique du sporocarpe chez la Pilularia glob ulif era, La Cellule, vol. iv, p. 325, 1887. 63. Campbell: The Structure and Development of the Ostrich Fern, Mem. Boston Soc. Nat. Hist., 1887. •. The Development of Pilularia globulifera , L., Annals of Botany, vol. ii, p. 233, 1888. — The Systematic Position of the Rhizocarpeae, Bull. Torrey Bot. Club, Oct., 1888. On the Affinities of the Filicineae, Bot. Gaz., Jan., 1890. Contributions to the Life-history of Isoetes , Ann. Bot., vol. v, August, 1891. 86 W or s dell. — The Structure and Morphology of the ‘ Ovule l 68. Campbell: On the Relationships of th Archegoniatae, Bot. Gaz., Dec., 1861. The Development of the Sporocarp of Pilularia americana , Bull. Torrey Bot. Club, April, 1893. On the Development of Azolla filiculoides , Ann. Bot., 1893. The Structure and Development of Mosses and Ferns, 1895. 64. Vines : On the Systematic Position of Isoetes , L., Ann. Bot., 1888. 65. Engler and Prantl : Die natiirlichen Pflanzenfamilien, 1889. 66. Busgen : Untersuchungen liber normale und abnormale Marsilienfriichte, Flora, vol. xlviii, p. 169, 1890. 67. Farmer : On Isoetes lacustris , Ann. Bot., vol. v, Dec., 1890. 68. WojENOWid: Beitrage zur Morphologic, Anatomie und Biologic der Selaginella lepidophylla , Inaug. Dissert., Breslau, 1890. 69. Giesenhagen : Die Hymenophyllaceen, Flora, 1890. 70. Bower: On the Structure of Lepidostrobus Brownii , Ann. Bot., vol. vii, 1893. ■ Studies in the Morphology of Spore-producing Members, I, Philosophical Transactions of the Royal Society, vol. 185 (Equisetaceae, Lycopodiaceae), 1894. — Studies, &c., iv ; The Leptosporangiate Ferns, Phil. Trans. Roy. Soc. B., vol. 192, 1899. 71. Scott : On Spencerites , a new Genus of Lycopodiaceous Cones from the Coal-Measures, &c., Phil. Trans. Roy. Soc., vol. 189, 1898. * The Seed-like Fructification of Lepidocarpon , a genus of Lycopodiaceous Cones from the Carboniferous Formation, Phil. Trans. Roy. Soc., vol. 194, 1901. 72. Johnson, D. S. : On the Development of the Leaf and Sporocarp in Marsilia quadrifolia> L., Ann. Bot., vol. xii, 1898. 73. Maslen : The Structure of Lepidostrobus , Trans. Linn. Soc., vol. v, 1899. 74. Smith, R. Wilson: The Structure and Development of the Sporophylls and Sporangia of Isoetes, Bot. Gazette, 1900. 74A. Dunn, L. B. : Morphology of the Development of the Ovule in Delphinium exaltatum. Proc. Amer. Assoc. Adv. Sci., vol. 49, 1900. 75. Worsdell : The Structure of the Female Flower in Coniferae: an Historical Study, Ann. Bot., vol. xiv, March, 1900. The Morphology of the ‘ Flowers’ of Cephalotaxus , Ann. Bot., Dec., 1901. On abnormal Flowers of Helenium autumnale, L. ; Journal Roy. Hort. Soc., vol. xxvii, part 4, April, 1903. 76. Balfour, I. Bayley : Address to the Botanical Section, British Association for the Advance- ment of Science, Glasgow, 1901. 76A. Angiosperms. Reprinted from the new volumes of the Encyclopaedia Britannica, 1902. 77. Hallier, H. G. : Beitrage zur Morphogenie der Sporophylle und des Trophophylls in Beziehung zur Phylogenie der Kormophyten, Beiheft zum Jahrbuch der Hamburgischen Wissenschaftlichen Anstalten, vol. xix, 1901. 78. Schumann, K. : Ueber die weibliche Bliithe der Coniferen, Abhandl. des botan. Vereins der Prov. Brandenburg, vol. xliv, 1902. 79. Oliver, F. W. : The Ovules of the older Gymnosperms, Annals of Botany, vol. xvii, June, 1903. 80. Coulter and Chamberlain : The Morphology of Angiosperms, p. 52 (sui generis theory), i9°3- On the Structure and Biology of Fegatella conica. BY F. CAVERS, B.Sc., Lecturer in Biology Municipal Technical School , Plymouth . With Plates VI and VII and five Figures in the Text. HE genus Fegatella Raddi ( Conocephalum Wiggers) is represented by X two species, of which one, F. conica , is very widely distributed in the North Temperate Zone, whilst the other, F . supradecomposita, is confined to Japan and China. F. conica , which is fairly common in Britain, is one of the largest of the thalloid Hepaticae, the broad dichotomously branched thallus sometimes reaching a length of six inches (1 5 cm.) or even more ; the older parts gradually die down as the new branches creep onwards over the substratum, which often becomes covered by a continuous layer of these plants, extending over several square feet. The plants grow chiefly in moist situations, especially on stones beside shaded streams, and sometimes become entirely submerged in water. This species was named Hepatica fontana by Micheli in 1729 ; the generic name, which was given on account of a supposed resemblance between the branched thallus and the lobes of the liver, was after Micheli’s time applied to the whole group ( Musci hepatici , Hepaticae), The upper surface of the thallus is marked by lines dividing it up into polygonal (mostly hexagonal) areas and forming an extremely regular network. Each of these areas corresponds with an underlying air-chamber, and has in its centre a light spot which stands out in sharp contrast with the dark green colour of the thallus and marks the position of a pore. The central portion of each area is raised so that the pore occupies the summit of a low cone. When water is placed on the upper surface of the thallus, it collects along the lines separating the air-chambers and is quickly drained off by means of this network of channels. In neither Marchantia nor Lunularia , which have a general external resemblance to Fegatella , is the areolation of the thallus so definite as in the latter genus, which is therefore readily distinguished from the other commonly-occurring genera of the Marchantiaceae, even when no repro- ductive structures are present. Another characteristic feature of Fegatella [Annals of Botany, Vol. XVIII. No. LXIX. January, 1903.] 88 Cavers. — On the Structure and is its peculiar fragrant odour, which resembles that of oil of bergamot, and becomes very pronounced when the thallus is crushed. The thallus is differentiated into a thicker median cylindrical portion (midrib) and a much thinner portion (lamina) on either side of it. This is more conspicuous on the lower than on the upper surface, since the midrib projects strongly below. On its lower surface the midrib bears two rows of scales, one row on either side of the middle line, together with numerous tufts of rhizoids, which are also arranged in two rows immediately outside of the scales and pass downwards into the soil. The ventral scales are relatively small, being confined to the sides of the midrib and hardly reaching the lower surface of the lamina. Each scale (Plate VI, Fig. 1 6) is made up of an oval or kidney-shaped free anterior portion, red or violet in colour, and a colourless posterior portion which has the form of a long narrow band, attached by its outer edge to the side of the midrib, whilst its inner edge reaches and is parallel with the middle line of the thallus and overlaps the hinder portion of the next scale in front. On tracing the scales forwards and removing them one by one, it is found that the coloured appendages are, at the anterior margin of the thallus, curved upwards and backwards so as to cover the apical growing-point, which lies in the notch observed at the end of each branch. As seen from above, this notch appears to be largely filled up by the overlapping appendages, which serve to protect the young tissues of the growing-point (Plate VI, Fig. i). When the scales have been removed the lower surface of the midrib shows a median furrow, containing a strand of rhizoids which do not, like those arising in tufts from the sides of the midrib, at once grow downwards into the soil, but pass backwards in a compact bundle. Apical Growing-point. The growing-point of each branch is occupied by a transverse row of actively-dividing cells, lying nearly in the plane of the ventral surface of the thallus. These initial-cells are wedge-shaped, appearing oblong and rectangular in surface view, triangular in a longitudinal vertical section through the growing-point (cf. Plate VI, Figs. 8, 9, 18). From each initial- cell segments are cut off by walls parallel with its upper, lower, and lateral walls. The growth of these segments is less rapid in the region immediately behind the apex than it is on either side, so that the apex itself comes to occupy a deep notch. Whilst a large portion of the tissue formed from the segments consists of cells that remain closely packed together, the superficial cells of both the dorsal and the ventral segments show a specialized form of growth, the former giving rise to the dorsal layer of air-chambers, and the latter to the ventral scales and the rhizoids. Biology of Fegatella conic a. 89 Apical Branching. Under normal conditions branching only takes place at the apex of the thallus. In autumn, usually during October, resting-branches are laid down at the apex of each lobe of the thallus. During the winter months vegetative growth is practically at a standstill, but towards the end of January or beginning of February the resting-branches resume their growth. When branching is about to occur the initial-cells grow in breadth and undergo repeated divisions by vertical walls, thus becoming increased in number, whilst the row becomes laterally extended. The central cells for some time grow more rapidly than those on either side, and form a projecting lobe, but their growth soon ceases, and the cells on each side of the lobe, continuing to grow and divide, form the apical growing-points of the two branches (Plate VI, Fig. 9). This process is repeated, so that four growing-points occupy the end of each of the club- shaped outgrowths formed in autumn. These outgrowths, one of which is nearly always to be found at the end of each of the thallus-lobes, remain dormant during the winter, the four growing-points being protected by the coloured scale-appendages (Plate VI, Fig. 2). In early spring growth is resumed, and each growing-point may then give rise to a broad expanded lobe, or one of them may have its growth in length arrested by the development of a male or female receptacle (Plate VI, Figs. 3-7). The thallus acquires a more or less jointed appearance, the earliest part of each season’s growth giving rise to a narrow stalk-like portion which then spreads out to form the broad lobes of the thallus. Occasionally the branching deviates from the type just described, owing to suppression or repetition of some of the divisions that normally occur. When plants are cultivated indoors the branches remain narrow and often cylindrical, and grow upwards instead of horizontally. This upward growth is especially marked when the plants are cultivated in a deep vessel illuminated from above. Plants kept in darkness present a somewhat similar mode of growth, the etiolated branches being long, narrow, and cylindrical. Air-chambers. Immediately behind the growing-point, each dorsal segment of the initial-cells becomes divided into an inner cell, which contributes to the formation of the compact tissue of the thallus, and an outer cell which gives rise to part of the spongy tissue. The outer cells soon begin to grow out and to become separated from each other so as to leave narrow air-spaces between neighbouring cells. As the tissues behind the growing- point become extended in area in consequence of active growth these spaces become widened, whilst the superficial cells grow up and remain attached laterally so as to form a series of vertical plates. The uppermost 90 Cavers .- — On the Structure and cells then grow out horizontally and form a roofing layer (epidermis), the cells of which divide only by vertical walls. The epidermis thus forms a single layer of cells, a pore being left above each chamber (Plate VI, Figs. 17, 18). Each pore is at first surrounded by from five to eight cells, each of which undergoes repeated division by walls tangential to the pore, giving rise to five or six concentric rings of cells (Plate VI, Fig. 10). The dorsal region of the thallus thus comes to be occupied by a series of wide air-chambers, forming a single layer and separated from each other by vertical partitions, which are for the most part one cell in thickness and are united with each other so as to form a network. It is of course to these vertical walls between the chambers that the areolation of the upper surface of the thallus is due. The concentric rows of cells surrounding each pore do not all lie in the same plane, the inner rows being higher than the outer, so that the pore becomes situated on the summit of a dome-like elevation. Immediately around the pore there is a thin but fairly wide membrane of cellulose, which shows in surface view a coating of granules ; these are readily dissolved in alcohol, and probably consist of resin or wax. A similar coating of resinous or waxy grains is found on the cells surrounding the barrel-shaped pores of Marchantia , and it was suggested by Kny (1890, p. 369), that these serve to prevent the entrance of water into the chambers. Whilst the chambers have been growing in width the cells forming the floor of each chamber grow out and divide by transverse walls, so as to form short rows of cells containing abundant large chloroplasts ; the latter are also found in the cells forming the floor and sides of the chambers, and in much smaller numbers in the epidermis itself. At the sides of the chamber the filaments are generally attached both above and below, but the central filaments, lying below the wide pore, have a peculiar form, the terminal cell of each filament being elongated and produced into a narrow tapering point. In each of these pointed cells the chloroplasts are confined to the swollen basal portion of the cell, the upper portion containing only clear sap with a thin lining layer of protoplasm (Plate VI, Fig. ix). The writer has repeated the experiments of Kamerling (1897, p. 50), which show that these pointed terminal cells have the function of giving off relatively large quantities of water-vapour. A plant having its rhizoid-bundles intact is kept for several hours in a weak (i per cent.) solution of red prussiate of potash (potassium ferricyanide), and afterwards treated with alcohol so as to precipitate the salt. On next placing the plant in a solution of ferrous sulphate it is found that the blue precipitate is almost entirely confined to the pointed terminal cells of the chlorophyll-bearing filaments, the cells below showing hardly any precipitate. This clearly shows that the evaporation of water is Biology of Fegatella conic a. 91 localized in the pointed cells, and that the solution absorbed by the rhizoids had passed through the ventral tissue of the thallus into the green filaments, and had become concentrated in the terminal cells owing to continued evaporation. Ventral Tissue. The compact tissue underlying the air-chambers is well developed in the midrib, but in the lamina on either side is much thinner, becoming reduced to two or three layers of cells near the margin of the thallus (Plate VI, Fig. 18). This tissue consists chiefly of large colourless cells, containing starch-grains and having their walls thickened by anastomosing fibres, the unthickened portions remaining as slit-like pits (Plate VI, Fig. 21). Towards the lower surface of the midrib the cells are much smaller (excepting the large cells which grow out to form rhizoids), and have thick dark-coloured walls forming a kind of ventral cortex. The cells of this compact tissue sometimes contain coiled chains of Nostoc ; colonies of this Alga are frequently observed in the groove enclosed by the ventral scales and amongst the rhizoids (Plate VI, Fig. 20). Endophytic (perhaps symbiotic) blue-green Algae have long been known to inhabit specialized organs in Blasia and the Anthoceroteae, but apparently their occurrence in the tissues of the Marchantiales has only been recorded hitherto by Reinsch (1877, p. 234), for a Riccia, , and by Mattirolo (1888, p. 7), for Grimaldia dichotoma . Nostoc- colonies have been observed by the writer in the compact tissue of the thallus of Reboulia hemispherical Preissia commutata , and Targionia hypophylla , and it is probable that they are of general occurrence in the thalloid Hepaticae. We may here describe two tissue-elements which are especially well developed in Fegatellai and which occur chiefly in the compact ventral tissue, viz. (1) oil-bodies, (2) mucilage-sacs. The oil-bodies are spherical or ovoid in form, dark brown in colour, and from 20 {jl to 40 /x in diameter. They occur singly in special cells, each of which is nearly filled by the oil-body (Plate VI, Fig. 21, O. c.). Similar bodies occur in many other liverworts, and they were first carefully studied by Pfefifer (1874), who found that they contain water, a proteid substance, oil, and in some cases tannic acid, these constituents being mixed and forming an emulsion, held together by an envelope consisting of proteid matter. Previous to the appearance of Pfeffer’s memoir the oil-bodies of the Marchantiaceae had been noticed and described by various observers. Mirbel, in his classical work on Marchantia polymorpha ^ 835)3 first described them, and believed that they consisted of starch. Gottsche (1842, p. 287), showed that these bodies do not give the reaction of starch ; he found that on treatment with alcohol the contents were dissolved and a sac-like membrane remained behind, and concluded that 92 Cavers . — On the Structure and the bodies consisted of resin or wax. Kuster (1894) has recently confirmed and extended the results of Pfeffer’s work by his exhaustive study of the oil-bodies. He finds that each oil-body consists of a ground-mass in which the oil and other substances are embedded. It would appear that in the living cell there is no special envelope around the oil-body, and that the membrane which becomes visible on treatment with alcohol is simply an artifact due to the action of this reagent on the substance of the ground-mass. The oil-bodies are distributed throughout the compact ventral tissue, both in the thallus and in the sexual receptacles ; they sometimes occur in the walls of the air-chambers, and occasionally in the epidermis. The only parts from which they appear to be invariably absent are the rhizoids and the sporogonia ; the spores contain numerous small oil-drops, but no oil- bodies. It is evidently to the presence of the oil-bodies that the thallus of Fegatella owes its characteristic odour, which can no longer be perceived when the plants are soaked in alcohol and then washed in water. When once the oil-bodies are formed they appear to remain unchanged until the death of the cells containing them ; plants may be kept in darkness for weeks or even months, and the new parts formed are invariably found to contain oil-bodies, whilst those already present in the older parts remain unaltered. These bodies can therefore only be regarded as products of excretion, but they appear to play an important part in the economy of the plant. As shown by Stahl (1888, p. 49), they serve to protect the plant against the attacks of snails, which will shun fresh pieces of the thallus of Fegatella even when there is no other available food. If, however, pieces of the thallus are soaked in alcohol and washed in water, they will be readily eaten by snails which reject the fresh pieces. Mucilage-sacs. In a transverse section through the thallus (PI. VI, Fig. 17), the midrib is seen to contain a number (from one to six) of large rounded elements which stand out sharply from the smaller polygonal cells surrounding them. Comparison with longitudinal sections shows that these rounded cells are arranged in continuous rows which traverse the midrib longitudinally and which may be traced forwards to the growing-point. These cells contain a highly refractive granular substance, which shows a stratified appearance, the layers being concentric with the cell-walls. On treatment with water, these cells increase in volume, their contents absorbing water and becoming homogeneous and transparent ; the transverse walls between the cells of each row at the same time lose their definite outlines and become disorganized, so that these organs have generally been described as continuous tubes. The cells which will give rise to a row of mucilage-sacs are recognizable at an early stage, becoming differentiated just behind the growing-point (PL VI, Fig. 18, M. o.). These cells are distinguished from their neighbours by their dense contents and the absence of chloroplasts 93 Biology of Fegatella conica . and starch-grains. On tracing the cells of a row backwards, the nucleus is found to be broken up and diffused in the protoplasm, whilst layers of deeply-staining mucilage are successively deposited on the inner surface of the cell-wall. The mucilage appears, therefore, to be formed from the protoplasm of the cell, the remains of which can for some time be seen occupying the centre of the cell-cavity (PI. VI, Fig. 20). Besides these long coherent rows of mucilage-sacs, the thallus of Fegatella also contains large isolated mucilage-cells, which are especially abundant in the lamina, in the compact tissue below the chambers ; they also occur abundantly in the sexual receptacles. The development of these isolated sacs agrees exactly with that of the constituent cells of the rows in the midrib. The latter, which are peculiar to Fegatella , were first described by Nees von Esenbeck (1838, Vol. IV, p. 188), who overlooked the mucilaginous contents and described the organs themselves as con- tinuous air-canals traversing the midrib. Goebel (1879, p. 531) was the first to show that these supposed air-passages were in reality mucilage- organs, but his account of their development does not agree with the later description of Prescher (1882), with which the writer’s own observations are entirely in accord. Leitgeb (1881, p. 1 6) believed that the mucilage-organs might serve to confer rigidity on the thallus, and Prescher was also inclined to attribute a purely mechanical function of this kind to these organs, but it seems more reasonable to accept the view, suggested by Goebel, that they act as water- reservoirs. The writer has found that in plants growing submerged in water the midrib shows no trace of mucilage-sacs. Ventral Scales. The development of the ventral scales can be readily followed by examining series of longitudinal sections through the growing-point. Immediately behind the latter, some of the ventral superficial cells grow outwards and forwards, dividing rapidly so as to give rise to plates which remain one cell thick. A superficial cell first grows out and becomes divided by a transverse wall ; the outer cell then grows rapidly in length, forming a long club-shaped mucilage-hair, which curves upwards over the growing-point. This hair then ceases to grow, but the cell below it divides actively and forms a plate, the hair being thrust on to the upper surface of this new outgrowth, which becomes the discoid appendage of the scale (PL VI, Figs. 12-15). In the meantime a zone of cells at the base of the appendage begins to show active growth, keeping pace for a time with the growth in length of the thallus and thus maintaining the position of the coloured scale-appendage in the notch at the apex of the thallus. The fully developed scale may therefore be divided into three portions, each representing a distinct stage in its development, viz. (x) the original 94 Cavers. — On the Structure and mucilage-hair, at first terminal but now standing in the axil of (2) the appendage, which is inserted by a narrow neck on (3) the long and narrow basal portion of the scale. Rhizoids. Each of the rhizoids springing from the lower surface of the thallus is formed by the outgrowth of a single superficial cell, and remains throughout undivided, though often reaching a length of an inch or more. The rhizoids are of two kinds, some being smooth-walled, whilst in others the wall shows numerous peg-like thickenings which project inwards and are arranged in a fairly definite spiral line. The smooth-walled rhizoids spring chiefly from the sides of the midrib, immediately outside the ventral scales, and pass straight down into the substratum, where their ends often become branched (PI. VI, Fig. 25). The tuberculate rhizoids arise in small bundles, each bundle being borne in the axil of a ventral scale, and these unite to form the compact bundle which occupies the median groove on the lower surface of the midrib. The tuberculate rhizoids generally end freely in the median bundle and do not become branched. The superficial cell which will produce a rhizoid is recognizable at an early stage on account of its large size and densely granular contents. This cell projects from the surface and grows enormously in length, but no cell-divisions occur, the rhizoid being simply an elongated cell (PI. VI, Fig. 26). The young rhizoid shows strictly apical growth in length, the protoplasm and nucleus being found near the growing tip, whilst further back the rhizoid contains only cell-sap. The plain rhizoids give the reactions of cellulose, but in the tuberculate ones the tubercles themselves become altered in composition, and on treating the rhizoid with sulphuric acid the tubercles alone remain unaltered after the rest of the cell-wall has been dissolved. Leitgeb (1881, p. 20) suggested that the tuberculate rhizoids of the Marchantiaceae not only shared with the smooth ones the function of attaching the plant to the soil and of absorbing water, but might also have a third function, namely, that of strengthening the thallus. As pointed out by Kamerling (1897, p. 12), however, the turgor of the cells composing the ventral tissue of the thallus is quite sufficient to maintain the rigidity of the latter, which remains unaltered when the median bundle of tuber- culate rhizoids is cut through at various points. Kny (1890, p. 371) believed that the tubercles would serve to prevent the walls of the rhizoids from sinking in through lateral pressure or lack of water; Haberlandt (1896, p. 196), that they increase the area for absorption, these ingrowths causing the protoplasmic lining of the cell-wall to become spread out. There is little to be said in favour of either of these conjectures, for (1) neither the plain nor the tuberculate rhizoids collapse to any great extent when plants 95 Biology of Fegatella conic a. are kept dry, and (2) the fully developed rhizoids can be shown by plasma- lysis to have entirely lost their protoplasmic contents, whilst in younger ones the protoplasm is only found at the extreme apex, where tubercles are absent and where the absorption of liquids is shown by experiment to be most active. Kamerling demonstrated the existence of negative pressure in the rhizoids, by placing on a razor a drop of water which contained carmine- powder in suspension, cutting through a bundle of rhizoids, keeping the cut ends in the liquid for a few seconds, and then spreading out the upper part of the bundle (i. e. that attached to the thallus) on a slide and examining the rhizoids under the microscope. In the case of the plain rhizoids, the negative pressure is very marked, the liquids passing in for a distance of several millimetres, but in the tuberculate rhizoids the carmine particles are arrested by the tubercles immediately beyond the cut end. As already stated, the tuberculate rhizoids spring from the angles between the ventral scales and the surface of the midrib, the scales evidently serving to protect the rhizoids against evaporation and mechanical injury, besides forming narrow spaces in which water is retained by capillarity. The apical appendage of each scale, after it has fulfilled its function of protecting the growing-point of the thallus, becomes, in con- sequence of the active growth of the latter, thrust on to the ventral surface and then soon withers. The basal portion of the scale, however, persists and covers the ventral groove in which lie the bundles of tuberculate rhizoids (Pl. VI, Fig. 17). In the submerged aquatic form of Fegatella , a few rhizoids are borne in the axils of the reduced ventral scales, but though these rhizoids correspond in position with the tuberculate ones of terrestrial plants, the tubercles are almost entirely wanting. This observation lends support to the view that the tuberculate rhizoids are adapted for the storage of water, whilst the plain rhizoids serve to conduct water and to attach the thallus to the substratum. Mycorhiza. The thallus of Fegatella is frequently infested by the hyphae of a Fungus, regarded by Beauverie (1902) as a Fusarium. These hyphae are usually confined to a zone of cells underlying the air-chambers in the midrib, but are sometimes found also in the compact tissue of the lamina on either side (PI. VI, Fig. 22). They penetrate the cell-walls and often become branched and coiled up within the cells, bearing here and there swollen vesicles, which may be either terminal or intercalary in position, and in the latter case are often arranged in chains (PI. VI, Figs. 23, 24, Ves). The writer has found that these vesicles are of two kinds. Those formed during summer are thin-walled, usually aggregated in chains, and some- times become ruptured, so that the cells of the thallus become filled with 96 Cavers . — On the Structure and the densely granular fungus-protoplasm. The vesicles formed in autumn are usually terminal and isolated ; they have thick walls and evidently function as chlamydospores, from which new hyphae grow out in spring, when the growth of the thallus is resumed. The plants containing the Fungus are larger and show more vigorous growth than those free from hyphae, and there can be little doubt that we have here a definite sym- biosis, the Fungus forming a mycorhiza by means of which the life of the Fegatella- plant becomes to a certain extent saprophytic at the expense of the humus on which it is growing. Golenkin (1902) has recently studied the mycorhiza of various Hepa- ticae, including Fegatella , and states that the infected cells never contain starch or chloroplasts. This is not quite true in the case of Fegatella , according to the present writer’s observations, for here the fungal hyphae may be seen traversing cells which contain starch-grains, and in a few cases the hyphae were seen to have also invaded the chlorophyll-bearing cells that form the floor of each air-chamber. Golenkin suggests that the func- tion of the mycorhiza in the Marchantiaceae is that of storing water and enabling the plant to resist drought, but this explanation will hardly apply to a thoroughly hygrophilous form like Fegatella , which possesses in its mucilage-sacs a special tissue that may very reasonably be considered as fulfilling the requirements of water-storage in an exceptionally complete manner. Since the observations of Beauverie and of Golenkin are in many respects lacking in detail, and do not form a sufficient basis for conclusions as to the biological importance of the mycorhiza in the Marchantiaceae, the writer has made a series of cultures, the results of which, though in some respects incomplete, may be here briefly given. On sowing ripe spores in heat-sterilized soil it was found that a relatively small number of young plants was obtained, and that these were always poorly developed, the thallus being long and narrow; the ventral scales were small and remained green throughout, and only smooth- walled rhizoids were formed. In no case were any fungal hyphae to be found in the tissues. For comparison, an approximately equal number of spores (the entire contents of a ripe capsule) was sown in ordinary garden soil, under similar conditions as to light and moisture. A much larger proportion of young plants was obtained, and in many cases these plants showed broad thallus- lobes with well-developed air-chambers, scales with violet-coloured appen- dages, and tuberculate rhizoids. In the smaller plants the thallus was free from fungal hyphae, but these were found abundantly in the larger and more vigorous plants. A third lot of spores was sown in rich humus (peaty soil). The young plants were more numerous and showed more vigorous growth than in the Biology of Fegatella conic a. 97 other two lots. The thallus was larger and of a deeper green colour, and in every case examined there was a well-marked mycorhizal zone. Similar series of cultures were made with very young adventitious plants, removed from old plants that had been found to be free from fungus, and also with the bulbils to be described presently. The results obtained agreed exactly with those just described for the spores. In the case of the bulbils, however, it may be mentioned that these structures themselves were frequently found to contain fungal hyphae at the outset. Asexual Reproduction. By the dying away of the older parts of the thallus, the branches become separated from each other and then grow out into independent plants. Besides this simple means of propagation, the observations of Schostakowitsch (1894), which have been confirmed by the present writer, show that the thallus of Fegatella possesses in a marked manner the power of regeneration. A plant is cut up into small pieces, which are cultivated on damp soil. After four or five days, given sufficient warmth, several shoots are seen to grow out from the lower surface of each piece of thallus, along the cut edge which in the intact plant was nearest to the apical growing-point. Each of these adventitious shoots arises through the active growth and division of a single superficial cell, forming at first a cylindrical body which grows out and after a time begins to become differentiated in the usual manner. When the cultures were made in dark- ness, new shoots were formed quite as freely as in the light, but they were long and narrow and did not grow into normal plants with typical air- chambers and scales ; on being brought into the light, they gave rise to normal plants, otherwise they remained abortive after attaining a length of 2 to 3 cm. According to the present writer’s observations, the property of re- generation is much more restricted in Fegatella than was found by Vochting (1885) to be the case in Marchantia and L unularia. In all cases, the young plants were found to arise from the compact tissue underlying the air-chambers, and attempts to induce the formation of new shoots from the sexual receptacles and from the sporogonia gave negative results. As already stated, the older plants become covered up by the new shoots, so as to form a kind of turf consisting of several superposed layers of plants. Along the ventral surface of these old plants, which have become for the most part brown and withered, there are frequently found numbers of small spherical or ovoid outgrowths, the bulbils or tubers, which appear to have been first described by Karsten (1887). In these tuber-forming plants, the cells of the compact tissue immediately within the ventral superficial layer of the midrib grow and divide actively, giving rise to a mass of tissue which later breaks through the superficial layer H 98 Cavers. — On the Structure and and forms a stalked outgrowth. The tuber ultimately becomes detached and gives rise to a new plant, but it does not appear to be specially adapted for resisting drought, for tubers that have been kept dry (between sheets of filter-paper) for even a week were found by the writer to be very rarely capable of germination. The tuber bears numerous rhizoids, and its cells contain starch-grains ; the superficial cells, except those that grow into rhizoids, have their outer walls cuticularized. In several cases abundant fungal hyphae were found in the cells of the tuber ; the fungus- infested tubers were found to germinate freely, and the hyphae evidently belong to the mycorhiza of the thallus. Sexual Organs. Fegatella is strictly dioecious, and the sexual organs are borne on specialized portions of the plant, the receptacles. Each receptacle repre- sents a branch of the thallus, or a system of branches. When a resting winter-shoot grows out, the four growing-points may either all give rise to ordinary broad thallus-lobes, or one of them may produce a sexual receptacle. The receptacles are laid down in spring, but the sexual organs do not become mature until about the end of June. The earliest stages in the development of the sporogonium are observed about the middle of July, as a rule, but the spores are not set free until the spring of the following year. Male Receptacle. The antheridial receptacle is sessile, forming an oval cushion studded with the openings of numerous cavities, each of which is found, on examin- ing sections of the receptacle, to lodge a single antheridium, or sometimes two antheridia. Each of these antheridial pores occupies the summit of a conical prominence, and between them there occur numerous less con- spicuous pores, each opening into an air-chamber (PI. VII, Figs. 1 8, 31). At the anterior end of the receptacle are seen the appendages of the ventral scales, which curve upwards and backwards in the same manner as in the growing-point of a sterile branch. The receptacle is at first bright green in colour, but later the walls of the antheridial cavities, together with the prominences just mentioned, assume a deep red or purple colouration. The examination of serial sections through receptacles of different ages shows that the apex of the male branch undergoes repeated forking, so that several (usually from six to eight) growing-points are formed. Each of the latter then gives rise to a zig-zag row of antheridia, the oldest being found at the centre of the receptacle and the youngest at the margin (PI. VII, Figs. *9-31). Biology of Fegatella conic a. 99 Antheridium (See Fig. 28). The superficial cell which will give rise to an antheridium projects above the surface and becomes divided by a transverse wall. The lower undergoes little further development, whilst the upper undergoes division by a series of transverse walls, so that the young antheridium consists of a row of cells, four or five in number (Fig. 28, V). The lowest cell of this row forms the short stalk, the body of the antheridium being derived from the upper cells, in each of which there next occur two vertical divi- sions, the walls intersecting each other at right angles and thus dividing the cell into quadrants (II, VI). The next divisions are also vertical, but the walls lie parallel with the surface of the antheridium, so that each tier of cells becomes divided into four central cells (sperm-cells) and four peripheral cells (wall-cells). The sperm-cells (III, IV, VII) are distinguished from the wall-cells by their denser protoplasmic contents, they undergo repeated divisions and give rise to the antherozoids, whilst the wall-cells divide only by anti- clinal septa (i. e. septa perpendicular to the outer surface of the antheri- dium) and form the single-layered antheridial wall (VIII, IX). The sperm-cells divide with great regu- larity, remaining nearly cubical in form, and the first-formed walls suffer very little displacement, being easily traced in the nearly mature antheridium (IX). The writer hopes shortly to publish a detailed account of the development of the anthero- zoid in Fegatella. Each of the cells composing the wall of the antheridium contains a few large chlorophyll-grains, the colour of which usually changes from green to yellow or red during the ripening of the antheridium. The cells are flattened, except at the apex of the antheridium, where there is a pointed beak made up of elongated cells. The development of the antheridia is accompanied by that of air- H 2 Fig. 28. I-IV. Transverse sections of develop- ing antheridia. V-VI1I. Longitudinal sections of corresponding stages. IX. Longitudinal section of well- grown antheridium, showing the regular arrangement of the sperm- cells. I- VIII, x 160 ; IX, x 80. IOO Cavers.— On the S true hire and chambers, which arise in essentially the same manner as those of the thallus. The dorsal portion of the receptacle thus acquires a spongy character, whilst the ventral portion consists of compact colourless tissue with mucilage-sacs and oil-cells. The air-chambers are at first long and narrow, each opening above by a pore which is surrounded by three or four superposed rings of cells (PL VII, Fig. 31, /.). Pores of this type, generally termed ‘ compound ’ or ‘ barrel-shaped/ are also found on the female receptacle, where they are more highly developed. Each antheridium becomes sunk in a deep cavity, formed in essentially the same manner as the air-chambers (PI. VII, Figs. 29-31). As the antheridium grows in size, its cavity becomes flask-shaped, with a long narrow canal opening above on the surface of the receptacle by a very narrow pore (PI. VII, Fig. 30). In consequence of the lateral pressure exerted by the growing antheridia, the air-chambers between the antheridial cavities become greatly compressed below, but in the upper portion of the receptacle they remain as wide spaces, each opening above by a barrel-shaped pore. The cells lining the air-chambers grow out here and there into filaments resembling those found in the air-chambers of the thallus. The male receptacle is therefore well provided with assimilating tissue. Water is taken up by means of the numerous smooth and tuberculate rhizoids which spring from the lower surface of the receptacle (PI. VII, Fig. 30) ; part of this water is given off as vapour by the green cells lining the air-chambers, whilst large quantities can be stored up in the colourless cells of the ventral tissue. The margin of the receptacle is beset with scales derived from the growing-points, and the tissue of the thallus behind and on either side of the receptacle grows up to form a partial sheath. In several cases the writer has found two antheridia occupying a common cavity, the sides along which they were in contact being strongly flattened (Fig. 29). The two antheridia were found, in those cases where fresh sections were examined, to be so closely joined that it was impossible to separate them without injury, and in a few cases the appearance of young pairs of antheridia strongly suggested the possibility of their having arisen from a common mother-cell. Similar examples of paired antheridia have been described by Leitgeb (1879, p. 67, Taf. 8, Fig. 17) for Sphaerocarpus , and by Campbell (1896, p. 500) for Geothallus , but they do not appear to have been hitherto recorded in the Marchantiaceae. The cells forming the wall become mucilaginous, and on adding water to a ripe antheridium, these cells become swollen. The delicate walls of the antherozoid mother-cells also swell up in the same manner and ultimately become dissolved. Mucilage is also formed by a number of long club-shaped cells (paraphyses) which grow up from the bottom of each antheridial cavity, around the base of the antheridium. In consequence of the pressure set up, the antheridial cavities eventually come to be separated IOI Biology of Fegatella conica. from each other only by a few layers of flattened cells; the paraphyses are flattened between the antheridium and the walls of the cavity. The cells forming the beak of each antheridium are ultimately thrown off, and the semi-liquid contents of the antheridium are violently forced out of the opening at the top of the canal, arising from the surface of the receptacle in the form of jets of spray, as has already been described by the present writer (1903 A). A closely similar process had been described in the case of Asterella calif ornica by Peirce (1902), whose observations were not known to the writer at the time of penning the note just referred to. In Fegatella , the jets of spray containing the antherozoids continue to be emitted from the same receptacle for several minutes, and were mostly observed on warm, sunny days. In most cases the jets reached a height of about 5 cm., but in a few cases the distance was greater, up to about 10 cm.1 Similar explosive discharges of antherozoids have been observed by the writer in Reboulia hemispherical Marchantia polymorpha , and Preissia commutata, living plants of which were kept under cultivation in the laboratory, and it appears probable that this phenomenon will be found to be of general occurrence in the Marchantiales, where the antheridia are embedded in deep cavities sunk in the tissue of the receptacle, and com- municating with the exterior by narrow canals. The biological importance of this mode of discharging the antherozoids is readily seen when it is observed in the dioecious forms, e. g. Fegatella and Marchantia. Preissia is generally also dioecious, but sometimes monoecious, and the same is the case with Reboulia. In Fegatella and Marchantia , especially, the female plants are often found to be removed several inches from the nearest male plants, and it is possible that the antherozoids on being ejected explosively from the male receptacles are carried by air-currents to the female plants. No doubt the same result is secured by rain-drops falling on the ripe male receptacles and then splashing over the female plants, as suggested by Goebel (1898, p. 310). Female Receptacle (Carpocephalum). The female receptacle arises just below the anterior margin of the thallus as a rounded outgrowth, which increases in size and becomes pushed forwards and upwards until it stands on the dorsal surface of the thallus (PI. VII, Figs. 38-41). The archegonia are formed on the sides of this knob-like outgrowth, but are soon carried downwards on to its lower surface. A receptacle ready for fertilization has already attained the conical form to which the plant owes its specific name as well as the generic name ( Conocephalus ) used by some writers. It has a short stalk, See Postscript on p, 114. 102 Cavers. — On the Structure and though appearing to be sessile at the anterior margin of the thallus (PL VII, Fig. 32) ; it is protected in front by the overlapping ventral scales, behind and at the sides by a sheath formed by upgrowth of the surrounding tissue Fig. 29. Vertical longitudinal section of female plant, traversing a receptacle. A. C. air- chambers ; P. pores ; Rhiz. rhizoids springing from the lower surface of the receptacle and passing down the groove in the stalk ; V. S. ventral scales ; Arch, archegonium. X 15. One of the * barrel- shaped ’ pores is shown above, in vertical section, x 120. Fig. 30. Vertical transverse section of female plant, traversing a receptacle. Lettering as in Fig. 29. X 1 5. Biology of Fegatella conica. 103 of the thallus (Figs. 29, 30). After fertilization of one or more of the archegonia, the development of the sporogonia begins and is continued during the rest of the year. In the following spring, the stalk of the receptacle, which has up to this time remained very short, begins to grow in length and in a few days reaches a height of from 3 to 6 cm., carrying up the receptacle (PI. VII, Fig. 33). Then the capsules open and the spores are set free. The earliest stages in the development of the female receptacle are a little difficult to follow, but it would appear that here, as in the case of the male receptacle, the growing-point undergoes repeated branching. Leitgeb (1881, p. 94) leaves the question open, though bringing forward strong theoretical grounds for regarding the carpocephalum as a branch- system. The examination of numerous series of sections through young receptacles has convinced the present writer that branching does take place in the growing-point, repeated dichotomy giving rise typically to eight growing-points, which are separated by rounded sterile lobes, exactly as in the branching of the thallus. The carpocephalum of Reboitlia hemi- spherica does not differ greatly in external form from that of Fegatella , but microtome-sections of the former show clearly the single growing-point lying below and in front of the receptacle. This is not the case in Fegatella , and there can be no doubt that here the apex of the fertile branch is used up in the formation of the whole receptacle, not merely of the stalk, and that the receptacle itself represents a modified branch-system, essentially homologous with the male receptacle. The young receptacle first appears as a dome-like prominence, formed by active growth of the dorsal segments of the initial cells (PI. VII, Fig. 38). These cells undergo repeated vertical divisions, so that the tissue of the young receptacle presents in horizontal sections a nearly circular outline, the cells being arranged in radiating rows (PI. VII, Fig. 42). The next stage shows a slightly rounded lobe occupying the anterior end of the re- ceptacle and recalling the ‘ middle lobe ’ seen in ordinary apical branching of the thallus. The cells composing this anterior lobe grow and divide very slowly, but those occupying the slight depression on either side of it divide actively, especially by vertical walls, and obviously constitute two distinct growing-points (PI. VII, Fig. 43). Next, each of these growing- points becomes broadened, and in its centre there appears a secondary lobe (PI. VII, Fig. 44). The receptacle now shows four growing-points, but each of these rapidly grows in breadth, and its central cells grow out to form a tertiary middle lobe (PI. VII, Fig. 45). In most cases, the branching of the young receptacle does not proceed further than this, but sometimes one or both of the two anterior growing-points again becomes branched, so that the receptacle may show nine or ten growing-points instead of the eight which are normally found. On the other hand, one or more of the divisions 104 Cavers. — On the Structure and that normally occur may be suppressed, so that the receptacle may present only six or seven growing-points. The dorsal segments of each growing-point give rise at first to sterile tissue, in which air- chambers are formed in regular acropetal succession (PL VII, Fig. 46), whilst the ventral segments give rise to tuberculate rhizoids and ventral scales. Each of the latter consists simply of a club-shaped mucilage-hair terminating a short row of cells ; sometimes the cells below the club- shaped hair undergo longitudinal divisions, giving rise to two or three rows of cells. Whilst the branching has been taking place the whole receptacle has become hemispherical in form. The tissue of the receptacle grows much more rapidly in all directions than does the tissue which lies below and behind it and which gives rise to the receptacle-stalk, so that around the insertion of the stalk there is formed a deep but narrow annular groove. This is continuous in front with the rhizoid-bearing groove on the anterior face of the stalk, which in its turn becomes confluent with the ventral furrow of the thallus. In addition to tuberculate rhizoids, this groove bears two longitudinal rows of scales, which are smaller and simpler in structure as they are traced upwards, every possible transition being found between the ordinary ventral scales of the thallus, with their coloured discoid appendages, and the greatly reduced scales already described as being formed behind each of the growing-points of the receptacle. Each growing-point gives rise to a single archegonium. The first archegonium invariably arises on the hinder margin of the receptacle, and it is closely followed by the appearance of the second and third at the sides of the receptacle, the latest-formed one being nearest to the front of the re- ceptacle. Ultimately we find from six to nine archegonia in all, standing singly on the margin of the receptacle at approximately equal distances from each other. These early stages appear to be passed through very rapidly, and the archegonia are frequently found to be all in nearly the same stage of development, though the anterior ones are invariably less advanced than those occupying the hinder portion of the receptacle. Their development is obviously not simultaneous in any case, though it may sometimes be nearly so. On comparing the carpocephalum of Fegatella with that of M archantia polymorpha , the organization of which was long ago worked out correctly by Mirbel (1835), it will be seen that the only points of essential difference are (1) the absence in Fegatella of the long hollow sterile lobes which are so characteristic of M archantia ; (2) the development in Fegatella of a single archegonium from each growing-point, instead of the group of archegonia found in Marchantia between each pair of sterile lobes. The higher Marchantiaceae, i. e. those having a stalked female recep- tacle, were divided by Leitgeb (1881, p. 49) into three sections, to which he gave the names Astroporae, Operculatae, and Compositae, and this Biology of Fegatella conic a. 105 classification is adopted by Schiffner (1893) in Engler and Prantl’s ‘ Pflanzenfamilien.’ The female receptacle of the Compositae was shown by Leitgeb to consist of a branch-system, having a number of growing- points, from each of which archegonia are formed in acropetal succession ; to this group Leitgeb referred Marchantia , Preissia , Lunularia , Dumortiera , and, doubtfully, Fegatella . In the Astroporae and Operculatae, however, the female receptacle was believed by Leitgeb to be merely a dorsal outgrowth of the thallus, formed behind the growing-point, which might in some cases ( Clevea , Plagiockasma) continue to grow and to produce several receptacles in succession. Recent observations have shown, however, that the distinction drawn by Leitgeb between the receptacle of his * Compositae * and ‘ Operculatae ’ cannot be maintained. Thus Cryptomitrium and Fim- briaria have generally been regarded as belonging to the lower type, but Abrams (1899) has shown that in the former genus the receptacle is formed directly from the apex of the shoot, which divides to form five or six growing-points, each producing about five archegonia in acropetal succession ; and Campbell (1895, p. 57) found that Fimbriaria calif ornica also shows a typical composite receptacle, the apex branching to form four growing- points, each of which gives rise to two or three archegonia. Moreover, Solms-Laubach (1897) has shown that although the genus Exormotheca is referred by Schiffner (1893, P* 29) to the Astroporae of Leitgeb, its bilobed receptacle presents two groups of archegonia, each group containing as many as five, though the number may be reduced to one ; hence this form also would belong to the Compositae rather than to the lowest group of the Marchantioideae. From the considerations here brought forward, it would appear that Fegatella represents the simplest type of the Compositae, the female receptacle being a branch-system, in which each growing-point usually gives rise to a single archegonium. In this respect Fegatella forms an interesting member of the series of forms that bridge over the gap between the two extremes — the simple dorsal outgrowth of Clevea and Plagiockasma and the elaborate branch-system of Marchantia polymorpha . Archegonium (Fig. 31). Since the number of archegonia developed in each receptacle is so small, and their differentiation is practically simultaneous, Fegatella is not a very suitable form in which to study in detail the cell-divisions that occur in the young archegonium. However, sufficient stages were observed to show that in this respect Fegatella agrees closely with the descriptions given for M archantia by Strasburger (1870) and by Kny (1890), for Preissia by Janczewski ([872, p. 386), and for Targionia and Fimbriaria by Campbell (i895, p. 52). Each archegonium arises from a superficial cell which projects above the surface and becomes divided by a transverse wall ; the io 6 Cavers. — Oil the Structure and lower cell undergoes irregular divisions and gives rise to the short stalk, whilst the upper cell becomes divided by three vertical walls which intersect each other so as to separate a central cell from three outer cells (Fig. 31, I-IV). The central cell, which is triangular in cross-section, next divides by a transverse wall, cutting off a small upper cell (‘cover-cell’) from a large lower cell (V). The cover-cell then divides into four by vertical walls that intersect at right angles (VI, X), whilst each of the three outer cells divides into two by a radial longitudinal wall (XI). Transverse divisions then occur both in the large central cell and the six outer cells, so that the young archegonium now con- sists of two tiers of cells, the upper forming the neck and the lower the venter (VI). The axial cell of the neck undergoes repeated trans- verse divisions, giving rise to the row of neck-canal- cells, whilst the axial cell of the venter divides by a single transverse wall, cut- ting off a small upper cell (ventral-canal-cell) from the large egg-cell (VII, VIII). The four cover-cells either remain undivided or undergo longitudinal divisions, so as to increase in number to six or seven in the mature arche- gonium. The neck-cells di- vide repeatedly by transverse walls, giving rise to the long neck, which eventually consists of about twenty tiers of cells ; the neck-canal-cells do not divide so rapidly, and there are never more than eight or nine altogether (Figs. 30, 31). During its development the archegonium becomes carried downwards on to the lower surface of the receptacle, whilst the tissue around it grows out to form a sheath, the tissue of which contains air-chambers ; beyond the opening of this sheath hangs the long neck, which is curved upwards and outwards. In a paper by Gayet (1897), we find an account of the development of the archegonium in various Hepaticae which differs considerably from that here given for the archegonium of Fegatella , as well as from those Fig. 31. I- VIII. Longitudinal sections of developing archegonia. The small figures above, I-IV, show the cor- responding stages in transverse section. IX. Venter of a mature but unfertilized archegonium ; the upper portion of the egg-cell is hyaline, forming a ‘ receptive spot.’ X. Cover-cells in transverse section, corresponding to stage shown in VII or VIII. XI. Transverse of neck, showing six neck-cells and a neck-canal-cell, x 360. Biology of Fegatella conica. 107 of all previous authors who have made a special study of this organ in the Hepaticae (Kny, Strasburger, Janczewski, Campbell). Gayet states that in all the Hepaticae examined by him, the cover-cell of the young archegonium acts as an apical cell, at any rate for some time, segments being cut off from it which contribute to the growth in length of the neck. Gayet’s main conclusion is that both in Hepaticae and in Mosses the neck-cells are, at any rate to a large extent, derived from the segmenta- tion of an apical cell (the cover-cell), and that in both groups the neck- canal-cells arise by division of the primary canal-cell — in short, that there is no essential difference in the development of the female organ in the two groups, such as has been generally held to exist. The Marchantiaceous forms studied by Gayet were Targionia , Preissia , and Marchantia. Owing to the fact that the archegonia of Preissia and Marchantia are developed in large numbers in each group, it is much easier to obtain a good series of stages than in the case of Fegatella , where each receptacle shows only seven or eight archegonia altogether. The result of the writer’s observa- tions on Preissia commutata and Marchantia polymorpha , using stained serial microtome- sections, has been to entirely confirm the accounts given by Strasburger, Janczewski, and Campbell for these and other Marchanti- aceae. Gayet’s figures are not convincing, and his methods of preparation are, as has been pointed out by Campbell (1898), quite inadequate for the exact determination of the points in question, for unless recourse be had to modern methods of fixation, paraffin-embedding, and serial sectioning by microtome, it is almost impossible to make out the precise sequence of nuclear and cell divisions in a developing organ which consists of a solid aggregate of cells. The tissue of the receptacle, above and between the archegonia, contains numerous air-chambers, separated by thin partitions, consisting mostly of a single layer of cells (Figs. 30, 31) ; here and there we find large mucilage- containing cells, both in the compact tissue and in the walls of the chambers. The pores are of the compound or barrel-like type ; the cells forming the upper tiers are narrow, whilst the lower tiers, projecting into the chamber are broader, the lowest cells being nearly spherical in cross-section. If sections of a fresh receptacle be examined in water, it is found that the cells- of this lowest ring are so arranged as to leave a wide opening into the underlying air-chamber ; but on irrigating the section with a five per cent. KN03 solution, these cells become flaccid and the opening becomes very much smaller, or the pore may become completely closed, the cells of this ring coming into contact with each other. On adding water, the cells resume their former state of turgescence and the pore again becomes open. It appears from this experiment that the lowest tiers of cells surrounding these barrel-shaped pores act in the same way as the guard-cells in the stomata of higher plants, and have the function of regulating the opening io8 Cavers. — On the Structure and and closing of the pores. This is not the case with the pores on the thallus, which remain permanently open when similar experiments are tried. On the lower surface of the receptacle, immediately around the insertion of the stalk, there is a circular groove from the surface of which there spring numerous tuberculate rhizoids. Some of these rhizoids pass into a shallow groove on the anterior face of the stalk (PI. VII, Fig. 47). This groove is continuous with the ventral groove of the thallus, and the rhizoids traversing it become merged in the median bundle of the midrib. Many of the rhizoids springing from the receptacle remain outside of this groove and hang down over the surface of the stalk. The canal-cells of the mature archegonium become disorganized and converted into mucilage, and on the absorption of water the walls between these cells break down, the cover-cells being at the same time thrust aside and the mucilage being forced out of the open neck. The process of fertilization can be readily followed in Fegatella , owing to the comparatively large size of the antherozoids. A male plant with well-grown receptacles is kept dry for a few days, then a few drops of rain-water are placed on the upper surface of the receptacle ; on drawing the water off with a pipette, it is usually found to contain large numbers of swarming antherozoids. Longitudinal sections are then made through a female receptacle, the razor being kept dry. In most cases, one or more of these sections will be found to include a mature archegonium, and on mounting such a section in water, it is often possible to see the opening of the neck and the discharge of the canal-cells. On adding to such a preparation some of the water containing the antherozoids, the latter are seen to swarm around the open neck of the archegonium in large numbers, being evidently attracted by the extruded mucilage. Since the cells forming the archegonium neck and venter are usually fairly transparent, one can follow the passage of the antherozoids down the neck-canal. In some cases large numbers of anthero- zoids were seen to enter the canal, and when they reached the egg-cell they caused it to exhibit a rocking movement. Ultimately one of the antherozoids penetrates the egg-cell, whilst the others perish. Sporogonium (Fig. 32). The fertilized egg-cell secretes a cellulose wall around itself and then begins to grow and divide. The first wall is transverse (I), and it is some- times followed by a second transverse wall in the upper (epibasal) cell, but as a general rule the next walls are vertical and divide the embryo into regular octants (I, II). From the first, the embryo grows more vigorously in length than in breadth, and soon the lower (hypobasal) portion becomes separated from the upper by a constriction, the former giving rise to the foot, the latter to the stalk and capsule (VII). The foot becomes elongated and penetrates the compact tissue of the receptacle, the cells of which contain Biology of Fegatella conic a . 109 abundant starch-grains. In the upper portion of the sporogonium there is evident at a relatively early stage a well-marked outer layer which gives rise to the wall of the capsule, the inner cells forming the archesporium , from which arise the spores and elaters. Spores. The archesporial cells are at first all alike, but later they become differentiated into two kinds, which show no definite relation to each other Fig. 32. I. Venter of fertilized archegonium, with eight-celled embryo (octant-stage), four cells being seen in the section. II. Transverse section of a similar embryo. Ill, IV. Older embryos in longitudinal section. V. Part of longitudinal section of developing capsule, showing differentiation of archesporial tissue into long narrow elater-forming cells (el.) and nearly isodiametric sporogenous cells (sp.). VI. Part of longitudinal section of receptacle, showing a young sporogonium in outline. VII. Similar section, showing two mature sporogonia in outline, a. c. air-chambers ; cal. calyptra ; caps, capsule ; f. foot ; n. withered neck of archegonium ; p. pore ; rec. si. stalk of receptacle ; rhiz. rhizoids ; s. seta of sporogonium. I-V, x 360 ; VI, x 76 ; VII, x 20. beyond a tendency to become arranged in longitudinal rows (Fig. 33, V). Some of the cells grow in length and remain narrow, whilst the others grow almost equally in all directions, and are further distinguished from the long cells by containing more densely granular protoplasm. The larger cells eventually become rounded off and constitute the spore mother- cells. The nucleus divides into two, and then each daughter-nucleus divides I TO Cavers . — On the Structure and again, so that the mother-cell contains four free nuclei, which move apart and take up equidistant positions near the periphery of the mother-cell. As shown by Farmer (1895), a disc of cellulose next appears near the centre of the mother-cell, and this becomes joined by ingrowths from the periphery, so that the four spores become partitioned off. The spore is at first pyramidal in form, having three flat sides meeting at the truncated apex of the pyramid and marking the surfaces where the spore was in contact with the other three cells of the tetrad. The coat is thin, but is differentiated into two layers, the inner being thin and homogeneous, the outer thicker, light brown in colour, and bearing externally numerous small papillae (PI. VII, Fig. 53). The spores begin to germinate whilst still enclosed within the capsule, each becoming divided up so as to form an ovoid mass of cells. In this respect Fegatella differs from the remaining Marchantiales, so far as known, though the same thing occurs in Pellia and in some species of Dendroceros. Elaters. The elongated cells which are mingled with the spore mother-cells remain sterile and give rise to the elaters. They continue for a considerable time to grow in length, their tapering ends passing in between the neighbouring cells, and then part of the protoplasm becomes arranged in a double spiral band on the inner surface of the cell-wall, whilst the remainder of the protoplasm, together with the nucleus, occupies the axis of the elater. Starch-grains are at first present, but these, together with the nucleus and the axial strand of protoplasm, eventually disappear, whilst the spiral band becomes thicker and at the same time assumes a brown colour. On tracing one of the turns of the spiral from one end of the elater it is usually seen to branch on its way towards the other end, and branches later, becoming united again ; hence each end of the elater shows a small loop, whilst at the middle there may be from three to five parallel bands (PL VII, Fig. 52 a). The typical fully-grown elater is spindle-shaped, from 0.15 mm. to 0.25 mm. in length. As a rule, one end is blunt and rounded, the other pointed and tapering, and there is generally a slight bend near the middle and often also at each end. The elaters often assume very curious shapes owing to the occurrence of branching. Nearly every capsule examined by the writer showed a certain proportion of branched elaters. The branching takes place at a relatively late period, after the spiral bands have been deposited but before they have assumed their final deep brown colouration, and after the spore-tetrads have become separated. In the hundreds of young sporogonia examined by the writer not a single case of branching was observed prior to the rounding-off of the spore mother-cells and their Biology of Fegatella conica , 1 1 1 division into tetrads of spores. Moreover, the examination of longitudinal sections of the sporogonium shows that in nearly every case the branching really consists in a forking of that end of the elater that lies nearest the apex of the capsule, and that this is in the typical unbranched elater blunt and rounded, whilst the lower end tapers to a sharp point (PI. VII, Fig. 52 b , c). From these facts it is obvious that the branching of the elaters is brought about by the loosening of the spores, in consequence of which the pressure on the upper portion of each elater becomes diminished, whilst its lower portion remains wedged in between the spores. In the lower part of the capsule the elaters either remain unbranched or slight branching takes place at their upper ends ; towards the middle of the capsule branching takes place more freely, but still only at the upper end of the elater ; but nearer the apex of the capsule branching may take place at both ends of the elater, the branches insinuating themselves between the spores. The writer has observed branching elaters in Reboulia hemispheric a and Targionia hypophylla , where the relative arrangement of spores and elaters is much the same as in Fegatella , the elaters being relatively short and the spores being loosely packed, whereas in Marchantia and Preissia , where the spores and elaters are arranged in extremely regular longitudinal series, the elaters being very long and narrow for the most part, branching does not appear to occur. It is hardly necessary to describe and figure the various and often bizarre forms assumed by the branched elaters of Fegatella ; this has already been done by a previous writer (Tilden, 1894). In most cases they are Y- or V-shaped, though in a few cases there were found towards the apex of the capsule a few X-shaped elaters in which both ends had become forked. The young elater contains numerous small starch-grains, but these become fewer as the differentiation of the bands proceeds, being evidently used up in the formation of these thickenings and disappearing by the time the bands have become brown. The bands at first, while colourless, give the reactions of cellulose, but later become lignified. It is probable that the primary function of the elaters is to aid in the nutrition of the developing spores. They present a relatively large surface through which food-materials, passing up in solution through the stalk of the capsule, can readily be distributed to the spore-producing cells. This primary function of nutrition, the only one performed by the sterile cells in the lower Marchantiaceae (e. g. Corsinia ), is in the higher forms superseded by a second, namely, that of assisting in the dispersal of the spores. This, however, only comes into operation after the dehiscence of the capsule, and it is probable that even after the elaters have ceased to be living cells and have become thickened, they still serve for the distribution of water to the developing spores. I I 2 Cavers .— On the Structure and Structure and Dehiscence of the Capsule-Wall. The wall of the capsule remains one cell thick, except at the apex, where there is a small but well-marked cap projecting into the cavity of the capsule and consisting of several layers of cells (PL VII, Fig. 48). This cap appears to be derived from the uppermost portion of the archesporium. The cells of the capsule-wall are at first uniformly thin- walled, but thickenings appear later in the form of ring-fibres. At the summit of the capsule the cells are much shorter than in the lower portion, being cubical in form, whereas lower down the cells are elongated. The differentiation of the fibres begins in the upper part of the capsule, and we therefore find a gradual transition in passing upwards from the base, where the fibres are thinnest and about six occur in each cell, to the apex, where each cell contains a single broad ring, which is often so thick as to almost divide the cavity of the cell in two (PL VII, Figs. 49-51). Besides the ordinary free elaters which are mingled with the spores we find a few which are attached at one end to the lowest layer of the apical cap, and others which arise from the bottom of the capsule (PL VII, Fig. 48). These fixed elaters are shorter and thicker than the ordinary free elaters*, and are sometimes, especially at the apex of the capsule, intermediate in structure between the free elaters and the cells of the capsule-wall, bearing ring-fibres in addition to spirally coiled ones. It is easy to isolate a single ripe capsule and to observe the manner in which it opens. Dehiscence takes place on drying, and may be hastened by gently warming the capsule. A line of cleavage runs round the upper portion of the capsule just outside the apical cap. The fissure thus formed is irregular and wavy, but it corresponds on the whole with the junction between the apical cap and the rest of the capsule-wall. Longitudinal splitting then takes place along several (usually from four to six) lines, which extend about half-way down the capsule, dividing the wall into the same number of valves. The apical cap remains intact, either becoming loosened all round and falling off, or remaining attached to one of the valves. As the longitudinal fissures extend towards the base of the capsule the tips of the valves become rolled backwards, thus exposing the mass of spores and elaters. The latter show hygroscopic movements, twisting about as they become dry, and thus helping to loosen the spores, which may then be either blown by the wind or washed away by rain. After the archegonium has been fertilized its neck withers, but the venter grows actively and closely invests the sporogonium, forming the calyptra. Fertilization is immediately followed by the appearance of tangential walls in the cells of the venter (Fig. 32, I); the calyptra ultimately becomes five or six cells thick, and remains intact until the Biology of Fegatella conic a, 1 1 3 sporogonium Is mature. In early spring the stalk of the receptacle becomes suddenly elongated, the receptacle, which has hitherto been practically sessile, becoming in three or four days carried up to a height of from two to six centimetres. The writer has found that this sudden elongation is due to growth of cells already formed, and not to rapid cell-division. Should none of the archegonia of the receptacle be fertilized the stalk remains very short, and the whole structure soon becomes brown and withered, but fertilization of even a single archegonium is soon followed by active cell-division in the receptacle-stalk. This process continues during the autumn and winter, so that in early spring, before elongation has taken place, the stalk consists of longitudinal rows of very short but broad cells, densely filled with small starch-grains (PL VII, Fig. 35). During the elongation of the stalk the starch disappears, and when growth in length has ceased the stalk is found to consist of long cells which contain little but sap (PL VII, Figs. 36, 37). On comparing the length of the whole stalk and the average length of each of its cells, before and after elonga- tion, it is found that the whole of this remarkable growth in length may be accounted for by the simple elongation of the cells, the starch being used up In the formation of the cellulose required to maintain the thickness of the cell-walls as they become stretched out. The process just described is exactly similar to that observed in the elongation of the sporogonium-stalk in Pellia and other Jungermanniaceae, but does not appear to have been described previously in the case of the receptacles. In Marchantia and Preissia the conditions are entirely different from those here described for Fegatella . In these highest forms of the Marchantiaceae the growth of the receptacle-stalk is a gradual process, accompanied throughout by repeated cell- division. In its final elongated condition the receptacle-stalk is a slender, pale- green filament, nearly as delicate in texture as the elongated sporogonium- stalk of Pellia , and very different from the rigid receptacle-stalk of M archantia or Preissia , with its dorsal layer of air-chambers and its deep ventral furrows which contain rhizoids and are completely enclosed by the overlapping marginal sheaths. In Fegatella the receptacle-stalk is, after it has become elongated and has carried up the head of sporogonia, a purely temporary organ ; It only lasts until the spores have been shed. Soon after this has taken place the whole carpocephalum withers, its remains usually becoming covered up by the new thallus-lobes. Hence the botanist who does not begin his outdoor observations before Easter is likely to receive the impression that Fegatella rarely fruits, even when the very patches examined by the belated observer had been a short time before covered with abundant carpocephala. The writer has found Fegatella to be one of the most freely fruiting of the Hepaticae growing in various localities which were visited at regular weekly or fortnightly intervals I Cavers . — On the Structure and 1 14 throughout the entire year. The importance of systematic outdoor observations at all seasons of the year can hardly be over-estimated in the case of the Hepaticae and the Mosses ; even in the middle of winter these plants yield material which is absolutely necessary to the investigator who wishes to follow the successive stages in their life-histories. Before the dehiscence of the ripe capsule takes place its seta grows in length, owing to simple elongation of its constituent cells, the starch-grains originally present in these cells becoming used up in the process. The seta finally becomes as long as, or a little longer than, the capsule itself, so that the latter is thrust out through the ruptured calyptra and projects beyond the opening of the sheath (involucre). It frequently happens that the seta becomes broken off at the base, so that the entire capsule falls from the receptacle. Germination of the Spore. As already stated, the ripe spore begins its germination before the dehiscence of the capsule takes place, so that on being liberated it encloses an ovoid mass of cells, usually five or six in number. At either end of the longer axis there is a bluntly conical cell, distinguished from the middle cell by containing fewer chloroplasts. When the spore is set free and its germination is resumed, one or other of these terminal cells grows out to form a colourless rhizoid (PI. VII, Fig. 54), the exospore not showing any definite rupture but simply becoming stretched out. It appears that here} as in other cases that have been investigated (cf. Zimmermann, 1879 ; Heald, 1898), light is the chief factor in determining the point of origin of the primary rhizoids, which invariably spring from the end which is furthest from the light, whilst the illuminated end grows out to form the apex of the young plant. The latter at first appears to grow by a single initial cell, but soon the typical transverse row is established, and the differentiation of the tissues proceeds as in the adult plant. At first only smooth-walled rhizoids are formed, whilst the ventral scales are simpler in structure than in the fully-grown thallus, but soon the typical structure of the latter is attained with the appearance of tuberculate rhizoids, coloured scale-appendages, mucilage-sacs, oil-cells, and well-developed air-chambers. The spores will usually germinate promptly on being sown directly after the dehiscence of the capsule, but they are not adapted to undergo a resting period, and if allowed to become dry they soon perish. In darkness they usually remain unaltered, though sometimes cultures in the dark show the development of rhizoids from one or both ends of the mass. Biology of Fegatella conic a. ”5 Postscript. In a recent number of ‘Torreya’ (April, 1903), there appeared an interesting note by C. A. King on the explosive discharge of antherozoids in Fegatella , in which reference was made to the previously published accounts of a similar phenomenon in Asterella californica by Peirce, and in Fegatella by the present writer. At the time of writing the note which appeared in the 5 Annals of Botany,’ January, 1903, I was not aware of any previous accounts of such discharges, which are not mentioned in any of the numerous works on the structure and biology of the Hepaticae to which I have had access (see list of literature consulted). It appears, however, that the violent discharge of antherozoids in Fegatella was observed and described by the late M. Thuret nearly half a century ago. M. Ed. Bornet, who kindly wrote informing me of this observation of Thuret’s, which appears to have remained unnoticed by practically all subsequent writers on this group of plants, says in his letter : £ L’emission a distance des anthero- zoi'des du Fegatella conica a ete observee en 1 856 par G. Thuret. II en a donne la description dans une note inserde dans le tome iv des Memoires de la Soci^te des Sciences naturelles de Cherbourg, p. 21 6. J’ai rappele cette observation dans la Notice biographique sur G. Thuret qui se trouve dans les Annales des Sciences naturelles, Botanique, s^r. v, tome ii, p. 336, 1875. En general les ouvrages consacres a la systematique ne reproduisent pas les observations biologiques. Voyer Botanische Zeitung, 1858, p. 1 44, ou la note de M. Thuret est analysee.’ Thuret’s observations are also briefly referred to by M. Le Jolis in his c Remarques sur la nomenclature hepaticologique,’ 1894, p. 130. Summary. 1. Although the pores are of the simple type, as opposed to the complex or barrel-like type found in Marchantia and Preissia, the thallus of Fegatella presents a higher degree of internal differentiation than is found in any other Marchantiaceous form. The air-chambers are lined by long hyaline cells, from which localized evaporation of the water into the chambers takes place. The midrib contains highly developed mucilage- organs, arising as rows of large cells which are devoid of chlorophyll and starch, but contain dense protoplasm ; the concentric layers of mucilage are derived from the protoplasm, not from the cell- walls. 2. The ventral tissue of the thallus, underlying the air-chambers, is frequently infested by a Fungus forming a mycorhizal zone. The relationship between the Fungus and the host-plant may be regarded as symbiotic in character, enabling the host to assume a partially saprophytic mode of nutrition at the expense of the humous substratum. 3. The sessile cushion-like antheridial receptacle presents from four to eight growing-points, each producing rows of antheridia in acropetal succession ; it clearly represents a branch-system, as in Marchantia and Preissia. The receptacle contains air-chambers lined by inwardly-projecting hyaline cells, as in the thallus, and the pores are barrel-shaped. Cavers . — On the Structure and 1 1 6 4. The antheridia are usually sunk in separate cavities, but in some cases a single cavity contains two antheridia closely joined together and strongly flattened along the surface of contact. 5. The antherozoids are explosively ejected from the openings of the flask-shaped antheridial cavities ; the essential factor in the process is the absorption of water by the mucilaginous antherozoid mother-cells and those forming the antheridium-wall, leading to considerable pressure, which is relieved by the discharge of the antheridial contents in the upward direction, that of least resistance. The antherozoids are larger than in other Marchantiaceae ; each consists of at least two complete turns of a spiral, and the thicker posterior end often shows a vesicle, the remains of the mother-cell. 6. The archegonial receptacle, like the antheridial, represents a branch- system, but each of the (5-9) growing-points only produces a single archegonium. The stalk of the receptacle remains very short until after the sporogonia are ripe, and immediately before the dehiscence of the capsules it suddenly attains a length of 3-6 cm. ; this elongation is due solely to growth in length of cells already present, the starch contained in the cells of the stalk being used up during the growth in length of the latter. 7. In the development of the archegonium, the cover-cell becomes immediately divided by intersecting vertical walls and takes no part in the growth in length of the archegonium, as asserted by Gayet to be the case in Marchantia , Preissia, and other Hepaticae. 8. The young sporogonium usually shows an octant-stage, and does not grow by means of an apical cell, as stated by Hofmeister. 9. The large, green, thin-walled spores begin to germinate within the capsule, each forming an ovoid cell-mass. Beyond the occasional formation of short rhizoids no growth takes place in darkness. The spores are not adapted to resist desiccation. 10. The relatively short elaters are frequently branched ; the branching takes place at the time when the spore-tetrads become separated and loosened within the capsule. 11. The capsule opens by throwing off a thickened discoid apical cap, the rest of the wall then becoming split longitudinally into 4-8 valves. 12. Fegatella may be regarded as the lowest member of the Marchantioideae-Compositae, with which it agrees in the structure and development of the thallus, male receptacle, and sporogonium. In the organization of the female receptacle Fegatella approaches the Marchanti- oideae-Operculatae. It appears, therefore, to occupy an intermediate position between the two highest series of the Marchantiaceae. Biology of Fegatella conica. 1 1 7 Literature. Abrams, L. R. (’99) : Structure and Development of Cryptomitrium tenerum. Botanical Gazette, vol. xxviii, pp. 1 10-12 1. Andreas, J. (’99) : Ueber den Bau der Wand und die Oeffnungsweise des Lebermoossporogons. Flora, Bd. lxxxvi ; separate ed. of fifty-eight pages. Beauverie, J. (’02): Etude d’une Hepatique a thalle habite par un champignon filamenteux. Comptes-rendus de l’Acad. des Sci. de Paris, pp. 616-618. Campbell, D. H. (’95) : The Structure and Development of the Mosses and Ferns. London. (’96) : The Development of Geothallus tuberosus. Annals of Botany, vol. x, pp. 491-510- (’98) : Recent Work upon the Development of the Archegonium. Botanical Gazette, vol. xxvi, pp. 428-431. Cavers, F. (’03 a) : Explosive Discharge of Antherozoids in Fegatella conica. Annals of Botany, vol. xvii, pp. 270-274. (’03 b ) : On Saprophytism and Mycorhiza in Hepaticae. The New Phytologist, vol. ii, PP- 30-35- Czapek, F. (’99) : Zur Chemie der Zellmembranen bei den Laub- und Lebermoosen. Flora, Bd. lxxxvi, pp. 361-381. (’01) : Die chemische Zusammensetzung der Mooszellmembranen. Sitzungsb. deutsch. nat.-med. Vereins Lotos, Prag, Bd. xxi, p. 1. Farmer, J. B. (’95) : On Spore -formation and Nuclear Division in the Hepaticae. Annals of Botany, vol. ix, pp. 469-523. Gayet, L. A. (’97) : Recherches sur le developpement de l’archegone chez les Muscinees. Annales des Sci. Nat., Bot., 8e s^r., tom. 3, pp. 161-258. Goebel, K. (’79) : Zur vergleichenden Anatomie der Marchantieen. Arb. a. d. bot. Inst, in Wfirz- burg, Bd. ii, pp. 529-535. (’82; : Die Muscineen. Schenk’s Handbuch der Botanik, Bd. ii, pp. 315-401. (’95): Ueber Function und Anlegung der Lebermooselateren. Flora, Bd. lxxx; separate ed. of thirty-seven pages. (’98) : Organographie der Pflanzen ; Teil II, Heft 1 (Bryophyten). Jena. (’02) : Ueber Homologien in der Entwickelung mannlicher und weiblicher Geschlechts- organe. Flora, Bd. xc, pp. 279-305. Golenkin, M. (’02) : Die Mycorhiza-ahnlichen Bildungen der Marchantiaceen. Flora, Bd. xc, pp. 209-220. Gottsche, C. M. (’42) : Anatomisch-physiologische Untersuchungen fiber Haplomitrium Hooker i. Nova Acta kais. Leop.-Car. Akad., vol. xx, pp. 265-398. Guignard, L. (’89) : Developpement et constitution des Anth^rozoides. Revue gen. de Botanique, tom. i, liv. 1-4. Haberlandt, G. (’89) : Ueber das Langenwachsthum und den Geotropismus der Rhizoiden von Marchantia und Lunularia. Oesterr. bot. Zeitschrift ; separate ed. of four pages. (’96) : Physiologische Pflanzenanatomie. Auflage 2. Leipzig. Heald, F. de F. (’98) : Conditions for the Germination of the Spores of Bryophytes and Pterido- phytes. Botanical Gazette, vol. xxvi, pp. 25-44. Hofmeister, W. (’51) : Vergleichende Untersuchungen der Keimung, Entfaltung, und Frucht- bildung hoherer Kryptogamen. Leipzig. Janczewski, E. von (’72) : Vergleichende Untersuchungen fiber die Entwickelungsgeschichte des Archegoniums. Bot. Zeitung, Jahrg. xxx, p. 377. Kamerling, Z. (’97): Zur Biologie und Physiologie der Marchantiaceen. Flora, Bd. lxxxiv ; separate ed. of seventy-two pages. (’98) : Der Bewegungsmechanismus der Lebermooselateren. Flora, Bd. Ixxxv, pp. 157-169. Karsten, G. (’87) : Beitrage zur Kenntniss von Fegatella conica. Bot. Zeitung, Jahrg. xlv; separate ed. of four pages. 1 18 Cavers . — On the Structure and Kienitz-Gerloff, F. (’74) : Vergleichende Untersuchungen iiber die Entwickelungsgeschichte des Lebermoos-Sporogoniums. Bot. Zeitung, Jahrg. xxxii ; separate ed. of twenty-two pages. (’75) : Neue Beitrage zur Entwickelungsgeschichte des Lebermoos-Sporo- goniums. Bot. Zeitung, Jahrg. xxxiii ; separate ed. of six pages. (’86) : Ueber die Bedeutung der Paraphysen. Bot. Zeitung, Jahrg. xliv, p. 248. King, C. A. (’03) : Explosive Discharge of Antherozoids in Conocephalum . Torreya, vol. iii, No. 4, p. 60. Kny, L. (’79) : Ueber eigenthiimliche Durchwachsungen an den Wurzelhaaren zweier Marchantia- ceen. Verhandl. des bot. Vereins d. Prov. Brandenburg, Bd. xxi ; separate ed. of two pages. (’90) : Bau und Entwickelung von Marchantia polymorpha. (Aus dem Text der Abth. 8 der ‘ Botanischen Wandtafeln ’.) Berlin. Kuster, W. (’94) : Die Oelkorper der Lebermoose und ihr Verhaltnis zu den Elaioplasten. Inaug. Diss. , Univ. Basel ; forty-one pages. Lampa, Emma (’02) : Untersuchungen an einigen Lebermoosen. Sitzungsber. der kais. Akad. d. Wiss., Wien, Band cxi, pp. 477-489. Leitgeb, H. (’79) : Untersuchungen iiber die Lebermoose. Heft 4, Die Riccien. Graz. (’81) : Ibid. Heft 6, Die Marchantieen. Graz. (’80 a) : Die Athemoffnungen der Marchantiaceen. Sitzungsb. der kais. Akad. d. Wiss., Wien, Bd. lxxxi ; separate ed. of fifteen pages. (’80 b ) : Die Inflorescenzen der Marchantiaceen. Sitzungsb. der kais. Akad. d. Wiss., Wien, Bd. lxxxi ; separate ed. of twenty-one pages. (’83) : Ueber Bau und Entwickelung einiger Sporen. Berichte der deutsch. bot. Gesellschaft, Bd. i, pp. 247-256. Lindberg, S. O. (’84) : Sandea et Myriorrhynchus nova Hepaticarum genera. Acta Soc. pro Fauna et Flora Fennica, vol. ii, no. 5 ; separate ed. of nine pages. Mattirolo, O. (’88) : Sopra alcuni movimenti igroscopici nelle Epatiche Marchantieae. Atti d. R. Accad. d. Sci., Torino, vol. xxiii ; pub. separately, 9 pp. Mirbel, M. (’35) : Recherches anatomiques et physiologiques sur le Marchantia polymorpha. Mem. de l’Acad. des Sci., Paris ; separate ed. of 100 pages. Miyake, K. (’99): Mahinoa, eine neue Gattung der Lebermoose. Hedwigia, Bd. xxxviii, p. 201. Nees von Esenbeck, C. G. (’38) : Naturgeschichte der europaischen Lebermoose. Berlin. Peirce, G. J. (’02) : Forcible Discharge of Antherozoids in Asterella calif ornica. Bulletin of the Torrey Botanical Club, vol. xxix, pp. 374-382. Pfeffer, W. (’74) : Die Oelkorper der Lebermoose. Flora, Bd. lxii ; separate ed. of twenty-five pages. Prescher, R. (’82): Die Schleimorgane der Marchantieen. Sitzungsb. der kais. Akad. d. Wiss., Wien, Bd. lxxxvi, pp. 132-158. Quelle, F. (’02) : Bemerkung fiber die ‘ Rhizoideninitialen ’ in den Ventralschuppen der Marchantia- ceen. Hedwigia, Bd. xli, p. 176. Reinsch, P. F. (’77) : Contributiones ad floram Algarum aquae dulcis Promontorii Bonae Spei. Journal Linn. Soc., Botany, vol. xvi, pp. 232-248. Ruge, G. (’93) : Beitrage zur Kenntniss der Vegetation sorgane der Lebermoose. Flora, Bd. lxxvii ; separate ed. of thirty- eight pages. Schostakowitsch, W. (’94) : Ueber die Reproductions- und Regenerationserscheinungen bei den Lebermoosen. Flora, Erg.-Bd., pp. 350-384. Solms-Laubach, Graf zu (’97) : Ueber Exormortheca, eine wenig bekannte Marchantiaceengattung. Bot. Zeitung, Jahrg. lv, pp. 1-16. Stahl, E. (’83) : Ueber den Einfluss des sonnigen oder schattigen Standortes auf die Ausbildung der Laubblatter. Jena; thirty-eight pages. (’88): Pflanzen und Schnecken. Jena; 126 pages. Strasburger, E. (’70): Die Geschlechtsorgane und die Befruchtung bei Marchantia polymorpha. Jahrb. ffir wiss. Bot., Bd. vii, p. 409. Tilden, Josephine E. (’94) : On the Morphology ©f Hepatic Elaters, with special reference to branching elaters of Conocephalus conicus. Minnesota Botanical Studies, Bulletin 9, PP- 43-53- Biology of Fegotella conic a. 119 Vochting, H. (’85) : Ueber die Regeneration der Marchantieen. Jahrb. fur wiss. Bot., Bd. xvi; separate ed. of forty-eight pages. Voigt, A. (79) : Ein Beitrag zur vergleichenden Anatomie der Marchantiaceen. Bot. Zeitung, Jahrg. xxxvii ; separate ed. of nineteen pages. Zimmermann, A. (79) : Ueber die Einwirkung des Lichtes auf den Marchantieenthallus. Arb. d. bot. Inst, in Wurzburg, Bd. ii, Heft 4. EXPLANATION OF FIGURES IN PLATES VI, VII. Illustrating Mr. Cavers’ Paper on Fegatella conica. The drawings for Figs. 32 and 33 (PI. VII) were kindly made for the author by Miss Lucy M. Phillips. Fig. 1. Surface view of anterior portion of thallus, from above, showing the areolation of the surface ; each area has in its centre a pore situated on the summit of a dome-like elevation ; the growing-point lies in the notch at the margin of the thallus, and is covered by the overlapping appendages of the ventral scales, x 12. Fig. 2. Similar view, showing a new shoot arising in the notch at the anterior end of the thallus. x 4. Figs. 3-7. Successive stages in the growth of a winter-shoot, x 1. Fig. 8. Horizontal section of thallus, showing the transverse row of initial-cells (x, x, x ) : v.s. ventral scales, x 360. Fig. 9. Similar section, showing dichotomy of apex : x, x, the initial-cells of the two new growing-points ; v. s. ventral scales, x 360. Fig. 10. Surface view of a pore, showing the concentric rows of cells, and the granular rim immediately bounding the pore, x 360. Fig. 11. Vertical longitudinal section of thallus, showing an air-chamber, lined by chlorophyll- bearing filaments, those below the wide pore being produced into long hyaline processes : ep. epi- dermis, forming roof of chamber ; st. starch-grains in cells underlying chamber, x 360. Figs. 12-15. Stages in development of ventral scale, in surface view: p. p. primary papilla; app. appendage ; b. sh. basal portion of scale, in the axil of which spring tuberculate rhizoids (rh.), whilst its free margin bears mucilage-hairs (m. k.). x 200. Fig. 16. Fully grown ventral scale ; lettering as in Fig. 15. x 15. Fig. 17. Part of transverse section of thallus: a. c. air-chambers; lam. part of lamina; m. 0. mucilage-organ ; sm. rk. smooth-walled rhizoids ; tub. rh. tuberculate rhizoids, borne in bundles in the axils of the ventral scales (v. s.). x 75. Fig. 18. Vertical longitudinal section through growing-point of thallus : a. c. developing air-chambers ; ep. epidermis ; m. 0. mucilage-organ ; p. pores ; v. s. ventral scales ; x. initial cell, x 360. Fig. 19. Cells in ventral tissue of midrib, containing Nostoc-chains. x 360. Fig. 20. Part of a longitudinal section through midrib, traversing a row of mucilage-cells. X 360. Fig. 21. Cells in ventral tissue of lamina, showing pitted walls: 0. b. oil-body, x 360. Fig. 22. Part of a longitudinal section of midrib, showing cells of the ventral tissue traversed by branching fungal hyphae. x 360. Fig. 23. Cell with fungal hyphae bearing thin-walled vesicles ( ves .). x 540. Fig. 24. Cell with fungal hyphae, one of which bears a large terminal thick- walled vesicle (m.). x 54°. Fig. 25. Branched ends of two smooth-walled rhizoids. x 120. Fig. 26. Two young rhizoids, with nuclei and protoplasm, x 200. Fig. 27. Transverse sections of tuberculate rhizoids, showing the peg-like thickenings, x 360. 120 Cavers. — The Structure and Biology of Fegatella Conica. Fig. 28. Antheridial receptacle, seen from above ; v. s. ventral scales, x 10. Fig. 29. Part of horizontal section through antheridial receptacle : v. s. ventral scales. Two antheridia are shown occupying a single cavity, in lower part of figure, x 200. Fig. 30. Vertical transverse section of male plant, traversing an antheridial receptacle : a. c. air- chambers of thallus ; an. antheridia ; sm. rhiz. smooth-walled rhizoids ; tub. rhiz. tuberculate rhizoids ; v. s. ventral scales, x 10. Fig. 31. Median longitudinal section of young antheridial receptacle: a. c. air-chamber; an. 1, an. 2, developing antheridia; p. pore; v. s. ventral scales, x 360. Fig. 32. Female plant, with two receptacles (carpocephala). x 2. Fig. 33. The same, after elongation of the carpocephalum-stalk. x 2. Fig. 34* Carpocephalum as seen from below, showing cut end of stalk with its ventral furrow and the slit-like opening of the sheath around each of the six sporogonia. x 12, Figs. 35-37. Cells in tissue of carpocephalum-stalk, as seen in longitudinal section before, during, and after its sudden growth in length, x 360. Figs. 38-41. Longitudinal sections of developing carpocephala : for description see text, x 6. Fig. 42. Horizontal section of young carpocephalum, corresponding with the stage shown in Fig. 38. The cells occupying the anterior margin of the outgrowth show active divisions, constituting at this early stage a single growing-point, which may be compared with that of the thallus, shown in Fig. 8. Below the row of initial-cells there stand numerous mucilage-hairs, shown in cross-section ; two of the ventral scales which curve upwards over the young carpocephalum are also shown in section, x 200. Fig. 43. Similar section of a later stage, showing dichotomy of the apex, a ‘ middle lobe ’ having been formed between the two growing-points, as in ordinary dichotomy of the thallus (cf. Fig. 9). x 200. Fig. 44. A still later stage : here each of the two growing-points has again undergone dichotomy, so that four growing-points have been established ; between these are the projecting ‘ middle lobes.’ x 200. Fig. 45. Part of a horizontal section through a carpocephalum with eight growing-points, four of which are shown, x 200. Fig. 46. Part of a longitudinal section through a carpocephalum of about the same age as that in Fig. 40, showing a young archegonium : a. c. air-chambers ; p. pores ; rhiz. tuberculate rhizoids springing from the lower surface of the receptacle, near the insertion of the stalk (rec. si.). x 200. Fig. 47. Horizontal section of receptacle bearing six nearly mature sporogonia : a. c. air- chambers of receptacle, now reduced in size owing to the pressure exerted by the growing sporogonia ; cal . calyptra ; caps . w. capsule-wall; rhiz. rhizoids; si. stalk of receptacle. x 75. Fig. 48. Longitudinal section of ripe sporogonium, showing the foot (/. ), seta (j.), and capsule (c.) ; the spores and the free elaters are omitted, in order to show more clearly the apical thickening of the capsule-wall and the apical and basal tufts of fixed elaters. x 60. Fig. 49. Apical portion of capsule-wall, showing the strongly thickened cells of the apical cap, in surface view, x 200. Fig. 50. Cells in upper half of capsule-wall, as seen from outer surface, x 200. Fig. 51. Cells in lower half of capsule-wall, as seen from outer surface, x 200. Fig. 52. Fully formed elaters : a, normal form; b, c, branched elaters. X 200. Fig. 53. Median section of spore at time of dehiscence of capsule, x 200. Fig. 54. Germination of spore (optical section) : rhiz. rhizoid. x 200. <- Annals of Botany F. C . del . CAVERS, FEGATELLA. Vol. XVIII, PI. VI University Press, Oxford. . ■ F. C . ael . CAVERS. F EG AT ELLA Vol. XVJIl PI. VII. University Press, Oxford- Urinals of Bo touny Vol. XVII f PI. VII. F. C . ael . CAVERS. FEGATELLA University Press. Oxford- On the Occurrence of Cellulose in the Xylem of Woody Stems. BY M. C. POTTER, M.A., Professor of Botany in the University of Durham College of Science , Newcastle-upon-Tyne . With Plate VIII. THE appearance of cellulose in the xylem of various trees when attacked by certain wood-destroying Fungi has been noted, and has been attributed to the action of a delignifying enzyme secreted by the Fungi. Hartig (78), in his classic work upon the destruction of the wood of Conifers and of the Oak by certain Fungi, has given an account of the appearance of cellulose in the wood when under the influence of such attack. In the Oak, he describes in detail the special action upon the wood of Telephora perdix, Polyporus ignarius, P. dryadeus, and Stereum hirsutum , and states in each case that in the process of decay the lignified walls become converted into cellulose, attributing this change to the delignifica- tion by the hyphae. In the wood of the Conifers the same appearance is noted and ascribed to the influence of Trametes and other Fungi. Mayr (’84) attributes the presence of cellulose in the stem of Betula to the action of the parasitic Fungi Polyporus betulinus and P. laevigatus. Marshall Ward (’97'), studying the action of Stereum hirsutum upon the wood of Aesculus, has demonstrated the presence of cellulose in the xylem when infected by a pure culture of this Fungus. The swollen inner layers of the cell-wall become delignified and consist entirely or almost entirely of cellulose, as shown by the differentiating colour-tests of chlor- zinc-iodine, and phloroglucin. The hyphae attack the walls of the tracheides and other wood-elements, and, he considers, gradually delignify them from the lumen outwards. Marshall Ward assumes the presence of an enzyme which effects the delignification, but ‘ did not succeed in extracting the enzyme from his cultures.’ Biffen (’01) has published some investigations upon the biology of Bulgaria polymorpha and the effects of its action on wood. Sections from a block of Oak thoroughly permeated by hyphae showed a swelling of the [Annals of Botany, Vol. XVIII. No. LXIX, January, 1904.] 122 Potter. — On the Occurrence of Cellulose in the thickening layers of the woody elements, which, on treatment with colour stains, gave indications of the presence of cellulose, suggesting that the action of the Fungus was one of delignification. Biffen also observed a sharply defined zone of cellulose surrounding many of the pits, especially in the vessels. He considers that the action of Bulgaria polymorpha upon Oak-wood causes the dissolution and probably decomposition of the lignin, and has assumed the existence of a delignifying enzyme to explain these results. Czapek (’99) claims to have isolated a delignifying enzyme from Merulius lacrymans and Pleurotus pulmojiarius. From the fungus-extract, by precipitation with alcohol, he obtained a white precipitate, soluble in water, which destroyed the lignin of the woody cell-walls. This fungus- extract, he states, lost its wood-destroying properties when once boiled. Hermann von Schrenk (’00) has described a disease of the Red Cedar caused b y Polyporus Juniperinus^ which attacks the walls of the wood-fibres, and extracting the lignin substances leaves a basis of almost pure cellulose. The amount of wood-fibre reduced to cellulose is very considerable, and it is supposed that some very potent lignin-splitting enzyme is concerned in the changes which are brought about. Schrenk (’03) also described a delignification in the wood of Fraxinus due to Polyporus fraxin oph ilus , which is said to give off an enzyme which attacks the inner parts of the wood-cells, extracting the lignin, and leaving an impure cellulose. Some evidence may be gathered with reference to cellulose occurring normally in the xylem. Writing in 1 86o Sanio (’60) states that the tertiary thickening gelatin- ous layer of lignified cells was first pointed out by T. Hartig and confirmed by von Mohl, and finally claimed by Schacht as an element present in all woody cells, which never becomes lignified but consists always of pure cellulose. Sanio regards this latter statement as too sweeping, and maintains that the share which this tertiary layer takes in the thickening of the woody cells varies greatly, sometimes being very strongly pronounced, at other times hardly perceptible. He mentions Acer platanoides as a stem in which this layer, which is coloured violet with chlor-zinc-iodine, is most distinctly to be observed ; and adds a list of seventeen species in which this cellulose thickening is found in the wood-fibres. In his general considerations on the structure of the Oak (p. 94), Hartig, speaking of the wood-fibres, says that in the walls of these fibres two, often three, layers can be distinguished, differing in their chemical behaviour. The most external, always a very thin layer, which is united with the corresponding layer of adjacent organs, encloses a very thick inner layer. This is lignified, and only in isolated cases is bounded internally by a third layer, which colours blue with chlor-zinc-iodine. Hartig himself makes only slight further allusion to this occurrence, as when discussing Hydnum Xylem of Woody Stems. 123 diversidens , he mentions the most internal, somewhat gelatinous layer of many healthy Oak wood-fibres, which colours blue with chlor-zinc-iodine. He evidently regarded the character as so limited as to be negligible, and the fact as noted by him appears to have escaped attention. Strasburger (’91) in the ‘ Leitungsbahnen ’ mentions incidentally that the inner thickened layer of the wood-fibres of Robinia pseudacacia colours violet by potassium iodide ; and also that in Wistaria the wood-fibres show an inner thickened layer, sometimes almost completely closing the lumen, which on treatment with potassium iodide and with chlor-zinc-iodine gave a beautiful wine-red colour. In the case of Pinus sylvestris it has been shown by Schellenberg (’96) that the primary medullary rays remain unlignified until the third year1. He has also observed that among the wood-fibres generally the middle lamella is very often more lignified than the other layers, but that com- pletely unlignified inner lamellae are only seldom met with, as in Trio se turn perfoliatum . With these exceptions I can find no reference2 to the presence of cellulose in mature wood, except as the result of the destructive action of certain Fungi. Abromeit (’84) mentions that after the use of chlor-zinc- iodine he has been unable to detect the tertiary gelatinous layer in the wood-fibres of the Oak, and, so far as I have been able to trace, it has come to be very generally accepted that the elements of the wood become entirely lignified, using the term in its simplest acceptation. After an extended examination of a very large series of specimens I have detected the presence of cellulose in the xylem of many trees as a normal condition in perfectly healthy and vigorously growing stems. It occurs chiefly in the wood-fibres, and proves to be by no means an excep- tional occurrence. My attention was first drawn to this question when studying the action of a Bacterium ( Pseudomonas destructans) (’99) upon the cell-walls of storage tissues, such as in Turnip, Carrot, &c., and following up some investigation into the destructive action of this parasite upon the xylem. Small pieces of the xylem of Quercus and Fagus were sterilized by steam- ing in test-tubes in which they were half immersed in water, and afterwards were sown with a pure culture of this Bacterium. After an interval of fourteen days transverse sections showed a very marked presence of cellulose, as indicated by chlor-zinc-iodine and haematoxylin. As this seemed to suggest an action by the Bacterium corresponding to that of the Fungi above mentioned, it became necessary to determine whether this cellulose was present before inoculation. 1 I have observed cells in the medullary rays remaining unlignified for a much longer period. 2 De Bary (Comp. Anat., Engl, ed., p. 482), quoting Sanio, mentions the occurrence of this gelatinous layer, which remains unlignified. 124 Potter. — On the Occurrence of Cellulose in the On examination of pieces of wood from the same branch, before sterilization, I was surprised to find that in many sections a thick layer was present in several of the fibres which gave the violet colour character- istic of cellulose, when treated with haematoxylin ; and similar sections showed a layer in the fibres which refused to react to phloroglucin. Treat- ment with iodine and sulphuric acid also gave the characteristic blue. This unlooked-for result in apparently sound wood led me to examine a number of shoots of Oak of various ages, carefully selected for their vigorous growth and entire freedom from disease. For the purpose of this investigation it was specially important that fresh material only should be used. Pieces of felled wood left in the open, even for a short time, would be liable to the objection that any cellulose present in the xylem might be due to the process of decay. For this reason freshly cut branches were always used and the sections made generally, on the day following that on which the material was obtained, when it was still fresh and before any change could have taken place. Shoots of Oak were selected from trees growing near sheltered river banks and from those fully exposed in the open country. At first transverse sections were cut by a microtome through stems of varying ages and about 1.3 cm. in diameter. The sections thus obtained were stained with phloroglucin and anilin sulphate, and it was found that in certain areas the internal layer of the walls of the fibres refused to act to these stains and remained as a white layer inside the otherwise coloured layer, but when treated with Schultze’s solution, with iodine and sulphuric acid, or with haematoxylin, the characteristic blue colour was at once given by this layer. These tests are all confirmatory one of another, and leave no doubt that the substance of the inner layer was of the nature of cellulose. As haematoxylin alone sufficiently indicated these results permanent preparations were made by using Delafields haematoxylin and mounting in Canada balsam. In transverse section this cellulose layer appeared as a broad violet lining, gradually shading off towards the middle lamella, sometimes being partially detached and lying kinked in the lumen (see PL VIII, Fig. 2). Longitudinal sections confirmed the above observations, the cellulose lining being sometimes torn out by the razor and resembling an isolated prosenchymatous element. It was noticeable in longitudinal sections that around the bordered pits a layer of violet was often present when stained with haematoxylin or Schultze’s solution, showing that the bordered pits were also surrounded by a ring of cellulose. The fibres containing the cellulose lining occurred singly, in isolated groups, or in broad bands concentric with the annual rings and interrupted by the medullary rays ; but they seldom passed round the whole of the stem. It was necessary to cut across the entire stem, for, although the 125 Xylem of Woody Stems wood fibres in certain small sections might give no cellulose reaction, on taking complete transverse sections I never failed to detect this reaction in some area or another. When stems of a larger diameter were employed it became necessary to divide them into smaller areas, and for this purpose they were cut into blocks about 1.3 cm. square. These blocks were numbered, and by this means a transverse section of a large stem was obtained. In one instance the base of a young branch approximately sixteen years old, measuring 5.4 by 3.8 cm., was divided by a fine saw into twelve blocks ; sections were cut from each of the blocks, stained with Delafield’s haematoxylin and mounted in Canada balsam. Although not one of these block-sections could be found which did not contain some fibres with a broad violet inner lining, in some they were quite isolated, while in others (more especially on the lower side) they presented a very conspicuous feature, so much so that their presence could be distinctly discerned by the unaided eye. In another instance I specially felled a young Oak, approximately forty years old, the stem measuring 14 cm. by 12 cm. This was divided by a fine saw into eighty-eight blocks, all of which were numbered in order, and transverse sections were cut from each, stained in haematoxylin and mounted in Canada balsam. In this way a complete transverse section of this trunk was obtained. As in the case of the smaller branch some fibres with a violet lining could be detected in almost every one of the eighty-eight sections, but in certain portions of the stem they occurred in very conspicuous groups. These fibres were well marked in the blocks situated in the centre of the stem, that is around the pith, and containing the first few years of growth, which showed that the oldest wood was not completely lignified and contained a large proportion of cellulose. Its presence in the wood of the first years of growth was also noticed in many smaller stems. The cellulose was somewhat irregularly distributed, though it was evidently developed more markedly on one side of the tree than the other, and it is probable that the orientation has some influence in this direction (Fig. 1). The trees above-mentioned were growing in the neighbourhood of Newcastle-upon-Tyne. It became an interesting question to determine whether climatic conditions had any influence upon the composition of the xylem, and specimens of Oak were obtained from other parts of England, i. e. from Warwickshire, Yorkshire, and Surrey. Transverse sections from Oaks growing in all these localities presented the same appearance, when stained with haematoxylin, as has been already de- scribed. Fig. 2 represents a section of an Oak from Wimbledon, taken from a branch eight years old, diameter 1.3 cm., showing strikingly the inner cellulose layer of the fibre-walls. The cellulose extended about one-third round the stem. From Warwickshire I was fortunate in obtaining, 126 Potter . — On the Occurrence of Cellulose in the ten days after being felled, the base of a well-grown sound Oak, ap- proximately sixty years old, and in this specimen too a considerable quantity of cellulose was found in the first years of its growth, and also distributed throughout the duramen and alburnum. Branches of various other trees were next examined. In the wood of Fagus the occurrence of cellulose was equally as strongly pronounced (Fig. 3) as in Quercus , if not more so, in trees grown both in the north and central parts of England. It occurs chiefly in the broader annual rings which exhibit special development during a season of vigorous growth. As in the case of Quercus , the cellulose may be found at the very centre of the stem, and branches of twenty-five and thirty years both showed this peculiarity in the wood of the first and second years of their growth. In Aesculus the cellulose, as indicated by the reaction to haematoxylin, was not so immediately apparent, and in specimens I first examined it was found only in the very young wood, and then quite locally. A further study, however, of numerous transverse sections of different trees showed it to be fairly prevalent in the growth of other years. The softer wood of Aesculus is very different in character to that of the trees considered above, and the inner lining of the wood-fibres is represented by a much thinner layer than is the case in many of the harder woods. In certain parts of the sections a thin lining can be observed inside the fibre-walls, which is stained violet with haematoxylin, and this is the most characteristic appearance ; but in other regions, quite locally, the violet-blue layer may assume a thicker, more gelatinous appearance, and lie quite detached and often crumpled in the lumen. This thicker layer sometimes occupies a large part of the fibre-lumen, and shades into a paler colour towards the middle lamella, being edged inside with a darker line. In one instance I found this disposition of the cellulose along a radius extending throughout the third to the ninth years, though most noticeably present in the third and sixth years, which showed a more vigorous development, while in other parts of the annual rings it was entirely absent. In another section it was well marked in the spring wood of the current year. The cellulose is more generally to be found in the spring than in the autumn wood. In Salix the presence of the cellulose lining to the wood-fibres is often very pronounced. In one branch, with a diameter of 1.6 cm. and containing ten annual rings, cut from a willow growing at the edge of a large pond, hardly a single fibre could be detected which did not show a broad violet inner layer when treated with chlor-zinc-iodine. Sometimes this layer completely closed the lumen of the fibre. In another instance, in a branch with a diameter of 3.7 cm., with twelve annual rings, from a tree growing upon a hillside with northern aspect, in many of the fibres no violet colour due to chlor-zinc-iodine could be observed, and also where the edge of the lumen was coloured deep violet this colour gradually 127 Xylem of Woody Stems . shaded away toward the middle lamella. The difference in the manner in which the fibre-walls reacted to chlor-zinc-iodine was most marked in the two cases, and this may perhaps be attributed to the different situations. In longitudinal sections these appearances are well-marked, the cellulose lining of the fibres being easily recognized. The cellulose ring round the pits on the walls of the vessels was exceedingly striking when stained in chlor-zinc-iodine. Fig. 4 is taken from the branch of the tree growing at the water-side. The iodine and sulphuric acid (76 per cent.) reaction agreed with the results obtained from chlor-zinc-iodine. In longitudinal sections, especially, it was most beautiful to see the edges of the pits turn blue in turn as the acid gradually swept over them. Further confirmation was obtained by treatment with Congo-red and also with haematoxylin. In the wood of Ulmus each annual ring commences with large wood- vessels succeeded by wood-parenchyma and fibres. When stained with haematoxylin and mounted in Canada balsam the wood-fibres are especially conspicuous by their thick yellow walls, inside which, in many instances, a distinct blue-violet lining of cellulose may be detected. An Alnus stem obtained from the banks of a stream in Warwickshire, about ten years old, and having a diameter of 2 cm., showed the presence of cellulose very distinctly round one half of the stem, while in the other half it was entirely absent. The development occurred chiefly in the third, fourth, fifth, and sixth years, being especially prevalent in the very wide fourth year, as indicated by the inner thickened layers staining clearly blue-violet with haematoxylin. A branch of Bettda six years old, with a diameter of 1.4 cm., was also examined. This, upon treatment with haematoxylin, indicated a special prevalence of the swollen cellulose layers in the fourth and fifth years. In Fraxinus , unlike the other woods examined, no well-marked patches of wood-fibres, in which a broad lining of cellulose is present, could be seen. The specimen from which my sections were taken was a vigorous and healthy sapling, with a diameter of 2.7 cm. and twelve annual rings. The stem had therefore grown quickly. It was obtained in July, and the next day sections were cut, stained with Delafield’s haematoxylin, and mounted in Canada balsam. In none of the sections, however, could any characteristic development of the thickening gelatinous layer be found ; at most on the internal surface of some of the fibres there might be detected a very thin line of blue, and this was found to be true of all the trees afterwards examined. It had now become perfectly plain that a thickening layer of cellulose occurred quite commonly as a natural feature in the woody fibres of 128 Potter. — On the Occurrence of Cellulose in the a variety of trees in all situations, probably representing a stage of arrested development. As this fact has not hitherto been recognized the observation is important, and has some bearing upon the study of fungoid action in timber. It is well known that boiling water has a destroying and dissolving action upon the cell-wall. As long ago as 1882 Singer (’82) has shown that four substances can be extracted from xylem by boiling water, namely, vanilin ; a substance which shows the reactions of coniferin ; a gum, which is soluble in water ; and a substance soluble in water not identical with the other three ; these all entering into the composition of what is known as lignin, though in what relationship is not determined. Recently, Van Wisselingh (’98) has shown that by heating sections of the root of Beta vulgaris for six hours in distilled water, at a temperature of about I25°C., a pure cellulose wall remains behind, and the pectin substances are destroyed and removed. It seems probable that the operation of sterilizing by discontinuous boiling would have some such effect upon the lignified walls, and that some substances might be destroyed or removed from the cellulose matrix during the process ; in other words it would have a delignifying action. The wood of Fraxinus affords very suitable material for studying the action of boiling water upon the xylem ; since, as far as I have been able to determine from the examination of numerous sections taken from various trees, no broad inner lining of cellulose is present in the fibres. The xylem appears to undergo complete lignification. Transverse sections were taken from a stem ten years old and about 2-4 cm. in diameter, and some of these were examined at once with chlor-zinc-iodine. Fig. 5 is a good illustration of the appearance presented, the walls are thick and hard and show no trace of violet colour, the only effect of the chlor-zinc-iodine being to turn the walls yellow. The other sections were placed in a boiling tube half filled with water and steamed for about two hours on consecutive days, remaining immersed in water the whole time. Some of these sections were removed at intervals and subjected to treatment with chlor-zinc-iodine. After two days some delignification was slightly indicated ; after three days it was well pronounced. Fig. 6 represents the general appearance after four days: the middle lamella is a bright yellow, while the inner layers of the fibre-walls are swollen and stained violet, more deeply on the inner surface and shading away towards the middle lamella, showing almost complete delignification. A comparison of the sections before and after boiling represented in Fig. 5 and Fig. 6 shows in a most marked manner the delignifying action of the boiling water. It should be noted that in the sections steamed four times a delignification had already commenced in some of the wood-vessels. 129 Xylem of Woody Stems . The xylem of Fraxinus is somewhat resistent to the action of boiling water, the same effect being produced in other woods after a less severe treatment ; but it has been selected for illustration because of the certainty of obtaining sections not exhibiting incomplete lignification in any part before boiling. In the case of Aesculus the delignifying action of boiling water is well marked. A portion of a healthy stem of Aesculus was procured, with a diameter of 2-2 cm. and ten annual rings. Sections containing the whole surface were cut by means of a microtome and placed in a boiling tube half-filled with distilled water and steamed for two hours on three consecutive days. The sections were then treated with chlor- zinc-iodine and phloroglucin, and permanent preparations made by staining with Delafield’s haematoxylin and Congo-red (Fig. 7) and mounting in Canada balsam. (Similar preparations to serve as controls were made before steaming.) When stained with chlor-zinc-iodine the appearance was most striking and in strong contrast to the unboiled sections. Instead of only in certain areas, the fibres and wood-cells in nearly every part contained a swollen blue-violet-coloured layer, which was frequently broken away and lying in contorted shapes in the lumen (see Fig. 7), while the middle lamella remained yellow ; the walls of the vessels were yellow, swollen, and striated, but lined with a faint blue tinge. The action of phloroglucin fully confirmed that of chlor-zinc-iodine. The middle lamella of the fibres showed the characteristic red, while the inner layers were quite white and colourless ; the walls of the vessels also stained red, but they were swollen, and a white unstained lining might in some cases be detected. The annual rings stood out distinctly red with phloroglucin, but when stained with chlor-zinc-iodine the elements composing these rings showed the inner layers swollen, striated, and yellow, enclosing a blue lining. Prolonged steaming further emphasized these results. Sections subjected to discontinuous steaming for six days when stained with chlor-zinc-iodine showed the inner layer of the fibres a deep blue- violet, much swollen and in many cases obliterating the lumen ; the walls of the vessels and tracheides were much swollen and showed distinctly violet ; the elements forming the boundary of the annual rings too had a well-marked violet lining. With phloroglucin the walls of the fibres remained white, and in some cases even the middle lamella refused to stain ; in the vessels too there was a distinct inner white layer, the medullary rays staining red. Among the many stains for lignin, Hegler (’00) has shown that thallin sulphate stains the vanilin yellow and phenol the coniferin a blue-green. The action of these stains on steamed wood is of importance, and gives confirmation of the results already obtained with phloroglucin. Steamed transverse sections from the same piece of Aesculus treated with thallin K 130 Potter. — On the Occurrence of Cellulose in the sulphate showed the middle lamella of the wood-fibres a distinct yellow, but enclosing an inner layer remaining white and unstained. The walls of the wood-vessels and tracheides stained yellow. In the longitudinal sections the same peculiarities of staining were noted, but in the wood- vessels the borders of the pits remained unstained. Similar sections when stained with phenol showed the middle lamella of the wood-fibres green, but the inner layers of the walls white and unstained ; the wood-vessels and tracheides stained green. Again, in longitudinal section the borders of the pits were unstained. Phloroglucin stained precisely the same structures, and again the borders of the pits remained unstained. Similar investigations with Quercus , Alnus , and Ulmus gave the same results. Fig. 8, showing a transverse section of Quercus after steaming, depicts a stage in which the progressive action of the delignification may be observed. The violet colour given by chlor-zinc-iodine is seen staining deeply around the lumen, and suffusing itself gradually outwards into the lignified layers. The above experiments show that boiling water extracts some sub- stance of the nature of lignin from the wood of all these trees and that sterilization by steaming would have a similar effect. To determine this point further some sawdust was obtained from a piece of Aesculus (diameter 3-5 cm., annual rings 19), covered with distilled water, and steamed for half an hour on one afternoon and again for an equal time next morning. The sawdust was then removed by filtration through an ordinary filter paper. A small portion of the watery extract thus obtained reacted very feebly to phloroglucin, but the colour was sufficient to indicate that some substance was present which gave the lignin reactions. The remainder of the filtrate was next extracted with ether, the water drawn off with a separating funnel, and the ether evaporated in a small white crucible on a water bath at 6o° C. A brown deposit remained as a thin layer in the crucible after the ether had disappeared. This brown layer gave at once the characteristic red colour when a drop of phloroglucin was added. In order to avoid any risk of minute fragments of wood passing through the filter paper and hence into the ether, or by any chance some substance being extracted from the filter paper which might react to phloroglucin, a similar watery extract of Aesculus sawdust was passed through a Kitasato tube, extracted with ether, and the ether evaporated. The result on addition of a drop of phloroglucin to the residue was the same, the red colour was instantly given. It only remains to say that the ether and distilled water alone, after the same treatment, gave no brown deposit, and not the slightest reaction to phloroglucin. This shows conclusively that a substance which reacts to phloroglucin is extracted from Aesculus wood by boiling water, and confirms in a remarkable manner the reactions to the lignin and cellulose tests of the Xylem of Woody Stems . 13 1 boiled sections of this wood. In fact, the process of sterilization by steam- ing delignifies the xylem and produces appearances exactly similar to those described after the action of Fungi. As a confirmation of these results with Aesculus it seemed worth while to try other woods, and pieces of Elm, Ash, Oak, and Scotch Pine were tested. From each of these, after removal of the bark, sawdust was obtained and placed in small flasks containing water. These were then steamed on three days for two hours, the decoctions filtered, extracted with ether and the ether evaporated. In all these cases a brown deposit was left after the evaporation of the ether which at once gave the character- istic red colour with phloroglucin. The above experiments having so clearly shown the effect of boiling water in extracting from xylem a substance which reacts to phloroglucin and thallin sulphate, it seemed probable that cold water would have the same power if allowed to act for a longer period. From the stem of the forty years’ Oak already mentioned small chips of alburnum and duramen were carefully made and five gram, of each placed in flasks containing 150 c.c. of distilled water and 5 c.c. of chloroform. The flasks so prepared will be referred to as the ‘ extract ’ of alburnum and duramen respectively. In all eight of these flasks were prepared, four containing alburnum and four duramen, which were then placed in an incubator at 28° C. After three days one flask of alburnum-extract was taken, the water filtered and extracted with ether, and the ether then divided into three portions in three porcelain crucibles ; it was then evaporated at 6o° C. A thin brown layer was deposited inside these crucibles which reacted at once, in one case to phloroglucin and in another to thallin sulphate (both of these being tests for vanilin), while the third gave no result with phenol, the test for coniferin. A flask of duramen- extract was similarly treated and gave precisely the same reactions. 0 These results show that a substance answering to the vanilin tests is dissolved out in cold water. Thus cold water has also an extractive power, and by continued soaking in water the xylem undergoes a partial delignification. The effects of boiling and immersion in water here described were under the exaggerated conditions produced by the use of thin sections and fragments of wood, yet it must be allowed that some delignifying action would take place upon even small blocks of wood subjected to steaming in water for the purpose of sterilization and remaining under damp conditions for many months during cultural experiment. It will be noted that the coniferin, or rather a substance answering to the test for coniferin, was apparently not extracted from the xylem. This result appears curious and somewhat at variance with the fact that the steamed sections, when treated with phenol -H Cl, always gave an un- 132 Potter . — On the Occurrence of Cellulose in the coloured inner layer, which seemed to indicate that the ‘ coniferin ’ had been removed. Some explanation, however, was gained by a comparison with fresh sections treated in the same manner. These showed a very limited distribution of the ‘ coniferin,’ it being found chiefly in the vessels, tracheides, and elements surrounding them, and appearing to be absent in a great measure from the fibre-walls. This suggests an interesting point touching upon the distribution of the vanilin and coniferin in woody stems, which remains to be followed lip ; and it appears to furnish a clue to a problem which presents itself. Would not the water passing up the stem as the transpiring current have a tendency to dissolve the ‘ lignin ’ ? The walls of the vessels and other elements connected with the transmission of water are apparently impregnated with some substance not readily soluble in water. To return to the alburnum and duramen ‘extracts.’ Of the eight flasks six remained in the incubator. After eight days another pair of these flasks was similarly treated and the ether extracted, divided into three porcelain crucibles, and evaporated as before. To my surprise the residue, after evaporation of the ether, gave only a faint reaction to phloroglucin and thallin sulphate in the case of the duramen, and the reaction with these stains was even less for the alburnum. This result at first seemed inexplicable, but I noticed that the extract in the flasks, especially that of the alburnum had become somewhat turbid. This seemed to indicate the presence of Bacteria, although it should be remarked that some of the chloroform still remained at the bottom of the flasks so that the water in which the Oak chips were immersed would be saturated with chloroform. However, as I have shown in a previous paper (’00), no reliance can be placed upon chloroform, thymol, &c. as antiseptics, as these substances do not necessarily prevent the growth of micro-organisms even when used in considerable strength. To ascertain whether any micro-organisms were present stab-cultures from these flasks were made in sterile, plugged test-tubes containing beef-, Liebig-, and turnip-gelatine. These test-tubes were then incubated for five days, at the end of which time colonies of Bacteria were found in all the tubes. It may be mentioned here that stab-cultures subsequently taken from the remaining two pairs of flasks, after fourteen days and after thirty-two days, also developed colonies of Bacteria. The extracts from the flasks after these intervals gave much the same colour reactions as before, distinct though faint for the duramen and hardly perceptible in the case of the alburnum. No attempt has been made at present to isolate and identify these Bacteria, but it may be remarked that the Bacteria from the alburnum invariably liquefied the gelatine, while in the tubes sown from the duramen the gelatine remained unliquefied. 133 Xylem of Woody Stems . The inoculation of the gelatine tubes from the cold-water extract of the Oak duramen and alburnum and the subsequent development of colonies clearly showed that Bacteria were present in these solutions, but further proof was needed to show whether or no these Bacteria could live upon and destroy the substance extracted from the wood by water. To determine this point strong decoctions were made from splinters of the duramen and alburnum of the same forty years’ old stem. The splinters were boiled in water in two flasks, and after standing all night in the steamer were again boiled on the following morning. The decoctions thus obtained were freed by filtration from any particles of wood and drawn into smaller flasks, each containing 150 c.c. of decoction. In this manner eight flasks were prepared, four containing each 1 50 c.c. of alburnum decoction and four containing each 150 c.c. of duramen decoction. These will be referred to as alburnum- and duramen-* decoctions.’ These eight flasks were then plugged with cotton wool and sterilized by dis- continuous steaming. Of the eight flasks one of each kind was used for a control. Of the other three flasks of alburnum-decoction one was sown with Penicillium , another with Bacillus subtilis — a pure culture obtained from Dr. Krai — and the third from the alburnum-extract in which the Bacteria had developed. Similarly one flask of duramen-decoction was sown with Penicillium , a second with B. subtilis , and the third from the duramen- extract in which Bacteria had developed. These flasks together with the controls were incubated at 28° C. Flasks sown with B. subtilis. After twenty days stab-cultures were made on gelatine-tubes of Koch’s beef-bouillon and Liebig-extract, and the flasks were then extracted with ether and tested as before. With phloroglucin the duramen-residue gave a most distinct red, but the colour was much fainter in the alburnum-residue. With thallin sulphate the colour was indicated in both cases but more faintly in the alburnum- residue, and with phenol-HCl no reaction could be detected in either case. In the stab-cultures, colonies developed from the alburnum-decoction, but not from the duramen. This experiment seems to indicate that B. subtilis can grow in the alburnum-decoction and destroy the substance extracted from the wood, while it is incapable of living in that obtained from the duramen. Flasks sown from alburnum - and duramen-extracts. After twenty- one days stab-cultures were made from both flasks in tubes of beef-bouillon and Liebig-extract. The stabs from the alburnum-decoction developed colonies which quickly liquefied the gelatine, but the colonies from the duramen-decoction developed much more slowly and without liquefying the gelatine. The ether-extracts from the alburnum-decoction showed no colour with either phloroglucin or thallin sulphate, but in the case of the 134 Potter . — On the Occurrence of Cellulose in the duramen-decoction both phloroglucin and thallin sulphate gave the characteristic colours, but somewhat faintly. Phenol-HCl gave no colour- reaction in either case. Flasks sown with Penicillmm. After twenty-three days several colonies had developed in the alburnum-decoction and produced conidia, but in the duramen-decoction no signs of colonies could be detected. With phloroglucin and thallin sulphate the ether-extracts gave very distinct reactions in the case of the duramen-decoction, but with the alburnum the colour was very much fainter. Phenol-HCl gave no colour-reaction in either case. The control flasks were next examined (after twenty-five days). The ether-extracts from both the alburnum- and duramen-decoctions gave the characteristic colours most distinctly with both phloroglucin and thallin sulphate, but not with phenol-HCl, again showing that the vanilin is extracted but not the coniferin. This experiment was repeated with decoctions of the alburnum and duramen from the sixty years-old Warwickshire Oak, using o, gram, of wood to each ioo c.c. of water. The results were in all respects confirma- tory of those just described. These experiments show very clearly that the substance extracted from the Oak-wood by water, which reacts to phloroglucin and thallin sulphate, is destroyed by certain Bacteria and in some measure by Peni- cillinm. Both B. subtilis and the Bacteria which developed naturally in the extract, grew vigorously in the alburnum-decoction and destroyed the substance extracted from the wood. Bacteria being thus able to destroy the ‘ lignin ’ substances, this fact explains why the colour-reaction to phloroglucin was no longer given in the cold-water extract after eight days (p. 132). It is further demonstrated that the alburnum is more readily acted upon by these organisms than the duramen. This is an important consideration, and suggests that the duramen contains substances not possessed by the alburnum which are unfavourable to the growth of Fungi or Bacteria. These investigations throw some light upon the natural decay of timber, and suggest that one of the initial stages of decay is the extraction by water of some substance or substances from the xylem ; by this process the cellulose is exposed, and it is then liable to be attacked by vegetable saprophytes. The fact that these organisms do not grow so readily in the watery- extract from heart-wood is suggested as a reason why the heart- wood is more durable. In criticizing Singer’s work Czapek considers his conclusions to be erroneous. He, however, makes no mention of repeating his experiments ; while my results support Singer and prove that he was right in attributing i35 Xylem of Woody Stems . to boiling water the power of extracting certain substances from the xylem which give the so-called lignin reactions, and the extractive power would be greatly increased under Singer’s conditions owing to the very long process of boiling to which his material was subjected. I have shown further that this delignification of the xylem may be accomplished by cold water. The relationship of the substance extracted by water to the hadromal which has been extracted by Czapek by means of a boiling solution of zinc chloride needs to be determined. Czapek (’99) has isolated a delignifying enzyme, but the details given of his experiment are not very complete. He employed the mycelium of Pleurotus pulmonarius and Merulius lacrymans | obtained from wood decaying under the influence of these Fungi, presumably not pure cultures. From these a watery extract was obtained and to the filtrate a small quantity of wood-filings added, together with chloroform, and incubated at a8°C. The alcoholic-extract, tested with phloroglucin, after three days showed no reaction ; after eight days a positive but weak reaction ; and after fourteen days the reaction was tolerably strong, while the wood was coloured strongly violet with chlor-zinc-iodine. In considering these experiments it would be useful to know from what tree the wood-filings were made, and again whether the wood-filings were examined previous to being subjected to the Fungus-extract, in order to determine whether any cellulose was present. The subsequent blue colour given by chlor-zinc-iodine cannot be accepted as evidence that the wood-filings had been delignified by the action of the Fungus- extract. Water alone extracts substances from wood which react to phloroglucin, and chloroform is not efficient in preventing the growth of Bacteria which would inevitably complicate the experiment. Positive evidence in favour of the hadromase is given by the fact that the Fungus- extract lost its wood-destroying properties when boiled, and that a white precipitate was thrown down by alcohol which had the same destructive action upon lignified cell-walls. It does not, however, afford absolute proof that the hadromase is a product solely of the Fungus, as, unless pure cultures were used under the strictest precautions to exclude Bac- teria, the problem is complicated by the probable development of these organisms, which I have shown also possess the power of destroying lignin compounds. The important papers by Marshall Ward upon the biology of Stereum hirsutum , and by Bififen upon the biology of Bulgaria polymorpha , which treat of the action of these Fungi upon the xylem, afford strong evidence in favour of a delignifying enzyme, these authors describing a gradually progressive delignifying action of the Fungus. In the examination of his cultures of Stereum hirsutum upon Aesculus wood, with the reagents for differentiating lignified membranes from those 136 Potter . — On the Occurrence of Celhilose in the devoid of lignin, Marshall Ward finds ‘ during the first month no distinct reactions . . . and stains, such as Delafield’s haematoxylin, do not colour the walls blue or purple, but merely brown or yellowish ; but in some cases a thin lining layer is found to react in wood acted on by the Fungus for six weeks to a couple of months, and the altered layer gets more and more decided as the action progresses.’ The unequal distribution of cellulose make it possible that its presence might have been overlooked in the first blocks examined, and no mention of any examination of the blocks prior to sterilization is made. I have not had the opportunity of studying a quite freshly cut stem of Aesculus of more than about three inches in diameter, but, judging from numerous sections of small branches and some much larger pieces of Oak, the distribution of cellulose is always somewhat irregular, and hence certain of the blocks cut from a good-sized stem might be entirely lignified, while in others lignification would be by no means complete. In Aesculus this partial lignification occurs more especially in the spring wood, and this may explain why it was found that c in some transverse sections the spring wood is invaded much more rapidly than the autumn wood of the same annual ring ’ when attacked by Stereum hirsutum. Marshall Ward’s figure 17 corresponds exactly with many sections I have seen of normal wood not attacked by any Fungus. In his description of tangential longitudinal sections of cultures a month old which had been treated with gentian-violet and Congo-red, Biffen noted that it was easy to find, especially in the vessels, walls in which every pit was marked out by a well-defined, bright pink zone surrounding it, indicating that that portion of the wall had been delignified and a cellulose basis staining with Congo-red remained. It is significant that he observed, moreover, ‘that no hyphae passed through the majority of these pits, so that one has to assume the secretion of a delignifying enzyme in quantity by the fungus into the wood-elements.’ This appears to be an unnecessary assumption which cannot be allowed in face of the fact that this appearance is observed very commonly without any Fungus being present. That the margins of the bordered pits often remain unlignified, especially where the vessel crosses the medullary rays, has been demonstrated in the instances above quoted ; and that it is no mere optical illusion is shown from the fact of the borders of the pits staining distinctly blue upon treatment with iodine followed by sulphuric acid, as in Salix (compare Fig. 4). Among herbaceous stems also I have observed this ring of cellulose very beautifully shown by chlor-zinc-iodine in the large wood-vessels of Cucurbita. It has become clear that the presence of cellulose in the wood-fibres cannot be attributed entirely to the action of a delignifying enzyme, and it is now necessary that the enzyme in each case should be isolated. i37 Xylem of Woody Stems . Granted that the Fungus finds in the cellulose a necessary food-element and given the existence of layers of this substance in the wood-fibres, it follows that the hyphae would proceed in the direction of the source of supply, and the delignifying enzyme only comes into play where this supply is no longer available. The large amount of cellulose which is now shown to occur in the wood, at all stages, having been previously unrecognized, the conclusion becomes inevitable that much of the effect ascribed to the action of Fungus-hyphae is referable to conditions already present, and the action upon the xylem due to any process of sterilization must be taken into account when estimating the effect of the penetration of hyphae into the wood in culture experiments. Without denying the existence of a delignifying enzyme, it is probable that the Fungus first attacks the elements where cellulose is already present, which would account for the direction in which the destruction of the wood sometimes advances. Thus in the formation of partridge- wood by Telephora perdix the Fungus attacks certain areas which eventually become hollow. Possibly these areas are those in which cellulose is present and which, therefore, succumb first, and the hyphae penetrate along the line where it is to be found, and the same reason may explain why Polyporus sulphureus extends in the direction of the annual rings. In describing the process of decay in the Oak, due to Polyporus dryadeus , Fr., Telephora perdix , M., and Stereum hirsutumy Fr., Hartig distinguishes two methods of attack, one accompanied by delignification and another in which the conversion of lignin into cellulose does not take place. It is conceivable that Hartig may be describing the action of these Fungi upon elements already containing cellulose and upon those in which cellulose is absent. The fact that in the former method of attack the progress of decay is more rapid gives support to this suggestion. It appears probable that the occurrence of cellulose in certain areas represents a stage of arrested development. The largest distribution of cellulose was very frequently observed in the wider annual rings, in which the wood appeared to have grown rapidly in a favourable season, and the direction of orientation evidently has its influence also. There may be some connexion with the phenomenon known to gardeners as the ‘ ripening ’ of the wood, and in certain seasons when the wood is said not to ‘ripen’ it may be an expression of the fact of an incomplete lignification. I may mention that I imagine this condition of partial lignification may be found to be generally prevalent. Among herbaceous plants, for instance in Vicia faba and Oenothera biennis , in the older stem-internodes just above the surface of the ground the cellulose lining of the wood-fibres is very beautifully shown. 1 38 Potter . — On the Occurrence of Cellulose in the Summary. 1. It is found that a gelatinous thickening layer which reacts at once to the various colour-tests for cellulose occurs very commonly, though very irregularly, in the fibre-walls of the xylem as a normal condition in a great number of perfectly healthy trees, in all localities and situations. It may have a very partial distribution, or may occur very generally and conspicuously through the stem, and may be present only in parts of the same annual ring. Sometimes this innermost layer is represented only by a thin lining, at other times by a very broad band which appears swollen and occupies a large part of the lumen. A margin of cellulose is also often present round the bordered pits. 2. A delignification of the xylem is effected by the action of boiling water, which removes substances which impregnate the cellulose and react to the lignin stains, leaving a basis of cellulose as indicated by the reactions to the various colour-tests. This is shown by submitting thin sections of wood to the action of boiling water, and is confirmed by the fact that ether removes from the watery extract obtained from sawdust and fragments of wood a substance which reacts to lignin tests. Further, cold water, operating for a longer period, has a similar power in extracting from the xylem a substance which reacts to phloroglucin and thallin sulphate, and thus by continued soaking in water wood under- goes a partial delignification. A substance showing a blue-green colour with phenol- HC1, the test for coniferin, is apparently not extracted. 3. It is demonstrated that the ‘ lignin ’ substances extracted from the xylem by water are destroyed by certain micro-organisms, and that these flourish more vigorously in the sap-wood than in the heart-wood extracts. This latter point suggests that the heart-wood contains some substances not readily attacked by Fungi or Bacteria, and accounts for the fact of its greater durability. 4. The presence of this unlignified layer in the wood-fibres probably represents a stage of arrested development. Its general prevalence having been overlooked, the conclusion is inevitable that the occurrence of cellulose which has been attributed to the action of Fungi must to some extent be ascribed to conditions already present, and the effect of any method of sterilization must also be taken into account. The delignification cannot be entirely attributed to an enzyme secreted by Fungi. Postscript. Since the above was written I have examined the roots of Lupinus , Phaseolus, Polygala Senega and Aesculus, and have found the cellulose lining in the fibres very distinctly shown in all these cases. Doubtless fibres remaining partially unlignified are of quite common occurrence in roots. 139 Xylem of Woody Stems, In view of the examination of additional specimens of Fraxinus grown in a different situation, my statement with regard to the complete lignifica- tion of the fibres requires some modification. Recently I have had the opportunity of examining the wood of Fraxinus attacked by Polyporus hirsutiis , and found that when the xylem was treated with chlor-zinc-iodine the fibre-walls became an intense blue. In accordance with the observa- tions previously made, this seemed to indicate a delignification due to this parasite. But on examining perfectly healthy shoots from the same tree, and also others from trees in the same locality quite free from any fungoid attack, it was at once seen that in these cases many of the fibres were only partially lignified ; hence the delignification was not primarily due to the action of this Polyporus. Local conditions of soil and climate seem in some cases to retard the complete development of the xylem, and thus render such trees constitutionally weak and very liable to attack. For instance, in the particular locality now under consideration very few of the trees were free from the infection of Polyporus hirsutus . Literature. Abromeit, J. (’84) : Ueber die Anatomie des Eichenholzes. Prings. Jahrbiicher fiir wissensch. Botanik, Bd. xv, 1884. Biffen, R. H. (’01) : On the Biology of Bulgaria polymorpha , Wett., Annals of Botany, vol. xv, 1901. Czapek, F. (’99) : Ueber die sogenannten Ligninreactionen des Holzes. Zeilschr. fiir physiol. Chemie, Bd. xxvii, S. 141, 1899 ; Botanisches Centralblatt, Bd. lxxix, 1899. (’99) : Zur Biologie der holzbewohnenden Pilze. Berichte der deutschen Botanischen Gesellschaft, Bd. xvii, 1899. Hartig, R. (’78) : Die Zersetzungserscheinungen des Holzes der Nadelholzbaume und der Eiche. Berlin, 1878. Hegler, R. (’00) : Histochemische Untersuchungen verholzter Membranen. Flora, Bd. lxxiii, 1890. Marshall Ward (’97) : On the Biology of Stereum hirsutum Fr. Philosophical Transactions of the Royal Society, vol. 189, 1897, Series B. Mayr, H. (’84) : Zwei Parasiten der Birke, Polyporus betulinus, Bull, und Polyporus laevigatus , Fries. Botanisches Centralblatt, Bd. xix, 1884. Potter, M. C. (’00) : On a Bacterial Disease of the Turnip ( Brassica napus ). Proc. of the Royal Society, vol. lxvii, 1900. Sanio, C. (’60): Einige Bemerkungen iiber den Bau des Holzes. Botanische Zeitung, Bd. xviii, i860. (’63) : Vergleichende Untersuchungen iiber die Elementarorgane des Holzkorpers. Botanische Zeitung, Bd. xxi, 1863. Schellenberg, H. C. C96) : Beitrage zur Kenntniss der verholzten Zellmembran. Pringsh. Jahrbiicher fiir wissenschaftliche Botanik, Bd. xxix, 1896. Schrenk, H. VON (’00) : Two Diseases of the Red Cedar, caused by Polyporus juniperinus , N. Sp., and Polyporus carneus , Nees. U. S. Department of Agriculture, Bulletin No. 21, Washington, 1900. • (’93) : A Disease of the White Ash, caused by Polyporus fraxinophilus. U. S. De- partment of Agriculture, Bulletin No. 32, Washington, 1903. Singer, Max (’82) : Beitrage zur naheren Kenntniss der Holzsubstanz und der verholzten Gewebe. Sitzungsber. d. Wiener Akad., Bd. lxxxv, Abt. i ; Botanisches Centralblatt, Bd. x, 1882. Strasburger, E. (’91) : Ueber den Bau und die Verrichtungen der Leitungsbahnen in den Pflanzen. Jena, 1891. Wisselingh, C. van (’98) : Microchemische Untersuchungen iiber die Zellwande der Fungi. Pringsh. Jahrbiicher fiir wissenschaftliche Botanik, Bd. xxxi, 1898. 140 Potter. — On the Occurrence of Cellulose in Xylem. EXPLANATION OF FIGURES IN PLATE VIII. Illustrating Professor Potter’s Paper on the Occurrence of Cellulose in Xylem. Fig. 1. Transverse section of a young stem of Quercus, diameter 1-3 cm. The shading extending partially round the stem indicates the regions of cellulose distribution. Fig. 2. Transverse section of wood-fibres from a stem of Quercus, cut when fresh and stained with Delafield’s haematoxylin. The lignified layers are yellow and the enclosed gelatinous thickening layer violet. Zeiss D., Oc. 4. Fig. 3. Portion of a transverse section from the stem of Fagus, cut when fresh, showing the inner gelatinous layer present in some wood-fibres, coloured violet. Zeiss D., Oc. 4. Fig. 4. Longitudinal section from the wood of Salix , cut when fresh and stained with chlor-zinc- iodine. Showing the margins of the bordered pits stained blue. Fig. 5. Transverse section of wood-fibres from the stem of Fraxinus, cut when fresh and stained with chlor-zinc-iodine. The yellow walls show a complete lignification with no internal cellulose layer. Fig. 6. Similar section from the same stem of Fraxinus , after boiling on four consecutive days and staining with chlor-zinc-iodine. The lignified parts of the walls remain yellow, while the inner swollen layers are coloured violet, having undergone delignification. Fig. 7. Transverse section from a stem of Aesculus , steamed for two hours on three consecutive days, stained with Congo-red, and mounted in Canada balsam. The inner delignified layer of the fibre- walls is stained red and lies detached, swollen, and crumpled in the lumen. Fig. 8. Transverse section of the wood-fibres from a stem of Quercus after steaming three times, stained with chlor-zinc-iodine. Showing the advancing process of delignification due to the action of boiling water. r^/htsJLOils of Botany N . H . Po tter, del . POTTER. CELLULOSE IN WOODY STEMS. VolXVIIlPL VIII. University Press. Oxford. Studies in the Dictyotaceae. I. The Cytology of the Tetrasporangium and the Germinating Tetraspore. BY J. LLOYD WILLIAMS, Assistant Lecturer in Botany , University College , Bangor. With Plates IX and X. IN 1897 I published an account of the discovery of motile antherozoids in Dictyota1. At the Bristol meeting of the British Association (1898) the mode of fertilization was described. The cytology of the antheridia and oogonia and of the apical region of the tetraspore plant was also ex- plained and illustrated by means of figures. It was found that the full number of chromosomes obtained in the apical nuclei of the asexual plants, while in the single division of the oogonium and in the early divisions of the antheridium the reduced number prevailed. This necessi- tated a more careful search for the actual reduction division. Ultimately it was found to be the first division in the tetraspore mother-cell. Mean- time Mottier 2 in 1898 announced his discovery of centrosomes in Dictyota. Subsequently (in 1900 3), he published a detailed description of the reduction division, as well as of the second division of the tetraspore mother-cell and of the vegetative mitoses in the thallus-cells. I have worked out the cytology of all the various kinds of cells in the three forms of Dictyota , male, female and asexual, and a considerable body of facts has been gathered relating to the natural history of the living plant, both under culti- vation and in its natural habitat. As far as cytological evidence is concerned there seems to be no reason to doubt that we have here a clear case of alternation of generations. Publication of the results has hitherto been delayed in the hope of completing the evidence by actual cultivation of the plant from spore to spore. All the numerous attempts made to produce this result have hitherto failed, owing to the difficulty of getting the plant to fruit in cultivation, and it has been thought best not to further delay 1 Williams, Annals of Botany, xi, 1897. 2 Mottier, Ber. d. deutsch. Bot. Gesellsch., xvi, 1898. 9 Mottier, Ann. of Bot., xiv, 1900. All subsequent references to Mottier will be to this paper. [Annals of Botany, Vol. XVIII. No. LXIX, January, 1904.] 142 Lloyd Williams. — Studies in the Dirty otaceae. publication of the results hitherto obtained. The present paper deals with the sporophyte generation, and the first segmentation of the spore. The cytology of the gametophyte and of fertilization, the very remarkable phenomena associated with parthenogenesis, as well as the natural history of the plant, will be dealt with in subsequent papers. Mottier has described and figured karyokinesis in the vegetative cells of the thallus. Our results being in general agreement it is not proposed to deal further with them. As, however, he has not described the stalk- cell division or the earlier stages of the reducing division, a detailed account of them will be given here. The Stalk-cell Division. One of the elongated rectangular cells of the thallus begins to swell in all directions, but chiefly outwards, in the direction of least resistance. At the same time the nucleus increases greatly in size. The nucleolus appears granulate, there is a fine reticulum uniformly distributed through the nuclear cavity, and upon it are numerous granules. The axis of the nucleus is as yet parallel to that of the thallus. At each pole there is a curved rod-like centrosome, and radiations extend into, and become merged in, the cytoplasmic reticulum. The chloroplasts stain uniformly ; they are much paler than those of the vegetative cells, and at first they are uni- formly distributed. Soon, however, the basal portion immersed in the surface layer becomes greatly vacuolated. In the vertical walls separating the cell from the neighbouring ones there are large pits and the vacuoles lie opposite them. The lateral extension of the cell soon ceases, but it grows rapidly out- wards until it is several times the depth of the cortical cell. In cross section the projecting part is nearly spherical, but a longitudinal section shows an elongated, dome-shaped mass. The dense cytoplasm, containing the bulk of the chloroplasts and the nucleus, is aggregated in the swollen, free portion. The axis of the nucleus has now swung round until it is vertical to the thallus, and the distal pole has a very distinct curved cen- trosome and beautiful radiations ; the space occupied by the latter being quite free of chloroplasts (PI. IX, Fig. 1). A cap of cytoplasm covers the basal pole. From this a cytoplasmic cord extends downwards into the vacuole, where it sometimes branches. The centrosome here is often only distinguished with difficulty and the radiations are completely absent. Stages where the axis of the nucleus is oblique to the thallus are very difficult to find, the change of position probably taking place very quickly. The nucleolus is now large and stains deeply. Sections through it show that it is uniformly fibrillate, or minutely vesicular in texture. Before the spirem there is a stage where the nuclear reticulum shows Lloyd Williams . — Shi dies in the Dictyotaceae . 143 a tendency to draw away from the nuclear membrane (Fig. 1). It is true that this effect is probably due to the fixing reagent, but the readiness with which the effect is produced both in this and the corresponding stage of the next mitosis shows that the relation of the network to the membrane — probably to the cytoplasm — is much less intimate than it is during the spirem stage. The spirem is very coarse, the chromatin granules being unequal in size and the staining not 'deep. It frequently seems polarized, but at this stage no signs of splitting can be seen. The nucleolus now becomes vacuo- late. Frequently, but not as regularly as in the next division, there is a large central vacuole with deeply stained inclusions. Immediately before the segmentation of the thread (Fig. 2) the nucleolus becomes swollen, irregular in form, and distinctly fibrillar in texture, — an appearance which is also seen at this stage in most of the other mitoses. Soon the nucleolar membrane disappears and the free, fluffy ends of the fibrillae project from the general mass of nucleolar substance. This is identical with what occurs in the prophase of the next division (Fig. 19). At this time the nuclear cavity has no reticulum within it. The chromosomes now begin to appear and are often disposed in the neighbourhood of the nucleolus. The latter loses its coherence and becomes an irregular mass of granules and curved fibrils, some of them looking not unlike minute chromosomes, and, unless very carefully counter- stained, very similar in their staining reactions. Besides these, the nucleolar globule, which always accompanies the spindle, is already differ- entiated. This is much less deeply stained, and, except for a small vacuole which sometimes appears in it, is quite homogeneous. Two sheaves of fibres now project into the nuclear cavity, their extremities reaching about one-third of the way to the opposite pole, leaving the equatorial region free. Here are aggregated the fragmenting nucleolus and the chromosomes. The latter are considerably bigger than during the succeeding mature spindle stage; they are curved and distinctly split. Fig. 3 represents one of three oblique sections through such a nucleus. After making allowance for chromosomes that have been cut and consequently appear partly in this and partly in other sections of this nucleus, the number of chromo- somes is from thirty to thirty-four. The spindle when fully formed is intranuclear, the membrane per- sisting close up to the poles. The spindle proper is truncated at the poles, and in some cases (Fig. 4) is not greatly dilated at the equator. The space between the spindle and the membrane is occupied by a few mantle fibres and a large number of curved fibrils almost certainly derived from the disintegrating nucleolus. The nucleolar globule is also seen somewhere in the vicinity of the spindle — it now stains just like the cytoplasm. The chromosomes are uniformly distributed through the nuclear plate in a dense 144 Lloyd Williams . — Studies in the Dictyotaceae. flat disc. Those at the periphery are curved and have the free ends directed outwards. Careful countings of polar and other views give from twenty-seven to thirty-two chromosomes. The polar radiations and centrosomes are not nearly as clear as they are during the early prophase stage. When the daughter chromosomes are halfway to the poles the nuclear membrane is still intact, and the centrosomes are visible at both poles. Between the chromatic discs and the poles the cones of fibres are clear of granules and the fibres are fine and close set, while the interzonal fibres are, as usual, fewer, far coarser, and the space between them invaded by granules and fibres. When the chromosomes reach the poles, they form two flat discs, the distal one being much greater in diameter than the basal. Very often a chromosome lags behind the others and projects towards the equator (Fig. 5). The membrane has now disappeared, so that the nucleolar fibrillae and globule are excluded from the chromatin discs (Figs. 5 and 6). The coarse connecting fibres still remain and the distal centrosome and radiations are distinct. Sometimes the basal cen- trosome can be made out, but there are no radiations. The lower half of the figure projects into the vacuolated region while the upper half is imbedded in the denser cytoplasm. A membrane now surrounds each of the daughter nuclei, the diameter of which is much less than in either the preceding or succeeding stages. Instead of lying parallel in a flat disc, the chromosomes now form a tangled mass, but they are still distinct and preserve their curved form. In this condition it is often easier to estimate their number than during the late diaster stage. At this period nothing can be distinguished within the nuclear membrane besides the chromosomes. The chromosomes gradually fuse so as to form an irregular coil or reticulum with a very few meshes. This coil is very much thicker in the tetraspore mother-cell nucleus than it is in the sister nucleus. Each thick thread is at least double, the median line can frequently be seen to be lighter than the two edges (Fig. 6, upper nucleus). The strands of this irregular coil are often swollen at intervals, and there may be ten to sixteen of the swellings. Throughout this period of partial fusion the membrane of the upper nucleus, on the distal side, is very irregular, there being a marked projection towards the centrosome. A very faint reticulum now appears in the nucleus, and the condensation of the coil proceeds till very often it assumes the form of a thick open ring. Ultimately the chromatin mass becomes spherical in form and uniformly granular or finely fibrillate in appearance, in fact it is the nucleolus. It is worthy of note that of the four mitoses described in this paper this is the only one in the telophase of which there is no differentiation of nucleolar and chromatin masses. The diameter of the nucleus is now much greater, and there is a fine but very faintly staining reticulum. 145 Lloyd Williams. — Studies in the Dictyotaceae. The septum dividing the stalk-cell from the sporangium proper soon makes its appearance, and the separation of the tetraspore mother-cell is completed. The First, or Reducing Division of the Tetraspore Mother-cell. Before the initiation of the spirem the nucleus passes through a similar condition to that shown in Fig. i, where the reticulum easily separates from the nuclear membrane. Soon, however, there is a very thin convoluted spirem distributed through the nuclear cavity. The fine granules upon this stain but feebly as yet, but here and there a few larger, but still small, deeper-stained bodies appear between the threads (Fig. 7). The nucleolus preserves for a time the fibrillar appearance already described, but as the spirem develops a vacuole appears in the former with, at first, a few smaller ones. Soon, however, there is only one fairly large vacuole which has a firm, deeply staining outline, and there are darker staining granules within it. The axis of the nucleus is once more parallel with that of the thallus, but the poles can only be distinguished with great difficulty owing to the faintness of the radiations (Fig. 7), and the centrosomes, if they exist at all, are not distinguishable. No signs were seen at this stage of the division of the centrosome, nor could the change of position of the nuclear axis be followed. The spirem gradually becomes more prominent, and shows a tendency to aggregate near the poles. A deeply staining spherical or sometimes angular body, considerably larger than any of the granules hitherto described, now makes its first appearance. It is peculiar to this mitosis, and it persists till after the splitting of the spirem (Figs. 8-12). It is always present during these stages, and there is hardly ever more than one. Whether it directly separates from the nucleolus or is formed by the fusion of the smaller granules shown in Fig. 7 could not be decided. The body will be designated for the present the chromophilous spherule. It should not be confused with the nucleolar globule which appears during several of the mitoses in this plant. The faintly staining cloudy nucleoplasm which appears during the synapsis is absent from this stage. The above condition leads directly to the very interesting knot stage or true synapsis , which forms a most striking feature in the cytology of the reduction stage, but appears nowhere else in the whole history of the plant. Mottier 1 says : * During the prophase of both divisions in the tetraspore mother-cell the behaviour of the chromatin differs strikingly from that in the higher plants. There is not developed here a regular and continuous chromatin-spirem which segments into the chromosomes, but these arise as isolated masses often differing much in size. It gives the impression 1 1900, l.c., p. 188. L 146 Lloyd Williams . — Studies in the Dictyotaceae. that the quantity of chromatin is not sufficient to form a continuous spirem.’ This, however, is a mistake. It would be hard to find a more dis- tinct spirem than this, or one more interesting in its development. Many hundreds of them have been studied, and in some material they are far more numerous than spindles, showing that they persist for a considerable time. The main features of the spirem stage in Dictyota are so constant that no one after seeing the synapsis could possibly call it an artefact. After completing the work on Dictyota I began to study the cytology of Padina. This plant was found to agree with Dictyota in all the principal points, while it presented the great advantage of having the various stages sorted out as it were. In Dictyota the sporangia are isolated and the various stages are generally mixed up in a haphazard fashion. Thus the finding of any particular stage is a matter of chance, and in order to fill up some of the gaps several thousand sections have been examined. In Padina , however, the sporangia are collected together in concentric bands on the fan-shaped thallus, and in these bands many hundreds are so closely crowded together as to be in actual contact. Furthermore, all the primary sporangia in any one band are approximately in the same stage. Thus the first band behind the apical region is frequently found to have all the nuclei in the spirem stage, while the older band behind this has the sporangia in various stages of division. In very old sori, where there has been liberation of spores, new sporangia are initiated ; here the uni- formity above spoken of does not obtain. The failure to recognize this stage is probably due to the fact that the tetrasporangia in Mottier’s plants were too old, for it is a striking peculiarity of Dictyota and Padina that the spirem appears as soon as the mother-cell has been separated and while the sporangium is still young and small ; it then disappears, and the nucleus assumes the appearance of the resting stage. In the preparation for this mitosis there are then three well-marked stages : (1) A very precocious spirem, which lasts a con- siderable time and goes through several interesting changes. (2) A reti- culum stage, which is exactly like that of rest. It is with this that Mottier’s studies commence, and his Fig. 1 comes between Figs. 14 and 15 of this series. (3) A prophase stage which is passed through somewhat quickly. The Synapsis Stage. The spirem is now very long, thin, and deeply stained. It is closely coiled up into one or two clumps of angular loops (Figs. 8 and 9). The two knots appear closely pressed against the nuclear membrane, one opposite each centrosphere. The centrosome and radiations are still very indistinct, those shown in Fig. 8 being the best observed. The nucleus is very large, but the membrane is thin and ill-defined. The knot is so pressed against the membrane that it is often difficult to decide whether the spirem is entirely within it, and it i47 Lloyd Williams . — Studies in the Dictyotaceae. frequently seems as if the thread were directly continuous with, or attached to some of the cytoplasmic fibres. Very often, also, the long angular loops of the spirem seem to radiate from the vicinity of the centrosome towards the nuclear cavity. Although generally in two knots, the spirem is con- tinuous, one or several connecting threads stretching across the cavity from one knot to the other. These frequently pass over the surface of the nucleolus, and can be seen to be in contact with it (Fig. 8). The behaviour of the latter is interesting. It invariably becomes swollen, very irregular in shape, and greatly vacuolated, btit never fibrillar. Many cases are seen where a portion is drawn out towards the spirem, to which the apex of the projection is attached (Figs. 9, 10). Sometimes big roundish lumps appear to be on the point of separating from it, and at others the whole nucleolus seems to be fragmenting. This, however, it never does ; as soon as the synapsis is over the nucleolus resumes its spherical form. Fig. 9 shows a case where there is only one knot. It sometimes happens that the spirem is entirely on the basal side of the nucleus (i. e. on the side nearest the ‘ basal ’ or ‘ stalk-cell ’) and not in the vicinity of the poles. In Padina it is most frequently either on the basal or on the distal side, or both. This has probably some reference to the fact that here the sporangia are com- pressed laterally by contact with each other, so that the axis of the first division spindle is vertical to the surface of the thallus instead of parallel to it as in Dictyota. The other constituents of the nucleu^at this period are the spherule above described and a small quantity of cloudy nucleoplasm, which on closer examination resolves itself into a very fine reticulum which, however, stains but slightly. The staining reactions at this stage are as follows. The spirem, spherule, and the physodes stain deeply with the chromatin stain, the two latter often showing different shades of colour. The nucleolus, chloro- plasts, and cytoplasm take the basic stain, the two former more deeply than the latter. It is very evident that there is an intimate relation between the spirem and the cytoplasm at this stage, and that communication between them is chiefly localized at the poles. The frequent connexion between the nucleolus and the spirem, as well as the poverty of the former in chromatin, both seem to indicate that much, if not all, of the chromatin comes out of the nucleolus. The Beaded Spirem Stage. The thread now becomes thicker, and the chromatin discs can more easily be seen (Fig. 11). The knots loosen out and the thread becomes distributed more uniformly over the membrane on the basal side of the nucleus, more rarely over the upper side. The cloudy nucleoplasm is also generally aggregated along the basal side, so that the spirem seems embedded in it. The nucleolus gradually 1 48 Lloyd Williams. — St tidies in the Dictyotaceae . becomes more spherical, and a large vacuole appears in it containing minute fibrils similar to those shown in Figs. 11 and 12. After a time the spirem is more uniformly distributed over the whole surface. Fig. 11 does not show this very clearly, as the section is a median one, the rest of the spirem being in other sections. The nucleoplasm begins to diminish in amount ; some of it is seen as little fluffy threads attached to the spirem. The staining reactions are peculiar. With brazilin and picric- Hofifman-blue, for instance (the former being the chromatin and the latter the plasma stain), the spirem and the nucleolus stain reddish brown, the spherule yellowish red, while everything else, including the fibrils in the nucleolar vacuole, stain blue. Later on the spirem loses its chromatin reaction and colours blue like the chloroplasts. No definite conclusion respecting the chemistry of the nucleus can be drawn from this, but it is interesting to note how the nucleolus and the spirem reverse their respective colour reactions in the knot and the later spirem stage. The Split Spirem. The thread now instead of lying along the nuclear membrane becomes more evenly distributed through the cavity. Longi- tudinal fission can be seen here and there, and before long it is found to form two distinct threads which often twist round each other (Figs. 12 and 13). The granules are irregular in size and more distant from each other. As in the preceding stage they show no special affinity for chromatin stains. The nucleolus stains more deeply, and shows a very big vacuole with inclusions. The spherule is still present, and besides the nucleole it is the only nuclear structure that colours with chromatin stains. The nucleoplasm diminishes in amount and finally disappears, being probably incorporated in the reticulum formed out of the split spirem. The centro- spheres are still very indistinct, the mother-cell is bigger and more vacuolated than in the preceding stage. The 4 Resting Stage.’ The next stage in the history of the spirem is difficult to make out. Instead of a comparatively small number of threads which are split, one sees a large number of very fine threads with numerous granules, very variable in size and disposition, crossing the nuclear cavity in all directions, and in many places appearing to form a reticulum (Fig. 14). Whether this appearance is due to the mere separation of the halves of the split spirem, with a lengthening and thinning out of the threads, or of this with a second split superadded, it is very difficult to make out. The halves undoubtedly do separate widely, and the identity of the constituent threads is completely lost, but a careful examination of a very large number of examples failed to disclose any clear evidence of such a second split. On this account, and after a close study of the prophase stage (Figs. 17-19), I have come to the conclusion that there is no second split at all, and that the appearance seen in Figs. 14, 15, is due to an alveolation of the separated halves of the chromatin thread Lloyd Williams . — SHidies in the Dictyotaceae . 149 and a joining together of the very fine reticulum thus produced by fine cross threads, so that the nucleus appears to be in a state of rest. The spherule now stains paler than before, then breaks up into two or several smaller ones, and at the close of this stage completely dis- appears. After a time the sporangium presents several striking features. The cytoplasmic constituents when seen in section are arranged in evident zones, so that the term c zonated ’ might be appropriately applied to it. The centrosomes and radiations are now exceedingly distinct at both poles, and there is a narrow zone of kinoplasm extending round the sides of the nucleus. This area is quite clear of chloroplasts. Immediately outside, and in sharp contrast to it, there is a zone of dense protoplasm with crowded chloroplasts which stain very deeply, generally taking more or less of the chromatin stain (Fig. 15). The outer region is vacuolated, and here the chloroplasts generally stain like the cytoplasm. This is not included in the figure. The nuclear membrane is firm and clear. At the poles it is generally very irregular ; the example selected to draw Fig. 15 from does not show this very clearly. It generally forms projections towards the centrosomes, and this causes a number of folds and wrinkles. Frequently a conical extension of the polar membrane has its apex touching the middle of the rod-like centrosome, the two ends of which are free. Occasionally the projection is truncated at the apex, and is in contact with the rod along its whole length. When this is viewed from the side the centrosome looks almost like a fold of the nuclear membrane. In Padina there is the same arrangement of the cytoplasm into concentric spheres, though perhaps not quite so pronounced as in Dictyota . The centrosome and radiations are very much less distinct. The nucleolus is spherical and somewhat deeply stained, but spongy, with large and small vacuoles. The only other constituent is the reticulum already described. The Prophase Stage. Very soon thick cloudy strands appear in the nucleus. There seems to be little, if any, diminution of the reticulum, so that it is not clear whether these are derived from a further condensation of the denser strands of the reticulum or not. As the cloudy bands assume a more definite form, a longitudinal split makes itself evident in them (Fig. 16). When these bodies have still further condensed, and assumed the forms of chromosomes, they are seen to be either closed rings or looped-up bands with their free ends crossing. This is well shown in Fig. 17 from Taonia. The next figure (18) represents three rather thick sections of a similar nucleus from Padina. These include all the chromo- somes, and they are uncut with the exception of the one in the extreme left of c, the bend joining the two halves of which is shown in section b. 150 Lloyd Williams . — Studies in the Dictyotaceae . This shows clearly that the loop- figures have not been caused by cutting off the ends of elongated rings. The failure to find evidences of a second split of the chromatin thread, and a careful study of the development of the chromosomes, inclines me strongly to Farmer and Moore’s view of the origin of these bodies, as recently advanced before the Royal Society1. According to this the space inside a ring-chromosome, or that between the two limbs of a loop-figure, does not represent the first split, but rather the space between the two constituents of a bivalent chromosome bent over on itself so that the two ends meet, or the two limbs cross, or some modification of these forms is brought about. The split itself (the only split according to this theory), though partly obliterated by the condensa- tion of the chromosomes, is still evident in the free ends of several of the chromosomes in Figs. 17 and 18. During this stage the nuclear reticulum is somewhat dense and the nucleolus minutely vesicular. By the time the spindle cones begin to develop the chromosomes have still further condensed until they are small and deep-staining, but the ring form is very prevalent (Fig. 19). Thus, although a long time has elapsed since the spirem stage, and in the meantime all traces of it have been lost in the apparently formless reticulum, yet the chromosomes make their appearance already segmented and longitudinally split. The length of the spirem period and the comparative shortness of the prophase irresistibly suggest that the arrangements for mitosis are completed, or nearly so, in the former ; that the grouping of the various constituents is there determined, and however disguised in the reticulum of the post- spirem stage that it still exists. This is equivalent to saying that the spirem maintains its identity throughout, and that when the stimulus is given which produces the prophase there is a rapid condensation of the previously extended, unrecognizable spirem into compact, deeply stainable, split, bivalent chromosomes. This does not exclude the possibility of substance passing into the chromosomes from the nucleolus. From actual observation, however, we can determine that this must be small in amount, for the bulk of the nucleolar substance is otherwise accounted for. In Fig. 19 the nucleolar membrane has disappeared, and the contained fibrillae, as deeply staining as the chromosomes, lie together in an irregular mass, some of them with their free ends projecting. Other sections show, as in the case of the stalk-cell division, that the spherical globule sub- sequently seen near the spindle is also formed out of the disintegrating nucleolus. The Spindle Stage. Of this a great many examples have been studied, and they all agree in their main features. The spindle is intra- nuclear, very narrow, and nearly iso-diametric. The nuclear membrane is intact excepting at the poles, where there are evident gaps nearly as 1 Proc. Roy. Soc., vol. lxxii, p. 104, 1903. Lloyd Williams. — Studies in the Dictyotaceae. 15 1 wide as the spindle itself (Fig. 20). There are divergent mantle-fibres between the spindle and the membrane — these rarely reach as far as the equator. The radiations and centrosomes are much fainter than during the post-spirem stage. Fig. 22 shows a very good centrosome from an oblique section ; there is a very similar one in the section containing the other pole. In many of the figures the spindle-fibres seem tense and nearly straight, while the mantle-fibres are lax and wavy (Fig. 20). The former give one the impression of having contracted and shortened till in some cases (not so well shown in the one drawn) the polar edges of the membrane are curved inwards and the centrosome is sunk within the gap. Here, as in the preceding mitosis, there are nucleolar fibrillae, which, as observed by Mottier, often stain almost as deeply as the chromosomes, but when a deep counterstain has been employed there is no difficulty in differentiating them from the chromatin. The nucleolar globule is nearly always present, generally at the periphery in the equatorial region. It always stains like the cytoplasm. The chromosomes, as already stated by Mottier, are sixteen in number, and decidedly heterotype in character. Figs. 20 and 21 show their form, and it seems hardly necessary to describe them in greater detail. Their small size makes it difficult to decide how they are placed on the spindle ; and whether the division is such as to bring about true reduction or not is for the same reason equally difficult to decide. In sections parallel to the surface of the thallus it happens occasionally that the nuclear figure is flattened so as to widen out considerably at the equator. In these cases the chromosomes are widely separated, and consequently more easily counted. The Anaphase Stage presents no peculiar features — the description of the stalk-cell division anaphase would almost apply to this. Mottier, however, states that chromosomes on their way to the poles often fuse together into larger masses. I have seen no instances of this in healthy cells. From the post-spirem stage onwards a large number of tetra- sporangia become arrested in their development, and frequently this takes place while the nucleus is in mitosis. In such cases the chromatin shows various abnormalities of structure. As a rule, however, one finds it easy to recognize such abnormal examples from the peculiar appearance of the cytoplasm. Occasionally, in the earlier stages of degeneration, there may be difficulty in deciding whether a phenomenon is normal or not. Here one has to depend entirely upon comparison of a large number of cases ; from such a comparison I have come to the conclusion that the fusion of chromosomes above spoken of never takes place excepting in degenerating sporangia. The Telophase Stage is strikingly different from that of the stalk-cell 152 Lloyd Williams . — Studies in the Dictyotaceae. division, and though more like that of the first segmentation of the spore there is a difference in the condition of the chromatin mass. Mottier has good figures of this stage (see his Figs. 8, 9). Instead of the chromosomes forming one large, deeply stained body, two, sometimes more, are formed. One is very large, irregular in form, and composed of fine fibrillae and granules, while the other is smaller, denser, and more suggestive of a nucleolus (Fig. 23). The former is undoubtedly the chromatin. In this condition it is exactly like a stage in the fragmentation of the nucleolus. A little later the chromatin fibrillae are seen to be more widely distributed through the nucleus (Fig. 24) ; eventually they disappear altogether. Co- incidently with this the reticulum becomes more and more evident and the nucleus increases in size (Fig. 25). The Second Mitosis in the Tetraspore Mother-Cell. The axes of the two daughter-nuclei are now parallel to each other and to the surface of the thallus, but at right angles to that of the stalk-cell. Thus a section tangential to the thallus shows both nuclei longitudinally. In such a section a first division spindle would also be cut longitudinally but it would be parallel to the axis of the stalk-cell, whereas in the stalk- cell division the figure would be cut across so as to give a polar view. During the prophase stage the chloroplasts are densely aggregated in the vicinity of the two nuclei, while the median plane is sharply marked out by their absence or scarcity. The two nuclei are somewhat flattened on the outer surfaces, whence radiations curve round the ends of each nucleus towards those of the sister nucleus. The spirem is thick and irregular, and it soon segments, the nucleolus at the same time going through the usual process of disintegration into fibrillae (Fig. 26). In both nuclei ‘ kinoplasmic 5 activity is confined to the nuclear surfaces remote from each other — those nearest the periphery of the mother-cell. It follows from this that in the early spindle the cones of fibres are not in a straight line. This has been well shown by Mottier. In a nucleus similar to his Fig. 12 the scattered chromosomes have been counted without difficulty, and here again the reduced number has been found to obtain. The mature spindle (Fig. 27) is an interesting object, especially when viewed in profile. Owing to the apparent c kinoplasmic ’ repulsion above described the spindle is placed against the outer side of the nucleus, the sister nucleus on the opposite side of the cell presenting the same features but reversed as to direction. Even at maturity there is frequently a slight curve in the figure. The centrosomes are not easy to make out, but the radiations are clear, and as before curved towards those of the sister nucleus. The spindle itself is elongated and narrow, the chromo- somes are curved, and countings of both oblique and polar views leave no doubt that the number is the reduced one. The remaining constituents — ■ 153 Lloyd Williams. — Studies in the Dictyotaceae. mantle-fibres, nucleolar globule, and fibrils — are shown in Fig. 27 and call for no special comment. Mottier says that ‘ the invariable tendency of the chromosomes to collect and fuse into larger masses has made it impossible to determine accurately their number in this division. In the equatorial plate often only two large chromatin masses are seen (Figs. 14, 15V This description does not apply to the material examined by me, where the chromosomes remain distinct up to a late stage in metakinesis. Both in the prophase and spindle stages the chromosomes are clearly seen to be homotype in character. The diaster stage is of the usual type. The figure becomes very narrow and elongated as a rule (Fig. 38). The membrane often persists until the daughter chromosomes are nearly at the poles. (Mottier says it ‘ disappears soon after the spindle is mature, generally before metakinesis/) The pale-staining nucleolar globule is still visible at this stage. The formation of the daughter-nuclei shows the same phenomena as in the preceding division. Here, however, the nuclei are exceedingly small. The radiations completely disappear. In Padina the four spore- nuclei within the sporangium are bigger than those of Dictyota , their chromatin thread is prominent during the resting period, and the cytoplasm is differentiated into distinct concentric spheres. Karyokinesis in the Germinating Tetraspore. The tetraspore on being liberated, instead of assuming a spherical form as is the case with the oosphere, is nearly always elongated and almost cylindrical. The position of the nucleus is indicated by a lighter area. Division of the nucleus takes place in twelve to sixteen hours. The resting nucleus is much larger than it was in the tetrasporangium. The condition of the cytoplasm — the greater width of the meshes and the separation of the plastids — shows that this is due more to distension con- sequent on liberation from the confined space of the tetrasporangium than to growth. One consequence of this is that the parts of the mitotic figure are much clearer here than in the sporangium. Before the development of the spirem the usual reticular structure is shown. At the same time the nucleolus has the fibrillar structure so frequently referred to already. The spirem (Fig. 30) is coarse and unevenly beaded like that of the stalk-cell division: in this case, however, it is accompanied by a considerable amount of fine reticulum. The spindle (Fig. 31) is intranuclear like the preceding ones. It is, however, much broader at the equator and the centrosomes and radiations are more prominent, a result probably of diminished pressure in the cytoplasm. It is a rather striking fact that no nuclear globule has been observed at any stage of this mitosis. The chromosomes have been counted in a great many figures, and they 154 Lloyd Williams . — Studies in the Dictyotaceae. are always about sixteen in number. Their form is curved just like those of the stalk-cell division. During the anaphase stage the membrane persists up to a very late period (Fig. 33). The mantle-fibres, centrosomes, and radiations are very distinct ; in fact all the accessory structures are far more easily seen here than in the preceding mitoses. The curved chromosomes aggregate at the poles in the usual way with their free ends towards the equator ; a membrane is formed ; then, as in the stalk-cell division, they lose their polarity but remain distinct enough to be countable for some time longer. Here (Fig. 33) again it is seen that the number is the reduced one. The centrosomes and radiations, together with the connecting fibres, are still clear. A striking feature of this and of the next stage (Fig. 35) is the projection of the nuclear membrane towards the centrosome, already seen in some previous stages. Fig. 34 shows the chromosomes to be accompanied by a distinctively staining body which in all probability is the rudiment of the nucleolus. The fusion of the chromosomes goes on at first somewhat like that of the stalk-cell division. Here, however, two masses are formed (Fig. 35), of which one is larger and more irregular in form than the other. A reticulum also appears which at first stains but faintly. Later on the smaller of the two masses dwindles in size and finally disappears, the reticulum at the same time becoming more deeply stained and the other body (the nucleolus) assuming a spherical form. Abnormalities. There are various abnormal developments here which it might be profitable to study in greater detail : — (1) Undivided tetraspore mother-cells are sometimes liberated, and several instances have been observed of their nuclei in the prophase or even in the spindle stage. Most of these, however, were evidently arrested, and there is as yet no evidence to show whether karyokinesis was initiated before or after separation from the thallus. (2) Several instances have been seen of liberated sporangia containing two spore-like bodies, each, however, being invested with a cell-wall. In one case the nuclei seemed to be in second division prophase, with spirem and fibrillar nucleolus. In the other the late anaphase stage was seen but the chromosomes could not be counted. (3) This is an abnormality which is of greater interest, as it occurs more frequently and may possibly be of some utility to the plant. Towards the close of the season, instead of dividing in the usual manner to form tetraspores, the mother-cell, without greatly increasing in size or taking on a deeper colour, divides into two and then into a small mass of parenchyma. Richards 1 observed this phenomenon in D. ciliata , but says : 1 Proc. Amer. Acad., 1890, p. 83. x55 Lloyd Williams . — Studies in the Dictyotaceae . ‘ The significance of this multiple division could not be explained with only alcohol material ; it was too frequent, at least in the specimens I examined, to be an accident/ The parenchymatous mass may later on develop an apical cell and grow out into an elongated germling-like branch, and, as will be shown in another paper, it may remain alive when the rest of the thallus has decayed. The cytology of this stage has not been fully worked out, but what probably happens is that the conditions (temperature and light) being unfavourable to the development of the reduction stage (it is evidently very sensitive to external influence) the stalk-cell division is followed by 5 ordinary vegetative divisions. As far as my observation goes this mode of cell-multiplication never follows the reduction division, and when this division is abnormal the death and disintegration of the cell inevitably ensues. (4) In Padina ‘ twin 5 tetrasporangia are of very frequent occurrence. After separation from the stalk-cell the mother-cell nucleus instead of going into synapsis divides vegetatively, there is a longitudinal division of the cell, then the two nuclei go through the various phases of the reduction stage, but they lag somewhat behind the neighbouring sporangia, as if they had lost time by the. extra division. Conclusions. 1. Alternation of Generations. It has now been shown that in the thallus- and in the stalk-cell divisions of the tetraspore plant the nucleus has about thirty-two chromo- somes. The first division of the resulting mother-cell is different from all the others in its long preparation for division, its elaborate and distinc- tive spirem stages, the heterotype character of its chromosomes, and in the fact that here the number is reduced to sixteen. The succeeding division has the same number of chromosomes, and in the young plant produced from the tetraspore the earlier mitoses all show the reduced number. As will be shown in detail in a succeeding paper the thallus-cells of both male and female plants are probably characterized by the reduced number: it is difficult to be absolutely certain where the nuclei are so small. In the oogonial and antheridial divisions, however, there is no doubt about the number. Furthermore, in all these various mitoses there is not one that resembles the reducing division in its distinctive characters. In the interesting series of abnormal figures observed in the partheno- genesis of unfertilized eggs the chromosomes are always scattered, and consequently easily counted, and the number is invariably sixteen, whereas the fertilized oosphere in all its segmentations shows the full number. It 156 Lloyd Williams. —Studies in the Dictyotaceae. it difficult in the face of these facts to resist the conclusion that the germ- ling produced from the tetraspore, with its sixteen chromosomes, is a young male or female plant, and that the segmenting oospore with thirty-two chromosomes to its nucleus is a young tetraspore plant. There is a diffiulty which will have to be explained. In certain localities experience has shown that it is most difficult to find a sexual plant. There are extensive tracts along the coast where all the plants seem to be the form intricata , many of which have no reproductive cells of any kind, and the remainder are nearly always tetrasporic. Taonia has a similar reduction division to Dictyota , and from analogy one would expect to find alternation here. Since the year 1897, however, I have not succeeded in finding a single sexual plant, whereas the others are found without difficulty. This of course is merely negative evidence, and after all there may be a few individuals of the gametophyte generation even in these localities. Assuming that the above observation is correct, one has to account for two things — the failure of the tetraspores to produce sexual plants, and the perpetuation of the former in the absence of the latter. This cannot be fully discussed without going into the whole question of environment, which will be done in a later paper. It may, however, be suggested : — 1. That the same unfavourable conditions that give rise to the intricata form also account for the absence of sexual plants. 2. That one of the several modes of vegetative reproduction occasion- ally resorted to by the plant, including perhaps the curious abnormal development of the tetrasporangium-rudiment above described, may enable the asexual generation to perpetuate itself indefinitely. Some hundreds of culture experiments have been tried in order to test the validity of the conclusions arrived at from cytological evidence. Plants reared from tetraspores and others from fertilized eggs have been kept alive for months, but owing to their sensitiveness to external con- ditions it has hitherto been found impossible to get them to develop reproductive cells. Improved methods are being tried at present, and it is hoped that these will be successful. It may possibly be urged that the two generations in the higher plants are generally quite dissimilar in form and structure, whereas in this case they are identical in both respects. Where, however, the two genera- tions, as in this case, are both strictly aquatic, there seems to be no inherent necessity for a difference of form or of structure. 11. The Nucleolus. With regard to the nucleolus, it has been seen that in the earliest stage in all the mitoses this body is always fibrillar or granular. It is a sugges- tive fact that the same appearance is resumed during the prophase stage, Lloyd Williams,— Studies in the Dictyotaceae. 157 when it swells, becomes angular, and finally its membrane disappears and the fibrillae are allowed to disperse through the nuclear cavity. The stages of the ‘reduction ’ nucleolus are the following : — 1. The nucleolus is uniformly fibrillar, and probably contains the bulk of the chromatin. 2. It becomes vacuolate and the spherule forms. 3. It is much distorted and attached to the spirem ; the staining power becomes gradually less. 4. A large vacuole appears, with fibrillar inclusions. 5. It becomes spongy and more deeply stained. 6. It swells, becomes angular and fibrillar. 7. It disintegrates into fibrillae and a nucleolar globule, which are both excluded from the daughter-nuclei. The distorted nucleus accompanying the ‘knotted’ spirem, and its frequent actual attachment to the thread, irresistibly suggest that the nucleolus nourishes the spirem at this stage. This view is strengthened by the fact that about this time the nucleolus becomes less responsive to chromatin stains. In the succeeding stage the very large vacuole occupying the greater part of the diameter seems to show that the nucleolus has been deprived of much of its substance. The significance of the included fibrils is not clear. The spherule accompanying the reduction spirem is most probably a derivative of the nucleolus. It stains intensely with most chromatin stains, but shows some distinctive reactions, especially its retention of Carbol Fuchsin. Very often the spherule remains bright red when the colour has been extracted from all the other constituents. It is not homo- geneous— very frequently deeper stained bodies can be seen within it. It shows no definite relation to other nuclear structures and its role cannot at present be explained. The disintegration stage shows that the nucleolus contains two very different substances, the fibrillar chromatin-like constituent, and the globule — plasma-like in its colour reactions. If the former substance be not chro- matin, then a third constituent must be added, for it is exceedingly probable that part of the chromatin is stored up in the nucleolus previous to karyo- kinesis. It was shown that in the reconstruction of the daughter-nuclei the process in the stalk-cell division is different from that in all the others. In the tetrasporangium nucleus all the chromatin is at first stored up in the nucleolus, and there is no second body within it. This may be part of the elaborate preparation for the reduction of the chromosomes. In the telo- phase of the next division the chromatin is aggregated in a separate mass of fibrils which, however, become dispersed through the nuclear space and 158 Lloyd Williams . — Shi dies hi the Dictyotaceae. then disappear. The study of these makes it clear that in the karyokinetic stage the bulk of the nucleolus is cast out into the cytoplasm in the form of fibrillae and globule, and it does not seem that either substance is used in the formation of the spindle, for they coexist with the mature spindle up to a late stage in metakinesis. The appearance of the new nucleolus follows so quickly upon the stage when the nuclear membrane encloses nothing but the dispirem, that we are forced to the conclusion that the substance separates from the chromosomes. The centrosomes and other accessory structures will be discussed when the cytology of the sexual generation has been described. With regard to the development of the chromosomes it is clear that if the description given above be correct, whatever the details of the metaphase stage may be, a transverse division of the chromosomes must occur somewhere. No other theory seems to account so satisfactorily for the facts as that of Farmer and Moore, already referred to ; and I regard the figures of the prophase stage in the three Dictyotaceous genera ex- amined as strongly confirmatory of the hypothesis. Lloyd Williams . — Studies in the Dictyotaceae . 159 EXPLANATION OF FIGURES IN PLATES IX AND X. Illustrating Mr. Lloyd Williams’ Studies in the Dictyotaceae. All the figures have been drawn with the aid of the camera lucida and the apochromatic 4.0 mm., aperture 1-30 of Zeiss with ocular 6 (x 800). For convenience of comparison the same scale of magnification has been kept throughout, but a few of the figures having been drawn from tangential sections are apparently smaller than they ought to have been. Figures 2-19 have the stalk-cell on the lower side, and the terms ‘ basal ’ and ‘ distal ’ are used with reference to this cell. Stalk- Cell Division. Fig. 1. Tetrasporangium rudiment before stalk-cell division. Fig. 2. Prophase of the stalk-cell division. The coarse spirem is beginning to segment and the nucleus is swollen and irregular in form. Fig. 3. Early spindle formation. One of three sections, slightly oblique, showing a number of curved, split chromosomes, the fragmenting nucleolus with its globule, and the basal pole with a few spindle-fibres. Fig. 4. Equatorial plate stage. The spindle is intranuclear. The remaining chromosomes are in another section. There is nucleolar globule on the spindle, and nucleolar fibrils between it and the membrane. Fig. 5. Anaphase stage. The two centrosomes are visible, but radiations are confined to the distal pole. The nucleolar globule is on the connecting fibres. Fig. 6. Telophase stage. The chromosomes form an irregular coil the strands of which in places appear double. The nucleolar globule is seen to the right. The tetraspore mother-cell nucleus is bigger than that of the stalk-cell, and its membrane is drawn out towards the distal centrosome. The First or Deduction Division in the Tetraspore Mother- Cell. Fig. 7. Early spirem. The nucleolus is vacuolated, and there are a number of small, deeply staining granules which probably fuse together to form the ‘ chromophilous spherule * peculiar to this mitosis. The radiations, to the right and left, are but faintly suggested, and no centro- somes can be distinguished. Fig. 8. The ‘knot’ or ‘Synapsis’ stage. The spirem is in two knots near the very faint centrospheres. The spherule is seen to the left. Fig. 9. Another example, showing the distortion of the nucleolus and the intimate connexion between it and the spirem. Fig. 10. A slightly later stage. The spirem is spread out on the basal side. A strand is shown attached to the nucleolus. The axis of the nucleus is not median but nearer the basal side. Fig. 11. The spirem is more evenly distributed over the membrane, and is distinctly beaded, with fine fluffy threads projecting laterally. The nucleolus has a large vacuole with included fibrils and granules. Fig. 12. The spirem begins to split in two, and the staining is less deep. The section is transverse to the nuclear axis. Fig. 13. A later stage in the splitting. The nucleolus and spherule appear in one of the other four sections through the nucleus. The reticulum becomes fainter. Fig. 14. The identity of the chromatin thread has been lost, and the nucleus appears as if in a state of rest. The spherule is still present. Fig. 15. A later stage. There are denser strands in the reticulum, the spherule has disappeared, the membrane, centrosomes, and radiations are more distinct, and the cytoplasm is differentiated into separate zones. Fig. 16. Early prophase. Thick cloudy bands appear, with indications of longitudinal splitting. The section is not median, and consequently does not include the nucleolus and centrosomes. Fig. 17. ( Taonia ). The chromosomes appear as loops with their limbs crossed, or as closed rings. The free ends of the chromosomes frequently appear split. The nucleolus is minutely vesicular. i6o Lloyd Williams . — Studies in the Dictyotaceae . Fig. 1 8, #, <5, r ( Padina .). Three sections of the same nucleus, showing the reduced number of chromosomes, all of which except one are uncut. Fig. 19. The chromosomes reduced in size and ready for being placed on the spindle. A few spindle-fibres have already appeared, and the nucleolus is breaking up into fibrillae. Fig. 20. Intranuclear spindle with mantle-fibres and nucleolar fibrillae. (The globule is in another section.) The chromosomes are heterotype in character. Fig. 21. A slightly oblique section, showing more clearly the form of the chromosomes. The number is sixteen, but those in a lower focus have not been drawn. Fig. 22. One of the poles of the preceding figure, showing the curved rod-like centrosome. The section containing the other pole is very similar. Fig. 23. One of the two daughter-nuclei of the first division. There is a large mass of fibrillar chromatin and a small nucleolus. Fig. 24. A later stage. The chromatin becomes distributed through the nucleus and the nucleolus increases in size. Second Division in the Tetraspore Mother-Cell. Fig. 25. Early prophase. Nucleolus swollen, irregular, and fibrillar. The section is transverse to the nuclear axis. Fig. 26. A number of short curved chromosomes have appeared, and the nucleolus is irregular and fibrillar. Fig. 27. Spindle. The polar radiations are directed towards those of the sister-nucleus, the membrane on the inner side is intact, the spindle being on the opposite side. A nucleolar globule is present. Fig. 28. Anaphase stage. The nuclear cavity is much narrower and curved. The chromosomes though small show no signs of fusion. Fig. 29. One of the four nuclei of the sporangium before spore differentiation. The First Division in the Germinating Tetraspore. Fig. 30. Prophase. Very coarse spirem and fibrillar nucleolus. Fig. 3T. Spindle with curved chromosomes (one has lagged behind). There are two sections of this, and the number of chromosomes in the two together is sixteen. The radiations are distinct ; there is no globule. Fig. 32. Anaphase stage. Membrane still intact at the sides. Fig. 33. 1 Dispirem ’ stage. The nucleolar membrane projects towards the very distinct centrosome. Fig. 34. Polar view of daughter-nuclei before fusion of chromosomes. About sixteen may be counted. Near the bottom of the figure a small body staining differently to the chromosomes is seen. This is probably the beginning of the nucleolus. Fig. 35. A later stage than Fig. 26, where a chromatin mass, a nucleolus, and a reticulum have been differentiated in each daughter-nucleus. VoimiLPi./x. University Press, Oxford. iJfjmcds of Bolouny Vol XVII f Pl.JX. LLOYD WILLIAMS, D I CTYOTAC EAE . University Press, Oxford. Annals of Botany Vol XVIII, PI. X. f. * a University Press, Oxford LLOYD WILLIAMS, DICTYOTACEAE Telangium Scotti, a new Species of Telangium (Calymmatotheca) showing Structure. BY Miss M. Benson, D.Sc., Senior Lecturer in Botany at the Royal Holloway College, Englejield Green . With Plate XI and a Figure in the Text. MONG the numerous plant remains preserved for us as impressions on the Palaeozoic rocks are some digitate clusters attached to branching petioles devoid of lamina, and associated with, and sometimes attached to, leaves of the Sphenopteris type. They were first investigated and named by Dr. Stur 1. The species Calymmatotheca Stangeri , Stur, may be taken as the type of these impres- sions. Dr. Stur regarded the constituent parts of the cluster as indusial valves, but they were differently interpreted by Renault, who figured them in his ‘ Cours Fossile’ 2 as sporangia. M. Zeiller3 also supported the sporangial interpretation of the lobes. Several species of genuine sporangia have subsequently been included in the genus Calymmatotheca , e. g. C. affinis and C. asteroides . They were all founded on casts, however, and it was not until May, 1902, that petrifactions were obtained. Sections of coal nodules from the Gannister beds of Dules- gate and Hough Hill have recently been yielding a good many of these synangia, some of which have been beautifully cut in series by Mr. Lomax of Bolton. This led to a re-investigation 4 of Stur’s type-specimens, which has con- vinced me that he was right in his interpretation of his specimens, and that 1 Die Culm-Flora, 1875-77. 2 Cours d. Botan. Foss., troisieme ann^e, p. 198, 1883. 8 Bassin houiller de Valenciennes. Flore Fossile, 1888, p. 34. Sur quelques Fougeres houilleres d’Asie Mineure. Bull. Soc. Bot. de France, tom. xliv, p. 199. 4 Dr. Scott and Prof. Oliver tell me they have come to the same conclusion after a careful inspection of the specimens. The re- investigation was rendered possible by the kindness of the Director of the Geol. Reichsanstalt at Vienna, who, at the request of Dr. A. Smith Woodward, F.R.S., lent the valuable specimens in question to the Geological Department of the British Museum, so as to give English Palaeobotanists an opportunity of examining them. [Annals of Botany, Vol. XVIII. No. LXIX. January, 1904.] 162 Benson. — Telangium Scotti , a new Species of no sporangia can therefore be included in the genus Calymmatotheca , which he founded upon C. S t anger i, C. Haneri , C. S chimp eri, and C. minor. I have founded therefore the form-genus Telangium for the reception of such specimens as can be diagnosed as follows : — Fertile and barren pinnae dissimilar ; fertile pinnae represented by synangia only ; synangia borne at the extremity of the ultimate ramifications of rachis, composed of 6-12 sporangia which taper to the apex and are united primarily for almost their whole length to form a body which is continued into a sterile base of decreasing diameter through which runs longitudinally a single vascular strand. Each sporange ultimately becomes almost free from the others by septicidal dehiscence and liberates large spores from a ventral suture. As I have not been able to identify these new specimens with any so far described species, I have much pleasure in calling it Telangium Scotti , after Dr. D. H. Scott, F.R.S., whose work on Lyginodendron has added to the interest in this type of fructification. The first specimen that came into my hands was the longitudinal, tangential section represented in PI. XI, Fig. t. The longitudinal dimensions of the sporange on the left are 3-2 mm., but the full length of the sporange was probably somewhat greater. If one compares Fig. 8 one sees that the sporange has really a free narrow apex which brings its length up to 3-8 mm. The sterile base would probably have brought the length of the whole synangium up to at least 5 mm. The width of the synangium averages a little under 3 mm. before dehiscence, if it be measured at the widest part. Shortly after examining this preparation I was enabled by the kindness of Prof. F. W. Oliver to look through slides from a similar source that belonged to the Collection at University College, London. These yielded a beautiful series of four slides (C.N. K3 a-d ), cut from a block containing three synangia, and Figs. 2 and 3 have been drawn from them. Mr. Lomax has recently cut another excellent series, which is now in Dr. Scott’s Col- lection, and has been kindly lent to me with other slides for the purposes of this paper. One slide from the Manchester Collection, which I owe to the kind- ness of Prof. Weiss, has also been of service and is shown in Fig. 9. It will be seen from the drawings that the synangium has eight sporangial chambers arranged in two rows. If we take the transverse sections in the order of their position, beginning at the base, we should first examine Fig. 4 A, which is a transverse section of a synangium immediately below the insertion of the sporangia. The long diameter of the ellipse is 1*7 mm., and the short diameter is ^9 mm. in length. This section shows the vascular bundle in transverse section, the Telangium (Calymmatotheca) showing Structure . 163 outline of which is not very clearly defined, but can be seen to contain tracheides of narrow lumen, v. Fig. 4 c. These are best seen rather to one side of the section, and possibly occupy one arm of a V-shaped strand. The whole section is limited by the large-celled epidermis with blackened contents ; within this are groups of thin-walled cells which have broken down, and at / the section has passed through the lacunar tissue, which is visible again in Fig. 2 and partially in Fig. 1. Fig. 2 represents a section of another synangium at a level just above that of Fig. 4 A. We see that the dimensions have much increased. Two of the sporangia have been injured. The walls are thicker than they are nearer the apex, and show at least seven layers of cells. The epidermis is supported by a hypoderm which is not so regular as is the case at a higher level. The lacunar tissue appears at / and lv I have been unable to trace any vascular strand at this level. Fig. 4 B is a section at a slightly higher level. The septa have given way at two places which probably represent the basal parts of the fissures shown in Fig. 5 B, which is another section of the same synangium. Fig. 4 B shows the hypoderm to consist at this level of an interrupted layer of cells empty of contents. In many cases they show scalariform marking. This can be best seen in Figs. 1 and 2. They are elongated in the long dimension of the sporange, and are probably not continuous with the vascular strand but simply hypodermal cells differentiated for some special function. It is of course possible they were of use at first as water- conducting elements, but it seems more probable that their chief function was to secure dehiscence. They may be regarded as physiologically analogous with the fibrous layer in the wall of the pollen-sac of Angio- sperms. If we refer to Fig. $ hh' we shall see that on the outer wall between each sporange there is a group of thin-walled cells which tear on dehiscence. This would be brought about by the contraction of the convex, free portion of the sporangia! wall. This contraction may very well have been due to the hygroscopic structure of the membrane of these hypodermal cells. At intervals the living parenchyma interrupts these cells. This may be seen well in Fig. 3 and in 5 A, at x and x1, and it is possible that contraction may have been aided by loss of turgidity in these cells. The sections taken at a higher level where the sporangia have separated from one another are represented in Figs. 3, 5 B, 7 B, and 9. It is easily seen that part of the wall of each sporangium is composed of a segment of the peripheral wall and a thinner part derived from the partition which splits longitudinally, reminding one of the septicidal dehiscence of a syn- carpous fruit. There is, however, one exception to this, for the epidermis is complete all round the extreme apices which appear to be free primarily, M 2 164 Benson. — Telangium Scotti, a new Species of v. Fig. 5 B, ep. This should be compared with Fig. 8, ep., where the free apex may be seen still containing spores. The spores seem to have been ripe, and in most cases have been partially shed. In one sporange in Slide C.N. . M . 21 (a) from Sharney Ford, kindly lent me by Prof. Oliver, there are numerous spores. Fig. 6, a , b, c, d , are drawings of such spores made with the help of Zeiss’s Abbe Camera, and show the form and characteristic wall of the spore. The spores vary slightly in size and form. Many are somewhat elliptical, measuring from 5 to 6 ix in the longer dimension, and from 4 to 3*5 \i in the shorter. Many look circular, but this may of course be due to the elliptical forms being looked at end-on. They are reticulately marked, the ridges meeting at an angle of about 120°, and the thinner areas are approximately hexagonal. In form and in the character of the wall these spores agree very closely with the pollen-grains in the pollen-chamber of Lagenostoma ovoides , which are drawn in Fig. 6, e , f g. The size of the latter is, however, slightly greater. Those in the pollen- chamber of a Lagenostoma ovoides in my collection measure 6-75, and 7-2^ in the longer dimension and 5 /x in the shorter. An increase in size of a pollen- grain after entering the pollen-chamber is known to occur. In the interesting parallel case recorded by Renault of the pollen-grains of Cordai- anthus the shagreen-like coats of the spores were similar, but the size of those in the pollen-chamber showed a marked increase. We are now in a position to construct a diagrammatic figure if we superpose these sections upon one another, and the result is given to scale in Fig. 10. Septicidal dehiscence has advanced almost to the base of the synangium. It will be seen how closely this resembles some of the Calymmatotheca impressions, which are reproduced for comparison (Figs. 11 and 12 and Text-Fig. 33). Stur’s drawings of Calymmatotheca Stangeri (Fig. 11) show a form very similar to that of C. Scotti, except that the former is dis- tinctly larger, measuring 6 to 8 mm. in length, whereas C. Scotti , even if we take account of the sterile base, can hardly have reached more than 5 '5 mm. But the chief difference is due to the absence of relief in the impres- sion, which in this respect offers a sharp contrast to the form reproduced photographically in Text-Fig. 33. The many specimens of the latter I have seen lead me to wholly agree with Mr. Kidston’s interpretation of it as sporangial, and I shall, for convenience, therefore refer to this species as Telangium affine. The same may be said of the other British species which will be referred to under the names Telangium asteroides and T. bifidum. Telctngium ( Calymmatotheco ) showing Structure . 165 Count Solms-Laubach, in a recent review1 of my preliminary note to this paper in the Annals of Botany, 1902, stated that the agreement between the new species and C. Stangeri is not perfect owing to the absence of thorn-like emergences on the back of the sporangia in the new species. This was in reference to Stur’s Plate VIII, figs. 5 and 6, which show emergences on the lobes reminding one of the glands on the newly dis- covered outer envelope of Lagenostoma Lomaxi\ and it is of course possible that this may be shown by Drs. Oliver and Scott to be their nature. This valuable criticism is met by my present action in with- drawing the new synangium altogether from Stur’s genus and founding a new one. In 1877 a paper2 by the late Mr. C. W. Peach was read before the Geological Society describing some beautiful casts 3 of a smaller form of Telangium , hitherto known as C. affinis (see Fig. 12). He found them attached to fronds of Sphenopteris affinis , and suggested that they were parasitic upon them. c Each flower-like form ’ (Mr. Peach’s expression for the synangium) ‘ is about | in. over and fully that in height.’ Mr. Peach compared his specimen with that of C. minor as figured by Dr. Stur, p. 237 of his ‘ Culm Flora/ and it is not impossible that C. minor may be a Telangium. By the kind permission of the Council of the Geological Society I have been enabled to reproduce two of Mr. Peach’s figures (Fig. j 2 a and b). The form of T. affine is very like that of T. Scotti , but it is distinctly smaller. The dimensions as kindly given me by Mr. Kidston are as follows : — Length, 2*5 —3*5 mm. Breadth, 275 — 3 „ (after dehiscence). Those of T. Scotti are as follows : — Length, 4‘5— 5*5 mm. Breadth, 3 „ (before dehiscence). The figure in the text is reproduced from a photograph kindly made expressly for this paper by Mr. Kidston of a specimen of T. affine in his possession. It is noticeable that the synangia of T. affine are represented sometimes in approximation and in planes parallel to one another (Fig. t 2 a). This seems to have been the case also in T. Scotti , as is shown in Figs. 3, 4, and 5. These specimens of T. affine were found by Mr. Peach in the Calciferous Sandstone rocks of North Britain, and the two species thus belong to different horizons. Another British species the description of which we owe to Mr. Kidston 4 1 Bot. Zeitung, 60, Dec. 1902. 2 Quarterly Journal of the Geol. Soe. of London, vol. xxxiv, p. 131. 3 Admirable specimens of these are preserved in the British Museum. 4 Trans, of the R. S. Edin., vol. xxxiii, p. 140. 1 66 Benson . — Telangium Scotti , « Species of is T. bifidum . The synangia are still more markedly aggregated than those of T. affine . A careful examination of Mr. Kidston’s Figs, i -6, Plate VIII, would lead me to conjecture that the synangium is composed of not more than ten or twelve sporangia, and that the appearance of a greater number is due to the shortness of the ultimate ramifications and hence the almost capi- tate condition of the fructification. The dimensions of the synangium of T \ bifidum are given by Mr. Kidston as follow : — Length, 6-5 — 67 mm. Breadth, 3-75— 4 Mr. Kidston’s Fig. 6a, Plate VIII, depicts one synangium in which bipartition is very noticeable, and may be compared with Fig. 7, which represents a synangium of T. Scotti showing the same tendency. We see from this review that, as respects size, T. Scotti is intermediate between T. affine and T. bifidum , and that it shows many features in common with both species. The only species of Telangium recorded so far from the Upper Carboniferous1 is T. asteroides. This is generally repre- sented as having had but six sporangia in its synangium, but I have not been able to confirm this from the specimens preserved in the British Museum. The longitudinal dimension of the synangium is a little over 3 mm. The synangia are borne on branching petioles like those of other species. Owing perhaps to its imperfect preservation it does not seem to be so near T. Scotti, the new Upper Carboniferous form, as do several of the species already referred to, which belong to the Lower Carboniferous. As the attribution of Telangium Scotti to Lyginodendron was at the time2 partly based upon what Dr. Scott and I now consider to be a misinterpretation of Stur’s type specimens of Calymmatotheca Stangeri, it remains for me to discuss what evidence is still available in support of the view adopted in the preliminary note 2 to the present paper. Not only is internal evidence available owing to the preservation of the tissue of Telangium Scotti, but the recent announcement 3 on the part of Messrs. Oliver and Scott that the seed Lagenostoma Lomaxi grows attached to an envelope showing characteristic structural features of Lyginodendron Oldhamium has given unexpected opportunity for further comparison. The evidence may now be summarized under the following headings : — 1. Association and character of impressions or casts. 2. Association of petrifactions. 3. Character of tissue. 4. Correspondence between the spores of Telangium Scotti and the 1 Potonie’s statement in Engler’s Pflanzenfamilien, Teil I, 4. Abteilung, p. 449, that Calym- matotheca belongs to the 1 Ober-Carbon’ seems due to an error, as he does not refer to C. asteroides. 2 Benson, The Fructification of Lyginodendron (note), Annals of Botany, xvi, 1902. 3 Proc. R. S., vol. lxxi. Telangium ( Calymmatotkeca )' showing Structure . 167 pollen- grains germinating in the pollen-chamber of Lagenostoma Lomaxi and L . ovoides . 5. Correspondence in certain morphological characters between the synangium of Telangium Scotti and the seed Lagenostoma. We will deal with these subjects in succession, and the last will be found to involve a wholly new theory of the phylogeny of the inner integument. Firstly, association and character of impressions. Those who have had the pleasure of studying the numerous and beautiful plates the late Dr. Stur included in his ‘ Culm Flora’ and c Carbon Flora’ cannot but be impressed with the family likeness which seems to reign among the fronds, whether they are called Calymmatotkeca , Diplothmema , or Sphenopteris. Zeiller has expressed the view that they all belong to stems of the Lyginodendron type. The branching of the leaves may be dichotomous, or pinnate, or various combinations of both systems. These leaves are in one case found associated with one species of Telangium fructification. Thus T. minor is found associated with Sphenopteris (Diplothmemci) patentissima , and also with indusiate seeds which Stur calls Rhabdocarpus conchaeformis . Turning to records of British impressions of Telangium we have three — T. affine and T. bifidum from the Lower Carboniferous, and T . aster oides from the Upper, i. e. the Lower Coal Measures. T. affine is not only associated with but attached to leaves of Sphenopteris affinis , so much so, indeed, that Mr. Peach in the description of his beautiful specimens suggests that they were parasitic upon the leaf. The frond in this case, which is familiar to many as represented in the frontispiece of Hugh Miller’s e Testimony of the Rocks,’ dichotomizes freely, and thus exhibits a type of branching also found in Sphenopteris elegans , the leaf of Heterangium. T. bifidum is also found growing on leaves very similar in character to those of Sphenopteris affinis , as bifurcation is frequent. There is no reason to expect in such an advanced type as Lyginodendron an exact correspondence in size and form between the microsporophyll and the sterile frond, and with this interpretation of such fronds in view it is interesting to note in this latter British species the appearance of the synangia only on the more basal part of the leaf. Secondly, the association of the petrifaction T. Scotti with Lygino- dendron in the coal-nodules of the Gannister beds of Lancashire. Not much weight can be attached to the fact of the association with fragments of Lyginodendron owing to the great abundance of the latter in these nodules. But the value of the association is augmented by the fact that T. Scotti appears in sections of a nodule from Sharney Ford which is otherwise almost purely composed of the vegetative organs of Lyginodendron. It may also be stated that in several of the slides containing sections of T. Scotti there are also Lagenostoma seeds. Their close approximation e68 Benson. — ■ Telangium Scotti, a new Species of Is shown in one case in Fig. 9, which is from a slide kindly lent by Professor Weiss from the Collection of the Manchester Museum, Owens College. Thirdly, the character of the tissue. The tissue of the lower part of the synangium has much in common with the familiar sterile pinnae of Lyginodendron. We have a well-developed epidermis, a definite hypoderm, and lacunar tissue which is indistinguishable from the corresponding tissue of the sterile pinna. The vascular strand of the pedicel is composed of trachei'des of very narrow lumen, and thus resembles those of the petiole of Lyginodendron . The preservation of the tissue is unfortunately not good enough to show the form of the strand clearly. The group of trachei'des which is preserved (Fig. 4 c) may be the whole, but it is possible that another corresponding group may have occupied the other arm of a V- shaped strand, but has become opaque owing to the minuteness of the lumen of the tracheides. Fourthly, correspondence between the spores of Telangium Scotti and the pollen-grains germinating in the pollen-chamber of Lageno- stoma Lomaxi and ovoides. Ripe spores occur in five of the synangia already to hand, and have been measured by Prof. Oliver and myself. As already pointed out they agree with considerable exactness in form and in the character of the wall with the pollen-grains in the pollen-chamber of Lagenostoma, but the latter slightly exceed them in size. This comparison, already found of value in the magnificent work of Renault on Cordaianthus , is of great interest. The spores of Telangium average 5*5 [x in their longer and 37 /x in their shorter dimension. The spores when they are germinating, apparently in the very act of yielding antherozoids like those of Cycads and Ginkgo , measure 7 x 5 /x. The wall of both is thick and has thinner areolae, and thus may be described as reticulate, v. Fig. 6. Fifthly, correspondence of Telangium Scotti in certain morpho- logical characters with the seed Lagenostoma. The seed Lagenostoma (the three species of which were first named and partially described by Williamson) has since received a searching investigation at the hands of Prof. F. W. Oliver1. The connexion of one species, L. Lomaxi , with Lyginodendron has recently been announced2 by him and Dr. Scott. To quote from their account of this species: fIn the most general relations of its organization the seed approaches the Gymnosperm type in that the integument and nucellus are distinct from one another in the apical region only, whilst the body of the seed which contains the large single macrospore shows complete fusion of the integument and nucellar tissues. But in other respects the seed is remarkable. The integument, which is a simple shell where fused with the nucellus, becomes massive and com- 1 See ‘ Oliver, The Ovules of the older Gymnosperms/ Annals of Botany, xvii, 1903, PI. XXIV. Fig. 9* 2 Proe. R. S., vol. lxxi. Telangium [Calymmato theca) showing Structure . 169 plicated in its free part which corresponds to the upper fifth of the seed. In this region it is usually composed of nine chambers radially disposed around the micropyle. The whole structure from within is like a fluted dome or canopy, the convexities of which correspond to the chambers. The vascular system of the seed enters as a single supply bundle at the chalazal papilla and branches a little below the base of the macrospore into nine radially-running bundles. Each of these bundles passes without further branching to the apex of the seed, running outside the macrospore and a little distance below the surface. At the canopy the bundles enter the chambers and end at the tips.’ A somewhat lengthy quotation has been made, as it is necessary to understand the structure of the seed if the comparison with the micro- sporangial sorus is to be appreciated. The transverse section1 of the seed, if taken in the plane of the canopy, somewhat resembles a cartwheel, in which the nucellar apex forms the axle, the radial walls between the chambers the spokes, and the peripheral walls of the chambers the rim of the wheel. The comparison does not hold good, however, in well-preserved sections, as the chambers are seen each to contain large, thin-walled cells which support the delicate branch of the vascular bundle that is contributed to each. The correspondence which must have already suggested itself to the reader is between such a seed as Lagenostoma and such a synangium as Telangium Scotti. The chambers surrounding the nucellus seem to represent its sister sporangia, which have become sterile, the large- celled, thin-walled tissue and delicate vascular strand being all that repre- sents the ancestral sporogenous tissue ; while the micropyle corresponds with the original space between the tips of the sporangia. The seed in fact is assumed to be a synangium in which all but one of the sporangia are sterile, and form an integument to the one fertile sporange which has become a megasporange with one large megaspore. In Lagenostoma physoides 2 the integumental ridges are continued into tapering tentacles around the micropyle, and this still further accentuates the resemblance to a sorus. In L. ovoides the number of chambers is often seven instead of nine. Hence we have only to imagine that one of the sporangia of a sorus of eight or ten sporangia gradually evolved megaspory, and that the remaining seven or nine sporangia became a sterile envelope, — a correlation in develop- ment which has many analogies in the animal and vegetable kingdoms. As soon as one of the sporangia became a megasporange the symmetrical arrangement of the sister sporangia would become an advantage and naturally follow. At the remote period of time at which the seed was 1 Oliver, loc. cit., p. 461. a See Williamson, Phil. Trans., vol. clxvii, 1877, PI. XI, Fig. 77. A full account by Prof. Oliver of this seed will appear shortly. It should be compared with Telangium bifidu?n. 170 Benson . — Telangium Scotti , a Species of evolved, a period probably anterior to the Carboniferous epoch, it may be conjectured that the arrangement of the sporangia in the sorus was irregular, and that the more centrally placed sporange with its better vascular supply may have gradually attained the megasporangial condition. In Gleichenia and Oligocarpia some sori have, and others have not, a central sporange. As respects the vascular supply in the centre of each compartment of the integument, it is well known that in many of the Permo-Carboniferous seeds a vascular bundle entered the base of the nucellus, even passing from the chalaza to the pollen-chamber \ and it is hence easy to conceive of a vascular strand having early entered its sister sporangia. Again, if we take an example from a seed of very remote affinity, we find that in Castanea a vascular strand may be demonstrated running up the whole length of the nucellus, and is especially well developed in nucelli whose embryo-sacs have long remained unfertilized. I will now proceed to show that this interpretation of the integument of Lagenostoma is helpful in clearing away many of the difficulties that have beset the general problem of the integument hitherto. The more generally accepted interpretation of the inner integument is that it is due to a special development of the indusium. We are compelled to regard the integument of Lagenostoma as a single integument, firstly because of the primitive character of the seed, and secondly because of the existence in Z. Lomaxi of an exterior envelope. Hence it is probably safe to regard it as homologous with the inner integument, and conse- quently as hitherto accounted for merely by Celakovsky’s theory of the indusium, or by another theory to which I will allude later. But the cohesion of integument and nucellus which we know to be characteristic of the Cycadean seed receives no explanation on the indusial theory, whereas on the synangial theory the cohesion is seen to be due to the origin of the seed from structures already coherent. Moreover, as it is generally agreed that the heterosporous habit arose from the homosporous, it is a priori probable that there should be a cor- respondence between the microsporangial sorus and the primitive seed, and this correspondence seems best obtained by harmonizing the seed and the synangium. If it should be shown conclusively that T. Scotti is the micro- sporangial organ of Lyginodendron the homologizing of Lagenostoma with its £ synangium would simplify the problem of the integument in that we should then have but one envelope to account for in the seed over and above what was present in the male sorus. I will now refer shortly to another widely accepted view, which has been adopted by Strasburger, Treub, and Dr. Lang. Though their views Oliver, loc. cit., p. 454. Telangium ( Calymmatotheca ) showing Structure. 171 vary as to the homologies of the seed as a whole they agree in regarding the integument as a new formation. Dr. Lang’s conclusions are based on his own investigation into the morphology of the sporangia of Stangeria 1J and on the results of work by Warming and Treub on other genera of Cycadaceae. He points out that * with regard to the development considerable correspondence between the ovule and the sorus can be traced in the early stages. The differences between the development of the sorus of microsporangia and the ovule only become pronounced when active growth becomes localized around each archesporial group V He therefore homologizes the sorus and the ovule at the outset, but looks upon the ovular sorus as monosporangiate and the integument c as an annular upgrowth, around the apex of the nucellus, of the bulky sporangial wall or, which comes to the same thing, of the edge of the receptacle which had kept pace with the single spor- angium.’ Thus it would appear that owing to the relatively advanced type of seed investigated, Dr. Lang could not homologize the upgrowing ‘edge’ of the receptacle with sterilized sister-sporangia of the nucellus. He adds that his view is only put forward as a provisional statement, which will have to be tested ‘ in the light of the evidence obtainable from extinct forms.’ It is in the light of these extinct forms that the new theory of the integument is now being put forward. Whether T. Scotti be ultimately proved to belong to Lyginodendron or not, we may well bear in mind that the synangium is a very ancient type of fern fructification, for from the Culm onwards we have numerous examples of it recorded. Where the individual sporangia are not en- tirely coherent they generally form a sorus of bulky sporangia like those of the Filicinean class e Simplices ’ suggested by Professor Bower. The ancient sporange was very rarely solitary, and we have already undoubted evidence in Cycadeoidea of a seed-plant having synangia for its micro- sporangial organ. Among synangia which are found associated with Cycadofilicinean seeds are Hawlea and Scolecopteris 3. The latter I will shortly describe, as I believe a reference to it may make the comparison of seed and synangium more clear. Scolecopteris is a form-genus including several species of sorus, which have been described by a succession of palaeobotanists 4. It is sufficient 1 Lang, Annals of Botany, xiv, 1900. 3 Note the support that these observations give to the soral theory of the seed. 8 Kidston, ‘ On the Fossil Flora of the Radstock Series of the Somerset and Bristol Coalfield.’ Trans. R. S. Edin., 1888. Also, 1 On the Fructification of Carboniferous Ferns,’ Trans. Geol. Soc, Glasgow, vol. ix, 1889, Plates II and III. Further announcements bearing on this subject will shortly be made by Mr. Kidston. 4 Strasburger, ‘ Scolecopteris elegans, Zenk.,’ Jenaer Zeitschrift fur Naturw., vol. viii, 1874. 172 Benson. — Telangium Scotti , a new Species of for our purpose to refer to the drawings of Scolecopteris polymorpha in Engler and Prantl, Teil I, Abt. 4, p. 440. It will be seen that the sorus as a whole somewhat resembles T. Scotti , but the four or five sporangia, which here constitute the sorus, are inserted around a pedicel along which runs a vascular strand. If this were to become continuous with a strand of trachei'des developed in the sporogenous tissue, we should obtain the vascular supply which characterizes Lagenostoma. The beautiful plates in Brongniart’s ‘ Recherches sur les graines fossiles silicifiees 5 afford many opportunities of applying and testing the new theory, and amongst others I would suggest a reference to the following : — Plate IX, Fig. 4, showing a vascular bundle entering the nucellus in Rhabdocarpus subtunicatus. Plate XIII, Figs. 6 , 7, 17, 19, showing the contrasted tissue-systems of the integument of Sarcotaxus avellana and its septicidal dehiscence. Plate IV, Figs. 1 and 3, showing the sporangial appearance of the inner integument continued to the base of the nucellus in Cyclocarpus nummularis. (These figures should be compared with Telangium Scotti , Fig. 8.) Plate C, Fig. 9, in which the seed Codonospermum is shown to present a striking external resemblance to such a synangium as Asterotheca . The similarity of the inner integument of Pachytesta to that of Lagenostoma has been recently pointed out by Professor Oliver1, and a transverse section has been constructed which exhibits its compartmental nature at a level much lower than that in which it can be demonstrated in Lagenostoma. Professor Oliver adds : { The presence of vascular strands in the chambers of Lagenostoma is the most important difference.’ Much fuller details are to hand of another seed which seems to bear out this theory. I refer to Bennettites Morierei , Sap. and Mar. (spec.), which has been admirably worked out by Professor Lignier 2. This fructification, as is well known, belongs to a much later horizon, i. e. Mesozoic, and shows Cycadean affinities. If one consults Lignier’s Plate III, Figs. 35 and 37, one sees transverse sections of the upper part of the seed, showing the thick integument divided up into four compartments by radiating vertical walls of flattened cells, very comparable to those which separate the constituent members of a synangium. The interior of each compartment is described as succulent tissue, but offers an abrupt contrast to the walls. Plate III, Fig. 38, shows 1 Oliver, ‘ On some Points of apparent Resemblance in certain Fossil and recent Gymnospermous Seeds.’ New Phytologist, vol. i, p. 150, Text-figure 5. 2 O. Lignier, ‘Structure et Affinit^s du Bennettites Morierei , Sap. and Mar. (sp.).’ VegCtaux fossiles de Normandie. Caen, 1894. Telangium (Calymmato theca) showing Structure . 1 73 the constitution of the integument at a lower level. Here we find the peripheral epidermis of the integument lined as in Telangium and other synangia, with a layer of reticulately thickened cells within which lie the large thin-walled cells which seem to correspond with the sporogenous tissue, and this is limited internally by thick-walled fibres. Plate IV, Fig. 45, shows also on a smaller scale the compartmental structure of the integument. It is interesting to note that Bemiettites Morierei is in some respects evidently less reduced than Bennettites Gibsonianus , in which, as Dr. Scott says in his { Studies,’ the structure of the pericarp is a matter of inference. Nor is there any possibility of avoiding the conclusion that the inner tube of the micropyle is nucellar in origin if we accept the diagrams Prof. Lignier gives. I cannot but regard this example as very strongly confirmatory of the homology of the seed with the synangium. If we compare the peripheral epidermis of the integument with that of the microsporangial sorus of Cycadeoidea we obtain a possible explanation of the radially elongated epidermal cells 1 of the sunken seed. Is it possible to call in here the aid of a wholly hypothetical indusium and invest it in turn with so many points of similarity to the sister sporangia of the nucellus, sporangia which it cannot but be granted originally surrounded the ancestor of the mega- sporange ? Or, on the other hand, can we, with others, call in a * new formation ’ to account for an integument so obviously compartmental ? Thirdly, I wish to refer to the seeds which somewhat unfortunately go by the name of Gnetopsis elliptica 2, Ren. Although they are not yet worked out with the same detail as Lagenostoma and Bennettites Morierei there is considerable internal evidence in support of their synangial origin. They are figured (after Saporta and Marion) in the English edition of Solms-Laubach’s Fossil Botany on page 128, and come from the Upper Coal Measures of Grand’ Croix. The ovules occur in one or more pairs in the hollow of a cup-like envelope which bears long hairs. For convenience I will quote from Solms-Laubach’s description of this most interesting type : ‘ That portion of the integument which encloses the apex of the nucellus behaves in a very peculiar manner, and may be com- pared perhaps with Lagenostoma , Will. It attains a considerable thickness and separates (sic) into a compact outer lamina and a similar inner lamina, while the cell-layer between the two is formed of extended filaments which represent so many cells and traverse a broad intercellular space at some distance from each other. This looser tissue ceases of course at the micro- pylar canal, where the outer and inner layer are in connexion with one 1 Cf. Figs. 99, ioi, and 102 in Coulter and Chamberlain’s * Gymnosperms.’ 3 Renault, ‘Cours de Bot Fossile,’ T. 4, p. 179, Plates 20-22. 1 74 Benson. — Telangium Scotti, a new Species of another. The margin also of the orifice of the micropyle is formed of a cup-shaped expansion which is seen to be drawn out at two points into long filiform appendages. A vascular bundle enters at the base of the ovule and splits into four branches.’ If this account were translated into the language of this new theory we should say that each of the four abortive integumental sporangia contains loose elongated cells in its upper part, and that their extreme apices are prolonged much as in Lag. fihysoides , only that they remain adherent in pairs. The other two species of Gnetopsis , G. trigona and G. hexago7ia , are known only as impressions, and show four or five tentacles around the apex 1. If it should be contended that in the case of Lagenostoma and Gnetopsis this special development of the inner integument is merely of biological significance, I would point out that it is difficult to see then why this should also occur in a seed outtopped by interseminal bracts as e.g. Bennettites Morierei. Nor does this explain the form of the section of the seed — triangular, hexagonal, &c. — nor the radiating vertical walls dividing the integument into compartments. If, however, such internal evidence as I have brought forward appears inconclusive, it is satisfactory to find that there is a record in the literature of an exactly comparable transformation occurring in the sorus of a very ancient monostelic fern stock. I refer to the fact that Renault in his Autun Flora describes a specimen of Botryopteris sporangia in which a group was found to be surrounded by an envelope formed of sterile and highly modified sporangia 2. Renault figures some of these sterile sporangia in his c Flore fossile d’Autun et d’Epinac.’ When we consider that on anatomical grounds it has long seemed probable that the Cycadofilices arose from some ancestral Filicinean group such as the Botryopterideae , we see that such a case as Renault cites is peculiarly significant in any discussion as to the phylogenetic origin of the integument of. the seed. Hence any further confirmation of Renault’s observation would lend a strong support to the new theory. I will now refer to a few analogous cases which lend a general support to the claim for the sterilization of certain sporangia in a sorus during the evolution of the Seed. In Azolla I believe most morphologists would admit that the micro- sporangial and megasporangial sori were originally similar, and that the megasporangial has gradually lost by abortion a number of sporangia, retaining only one. If the development of the megasporange in Azolla involved the total loss of its free sister sporangia, are we claiming too much 1 Zeiller, Elements de Paleobotanique, p. 224. 2 Renault, Bassin houiller d’Autun et d’fepinac. Flore Fossile, ii, p. 54. Telangium ( Ccilymmatothcca ') showing Structure. 175 if we conjecture that in another Fern the sister sporangia, which were already adherent, were retained as a sterile envelope ? Turning to the Angiosperms, the modification and abortion of flowers in an inflorescence to construct the biologically interesting ‘ flag apparatus ’ is exceedingly common. The peripheral flowers in the capitulum of the Cynareae , in the thyrsus of Viburnum Optilus and Hydrangea , are among the most familiar examples. In Muscari comosum (var. racemosissimum ) a very remarkable modification follows the sterilization of the central flowers. In Rhus cotinus De Candolle noted an increased growth of trichomes on the pedicels of the sterile flowers, and it has hence become a classical. example of what he meant by the expression £ correlation of growth.’ Passing from flower to sporophyll we have no need to mention any of the innumerable instances of the change from stamen to sheathing organ which occurs commonly in Ranunculaceae, Scitamineae, &c. In Salvia we find that half the anther is sterilized to provide the lever which is to assist in the process of cross-fertilization. If a part of a sporophyll can be steri- lized and adapted for an accessory function, why should not some members of a synangium ? Summary of evidence in support of the view that a seed is a synangium in which the peripheral sporangia are sterilized and specialized as an inner integument : — 1. Ontogeny. It is shown that wholly independent testimony is borne to the fact that in the most primitive of existing Spermophyta, the Cycadaceae, a correspondence obtains both in position and development between the microsporangial sorus and the seed. 2. Phylogeny. General considerations would lead us to expect comparable characters in the microsporangial sorus and the primitive seed. A synangium is the only form of microsporangial sorus so far known among the Cycadofilices, and it is found also in Cycadeoidea. A special case is cited of sterilized sporangia in the tufted sori of Botryopteris. 3- Suggestions of sporangial origin in the inner integument of primitive seeds 1. It is frequently compartmental. 2. Each compartment contains large thin-walled cells as contrasted with the firmer peripheral layers. 3. The peripheral wall is constructed of the same characteristic layers as are met with in many synangia. 4. The form of the base and apex of each compartment is often very similar to those of members of a synangium. 5. In some cases there is considerable freedom between the constituent compartments whose apices form the so-called tentacles around the micro- pyle. 176 Benson . — Telangium Scotti , a new Species of 6. The compartments are comparable in size with the nucellus. 7. The compartments vary in number in the same way as the members of many Palaeozoic synangia. 8. The integument of many of the seeds undergoes septicidal dehis- cence like a synangium. 9. The integument is generally as concrescent with the nucellus as the members of a synangium are with one another. In conclusion, I may add that though I regard the theory as of wide- reaching importance, I am only concerned at present to put it forward as probably the true interpretation of the canopy of Lagenostomay and hence as adding support to the view that Telangium Scotti is the micro- sporangial sorus of Lyginodendron . The full exposition of the seed which we await from Prof. Oliver will be based on wider researches into primitive types and a more intimate acquaintance with the difficulties of the problem. I cannot conclude without expressing my indebtedness to Dr. Scott and Prof. F. W. Oliver. The stimulus of their example and criticism and their kindness in lending valuable slides have been most helpful. Telangium ( Calymmutotheca ) showing Structure. 177 EXPLANATION OF THE FIGURES IN PLATE XI. Illustrating Miss Benson’s paper on Telangium. Telangium Scotti, figs, i-io inclusive. h, h' = groups of cells (which tear on dehiscence), as seen in transverse section between the sporangia. /, /' — lacunar tissue. s , s' == fusiform empty cells with scalariform marking, which form the hypodermor fibrous layer. x, x' — parenchyma cells interrupting here and there the otherwise uniform fibrous layer. cp = free apex. Fig. 1. Longitudinal, somewhat tangential section of the synangium. x 33. A few spores can be seen in the sporange on the left. Slide M.B. Coll. 71. Fig. 2. Transverse section of the synangium but little above the insertion of the sporangia. Two of the latter are injured. The fibrous layer does not extend over the whole surface at this level. Slide lent by Professor Oliver. U.C.L. Coll. K 3 (b). Fig. 3. Transverse sections of two synangia at a higher level, x 50. The fibrous layer is now complete. The group of cells ( h ) remain undehisced. This preparation also shows the ex- tension of the parenchyma to the surface at x and x'. Slide lent by Professor Oliver. U.C.L. Coll. k3(4 Fig. 4. A and B are transverse sections of two nearly approximating synangia which are taken at different levels, x 33. Fig. 4. A is taken below the level of insertion of the sporangia and shows a vascular strand v, composed of very small elements. These are reproduced in a larger scale in Fig. 4 C. x 50. Slide lent by Dr. Scott, C.N. 1803. Fig. 5. A and B are transverse sections of the same two synangia represented in Fig. 4 A and B} but at a higher level. Some of the spores are well preserved. Slide lent by Dr. Scott, C.N. 1804. Fig. 6. Spores, x 260. Fig. 6, a , h, c, d, are taken from the sections of Scotti , and fig. 6, fy. •/■'{? jv, •. 1 : A A A A A '■= JiJ :0T,7 jpA i’Of.WH cult; 7. wtf. f'AA/) iUi iCv,;] : fl V. .. vj : A .7 ,:i A;. ' r ,y,0, / . ; . . . '.A , : a a : •••" : - A'- ...1 A AA - pv/T 4 q i'li i £ pv c . . • :i.t> J- ->i '■;/ 0 i ■ j v//o fAtAito ioi ut rri'/vjil i’ raa-. ■ M .... ... ■ ! . Oi.vj dofi i ; r cm J[« A „J,0.U ' .loviJO' 7. ii ik. 0 ;/l ip six-riota p i 7. " t/i (A) •, if cp 'ip or A ocxl v £ A .. u ,7' l 7: • ’ i:’:< ’ r lyt < ■ i.y.'f ,:f jo :... ,7 • Ho Jpvd fi.'Jt V,.' r 1 ! 1 Hi k 4 7' i .3 • :• A-: ! ;i 1.7;;:.' 7 r-i bvz’iiji <■. ; A. .. [ . liXi.:; 7 .dv lb j •:* • 3' - • A 1 t. il'j 1 : 7- - - - ' : ; ; ■ • :-r r. ■- . [ pf.iicf y, A yc v :< 7 A : . Ay;! ic -A. A': txb<\.r> .il.l.l .c.c .< . pr.o la * • '• ■■■ • ■ Sv . . ,v. . . : u . • hi. O ctu.vO .£ar?p5*H.M isjxAtyyM V-A^l * I* >' .>v.ol.po;j cuisvsiuj'J kyskii;- pj'I '/ 1 i::.. ;k :vj> p:t rs**^;_r.iCI .01 ! '•£ ; imipl ' ...... d'y.'d -f ■ ■ j .. . ; cpPop-jMump- = «t&pfx*j’vtojrr '• ■ ' ■ . .7 .- j., x ■ ) . ■ — NOTES ON THE GENUS CORYNOCARPUS, FORST. SUPPLEMENTARY NOTE. — I much regret that both Dr. Fritsch and I overlooked a paper1 by Professor P. Van Tieghem, in which he discusses the anatomy, floral structure and systematic position of Corynocarpus. The Author courteously sent me a copy of the paper in question, and I hasten to make some amends for the omission of all reference to it in my account of the genus 2 by giving here the principal results of his investi- gations. His description of the anatomy of the stem and leaf agrees in the main with that of Dr. Fritsch ; but the latter omitted to mention the isolated crystals, which one cannot possibly overlook in a longitudinal section. His part, however, was hurriedly done just before his departure for Ceylon. I wished merely to show that there are no resin- ducts. Van Tieghem agrees with Engler in treating Corynocarpus as the type of a natural order or rather ‘ famille autonome/ as he designates it ; but his views on the immediate affinities of this genus are wholly different. He lays great stress on the structure of the ovule in classification, and ranges Corynocarpus ‘dans Tordre des Pernucelldes bitegmin^es . . . qui a pour famille type les Gdraniac^es/ I do not underestimate the value of anatomical characters, but I do not attach so much importance to them for purposes of classification as the learned author, mainly because it leads to refine- ments and generic subdivision impracticable in applied or daily botany. I would none the less recommend a perusal of his interesting paper 3. Van Tieghem describes the ovule of Corynocarpus as having a large, persistent nucellus, surrounded by two integuments. The external integument is very thick, consisting of from twenty to thirty layers of cells; whilst the internal is thin, having only three layers of cells, of which the outer has larger cells elongated radially. At the micropyle the internal integument extends up the exostome but not above it, and the nucellus also projects into the endostome ‘so that the pollen-tube comes directly upon the nucellus, without having to traverse the micropyle, as directly, indeed, as though the two integuments did not exist/ I must confess that I do not quite understand this description, because the pollen-tube must traverse the micropyle, whether open or narrow, unless the nucellus actually projects to the very top of the combined endostome and exostome. All the writers whom I have consulted, except Van Tieghem, describe the leaves of Corynocarpus as exstipulate. He says they are furnished with broad, caducous stipules. In flowering or fruiting herbarium specimens, no stipules are 1 Sur le genre Corynocarpe, consider^ comme type d’une famille distincte, les Corynocarpacdes. Tournal de Botanique , xiv (Jnillet, 1900), pp. 193-7. 2 Annals of Botany , xvii (1903), pp. 743-60, pi. 36. 3 See also his paper : L’Oeuf des Plantes considere comme base de leur classification, Ann. Sc. Nat., 8 me. sdrie, xiv (1901), pp. 213-390. [Annals of Botany, Vol. XVIII. No. LXIX. January, 1904.] i8o Notes . to be found, but there are lines running from the base of the petioles, partly round the branches, which might be taken for the scars of fallen stipules or the decurrent margins of the petiole. Fortunately I have been able to examine young shoots from a plant growing at Kew, and I suspect that the organs termed stipules by Van Tieghem are really bud-scales, but I have not been able to examine them so thoroughly as to give a positive opinion, though possibly he has. They are deltoid, acute bodies 2-3 lines long, situated more or less within the axils of the leaves, but falling away as the leaves above them develop. Occasionally they lodge in the axil of a leaf long after organic connexion has ceased to exist, sometimes even until the leaf above is fully developed. I may add here that Mr. T. F. Cheeseman, Curator of the Museum at Auckland, New Zealand, has obligingly sent to Kew several flowering specimens and a quantity of flowers in alcohol of Corynocarpus laevigata , in order to give me further oppor- tunities of examining the staminodes. It may not be remembered, perhaps, that there was a doubtful point in this connexion. I explained in my former paper, p. 744 (and Fig. 7, PI. 46), that Banks and Solander described and figured the staminodes as three-toothed at the apex, whereas all the staminodes of C . laevigata examined by myself and others, including Cheeseman, were irregularly and minutely toothed or fringed from about the middle upwards and around the top. I have examined a number of the flowers sent by Mr. Cheeseman and they all presented the same kind of staminode, except that the toothing or fringe in some instances extended almost to the base. But in none of these did I find a second carpel, or traces of a second, though there was a slight obliquity in the one present. With regard to the Maori name karaka , I have the authority of Mr. Cheeseman1 that it is applied to Elaeocarpus rarotongensis , Hemsl., by the natives of Rarotonga. What connexion there may be, I cannot even suggest, but the foliage is somewhat similar to that of Corynocarpus and the fruit is a drupe. As mentioned before, some writers have endeavoured to prove that the Maoris migrated from ‘ Hawaiki ’ by way of Rarotonga to New Zealand. W. BOTTING HEMSLEY, Kew. THE VASCULAR SUPPLY OF STIGMARIAN ROOTLETS.— In Vol. XVI of these Annals I described the course and termination of certain vascular branches of Stigmarian rootlets, which had first been observed but wrongly inter- preted by M. Renault. Instead of supplying lateral roots as Renault had supposed, the vascular branches terminate in the outer cortex in wide spirally thickened cells, resembling in appearance the transfusion-cells of leaves. These cells I figured in transverse and longitudinal sections of rootlets (PI. XXVI, Figs. 4 and 5), and also in surface view (Fig. 2 c). But this latter figure was not very clear, as the section was somewhat oblique and slightly compressed at c, and, as I stated in my paper, t it is difficult to ascertain what was the size and distribution of these patches 1 Trans . Linn . Soc.} 2nd series, Bot. vi, p. 275, t. 31. [Annals of Botany, Vol. XVIIL No. LXIX. January, 1904.] Notes . 181 of spirally marked cells of the outer cortex.’ The patches were apparently of some breadth, but only a tangential section near the surface of the rootlet could show clearly the extent and arrangement of these cells. Such a section I have found on a slide (prepared by the late Mr. J. Spencer) kindly lent me by Dr. Scott from his collection. Its cabinet number is 1527. This slide has two rootlets cut tangentially through the outer cortex, and from the better one of the two the accompanying drawing has been made with the camera lucida. The rootlet could hardly have been cut in a better direction or at a better depth for revealing the details of the vascular Fig. 34. elements of the cortex. As will be seen from the Fig. 34, they form a complete net- work, over a width of six or seven cells, and greatly resemble the termination of the vascular bundles in the leaf. Between the spirally thickened cells are found wide thin- walled elements from which water could readily pass into the spiral elements and thence through the vascular branch into the stele of the rootlet. At one or two points where the spiral elements are shown in the drawing in an incomplete condition, it is obvious from the difference in focussing that the spiral cells were there connected with a vascular branch lying at a different level and not in the plane of the section. In the other rootlet seen on the slide there seemed to be a slight difference between the cells at the ends of the ramifications and the connecting tracheids. The latter were slightly narrower and had a closer spiral marking than the former which were of greater width. I must express my indebtedness to Dr. Scott, who has placed at my disposal this excellent preparation which throws further light on the curious vascular supply of the Stigmarian rootlets. F. E. WEISS. Owens College, Manchester. ROOT-PRESSURE IN TREES. — The following observations made upon the Wych Elm ( Ulmus Montana) appear to be of some interest. The tree used was over thirty feet high, and branched at the base into two main trunks. Feb. 20. The larger trunk was sawn across. No bleeding now or sub- sequently. Mar. 15. The second trunk was ringed, from eight to ten annual rings of wood being removed, which formed one half of the alburnum. Flowering and foliation in April were hardly at all delayed. [Annals of Botany, Vol. XVIII. No. LXIX. January, 1904.] i82 Notes . Ap. 19. A root 2-5 cm. diameter was cut through and mercury manometers attached to both ends. A rapid escape of sap took place from the end attached to the stem, the pressure varying from ten to fifteen feet of water until the fourth day when it began to fall. Practically no escape of sap took place from the end of the severed portion on the first day, but on the second, pressures equivalent to between two and three feet of water were shown, rising on the fourth day to nearly six feet, but distinctly falling on the sixth and seventh days. Ap. 27. The second trunk was cut across completely. No bleeding now or at any time. T wo anomalies will be noticed here. Firstly, that a higher pressure was shown by the attached end of the root 1 than by the severed portion, as though the ‘ root- pressure ' were driving the sap downwards instead of upwards; and secondly, that the pressure in the attached portion of the root was much higher than was re- quired to raise water to the cut surface of the stump, and yet no bleeding took place from it, although an active exudation was shown by the root. Even when the second trunk was cut and covered with indiarubber no actual drops of water exuded, although both the duramen and the remaining alburnum were quite moist. The explanation appears to be that, early in the year, the wood of the intact trunk offers a higher resistance to the passage of water than it does later on, although to obtain direct evidence of this is by no means easy. Furthermore, different portions of the root-system appear to awaken to active absorption at dissimilar times. Even although the pressure in the intact root-system was nearly uniform throughout, the maximal pressures shown by manometers attached to severed portions of it might vary considerably, according to the amount of absorbing surface and the relative activity of absorption. This is because the maximal pressure in a severed root always decreases sooner or later, so that the height of the pressure shown by an attached manometer will depend upon the rapidity with which the maximal pressure is attained, which again depends upon the rapidity of escape of sap, and this upon the activity of absorption. It would, in fact, be more accurate to test the pressure of absorption in a severed root by applying increasing pressures of Mercury until sap neither escapes nor is driven backwards. To perform an extended series of observations of this kind, each of which demands the sacrifice of a large tree, is however possible only to a privileged few. The above observations are therefore merely put forward as suggesting the need of solving the following questions : — (1) Does the total resistance to the flow of water in the trunk of a deciduous tree vary and show an annual rhythm or periodicity ? (2) Is the root-pressure comparatively constant throughout large root-systems, and do all regions of such systems awaken to active absorption at the same period of time? ALFRED J. EWART. 1 The root was subsequently traced to its junction with the parent-tree. ANNALS OF BOTANY, Vol. XVII. Contains the following Papers and Notes: — Sargant, Miss E. — A Theory of the Origin of Monocotyledons, founded on the Structure of their Seedlings. With Plates I-VII and ten Figures in the Text. Darwin, F., and Pertz, Miss D. F. M. — On the artificial Production of Rhythm in Plants. With a note on the position of maximum heliotropic stimulation. With four Figures in the Text. Salmon, E. S. — A Monograph of the Genus Streptopogon, Wils. With Plates VIII, IX, and X. Marloth, R. — Some recent Observations on the Biology of Roridula. With a Figure in the Text. SPRAGUE, T. A — On the Heteranthus Section of Cuphea (Lythraceae). With Plate XI. Barker, B. T. P. — The Morphology and Development of the Ascocarp in Monascus. With Plates XII and XIII. Vines, S. H. — Proteolytic Enzymes in Plants. Allen, C. E.---The Early Stages of Spindle-Formation in the Pollen-Mother-Cells of Larix. With Plates XIV and XV. Willis, J. C., and Burkill, I. H. — Flowers and Insects in Great Britain. Part II. Observa- tions on the Natural Orders Dipsaceae, Plumbaginaceae, Compositae, Umbelliferae, and Cpr- naceae, made in the Clova Mountains. Miyake, K. — On the Development of the Sexual Organs and Fertilization in Picea excelsa. With Plates XVI and XVII. Howard, A. — On some Diseases of the Sugar-Cane in the West Indies. With Plate XVIII. Hill, T. G., and Freeman, Mrs. W. G. — The Root-Structure of Dioscorea prehensilis. With Plate XIX and a Figure in the Text. Arber, E. A. N. — On the Roots of Medullosa anglica. With Plate XX. Thiselton-Dyer, Sir W. T.— Morphological Notes. IX. A Kalanchoe Hybrid. With Plates XXI-XXIII. Oliver, F. W.— The Ovules of the older Gymnosperms. With Plate XXIV and a Figure in the Text. f Davis, B. M. — The Origin of the Archegonium. With two Figures in the Text. Tansley, A. G., and Chick, Miss E. — On the Structure of Schizaea malaccana. With Plates XXV and XXVI and a Figure in the Text. Boodle, L. A. — Comparative Anatomy of the Hymenophyllaceae, Schizaeaceae and Gleicheniaceae. IV. Further Observations on Schizaea. With three Figures in the Text. Willis, J. C., and Burkill, I. H. — Flowers and Insects in Great Britain. Part III. Observa- tions on the most Specialized Flowers of the Clova Mountains. Dale, Miss E. — Observations on Gymnoascaceae. With Plates XXVII and XXVIII. Vines, S. H. — Proteolytic Enzymes in Plants (II). Fritsch, F. E. — Further Observations on the Phytoplankton of the River Thames. Two Fungi, parasitic on species of Tolypothrix (Resticularia nodosa, Dang, and R. Boodlei, n. sp.). With Plate XXIX. Campbell, D. H.— Studies on the Araceae. The Embryo-sac and Embryo of Aglaonema and Spathicarpa. With Plates XXX, XXXI, and XXXII. Gwynne-Vaughan, D. T. — Observations on the Anatomy of Solenostelic Ferns. Part II. With Plates XXXIII, XXXIV, and XXXV. Hemsley, W. Lotting.— On the Genus Corynocarpus, Forst. With Descriptions of two new Species. With Plate XXXVI and two Figures in the Text. Scott, Rina. — On the Movements of the Flowers of Sparmannia africana, and their Demonstra- tion by means of the Kinematograph. With Plates XXXVII, XXXVIII, and XXXIX. Thiselton-Dyer, Sir W. T. — Morphological Notes. X. A Proliferous Pinus Cone. With Plate XL. NOTES. Hill, A. W. — Notes on the Histology of the Sieve-tubes of certain Angiosperms. Crossland, C. — Note on the Dispersal of Mangrove Seedlings. With a Figure in the Text. CAVERS, F. — Explosive Discharge of Antherozoids in Fegatella conica. With a Figure in the Text. Fritsch, F. E.— Algological Notes. IV. Remarks on the periodical Development of the Algae in the artificial Waters at Kew. Bower, F. O. — Note on abnormal Plurality of Sporangia in Lycopodium rigidum, GmeJ. With a Figure in the Text. Hemsley, W. B., and Rose, J. N. — Diagnoses Specierum Generis Juliania, Schlecht., Americae Tropicae. Hope, C. W.— Note to Article in the Annals of Botany, Vol. xvi, No. 63, September, 1902, on * ‘ The “ Sadd ” of the Upper Nile.’ Pearson, H..H. W.' — The Double Pitchers of Dischidia Shelfordii, sp. nov. Bower, F._ O. — Studies in the Morphology of Spore-producing Members. No. V. General Comparisons, and Conclusion. Oliver, F. W., and Scott, D. H. — On Lagenostoma Lomaxi, the Seed of Lyginodendron. Wigglesworth, Miss G. — The. Cotyledons of Ginkgo biloba and Cycas revoluta. With a Figure in the Text. Stopes, Miss M. C. — The ■ Epidermoidal * layer of Calamite roots. With three Figures in the Text. CONTENTS OF VOL. XVII, INDEX. The JOURNAL of BOTANY, British and Foreign. A Monthly Journal. Edited by James Britten, F.L.S. Usually a Plate with each number, is. 8 d. Contains original articles by leading Botanists, Notices of Books, and botanical news generally. January , 1904, contains ‘ R. Brown’s List of Madeira Plants,’ by James Britten, F.L.S.; ‘ South Devonshire Plants,’ by G. Claridge Druce, M.A., F.L.S. ; ‘ Inula grandiflora, Willd.,’ by A. B. Rendle, D.Sc. ; ‘ North-East Highland Plants (1903),’ by the Rev. W. Moyle Rogers, F.L.S. ; ‘Wayfaring Notes from the Transvaal,’ IV, by R. Frank Rand, M.D., F.L.S. (concluded); * Epipactis atroviridis’ by Arthur Bennett, F.L.S. ; Short Notes; Notices of Books ; Book Notes ; News, &c., and Supplement Notes on the Drawings for English Botany , by F. N. A. Garry (continued). Subscription for the year, including postage to any part of the world , 16s. London: WEST, NEWMAN & CO., 54 Hatton Garden. OXFORD UNIVERSITY PRESS. • The Physiology of Plants. A Treatise upon the Metabolism and Sources of Energy in Plants. By Dr. W. Pfeffer, Professor of Botany in the University of Leipzig. Second fully Revised Edition, translated and edited by Alfred J. Ewart, D.Sc , Ph.D., F.L.S. With many Illustrations. Royal 8vo, half-morocco, Vol. II, 16s. Previously published, Vol. I, 28.L Schimper’s Plant=Geography. A Summary of the Existing State of Knowledge about the Geographical Distribution of Plants so far as dependent on Conditions now present; with Special Consideration of the Principles determining the Habits and Structure of Plants, the Evolution of their Adaptive Characters, and the Congregation of Plants to form Communities. Authorized English Translation by W. R. Fisher, M.A. Revised and Edited by P. Groom, M.A., D.Sc., F.L.S., and I. Bayley Balfour, M.A., M.D., F.R.S. With a Portrait of the late Dr. A. F. W. Schiinper, Four Maps, Four Collotypes, and 497 other Illustrations. Royal 8vo. In Four Parts, at gs. net per Part. Complete Volume, half-morocco, 42 s. net. , Scotsman : — ‘This work on Plant-Geography is undoubtedly Professor Schimper’s magnum opus. It is already recognized on the Continent as the standard work on oecological botany, and its translation into English will give an immense stimulus to a fascinating branch of botanical study in this country.’ On the Physics and Physiology of the Proto= plasmic Streaming in Plants. By Alfred J. Ewart, D.Sc., Ph.D., F.L.S., Lecturer on Botany in the Birmingham Technical Institute. Communicated to the Royal Society by Francis Gotch, D.S6. (Oxon.), F.R.S. Royal 8vo, half-morocco, with 17 Illustrations. 8.r. 6d. net. Plant World (Washington) : — ‘The present volume not only contains a summary of what has already been learned about protoplasmic streaming, but also the results of many experiments and observations made by the author. The work is comprehensive, and will be indispensable to all physiological botanists.’ REPRINTED FROM THE ‘ANNALS OF BOTANY.’ A Sketch of the Life and Labours of Sir William Jackson Hooker, K.H., D.C.L., Oxon., F.R.S., F.L.S., &c , late Director of the Royal Gardens of Kew. By his son, Joseph Dalton Hooker. Royal 8vo, paper covers, 3^. 6d. net. LONDON: HENRY FROWDE, OXFORD UNIVERSITY PRESS WAREHOUSE, AMEN CORNER, E.C. Vol. XVIII. No. LXX. April, 1904. Price 14s. Annals of Botany EDITED BY ISAAC BAYLEY BALFOUR, M.A., M.D., F.RS. KING’S BOTANIST IN SCOTLAND, PROFESSOR OF BOTANY IN THE UNIVERSITY And keeper of the royal botanic garden, Edinburgh D. H. SCOTT, M.A., Ph D., F.R.S. HONORARY KEEPER OF THE JODRELL LABORATORY, ROYAL BOTANIC GARDENS, KEW AND WILLIAM GILSON FARLOW, M.D. PROFESSOR OF CRYPTOGAMIC BOTANY IN HARVARD UNIVERSITY, CAMBRIDGE, MASS., U.S.A. ASSISTED BV OTHER BOTANISTS London HENRY FROWDE, AMEN CORNER, E.C. CLARENDON PRESS DEPOSITORY, 116 HIGH STREET 1904 Printed by Horace Hart, at the Clarendon Press, Oxford . PAGE CONTENTS. Williams, J. Lloyd— Studies in the Dictyotaceae. II. The Cytology of the Gametophyte Generation. With Plates XII, XIII, and XIV ........... 183 Bower, F. O. — Ophioglossum simplex, Ridley. With Plate XV . 205 Parkin, J. — The Extra- floral Nectaries of Hevea brasiliensis, Miill.- Arg. (the Para Rubber Tree), an Example of Bud-Scales serving as Nectaries. With Plate XVI . . . . . .217 Church, A. H. — The Principles of Phyllotaxis. With seven Figures in the Text ........... 227 Mottier, D. M. — The Development of the Spermatozoid in Chara. With Plate XVII ......... 245 Weiss, F. E.— A Mycorhiza from the Lower Coal-Measures. With Plates XVIII and XIX and a Figure in the Text . . . 255 Reed, H. S.— A Study of the Enzyme-secreting Cells in the Seedlings of Zea Mais and Phoenix dactylifera. With Plate XX . , 267 Vines, S. H. — The Proteases of Plants 289 NOTES. Massee, G.— On the Origin of Parasitism in Fungi . . , .319 Salmon, E. S. — Cultural Experiments with ‘ Biologic Forms’ of the Erysiphaceae 320 Oliver, F. W., and Scott, D. H.— On the Structure of the Palaeozoic Seed Lagenostoma Lomaxi, with a Statement of the Evidence upon which it is referred to Lyginodendron . . . . 321 NOTICE TO SUBSCRIBERS. The subscription-price of "each volume is thirty shillings, payable in advance : the Parts, four in number, are* supplied as they appear, post free to subscribers in the United Kingdom, and with a charge of is. 6d. per annum for postage to subscribers residing abroad. The price of individual Parts is fixed at a higher rate. Intending subscribers -should send their names, with subscription, to Henry Frowde, Oxford University Press Ware- house, Amen Corner, London, E.C. As the earlier volumes of the Annals of Botany are becoming scarce, Vol. I will only be sold as part of a complete set; and Parts will not as a rule be sold separately, after the publication of the volume to which they belong. A few extra copies of particular Parts at present remain on hand, for which special application must be made to the Editors, Clarendon Press, Oxford. NOTICE TO CONTRIBUTORS. Contributors in America should address their communications to Professor Farlow, Harvard University; and all other contributors, to the Editors, at the Clarendon Press, Oxford. Papers sent in with a view to publication must be type-written ; and the illustrative figures should be planned so as to properly fill a 410 or an 8vo plate. Attention to this will conduce to the rapid publication of the papers if accepted. Each contributor to the Annals of Botany is entitled to receive gratis twenty-five separate copies of his paper, and may purchase additional copies if he informs the Editors of his wishes in this respect when he returns corrected proofs. The price of these additional copies will depend upon the amount of text and the number of plates in the paper. Studies in the Dictyotaceae. II. The Cytology of the Gametophyte Generation. BY J. LLOYD WILLIAMS, Assistant Lecturer in Botany , University College , Bangor . With Plates XII, XIII, and XIV. IN a previous paper the development, germination and cytology of the tetraspore in Dictyota and other members of the group were described. In the present paper it is proposed to deal with the development of the oosphere and antherozoid, the fertilization of the ovum and its subsequent segmentation, together with the parthenogenesis of unfertilized eggs — the observations in this case applying to Dictyota dichotoma only. In this plant the male and female gametes are borne upon different individuals. Excepting so far as the reproductive cells themselves are concerned, there is no difference either of form or of structure between the tetrasporic plants on the one hand and the sexual ones on the other. It is true that the latter generally have the branches of the thallus broader than those of the former, but to this rule there are frequent exceptions. As already described, the discovery of motile antherozoids, first made in 1896, was not confirmed till nearly a twelvemonth after. Subsequently the astonishing fact was established that, unlike the tetraspores, both the oogonia and antheridia are developed simultaneously in fortnightly crops, each crop being initiated a little before the lowest neap tide, and arriving at maturity about the period of the highest succeeding spring tide. The gametes thus produced are liberated during two or three days while the high tides are on the wane. A regular succession of crops continues thus from July to the end of October. So far as I am aware nothing like this remarkable periodicity has hitherto been observed in the case of any other Alga. During the last six summers careful records have been kept of the appearance of these crops, and their relation to the conditions of their environment. The data thus collected, together with a discussion of the factors concerned, will be published in a separate paper. These observa- tions only apply to the plant as growing on the North Wales coast ; it [Annals of Botany, Vol. XVIII. No. LXX, April, 1904.] O 184 Lloyd Williams . — Studies in the Dictyotaceae. would be instructive to know whether they hold good for other localities also. As the sexual cells pass through the various stages of development at approximately the same dates, it is far easier to find any required mitosis here than in the gametangia of the Fucaceae or in the tetrasporangia of the Dictyotaceae. After the preliminary trials one has only to consult the tide tables for the particular locality to know within two or three days when certain stages may be found. Although Thuret and Reinke have given excellent figures of the sexual cells of Dictyota no description of their cytology has as yet been published. I. The Development of the Oogonia. The oogonia instead of being isolated, as is the case with the tetra- sporangia, are grouped together in sori, each sorus consisting of twenty-five to fifty oogonia, closely packed together without paraphyses or any other accessory cells. In the case of weak plants, or under adverse conditions, the number may be reduced to half a dozen or fewer. The sori are scattered all over the two surfaces except the basal portion, the apices, and a very narrow band at the margins of the thallus. The sori of the new crop appear between the scars of the older ones, and, in the case of an elongating plant, in acropetal succession on some of the surface cells which have sufficiently matured since the preceding crop. This process may go on till the whole of the available surface has been used up, then the plant dies. In no case are gametangia produced by the stalk-cells left by preceding sori. As in the case of the tetrasporangium the rudiment of the oogonium is one of the small cells of the assimilating layer, which increases in dimensions till the free part is about three or four times the height of the original cell (PI. XII, Fig. 3). The nucleus then divides, and a stalk-cell is cut off. The oogonium mother-cell increases somewhat in size, but it undergoes no further division, the nucleus becoming that of the ovum without an additional mitosis. This is essentially different from the state of affairs in the oogonium of the Fucaceae, where eight nuclei are produced even in the cases where only one egg ultimately matures. In the Fucaceae also, where there is no asexual generation, reduction takes place in the first of the three oogonial mitoses ; in Dictyota , as shown in the preceding paper, it is brought about in the first of the two tetrasporic divisions. In a few days after the separation of the stalk-cell the apex of the oogonium ruptures and the naked oosphere is liberated into the water, where fertilization takes place. The stalk-cell mitosis is accomplished from one to four days before the highest spring tide, the variation in the time being dependent on the length of the interval between two spring tides, the conditions with respect II. The Cytology of the Gametophyte Generation. 185 to temperature and light prevailing at the time, and the period of the year. The whole process is gone through with considerable rapidity, and all the stages may be found in a single sorus. In the case of the tetrasporangium the stalk-cell separates while the rudiment is still comparatively small ; here, however, it is delayed till the cell is nearly fully grown, and liberation follows in about three to five days. Of the oogonia in a sorus the central ones only are vertical to the thallus ; in consequence of their great increase in diameter as compared with the stalk-cells the others lean outwards, the angle being greater as the periphery is approached ; this makes it difficult to obtain median sections through the dividing nuclei. It is generally stated that there are no borders of elongated sterile cells to the oogonial sori, as there are to the antheridial ones. This, however, is not strictly correct ; borders, partial or complete, are frequently found and may arise in one of two ways. 1. Some of the outer cells of a sorus may at a comparatively early stage in their development be compressed by the more actively growing inner ones, and prevented from developing further. These are visible in sections as narrow elongated cells, but in a surface view are completely hidden by the outermost of the fertile oogonia. 2. Towards the close of the season, when unfavourable conditions supervene, most of the outermost oogonia fail to mature even after attaining their full size; the cytoplasm is smaller in amount and the nucleus does not divide. These facts suggest the mode in which the antheridial borders may have arisen, and the manner in which borders may also in the course of time be acquired by the oogonial sori. In this connexion it is instructive to notice that the central portions of both male and female sori are more active than the peripheral ones ; in the early stages the nuclei are more advanced, liberation generally starts in the middle of the sorus, and sterili- zation, although it sometimes occurs there, is not nearly as frequent as it is in the marginal regions of the sori. Fig. 1 shows the structure of the oogonium-rudiment shortly before the separation of the stalk-cell, as seen in a section vertical to the surface of the thallus and transverse to the branch. As in the case of the tetra- sporangium-rudiment the basal part of the cell has very large vacuoles, while the remainder is occupied by denser cytoplasm, in which the chloro- plasts are much more crowded. The curved, rod-like centrosomes and radiations are very distinct, not only at the distal pole but most commonly at the basal pole as well. The physodes as usual are chiefly located in the peripheral layers of the free surface. The early prophase stage is very much like that of the stalk-cell division in the tetrasporangium ; there is the same coarse chromatin thread, irregular in the arrangement and thickness of its granules, and soon 1 86 Lloyd Williams . — Studies in the Dzctyotaceae . segmenting into the chromosomes. When, however, the latter make their appearance, they exhibit the longitudinal split far more clearly than do the chromosomes in any other division (Figs. 2, 3). At first the nucleolus has one or more fairly large vacuoles with a number of smaller ones ; soon, however, the nucleolar membrane disappears, and the bulk of the nucleolus is seen to consist of a large number of deeply staining fibrillae, together with a single, slightly staining globule (Fig. 3). During this period the amount of nucleoplasm is very small, and the centrospheres are less distinct. The spindle (Fig. 4) is intranuclear and narrow. There are a few mantle-fibres extending outwards, the spaces between which are occupied by a large number of the deeply stained fibrillae derived from the dis- integrating nucleolus. The globule is generally present either on the spindle- fibres or in the immediate vicinity. The polar view represented in Fig. 5 shows the chromosomes to be curved in form and clearly sixteen in number. The nucleus at this stage is somewhat nearer the base of the cell, and the cordate appearance shown in Fig. 4 is very frequently met with ; it is almost certainly due to a contraction of the spindle after fixation. In preparations of segmenting eggs depressions are frequently found at both poles of the nucleus. These facts confirm the observation that the spindle-fibres at this period are in a state of tension and that they contract on fixation. During the anaphase stage the spindle is still narrow, but the nuclear cavity is wide and the membrane is intact except at the poles. The centrosomes are clearer than during the preceding stage, and in most cases the nucleolar globule is present. When the chromosomes reach the poles of the spindle the membrane disappears, the lower nuclear mass remains in the same position or descends slightly while the upper ascends, thus causing the whole figure to become greatly elongated. In the telophase stage (Fig. 6) of the oogonial nucleus the chromosomes, now greatly increased in size, but still preserving their curved form, are seen to be quite separate from each other and sixteen in number. When they fuse they frequently form a thick deep-staining irregular ring in which at one particular stage of the fusion about eight deeply stained masses may be distinguished (Fig. 7). Something similar was described for the stalk-cell division of the tetrasporangium, but was not observed in any of the other mitoses. The two divisions are strikingly similar also in the fact that only one deeply staining mass is formed in both cases, which ultimately becomes the nucleolus. If Figs. 1, 10-13 be compared certain peculiarities of cell- wall structure may be observed. Although the drawings are made from preparations fixed and stained by different methods they all agree in showing a distinct difference between the membranes of the parts of the cells embedded in II. The Cytology of the Gametophyte Generation. 187 the thallus and those of the free, reproductive portions. For want of space this topic must be deferred for discussion in another paper. It may be pointed out that we have a provision here for the rapid elongation of the gametangial walls, for the early splitting of the lateral walls, and for their dissolution when the gametes arrive at maturity. It is evident that the basal cells have their walls of different chemical composition, and of more enduring character. When quite mature the degeneration of the apical walls and the increasing turgescence of the ovum causes the former to burst and to liberate the oosphere, which now lies a naked spherical mass of protoplasm. Fig. 19 shows a section through such an ovum. The nucleus now occupies nearly one-third of the diameter of the egg. It is surrounded by a zone of 4 kinoplasm,’ outside of which is the somewhat vacuolated cytoplasm with its numerous chloroplasts. In well fixed material these are generally oval or rounded, but when much contraction has taken place they are very narrow and elongated. At the periphery of the egg there is a zone of granular cytoplasm in which are embedded numerous intensely staining physodes. Very soon after liberation of the oospheres in a sorus the oogonial walls are completely dissolved, leaving only the basal cells, which are marked from the other surface cells of the thallus by their slightly greater width, and particularly by their paler colour. The latter pecu- liarity is due to their having fewer and paler chloroplasts. As soon as they are exposed to the light by the liberation of the overlying eggs, the chloroplasts begin to multiply and acquire a deeper colour, but the scars of two or three of the most recent crops can be easily distinguished from each other by their different shades of colour. II. The Development of the Antheridium. If a fresh plant of Dictyota be examined one to five days before the lowest neap tide the rudiments of the antheridia may be distinguished by their slight increase in diameter and their deeper colour, but chiefly by the pale circular area in the centre, indicating the position of the nucleus. Material fixed about this period would show the various stages in the stalk- cell mitosis, and perhaps also in the first division of the mother-cell of the antheridium. The antheridia, like the oogonia, are arranged in sori, but they are much more numerous than the former, the number in a fair-sized sorus being from one to two hundred. Each sorus has always a well-defined border of sterile cells consisting of three or more rows, the innermost of which are of about the same length as the antheridium, while the others successively diminish in height (Figs. 12-14). Some time after the libera- tion of the antherozoids the innermost border-cell cuts off a cell near the 1 88 Lloyd Williams . — Studies in the Dictyotaceae . top ; occasionally this happens also in the case of some of the cells of the next row. The stalk-cell division takes place when the cell is still very small, consequently the nucleus, though much larger than that of one of the vegetative cells, is considerably smaller than the nucleus of the tetra- sporangium or of the oogonium. The result is that the various processes are more difficult to follow, the counting of the chromosomes being in some stages particularly difficult. This has necessitated the comparison of a very large number of figures. It is found that the mode of division closely resembles that which obtains in the stalk divisions already described. While the increase in size is as yet exceedingly slight and there has been but little accumulation of protoplasm, the nucleus and nucleoli are greatly enlarged though the nuclear network is still inconspicuous. Even at this stage the centrosomes are perfectly clear at both poles, and their form is that already described in other cases. As usual the distal pole is dis- tinguished by beautiful radiations, which are nearly always absent from the other pole. There are no new points to be noted with regard to the spirem, the chromosomes, or the spindle. In many instances the lower pole seems to project into the vacuolated part of the cell. Sometimes a few strands of protoplasm bridge over the space, and connect the nucleus with the thin lining of protoplasm at the base. Frequently thick sections show strands that end blindly in the vacuole. These recall the appearance of the cords of protoplasm described by Phillips and seen in the living cells of certain Florideae, and which are in a state of continual motion within the vacuole, either extending or retracting or bending upon themselves. The absence of radiations and the position of the lower pole with regard to the vacuole preclude the possibility of a pull being exerted by any kino- plasmic structures at the base. During the metaphase and anaphase stages the nuclear membrane is present, in fact it persists in some cases till the very commencement of the telophase stage, then, as in the oogonium, the membrane disappears and the lower nucleus descends into the middle of the vacuolated region. The reconstruction of the nucleolus sometimes follows the same course as in the two stalk-cell divisions already described. In other cases several masses may be seen which differ greatly in size. These soon coalesce to form the spherical nucleolus, which at first presents a curious appearance, as if it consisted of a globule of pale-staining substance with a number of chromosome-like bodies embedded, chiefly though not solely at the peri- phery. This appearance, together with the very scanty, slightly staining nuclear network, confirms the idea that in the daughter-nuclei resulting from these three stalk-cell mitoses the whole of the chromatin is located in the nucleolus. The partitioning of the antheridium into antherozoid mother-cells II. The Cytology of the Garnet op hyte Generation . 189 is completed a few days before liberation, the whole process occupying from six to ten days. Particular attention was paid to the later cell- divisions to see if any evidence of amitosis could be observed. As far as can be seen, however, all the usual stages of karyokinesis are repeated at each step, and after each cell-division the nucleus assumes the appearance of the resting condition. Several reliable countings of chromo- somes were obtained in the earlier mitoses, both in the prophase stage and in polar views of the equatorial plate stage (Fig. 11). All these agreed in giving the number as sixteen. Occasionally a larger number was indicated, but in all such cases it was easy to see that some of the bent chromosomes had their angles cut off, thus giving two chromatic segments instead of one. The first division-wall of the antheridium mother-cell is vertical to the thallus and at right angles to the long axis of the branch, the second is also vertical but parallel to the axis, so that in surface view four cells are seen. The succeeding divisions are not so regular, but a number of tangential division-walls appear before any further vertical ones are formed. Ultimately the antheridium in surface view is seen to consist of sixty-four cells arranged in four groups of sixteen. In vertical section there is great divergence between different antheridia, the number of tiers in some cases being twenty-four (Fig. 14), in others about twenty. The number of successive nuclear divisions must be about twelve, but it is clear that many of the cells fail to divide and others abort. It is instructive to note in this connexion that in many cases when the circumstances are unfavourable it is the lowermost cells in an antheridium that are first retarded ; this fact will be referred to again when discussing the importance of light as a factor in the development of the gametangia. Careful countings of the antherozoid mother-cells in a large number of antheridia show that the average number of sperms produced in an antheri- dium is about 1,500. It has already been shown that the number of oogonia in a sorus ranges from about twenty-five to fifty, while the number of antheridia is generally from 100 to 300. Taking the lower number in each case we see that for every egg in an oogonial sorus there will be 6,000 antherozoids in a sorus of antheridia. In order to form an approximate idea of the number of gametes produced by a single plant the sori were counted on a specimen twelve inches long. They were found to be over 3,500 in number. Allowing an average of 1,500 antherozoids for every antheridium this would give us a total of over 500 millions for the whole plant, a number which is certainly much too low for a vigorous, full-grown plant. If we consider a single locality, the Menai Straits for instance, where Dictyota flourishes in great abundance, and the vast multitude of sexual plants which on certain days every fortnight throughout the summer months fill the water in their vicinity with untold myriads of swarming gametes, we cannot but be filled with astonishment at the 190 Lloyd Williams. — Studies in the Dictyotaceae . prodigious numbers produced, and yet so far as the production of new plants is concerned nearly the whole of the energy expended and the elaborate mechanism employed has been wasted. This, of course, is an old story, but every new exemplification of it comes with a fresh surprise. The antherozoids were described by me in a short paper published in 1897. Since then I have had many opportunities of examining these bodies in the active condition, of fixing and staining them by better methods, and of studying their structure by means of superior objectives. Some of the conclusions first arrived at are now found to be incorrect in certain particulars, and several points which were then obscure have now been satisfactorily cleared up. Taking first the antherozoid as it appears in the living condition, we find that it is pear-shaped while active, but spherical when it comes to rest (Fig. 18). At the broader, posterior end there is a globular colourless part which with chromatin stains takes on an intense colouration, showing it to be the nucleus. The pointed end is occupied by granular protoplasm in which there is always a very minute red eyespot, sometimes two as in the lowest example in the figure. Instead of being situated close to the attachment of the cilium as is usual in other antherozoids and in zoospores, the eyespot here is always at a distance from it, frequently at the pointed anterior end. The cilium is not terminal as was stated in the ’97 paper, but lateral as in other Phaeophyceae ; this fact can be clearly seen only while the anthero- zoid is in motion. The cilium is attached at the posterior edge of the cytoplasmic cone, near its junction with the nucleus. Very often there seems to be a slight thickening at the very base, but on this point it is not easy to speak with certainty. Of the fixed preparations one of the most successful is that shown in Fig. 1 5, fixed with potassium-iodide iodine in sea- water ; this preserves the form and structure very faithfully. Those in Fig. 16 were fixed in dilute Flemming solution, and stained by the iron-alum method. Here the anterior part is much wrinkled, but the nuclei and physodes are very distinct. When fixed with dilute Hermann’s solution (Fig. 17) and stained with gentian violet the nuclei stain very deeply, the cytoplasm slightly, and occasional physodes are seen. At the upper right-hand corner is shown a monstrous antherozoid with two nuclei. If doubly stained with brazilin and Hoffmann’s blue the nucleus takes on the red colour, while the anterior part stains blue. The number of physodes present varies greatly, as also does their reaction to a solution of vanillin in HC1. When fresh plants are tested with the above reagent the sori frequently give no trace of the red colour (phloroglucin ?) usually given by the physodes of mature cells. At other times every sorus is coloured a deep red ; this happens generally when there has been some delay in the liberation of the antherozoids. When II. The Cytology of the Gametophyte Generation. 19 1 the reaction is partial it generally appears in the central antheridia only, an additional confirmation of the greater vigour of this region. Much time and trouble have been expended in trying to find whether there is a second cilium or not. In the Fucaceae, as is well known, the posterior cilium is very long but very much thinner than the anterior one. The possibility had to be considered that in Dictyota the second cilium is so reduced in length as to be quite rudimentary, or that it is attenuated to such an extent as to be indistinguishable. In the upper left-hand corner of Fig. 17 is represented an apparently biciliate antherozoid, and several more such cases have been observed. After a careful comparison with fixed and stained antherozoids of the Fucaceae I have failed to satisfy myself that these appearances are not due to strands of mucilage. In the living condition I have only once seen anything suggestive of a second cilium. On one occasion, while examining a number of antherozoids by means of Zeiss’s D* water-immersion objective, one was observed which had got stuck fast at the bottom of the dish. In almost all cases of the kind it is the end of the long cilium that becomes attached ; in this instance, however, the cilium was quite free, and it could be clearly seen violently lashing about in the water. Careful observation showed that the antherozoid was attached to the glass by a filament of much greater tenuity than the cilium, and of about the same length as the body of the sperm. This is the strongest evidence yet obtained for a second cilium, but even this is quite insufficient to enable one to come to a positive decision on the point. III. The Fertilization of the Egg. Thuret, Re'inke*, and other investigators tried to solve the question of the mode of fertilization in Dictyota. Their failure to obtain active antherozoids for their experiments prevented them from arriving at a correct solution of the problem. It was further complicated by the fact that many of the eggs segmented in the total absence of antherozoids. Among the various suggestions advanced from time to time in explanation of the difficulty were the following : — (1) That fertilization is effected before the liberation of the eggs from the oogonia ; this seemed to explain the apparent germination of liberated oospheres in the absence of sperms. (2) That there are two kinds of asexual spores. (3) That the antherozoids have become functionless and that the eggs are habitually parthenogenetic. To Johnson, in his study of Dictyopteris , belongs the credit of having first pointed out the true answer to the question, though his somewhat tentative pronouncements were never con- firmed by him. The present investigation has now conclusively demonstrated that 192 Lloyd Williams. — Studies in the Dictyotaceae • fertilization is entirely external as it is in the Fucaceae, that unfertilized eggs will pass through a few of the earlier segmentation stages by an abnormal process of parthenogenesis, but that normal germlings can only be produced from fertilized eggs. The following is the method employed for securing material for this stage. Oogonial and antheridial plants are collected when the sori are quite mature, if possible on the evening preceding the day of maximum liberation. The plants are kept separate in moist chambers till the follow- ing morning, pieces are then immersed in glass dishes containing clean sea- water. The oospheres slowly burst through the walls of the oogonia and gradually sink to the bottom of the vessel. Before a sufficient number have been liberated to make it profitable to experiment with, half an hour or more will have elapsed. When the antheridial plants are immersed the antherozoids come out of the restraining cells very quickly, and soon the water is cloudy with swarming sperms. If now some of the latter are added to the oospheres by means of a pipette, it will be seen that instead of impartially surrounding all the eggs, as do Fucus antherozoids, many of the eggs are here passed over and entirely ignored, whereas others are covered with antherozoids, which in some cases lie three to twelve layers deep. If the same material be fixed after an interval of twenty-four hours and sub- sequently sectioned it will be found that the antherozoid-covered eggs have segmented in a perfectly normal manner, whereas the others show the abnormal mode of division described below, a method which is character- istic of the so-called parthenogenesis already referred to. When pieces of oogonial plants have been immersed for about an hour the eggs set free will have been liberated at various intervals of time, some nearly an hour, others only a few minutes. If now instead of adding antherozoids we test the eggs with distilled water, we find that those that have been liberated for half an hour or more have already acquired walls. This partly explains why they are no longer capable of attracting the antherozoids, and why they cannot possibly be fertilized. From this cir- cumstance it is very difficult to get uniformity of stages in cultures of Dictyota oospheres. Although the mode of fertilization in Dictyota is very similar to that in Fucus there are important differences. In the former the liberation of the gametes is simultaneous ; this tends to give it a great advantage over Fucus , where there is no regular periodicity in the production of the sexual cells. On the other hand, the eggs in Dictyota very soon lose their capacity for fertilization, whereas those of Fucus may retain it for days. This would seem to act as a disability, and so to neutralize the advantage obtained by the simultaneous emission of gametes. As already explained it is my intention to deal with physiological con- siderations in a succeeding paper ; two points, however, demand mention here. 193 II. The Cytology of the Gametophyte Generation. The first is the relation of the antherozoids to light. It was stated in a former paper that light is essential to their mobility. It is true that light has a very great influence upon the antherozoids, as will be shown again, but if they are really ready for liberation they will on immersion display their activity at any hour, night or day. The second point relates to chemotaxis in fertilization. Buller has recently performed some experiments on the eggs of certain animals, which lead him to infer that in external fertilization there is no such thing as chemotactic attraction of the spermatozoids. He seems inclined to follow Bordet, and to extend this generalization so as to include the Fucaceae. Thus he says in explaining Bordet’s conclusions : ‘ According to this observer it is simply the ability of the spermatozoa to adhere to surfaces by the tip of one of the cilia which leads to their collection upon an egg, while their meeting with it is a matter of chance. A few observations of my own at Naples upon the fertilization of Cystoseira barbcUct did not reveal to me any certain attraction of the spermatozoids from a distance, but the collection of the spermatozoids upon the eggs in consequence of their ability to cling to surfaces was clearly seen. Nevertheless in view of the positive statement of Strasburger a careful re-investigation of the question seems to be desirable.5 It is not my intention at present to discuss the question of chemotaxis in the Fucaceae. As regards Dictyota , however, it would be difficult to find anywhere a stronger suggestion of attraction at a distance than is offered by the phenomenon described above, where, out of several hundred eggs crowded together in a small dish, a few here and there are surrounded by several layers of spermatozoids, while the great majority of the eggs have none attached to them. Nor can it be said that in the former case the crowd of antherozoids are entangled in a mucilaginous envelope, for in the first place their movements are absolutely unrestrained, and furthermore, staining with gentian violet fails to demonstrate the presence of any such substance. During the process of fertilization the eggs of Dictyota do not revolve as do those of Fucus , but the antherozoids, unless too crowded, slowly gyrate with the end of the cilium resting upon the surface of the egg. At the same time the body of the antherozoid displays a rapid vibratory motion. Frequently the cilium, instead of being straight, presents a wavy appearance (figured in the ’97 paper). When the sperm is becoming less active waves can be seen travelling along the length of the cilium. It has been far more difficult to trace by means of sections the path of the spermatozoid through the cytoplasm to the nucleus than it was in Fucus , and the reasons for this are obvious from the description given above. The examples observed are too few in number to enable one to base any conclusions upon them. The case shown in Fig. 20 is interesting for various reasons. The dark object seen to the right 194 Lloyd Williams. — Studies in the Dictyotaceae. of the egg-nucleus is almost certainly an antherozoid. When compared with those that are lying outside the membrane it is seen to have greatly increased in size and to be somewhat fibrillar in texture. The egg-nucleus is in a prophase stage, but its appearance is more like that of partheno- genesis than of normal segmentation (see Figs. 22, 23). For some reason or other this particular spermatozoid, although it has entered the egg, has failed to reach the nucleus. Fig. 29 is still more curious. Here I regard the egg-nucleus, the one to the right, as having divided partheno- genetically, while the left-hand nucleus is probably an antherozoid nucleus which has failed to fuse with the former, but, at the expense of the egg cytoplasm, has greatly increased in size. On this view this is a stage further than the preceding one. IV. The Segmentation of the Fertilized Egg. The first reliable evidence of fertilization having been accomplished is the presence in the egg-nucleus of a second nucleolus, which is nearly always smaller than the original one (Figs. 21, 22, 23) ; this undoubtedly represents the chromatin brought in by the spermatozoid. Very soon the nucleoli present an appearance suggestive of their containing a number of deeply stained granules or chromosome-like masses. At the same time the chromatin thread is very fine and distributed through the whole of the nuclear space. After this the chromosomes are seen as thick curved rods, coarsely and irregularly beaded (Fig. 22). These after a slight increase in size are longitudinally split, and the chromatin disks become more distinct. The crowded appearance of the chromosomes at once suggests that they are more numerous than in the preceding stages : when they are counted the number in each case is found to be thirty-two. The nucleoli at this period are very irregular in form, but though they seem as if going to fragment they keep their coherence till the spindle stage. Returning now to a consideration of the extra-nuclear structures we find that in the unfertilized newly liberated egg, there is a zone of ‘ kino- plasm ’ surrounding the egg, and that outside this the chloroplasts tend to place their axes radially. At this period it is most difficult to find any recognizable centrosomes. In the stage represented by Fig. 21 there is a centrosome and radiations on one side of the nucleus ; everywhere else the chloroplasts seem to be disposed without definite order, and in many instances they abut directly on the nuclear membrane. It is difficult to be quite certain about the number of centrospheres, but from all I have hitherto seen I am strongly inclined to think that there is only one. Fig. 24 shows an appearance which is so common as to lead me to think that it represents the normal mode of development of the seg- II. The Cytology of the Gametophyte Generation. 195 mentation spindle, as it undoubtedly does also in the second division of the tetraspore mother-cell. Two sheaves of fibres make their appearance at points in the periphery not far from each other ; the whole figure forming a kind of angular spindle with the fibres not continuous. At the ends of the fibres are the chromosomes and a nucleolar mass. The spindle gradually straightens out till it assumes the form of a normal spindle (Fig. 25). If this view is correct then it furnishes a corroboration of the theory that the spermatozoid brings into the egg something (the centro- some, or the influence which produces the centrosome) that determines the polarity of the spindle. First we have the newly liberated egg without centrosphere or any evidence of polarity. From the description given below we find that in parthenogenetic figures there is a total absence of directive influence, the figures are always multipolar, and very irregular. In the prophase stage of the fertilized egg there is a single centrosphere which divides into two ; as the two separate the angular spindle straightens out until the two poles are exactly opposite each other. Of the mature spindle it is unnecessary to speak at any length. It presents the usual features, as may be seen from Fig. 25, excepting that there is always less nucleolar matter. An interesting question, which I have not yet been able to solve, is the fate of the male nucleolus ; is it all used up in chromosome-formation or does it ultimately fuse with the other nucleolar mass? I have up to the present been unable to identify the two masses during the spindle stage. The next figure shows a good polar view of the equatorial plate, which conclusively shows the chromosomes to be thirty-two in number. The telophase stage (Fig. 27) is much like that of tetraspore-segmenta- tion, but there is more frequently the rudiment of a nucleolus present ; sometimes two may be distinguished. Later on the daughter-nuclei seem to contain several nucleolar masses. Very soon, however, the nuclei assume the appearance of the resting stage, in which there are most frequently two nucleoli. These appear very similar to those shown in Fig. 35 of the tetraspore paper ; there, however, only one of the two bodies represents the true nucleolus, the other disappears soon after; here the two bodies are almost certainly the male and female nucleoli. The centrosomes and radiations are very clear at this stage, and they remain distinct during the early stages of the succeeding division. Many of the later mitoses have been studied, but they present no new features, so it is useless to describe them. The time occupied by the segmentation process in Dictyota is much less than it is in the Fucaceae. In the latter it averages sixteen to twenty-four hours : if fertilized eggs of Dictyota be left for this length of time it will be found that many of the germlings are already two to 196 Lloyd Williams. — Studies in the Dictyotaceae. four-celled. In material which was left for nine hours after the addition of antherozoids the following stages were found : (1) Nuclei with two nucleoli andspirem. (2) Nuclei with two nucleoli and chromatin segments. (3) Spindles. (4) Diasters. (5) A few binucleate stages. In addition to these, however, there was a large number of eggs which had been liberated from half an hour to an hour before the addition of antherozoids, and so had become ‘ stale ’ and incapable of being fertilized ; these showed various stages of parthenogenesis. V. The Parthenogenesis of Unfertilized Eggs. A very large number of germination experiments was carried out upon the oospheres of Dictyota . Most frequently antherozoids were added, but control experiments were also performed with the antherozoids omitted. In all the former cases there was a mixture of germlings which had followed the normal course of segmentation as detailed in the preceding section, together with others which showed the characteristic phenomena of parthenogenesis described below. There is no difficulty in distinguish- ing between the two sets of figures. Those of the fertilized eggs can easily be recognized by their resemblance to ordinary karyokinetic figures ; those of the parthenogenesis are quite abnormal in appearance. That the latter really are parthenogenetic figures is confirmed by the fact that they are very numerous in the control cultures, to which no sperms were added, and that these never show any signs of normal mitotic figures. Similar irregular multipolar figures are obtained in certain animal eggs, chiefly as the result of artificial stimulation or of polyspermy, and also in malignant tumours. The cases described by Loeb, Morgan, the Hertwigs, Wilson, Galeotti, and a great many others are too well known to need recapitulation. It is important, however, to bear in mind that parthenogenesis in Dictyota is not the result of any such unusual conditions ; the very fact that they occur side by side with perfectly normal karyo- kinetic figures is sufficient to exclude any such supposition. As it is very important to ascertain the exact number of chromo- somes in the figures, and to know whether two nucleoli are present or only one, it is evidently essential to compare together all the sections of a nucleus before drawing any conclusions. I have throughout adhered rigidly to this practice, but in dealing with cultures of oospores and tetraspores there are certain difficulties which do not confront one when sectioning the reproductive cells on the plant itself. It is very difficult to get cultures of Dictyota eggs free from sand, diatoms, &c., as in II. The Cytology of the Gametophyte Generation. 197 order to get a sufficient number of the ova large pieces of the plants have to be placed in the culture dishes. When such siliceous particles are present they often interfere with the microtoming, and cause the tearing of the ribbons and the consequent loss of sections. Even when the cutting has been successful there are so many eggs in each slice that the identification of the successive sections of any particular one is often a matter of some difficulty. To obviate this trouble it is often a convenient practice to include in the material some other easily cut, and sufficiently distinctive, object to serve as ‘landmark’ so to speak, by reference to which to locate the others. There are two ways in which an unfertilized egg-nucleus of Dictyota may initiate its division. In both cases the single nucleolus increases in size, and shows a tendency to break up into chromosome-like masses. In the mode shown in Fig. 30 the nuclear membrane disappears, though the limits of the nuclear cavity are still recognizable. A globular mass is found in the centre, which consists probably of both nucleoplasm and nucleolar substance. This is the least common method, but a fair number of examples has been met with. Fig. 31 shows a far more frequent case ; here the membrane is still intact, but the nucleolus seems to be fragmenting. The considerations that incline me to the conclusion that the nucleolus breaks up directly into chromosomes are the following : (a) In many hundreds of cases examined I have never yet seen a spirem. (b) As soon as the chromosomes are formed the nucleolus com- pletely disappears ; neither nucleolar globule nor chromatin fibrillae remain. While the change in the nucleolus is going on the nuclear network preserves its usual appearance. Outside the nucleus there is nothing suggestive of either centrosome or polar radiation, while the fertilized egg- nucleus on the other hand has a distinct centrosphere (Fig. 31). The cytoplasm and chloroplasts frequently retain the faintly radiate structure shown in Fig. 19. The stage where the chromosomes have been differen- tiated, and the nucleolus has very nearly disappeared, is seen in Figs. 30 and 33. Not only is this stage distinguishable by the absence of nucleoli, but the chromosomes are only sixteen in number, and they are different from those of fertilized eggs in being thinner and less regular in form. Fig. 33 shows the only case observed where any fibres had appeared before the dissolution of the nuclear membrane. It is not unlike Fig. 34, but the very rudimentary appearance of the figure, and the fact that there are only sixteen chromosomes in the three sections of the figure, prove that it is merely an exceptionally regular case of parthenogenesis. It is a very striking fact that while in nearly all the normal mitoses described in these two papers the nuclear membrane persists until a very 1 98 Lloyd Williams . — Studies in the Dictyotaceae. late period of the anaphase stage, in the division figures of parthenogenetic nuclei the membrane is invariably absent. The disappearance must be very rapid., for in all the hundreds of cases examined I have never met with any where the dissolution was partial. Coincidently with the disappearance of the membrane the nuclear space, now sharply delimited by the crowded chloroplasts and coarser reticulation, is seen to be occupied with very fine meshed or finely fibrillar ‘ kinoplasm,’ in which long fibres can be distinguished. No two figures are alike in the arrange- ment of the threads. Some of the most regular are shown in Figs. 33-36. In many the longer threads cross each other in inextricable confusion, so that no two can be seen to converge. In such cases it almost seems an abuse of language to call them spindle-fibres. A point in which these figures present a sharp contrast to the artificially produced multipolar figures of animal eggs is the total absence of both centrosomes and radiations. The figure is entirely confined to the nuclear space, and no cytasters or any other radiate structures are seen outside. This is the more remarkable as centrospheres are such prominent features in normal mitoses. The subsequent history of the dividing nucleus is rather peculiar. In the dense ‘ kinoplasmic ’ mass a number of approximately spherical lighter areas appear, and in each area one or several curved chromosomes. Very soon after differentiation each area is delimited by a membrane, and the interior of the dense felt of kinoplasm is occupied by a cluster of nuclei, some large and some small, the size generally depending upon the number of the contained chromosomes (Figs. 37, 39). I have been unable as yet to decide whether as a preparation for this stage there is a division of chromosomes. I am inclined to think there is, but the supposed instances of it are not clear enough to enable one to come to a final decision. There is undoubtedly a subsequent increase in the number of nuclei, but not by karyokinesis. The number of eggs in a group varies : comparing the several sections of a germling together, I found it to contain twenty nuclei ; in others the number would not be more than five or six. After a time the several chromosomes in a nucleus become spherical in form, then fuse together to form a nucleolus ; at the same time a faint reticulum appears (Fig. 40), which very often contracts away from the membrane. It was stated above that no cytoplasmic radiations were ever associated with the multipolar spindles described in the preceding section. Fig. 41 represents a later stage in which radiations occurred, and where also one of the nuclei had a well-formed centrosome and the usual distortion of the membrane opposite to it. Whether this is an instance of the formation of these structures de novo , or whether we have here one of those not II. The Cytology of the Gametophyte Generation. 199 infrequent cases of polyspermy already referred to, could not be decided, as I have not yet seen the phenomenon in any of the control cultures. After a time the nuclei separate into two or more groups. As a general rule there is nothing more to be seen than a constriction of the kinoplasmic mass containing the nuclei, followed by the appearance of two groups instead of one ; these groups taking up the positions which in a normal segmentation would have been occupied by the two daughter- nuclei (Fig. 42). The plane of division of the cell becomes almost clear of chloroplasts, and after a time the dividing wall makes its appearance ; the early stage of this process is shown in Fig. 42. In a solitary instance nuclei were observed with connecting fibres between them (Fig. 38) ; these presented the additional peculiarity of being in the dispirem stage. Although it is very evident that the nuclei increase in number, the chromatin does not increase proportionately in amount. There is always a great difficulty in staining the nuclei, and in the later stages many of them are undistinguishable from cytoplasmic structures, excepting by their form. The multiplication of nuclei must be accomplished by direct division as, after the original multipolar figures, neither chromosomes, fibres, nor any other signs of mitosis are ever seen. Instead of first dividing into two cells each with its group of nuclei, the original cluster in rare cases separates into three or even four groups. The largest number of cells seen in a germling was six ; they were very- unequal in size, but every cell contained a cluster of nuclei. After a few divisions the germlings invariably died. It may be suggested that some of them may develop into mature plants in their natural habitats ; this is very unlikely, for the fertilized eggs in the same culture grow vigorously. General Considerations. A few general questions suggest themselves for consideration. At the present stage I do not intend to discuss them fully ; this will be done when the investigation of other Dictyotaceae has been completed. 1. The Nucleolus. That this structure in all the cases discussed con- tains the bulk of the chromatin is fairly obvious. In most cases there is a quantity of substance which is not employed in building up the chromosomes. How is it that in unfertilized eggs the nucleolus breaks up directly into chromosomes, and how is it that no residual nucleolar substance is ever left over after the differentiation of the chromosomes? Can it be that in the mature egg the nucleolus has lost some of its capacity for metabolism, and that this can only be restored by the advent of the antherozoid ? 2. The centrosome and radiations. It is clear from the foregoing description that, while the centrosome cannot be regarded as a mere condensation point at the focus of the system of radiations — its peculiar P 200 Lloyd Williams.- — Studies in the Dictyotaceae. curved rod-like form precludes that possibility-— it cannot, on the other hand, be regarded as a permanent cell-organ, for at several stages it disappears completely. At the same time, the comparison of normal with parthenogenetic segmentation strongly supports the idea that the egg after maturation is far less capable of giving rise to a new centrosphere than before. The unfertilized egg shows no indication of its presence, while in the fertilized ovum there is apparently only one, which subsequently divides to form the two centrosomes of the angular spindle. It is difficult to resist the conclusion that the antherozoid introduces into the egg something which enables it to form a centrosphere afresh. If this be correct, the imported substance can hardly be a centrosome, for no such structure can be recognized in the antherozoid before entry. With regard to the division of centrosomes, Mottier has described and figured the phenomenon in the early prophase of the second division of the tetrasporangium. I have never been able to satisfy myself that I have actually observed the splitting in the way that he figures it. In certain stages it is quite common to see the two chromosomes close to each other, and to find that during the later stages they travel farther from each other until the spindle, if already initiated, from being angular as at first, becomes quite straight or only very slightly curved. This occurs in the prophase of the first mitosis of the fertilized egg, in the newly formed daughter-cells of both tetraspore and oospore segmenta- tion, as well as in the case described by Mottier, but not in the newly formed nucleus of the tetrasporangium or in the first division of the tetraspore. The peculiarly irregular appearance of the membrane in the neighbour- hood of the centrosome in some stages tempts one to subscribe to the idea advanced by Mathews and others, that the centrosome is a liquefying enzyme, and that in virtue of this characteristic it may thus act not only upon the cytoplasm but upon the polar regions of the membrane also. This, however, is exceedingly doubtful, for not only have investigators failed to extract an enzyme from cells containing centrosomes, but in the parthenogenesis of eggs of Dictyota it has been shown that where the centrosome is absent the dissolution of the membrane is sudden and complete, and takes place at an early stage. 3. The nuclear membrane. Whatever may be the structure and consistency of the membrane the phenomena just described support the hypothesis that its formation is determined by the metabolic processes going on in the chromatic mass. Evidently the reason for the formation of a number of nuclei in the parthenogenesis here described is that the chromosomes are too widely scattered, so that when a chromosome is isolated it directly or indirectly initiates the formation of a membrane round itself. Juel, in his paper on Hemerocallis , describes a stray chromo- II. The Cytology of the Gametophyte Generation. 201 some surrounding itself with a membrane. It is a curious fact that in Dictyota a stage like Fig. 5 in the Tetraspore paper (PI. IX) is followed by one similar to Fig. 33 of the same series, or Figs. 6, 27 accompanying the present paper, where the chromosomes are less crowded than before. Whatever be the cause of the former condition, it probably facilitates the formation of a common membrane, and so prevents the separation of the chromatic elements. It has been shown that the thickness of the membrane varies con- siderably in different stages : we find it to be thinnest, for instance, during the synapsis stage of the tetrasporangium nucleus. This has probably some significance from the point of view of nutrition. The further fact that at certain periods, in some newly formed nuclei for example, the nuclear network contracts away from the membrane far more easily than at other stages, also suggests that the relation of the network to the membrane, and probably also to the cytoplasm, varies from time to time. 4. Segmentation in parthenogenetic germlings. Is the separation of the nuclei into two groups due to inherent repulsion between the nuclei of the separating groups, or is it due entirely to the action of the cytoplasm ? If the former, then there must have been some sort of division of the chromosomes, or, failing that, subsequent differentiation in the characters of the nuclei must have taken place. At present the matter is obscure, and must be left over for further investigation. Summary. 1. The sexual cells, unlike the tetraspores, are produced and liberated simultaneously in fortnightly crops. Fertilization is external. Eggs not fertilized within about half or three-quarters of an hour after liberation become invested with walls and germinate parthenogenetically. Freshly liberated oospheres strongly attract the antherozoids, are fertilized and segment in a normal manner. 2. The oogonium and antheridium are produced by the increased growth of surface cells, which, after cutting off a stalk-cell, form re- spectively a single egg, or over 1,500 antherozoids. 3. There is no division of the nucleus in the oogonium as there is in that of Fucns. All the divisions of the antheridium as well as the stalk-cell division of the oogonium are homotype, and are very similar to the stalk-cell division of the tetraspore, except for the fact that there are only sixteen chromosomes. 4. The antherozoid has the cilium lateral. There may be a second very much reduced cilium, but this is difficult to demonstrate. The nucleus is in the thicker end of the pear-shaped antherozoid ; the eye- spot is very small, and instead of being at the base of the cilium is generally near the anterior end of the ‘ beak.’ The antherozoids crowd 202 Lloyd Williams . — Studies in the Dictyotaceae . round the newly liberated eggs in preference to the others, probably in consequence of a chemotactic attraction. 5. When an egg has been fertilized the nucleus generally has within it a second smaller (male) nucleolus. Later on thirty-two chromosomes appear, but the two nucleoli are still present. There is at first a single centro- some which divides into two ; as the two separate the two spindle-cones also diverge, till ultimately they form a normal spindle. 6. When an egg has not been fertilized the nucleolus breaks up into chromosomes ; leaving no residual nucleolar matter to be extruded into the cytoplasm as occurs in other mitoses. The mitotic figure is very irregular and multipolar ; there is no nuclear membrane, and a cluster of nuclei is formed each containing sometimes one, sometimes several chromosomes. These separate into two or more groups, and partition walls are formed between them. The process may go on a little further, but very soon it stops and the germlings die. 7. When normal germination is compared with parthenogenesis we find that the entry of the antherozoid into an egg produces the following results : — (a) It causes a centrosome and radiations to appear in the cytoplasm. (h) Sixteen additional chromosomes are introduced into the nucleus. (c) The metabolism of the nucleus is far more active, as is shown by the greater amount of nucleolar matter and the increased size of the chromosomes in the prophase stage. (d) In the mitosis it introduces a directive influence which is completely absent from the parthenogenetic figure. This prevents the scattering of the chromosomes and the consequent formation of a large number of nuclei. (e) It prevents the early disappearance of the nuclear membrane. II. The Cytology of the Gametophyte Generation. 203 Literature. Bornet et Thuret, 1878 : Etudes phycologiques. Buller, 1902 : Quart. Jour. Micr. Sci. xlvi. Farmer and Williams, 1896 : Proc. Roy. Soc. lx. 1898 : Phil. Trans, xc. Galeotti, 1893 : Beitrage zur patholog. Anat. u. z. allg. Pathol, xiv, 2. Hertwig, O., 1893 : Die Zelle und die Gewebe. Hertwig, R., 1896 : Ueber die Entwicklung des unbefruchteten Seeigeleies, Leipzig. Johnson, 1891 : Lin. Soc. Journ. xxvii. Juel, 1897 : Jahrb. f. wiss. Bot. xxx. Loeb, 1902 : Archiv f. Ent wickelungsm echanik, Bd. xiii, Heft 4. „ „ ,, Bd. xiv, Heft 4. Mathews, 1901 : Amer. Jour, of Physiology, vi. Morgan, 1899 : Arch. f. Entw.-mech. viii, 3. Mottier, 1898 : Ber. d. D. Bot. Gesellsch. xvi. 1900 : Annals of Botany, xiv. Phillips, 1898 : Rep. Brit. Assoc., Bristol. Reinke, 1878 : Nova Acta Leop.-Carol. Bd. xl, No. 1. Thuret, 1855 : Ann. d. Sci. Nat., 4® ser., Bot., t. iii. Williams, 1897 : The Antherozoids of Dictyota and Taonia. Ann. of Bot. xi. 1898 : Reproduction in Dictyota. Rep. Brit. Assoc., Bristol, 1898. • 1903 : Alternation of Generations in the Dictyotaceae. New Phytologist, ii. 1904 : The Cytology of the Tetrasporangium, Annals of Botany, xviii. Wilson, 1901 : Arch. f. Entw.-mech. Bd. xii, Heft 4. „ ,, ,, Bd. xiii, Heft 1. explanation of figures in plates XII, XIII, AND XIV. Illustrating Mr. Lloyd Williams’s paper on the Dictyotaceae. For convenience of comparison the same scale of magnification has been preserved throughout, all the figures having been drawn with the camera lucida under the Zeiss apochromatic oil-immersion objective 3.0 mm., aperture 1-40, with ocular 8. The development of the Oogonium. Fig. 1. The rudiment of the oogonium before stalk-cell division. Fig. 2. Prophase. The spirem has segmented, and the nucleolus is becoming vacuolated. Fig. 3. A stage further; the chromosomes are longitudinally split, and the nucleolus is breaking up into fibrillae and the usual globule. Fig. 4. Stalk-cell division spindle. The cordate form is due to contraction, the nucleolar globule and fibrillae are present. Fig. 5. Polar view of the equatorial plate, showing the curved form of the chromosomes ; the number is sixteen. Fig. 6. Dispirem of the nucleolus of the oogonium, showing the increase in size of the chromo- somes before fusion. Fig. 7. Telophase stage of the nucleus of the oogonium, showing a number of deeply stained masses embedded in lighter staining substance. The development of the Antheridium. Fig. 8. Rudiment of the antheridium, the nucleus in the prophase of stalk-cell division. Fig. 9. Telophase of the stalk-cell division ; the stalk-cell nucleus has descended into the vacuolated basal region. 204 Lloyd Williams. — - Studies in the Dictyotaceae . Fig. io. Vertical section parallel to the long axis of the thallus. The antheridium has already divided transversely. Fig. ii. Vertical section transverse to the axis of the thallus. First division of the antheridium, polar view of the nuclear plate, showing the number of chromosomes. Fig. 12. Vertical section showing border -cells. Fig. 13. Vertical section transverse to the thallus. Division more advanced. The inner border-cell is seen to the left. Fig. 14. Vertical section through a mature antheridium ; the inner border-cell is still undivided. Fig. 15. Antherozoids fixed with potassium-iodide iodine in sea- water and stained with methylene blue. Nuclei and physodes shown. Fig. 16. Antherozoids fixed with dilute Flemming’s mixture and stained by the iron-alum- haematoxylin method. Fig. 17. Antherozoids fixed with dilute Hermann’s solution and stained with gentian violet. At the top left-hand corner is an apparently biciliate, and at the right-hand side of the figure a binucleate antherozoid. In Figs. 15-17 the small black dots are physodes. Fig. 18. Antherozoids drawn while motile. The small black dot is the red eye-spot, the colourless part is the nucleus ; some show the presence of physodes. Fertilization and segmentation of the Egg. Fig. 19. Section of newly liberated egg fixed with dilute Flemming. There is a zone of ‘ kino- plasm ’ round the nucleus, and a radiate arrangement of the chloroplasts. The physodes are at the surface of the egg. Of the numerous antherozoids in the neighbourhood three have been drawn. Fig. 20. Delayed fertilization. The antherozoid, greatly increased in size, is seen to the right of the nucleus, a few are shown outside the membrane. The egg-nucleus is in the prophase of parthenogenesis. Fig. 21. Fertilized egg-nucleus with male and female nucleoli and a single centrosphere. Fig. 22. Prophase of the first segmentation mitosis ; the chromosomes are thirty-two in number, the remainder being in two other sections. Fig. 23. A later stage ; the chromosomes are split and the nucleoli very irregular in form. Fig. 24. Early spindle showing two cones at an angle with each other. The faint radiations are better seen in the succeeding section. Fig. 25. The completed spindle. Fig. 26. Polar view of the nuclear plate, showing the thirty-two chromosomes. Fig. 27. The two daughter-nuclei in the dispirem stage ; the nucleoli beginning to differentiate. Fig. 28. The daughter-nuclei completed; each has two nucleoli as in the prophase of the preceding mitosis. Fig. 29. Abnormal fertilization ; a stage later than Fig. 20. Parthenogenesis of unfertilized Eggs. Figs. 30, 31. Two modes of initiating the parthenogenetic figure. In both the nucleolus seems to be directly converted into chromosomes. Fig. 32. An exceptionally regular early spindle. Here again the chromosomes seem to be directly converted into chromosomes, of which there are eleven in this section and five in the next. Figs. 33-36. Four examples of the parthenogenetic ‘spindle.’ It is very multipolar and exceedingly irregular, there is no trace of nuclear membrane or of nucleolar substance, and the chromosomes are invariably sixteen in number. Fig. 37. Vesicles (nuclei) containing one or several chromosomes beginning to differentiate in the kinoplasmic mass. Fig. 38. An exceptional figure showing the separation of two groups of nuclei with a few connecting fibres. Fig. 39. Newly formed nuclei with chromosomes. Fig. 40. A few nuclei out of a group of twenty ; nucleoli beginning to appear. Fig. 41. The only instance observed of a group of nuclei with centrosome and radiations. Fig. 42. Parthenogenetic germling in which the nuclei have separated into two groups; a septum is beginning to form between the two halves of the cell. i/bmals of Botany. Fig-11- Fig- 2. Fig. 9. Fig-1- J. Lloyd Williams del. LLOYD WILLIAMS.— DICTYOTA . Voi.xvm}pi.xiL Umvei sity Press, Oxford, t/bmals of Botany. \ ®, 1 A - 1 V* u' &] Fig-15 ) % (Sfa S'' j ^ Fig. 16. . - • . .- . - ■■ ;• #4 <^#L \ ■ " • ' i 1 $f L ■ . >. \ \, ’ $ ■: ' '■ Fig. 20. / & ,?*£ A $'T .’*&■** W ■ f 1 . v .t::V'. ¥ 1 •• ■ >^k V Fig-23. ® f ■ A->V# \ ^ ft* ■ . ;>. Fig. 24. # % ( X* f f /*' ) tx 1 —xJ ''-I. <•) <|» Fig. 21. Mfi P .v\ :/* V i -. A A .- . >.'• - SBfe, Iff # % Fig. 2 5 f-fy ff-N- x 1 4 ptt .- ' . -S' . »•... 'X .- I . ... > W* \k$$gMiA,t Ax* ■ ' t*v #«^ i k ■ X • ,' '-:::; ' " - •' ’ 'V Fr< j0> Fig. 2 9. J. Lloyd. Willi eims del. Fig. 30 LLOYD WILLIAMS.— DICTYOTA. Voixvinpim. Fig. 18. Fig.22. Fig. 2 6. Fig.19. Fig. 2 7. Fig. 31. Fig.32. University Press, Oxford. ■Annals of Botany: if % Fig-15 § \ ® Fig. 16. . Afy*\ * /-V J» *V . UMm i' : fVw! Fig. 20, ,4# i&s*i i'.T - n . ; t » % Fiji.23. ?M>|| >v M1* • '■ ;*> /j» ■'% Fig. 2.4. 'aVV o 4 ts -c %> & Fig. 17. »■& J Gm s. wm. « ■ ■'■aa e\ ^7: A« c'; ; ■ ; i S e ■;« Fig. 21 Fig. 2 5 ■A§0 * - rf t , * ' i ■f ' i Fig. 2 8. I « par ?-, % . /* Fig. 3 3 . J^T University Press, Oxford. LLOYD WILLIAMS.— DICTYOTA. Jbuicds of Botany. Vol.Xm,Pl.XIV. Fi£-37- Fig. 38. Fig. 3 6. Fig. 3 9 . Fig. 41. J.Lloyd Williams del. Fig 42. Univer sity Pr e ss , 0 xf o r d . LLOYD WILLIAMS. -DICTYOTA. Ophioglossum simplex, Ridley. BY F. O. BOWER, F.R.S., Regius Professor of Botany in the University of Glasgow. With Plate XV. IN May, 1900, I received from Professor Groom, to whom my grateful thanks are due, a specimen of a plant which had been collected in Sumatra, in 1897, by Mr. Ridley ; his description of it is as follows: — ‘A New Species of Ophioglossum. ‘The very remarkable little species of Ophioglossum I am about to describe was found by me in a dense wet forest on the banks of the Kelantan River, near Siak, in Sumatra, in December, 1897. I was only able to find three plants, for it is very inconspicuous, and as I was still far from camp, and it was very late when I perceived the plants, there was not sufficient time to search thoroughly the locality, which is one by no means easy of access. I would describe it as follows : — ‘ Ophioglossum simplex, n. sp.1 Terrestrial, rhizome short, and tuberous, with few roots. Fertile fronds solitary, or two together, slender flattened, with a blunt apex, 4 to 6 inches long, |th inch wide, dark green, sterile division represented by a very small lateral process, or quite absent. Fertile portion about an inch long. ‘ Hab. — Dense wet forests on the Kelantan River, Siak, Eastern Sumatra. ‘ The affinity of this curious plant is with O. Bergiana , Schlecht, of South Africa, and with O. pendulum , L., an epiphytic plant common in Eastern Asia. The almost complete suppression of any trace of a sterile portion of the frond, and the consequent reduction of the plant to the very simplest elements, is the most peculiar feature in this species. Its habitat is very peculiar for the genus, for though there are two indigenous species of terrestrial Ophioglossum , viz. O. nudicaule and O. reticulatum , to be found 1 This name has already been used by Rumphius in 1750 (Herb. Amb. vi, p. 152, Tab. 68, Fig. 1), but the plant so described falls under 0. pedunculosum (Prantl, Beitrage z. Syst. d. Ophio- glosseen, p. 329). [Annals of Botany, Vol. XVIII. No. LXX. April, 1904.] 206 Bower. — Ophioglossum simplex , Ridley. here, besides the epiphytic O . pendulum, L., none of these are forest plants, occurring only in open grassy spots ; while even the epiphytic one inhabits open woods, or grows on trees where there is plenty of light.’ Professor Groom examined the plant himself, and found its characters, so far as observed, to be those of Ophioglossum ; he then kindly forwarded it to me for further analysis. The description now given is based on only the one specimen ; two others were collected by Mr. Ridley, but they are said to be incomplete, and are at Singapore. The whole description must therefore be considered as provisional, and open to amendment when fresh material is available. The specimen sent by Mr. Ridley was figured by Professor Groom in its unaltered state ; his drawing is shown in Fig. i. The plant consists of a short stock, bearing three appendages of different ages. The largest is almost mature, and shows a structure like a fertile spike of Ophioglossum , with the lateral rows of sporangia almost mature. The second is similar in outline, but the sporangia are so young as to be barely perceptible ; still there can be no doubt that it is also of the nature of a fertile spike. The third is a small conical body of oblique position, and forced out of shape by the pressing. A careful external examination of the two largest appendages discloses no part which could be compared with the sterile lobe, or lamina of a subtending sporophyll, of other Ophioglossaceae. The leaf-stalks were examined throughout their length for scars or other traces of the insertion of a sterile lobe, but none was found. It is true there were scars of a reddish and rough appearance at various points on the stalks, but they are very irregular in form and position, and are not to be taken as scars of insertion of a withered sterile lobe, notwithstanding that one of them is near the base of the oldest leaf. This is clearly seen in Fig. 2, which shows the base of the plant on a larger scale ; it would seem probable from the small size of the stock, its form, and from the entire absence of roots (with the exception of the small conical body which will be shown below to be a root) that the under- ground parts had been partially broken away in removing the plant from the soil. This was shown in the sections subsequently cut from the stock ; the insertions of roots which had been broken away while still actively functional were found. After the specimen had been soaked out in water, the stalks became sufficiently transparent for the vascular strands to be visible ; and it is clear that the course of the strands does not countenance the idea that the large scar above noted is of the nature of the insertion of a sterile lamina with a vascular supply. Moreover, scars of similar apparent texture, but of different outline, are found most irregularly scattered on the leaf-stalks, and are probably due to some pathological state. The fact that the scars do not correspond in position on the several leaves shows also that they are not constant morphological features. Thus 207 Bower . — Ophioglossum simplex , Ridley. the external observation of the mature parts of the specimen affords no evidence of any sterile lobe or lamina, as in the known Ophioglossaceae. Mr. Ridley speaks of the ‘ almost complete suppression of any trace of a sterile portion of the frond 5 ; I find in the specimen sent to me no need for the qualifying word c almost.’ The small appendage at the base of the plant was removed with care, soaked out, and examined. It appears to have been compressed in drying, and after soaking the form is not recovered. As far as form is concerned there is nothing distinctive, while its oblique position on the specimen as received would allow of its being either a root or a young leaf displaced in the pressing. To decide the point, the whole appendage was removed and embedded ; sections then showed that it is a root, and the following structural points were observed. The stele appears to be of the usual type ; it is diarch, and the xylems may remain separated by parenchyma-cells, which occupy the centre. The phloem forms an arc on either side, while even in these sections, which only partially recover from the pressing and drying, an endodermis can sometimes be traced. Outside this is a broad zone of cortical parenchyma, rather thin-walled ; this merges into a peripheral band which contains the c gru- mous ’ masses characteristic of mycorhiza : notwithstanding the only partial recovery of the section from drying, there is no room for doubt that the root has been mycorhizal, evidences of the presence of the fungus being seen in some four or five layers of the outer cortex. The periphery of the root is occupied by a layer of cells with their outer walls much thickened ; the thickening sometimes extends to the inner walls as well : this layer appears to be of the nature of an exodermis, for at some points remains of an outer layer are still to be seen. Comparison with O. pendulum shows that the details thus described correspond in the main to those there seen : the stele may, it is true, have more than two protoxylem groups in O. pendulum , but that character is known to be variable in that species. The characters of the cortex are very closely matched, including the mycorhizal band and the layer (probably exodermis) with the thickened outer wall. These details serve to strengthen the comparison of our plant with O. pendulum 1. Since the external form of the mature appendages gives no indication of the existence of a sterile lamina, it will be well at this point to consider what views are open to us as to their nature in this remarkable plant. Two alternatives are possible, (i) that the appendages are spikes pure and simple, without any structural evidence of the subtending lamina 1 Compare Atkinson, Bull. Iowa Bot. Club, xx, 1893, p. 356, on Symbiosis in the roots of the Ophioglossaceae: also Janse, Ann. Jard. Bot. de Buitenzorg, xiv, p. 65, PI. IX, Figs. 11, 12, I have not however in either species noted the infection as local ; it appeared to me to extend all round the cortex. 208 Bower . — Ophioglossum simplex, Ridley. having figured in their past history, or (e) that they are leaves of the ordinary Ophioglossaceous type, in which the sterile lamina has become entirely abortive ; in the latter case the whole appendage would be of a composite nature, the upper part being the spike, the lower part being of the nature of a leaf-stalk. A decision can best be approached on an anatomical basis, for there is between the sterile leaf and the fertile spike the general difference of orientation of the vascular strands : the xylem in the former being on the adaxial side, in the latter it is inverted, and facing the abaxial side. The following notes on the distribution of the vascular strands in the leaves of certain species of Ophioglossum may be of use, as a basis for comparison. In O. Bergianum , Schlecht, the leaf-trace originates, as in all the Ophioglossaceae hitherto examined on that point, as a single strand ; in most species this single strand branches early, but in O. Bergianum it remains at first unbranched. The vascular supply for the fertile spike comes off as two lateral bundles from the margins of the leaf-trace ; these fuse together to form the single bundle of the base of the fertile spike. This single strand may branch as it passes upwards, so that the transverse section of the fertile spike may show two, three, or even four strands, but they are always orientated with the xylem directed to the abaxial side. In the upper region there is only one strand, which passes through the rows of sporangia to the extreme tip1. The branchings in the sterile leaf are few and irregular, but the number of strands commonly seen is three, with the xylems on the adaxial side. This is the simplest species of the genus structurally, but the leading features are the same in the more complex. A slightly increased complexity is seen in O. lusitanicum , where the vascular supply of the sterile leaf branches into three, and the supply from the spike comes off from the lateral strands (Prantl, loc. cit. PI. VII, Fig. 1). In O. vulgatum the origin of the vascular supply of the spike is here again as two lateral strands, one from either marginal bundle of the sterile leaf 2. These strands branch again as they pass upwards, and five strands usually appear in the transverse section, an arrangement which may be very nearly matched by some sections near the base of the leaf-stalk ; but the distinction is always easily drawn by their inverted position : the xylem of the fertile spike being on the abaxial side. In O . reticulatum , L., the arrangement is similar to that in O. vulgatum , as regards the fertile spike, and it is probably the type general for the ordinary ground-growing species. The similarity of O. simplex to O. pendulum in certain external points, 1 See Prantl, Beitr. z. Syst. d. Ophioglosseen, Jahr. d. k. bot. Gart. Berlin, iii, p. 297. Also Bower, Studies in the Morphology of Spore-producing Members. II. Ophioglossaceae, p. 68. * See Prantl, 1. c., Taf. vii, Fig. 2. 209 Bower . — Ophioglossum simplex , Ridley. and in the structure of the root, suggested a fresh examination of the vascular supply to the leaf and spike in that species. Prantl 1 in his diagnosis of the three sections of the genus gives for \Euophioglossum , ‘ petioli fasciculi basi tres * : this section includes the bulk of the species of the genus. For \Ophioderma (including O. pendulum ), and § Cheiroglossa (including O. palmatum), ‘petioli fasciculi numerosi.’ It is true that there are three bundles at the base of the leaf-stalk in E uoph ioglossu m , but on entering the axis they fuse to one, before insertion on the main system 2. Hitherto this insertion as a solitary strand has been found uniform in the Ophioglossaceae examined : but in O. pendidum it is not so, as the following description of a definite example will show ; but it is possible that in a species which varies so greatly in size, the vascular complexity may vary also. The leaves in this species differ in vascular supply accord- ing as they bear spikes or are sterile. In the latter case the transverse section of the leaf-base in the specimen examined shows an open arc of some seven bundles (Fig. 3), which are reduced by fusion to five ; these remain as distinct strands, till they insert themselves individually upon the system of the axis (Figs. 6 to 13, leaf to the left). In the case of leaves which bear spikes, the transverse section of the leaf-base shows a complete ring of bundles with their xylems facing inwards (Fig. 4) ; fusions on the adaxial side show that the distinction of the two margins near the base is not always maintained. As the base of the leaf is approached this foliar ring opens on the adaxial side — with or without some previous fusions on that side (compare Figs. 6 to 13). The strands which are reduced in number by further irregular fusions are then inserted individually upon the system of the axis. This, which is at times in form of an almost complete ring (Figs. 6, 7, 8), opens to receive the foliar strands upon the margin of the gap. This vascular system is in itself not very uniform in detail, and differs from that of other Ophioglossaceae in the leaf-trace not being united at the base to a single strand. In the sterile leaf the strands pass upwards from the system of the axis as a single curved series (Figs. 2-17, leaf a ), showing occasional reticu- lations : the series is open on the adaxial face. As the strands enter the flattened region of the lamina the curve is flattened out also, with the xylems directed upwards. In the leaves which bear spikes the strands on passing from the system of the axis form a circular series, closed on the adaxial face (Fig. 4) : irregular branchings and fusions are often seen between the bundles at the opposite margins of the curve, so that those strands which occupy the adaxial side are connected indifferently with both 1 Beitr. z. Syst. d. Ophioglosseen, p. 299. 3 Holle, Vegetationsorgane der Ophioglosseen, Bot. Zeit., 1875, p. 269 ; Rostowzew, Recherches sur V Ophioglossum vulgatum , p. 21, PI. I, Fig. 5. 2 lO Bower . — Ophioglossuni simplex , Ridley. margins. Proceeding up the leaf the ring flattens, and as the margins of the lamina become defined strands pass right and left from the ring ; the circular series thus becomes broken up into the supply for the sterile lamina on the one hand, and the supply for the spike on the other. The latter consists of five or more strands, with their xylems directed abaxially, while the strands of the former are more numerous, and have their xylems directed adaxially (Fig. 5). Occasional connexions are found higher up, between the strands of the two systems, after their separation. The vascular supply of the spike may thus be held to be mainly, though not always exclusively, a product of marginal branching from the original vascular supply of the leaf-base, as in Euophioglossum. Prantl’s section Cheiroglossa , including only O. palmatum , is described as having ‘ petioli fasciculi numerosi.’ The following observations on their relations to the spikes were made on a specimen sent to me by Mr. Fawcett from Jamaica. Transverse sections about the middle of the stalk show the vascular strands arranged, as in O. pendulum , in a ring, with their xylems directed centrally : they number about fifteen, and are of very variable size. Material was wanting for tracing the system downwards into the axis. Following it upwards no marked change took place at first, there being no obvious distinction of those strands which will enter the spikes till immediately below their insertion. Where the spikes are large, as in my leaf from Jamaica, and apparently also in that investigated by Professor Bertrand 1) a number of strands enter each spike. Immediately below the insertion of the lowest spike, though even the outline of the section shows where the stalk of the spike will be inserted, the ring of strands in the leaf-stalk remained undisturbed (Fig. 14) ; further up the ring opened, and with sundry branchings, four strands — subsequently reduced by fusion to three — passed out into the spike. Here as in other Ophioglossaceae the vascular supply appears to originate from both margins of the parent leaf, and not from one margin only (Figs. 14, 15, 16). The ring of strands in the leaf-stalk having thus opened, it did not again close (Fig. 17), but flattened out with the flattening expansion of the lamina into a wide arc ; and in the leaf in question the vascular supply for the higher spikes came off from the margins of the arc, as shown in Figs. 18 to 23 (compare Bertrand’s Fig. 97, 1. c.). In those specimens of O. palmatum where the spikes are numerous and small, their vascular supply appears to be only a single strand ; this origin- ates as a branch from one of the vascular strands, which may subsequently take a course distinctly intra-marginal in the lamina 2. The characteristics of the vascular arrangements in the appendages 1 Travaux et Memoires de l’Univ. de Lille, Tome x, p. 189, Fig. 97. 2 Compare Studies in the Morphology of Spore-producing Members. II. Ophioglossaceae, PI. VIII, Figs. 120, 121 ; PI. IX, Figs. 126, 127. Bower. — Ophioglossum simplex , Ridley. 21 1 of the species of Ophioglossum investigated may then be summed up thus : — (1) The xylem in the strands of the leaf -stalk at first faces directly or obliquely to the adaxial surface : in sterile leaves they constitute a more or less extended arc, open on the adaxial side ; but in the fertile leaves 1, and especially clearly in O. pendulum and palmatum , as the strands pass upwards from the base the arc closes in, and the strands together constitute a ring , the margins of the arc being indistinguishable, and they may be related to one another by fusions. In the lamina the ring again opens out into a flattened arc, the opening taking place at the insertion of the spike, or of the lowest spike where there are several. (2) In the spike the strands are always arranged with their xylems directed abaxially , in a flattened arc , never in a closed circle. Thus diagnostic characters, though of a somewhat imperfect sort, exist, marking off the spike from the leaf-stalk. It remains to attempt the appli- cation of this diagnosis in the case of O. simplex , with a view to deciding the question of the morphological nature of the appendages which it bears. The following considerations as regards probable affinity may help towards the concentration of the problem down to a definite issue. Mr. Ridley suggests that the affinity of our plant is with O. Bergianum and with O. pen- dulum : the former affinity I should regard as doubtful, on the grounds of form as well as of anatomical character ; the latter affinity seems for similar reasons more natural. But the nearest similarity in form is with the ground- growing plant designated O. intermedium, Hook., found by Lobb, near Sarawak, Borneo 2. This plant was doubtfully included by Prantl in O. pendulum 3. I think that these three plants constitute a natural group, or section of the genus, which may be designated with Prantl, § Ophioderma, and be held to consist of three species, viz. O. pendulum , L., O. intermedium , Hook. 4, and O. simplex, Ridley. If the probable affinity of our plant be, as suggested, with O. pendulum 9 the anatomical issue assumes some degree of definiteness ; for we know that in O. pendulum the vascular supply in the spike is in the form of a flattened arc, with the xylems directed abaxially, while that of the leaf- stalk shows the strands arranged in a complete ring, with their xylems directed centrally. The question may therefore be put thus : if we find in O. simplex that the strands are in a flattened arc and the xylems 1 Compare Rostowzew, 1. c., Text fig. 2, p. 7. 2 Hooker, Century of Ferns, Tab. xcv. 3 Prantl, 1. c., pp. 331, 332. 4 I see no sufficient reason for sinking this species, as Prantl does doubtfully. The discovery of 0. simplex seems to me an additional reason for its retention as a valid species. I have compared the type specimen of 0. intermedium , at Kew : the habit of the lower part of the plant is very similar to that of 0. simplex ; the difference chiefly lies in the presence of the small sterile lamina, and winged stalk below its insertion in 0. intermedium, and in a narrower tract of tissue between the rows of sporangia of the spike than in 0. simplex , but the latter character is variable in 0. pendulum. 212 Bower . — Ophioglossum simplex , Ridley . abaxial, the part in question will be of the nature of a spike ; if we find that the strands are in a complete ring with the xylems central, the part where that occurs will probably be of the nature of a leaf-stalk. Further, if the leaf-trace continues downwards as separate strands, till the insertion of those strands separately upon the vascular system of the axis, then not only will the affinity with O. pendulum be confirmed, as apart from the rest of the genus, but also there will be a strong presumption that the appen- dages of O. simplex are the true correlatives of those of O. pendulum , not- withstanding the absence of the sterile lamina, as above noted for our plant. It may be said at once that the question of orientation is difficult in the upper region of any elongated succulent part, having approximately cylin- drical form ; and especially will this be the case when, as here, the only specimen has been pressed and dried. Certain results as to orientation can, in the present case, be obtained only at the base, and especially at the point of insertion of the appendage upon the axis. With this caution the facts available may be stated as follows. A transverse section of the stalk at point A, Fig. i, shows the vascular strands five in number, forming an arc open on the flattened side, which will presumably correspond to the abaxial side in a normal Ophioglossum (Fig. 24) ; it is however impossible in this dried specimen to be certain that its direction is actually abaxial. The xylems are directed towards this flattened side, and the structural arrangements are such as to indicate that here we have a part corresponding to that of a normal spike — a conclusion which the external form with its rows of sporangia amply bears out. It seemed useless in the absence of precise knowledge of orientation to pursue the details continuously downwards throughout the stalk, so the next section was taken at level B , in Fig. 1. Here the transverse section was more rounded, and the number of strands was found to be as high as seven, arranged in a regular circle, with their xylems directed centrally (Fig. 25) ; this result coincides fairly with the observations on the stalks soaked out (Fig. 2) : the number seven is slightly in excess of the number of strands there shown — possibly in the crushed stalk one of the strands may have overlain another. Comparing this result with the transverse sections near the base of the stalks of fertile leaves of O . pendulum (Fig. 4) and of O. palmatum (Fig. 14), it appears that the form of the section and the vascular arrangement are similar, though of a rather simpler type, as shown by the smaller number of the strands ; and thus the structural evidence points to the lower part being of the nature of a leaf-stalk. It remains to trace these strands downwards to their insertion on the system of the axis. This is shown in the successive sections (Figs. 27, 28, 29) for O. simplex , and it is clear from these that though there may have been reduction in number by fusions, the strands do not unite as in most Ophioglossaceae into a single strand, but remain as some four or five Bower. — Ophioglossum simplex , Ridley. 213 separate strands ; these insert themselves individually upon the rather irregular ring of bundles of the stock, which opens by a lateral gap to receive them. A comparison with O. pendulum (Figs. 6-13) shows that the arrangement in O. simplex is virtually the same, though somewhat smaller and simpler — a fact which still further accentuates the affinity with that species. We are now in a position to discuss the morphology of this curious plant, and the bearing which its existence may have upon the general theory of the Ophioglossaceous form. It seems clear that the lamina, commonly present, is practically absent here : not even any vestigial trace of it was seen, though if young specimens were available it is possible that such might be found. So far as observation has been possible the appendi- cular organ appears simple, and to be terminated by a normal fertile spike. Either of two possible views may be based on these facts : (1) that the appendages are simple spikes, which have not, and never had, a subtending sporophyll ; (2) that they are leaves of the ordinary Ophioglossaceous type, in which the sterile lamina is entirely abortive. If the former be their real nature, then we see in O. simplex some support for the view put forward some years ago by Campbell 1. He com- pared the Ophioglossaceous spike with the sporogonium of Anthoceros , and remarked : ‘ If we could imagine such a sporogonium to develop a root fastening it to the ground, and thus rendering it entirely independent of the oophyte, we should have the simplest possible form of a Pteridophyte.’ The plant of O. simplex seems to consist of little more than two such ‘ sporogonia,’ with root, or roots ; in fact it would almost realize Campbell’s forecast, and suggest more strongly than before that the Ophioglossaceae are, as he held, the most primitive Pteridophytes. If it were thus primitive it would point to the spike being the prior existent part, and the lamina a mere subsequent appendage upon it. But in Celakovsky’s view, which I have found reason to support elsewhere2, the Ophioglossaceae are regarded as derivative forms from a Lycopodinous type of construction, in which a constant relation of the spore-bearing organ to the lamina existed throughout the descent. With such a type the condition of O. simplex could only be brought into con- formity on a theory of abortion of the lamina, which subtends the spike in the usual Ophioglossaceous type: this is the alternative theory above suggested. Against this theory there can be no a priori objection, for in the Ophioglossaceae we see various degrees of abortion of the fertile spike, which lead to the condition of its complete absence ; and there seems no reason to hold that what may happen to the spike may not equally 1 On the affinities of the Filicineae, Bot. Gaz. xv, No. i, Jan. 1890 ; also Mosses and Ferns, pp. 296, 297. 2 Studies in Morphology of Spore-producing Members. II. Ophioglossaceae. Dulau and Co., 1896. 214 Bower. — Ophioglossum simplex , Ridley . happen to the subtending leaf. As supporting this theory there is the fact of mycorhiza in the root. We have at present no means of measuring from mere structural evidence the nutritive powers of mycorhizal roots, or how far in any given case they may supplement or supersede the chlorophyll-nutrition. In proportion as the mycorhiza is more effective a reduction of the assimilatory system may be anticipated, such as our theory demands, while the spore-producing parts would retain their dimen- sions, provided that the efficiency of nutrition be not diminished 1. A second point is the habitat, which Mr. Ridley specially describes as £ dense wet forests,’ and pointedly compares it with the c open grassy spots ’ where other ground-growing Ophioglossums are found. This seems to throw the onus of nutrition upon the mycorhizal roots. Applying these con- siderations to our present case, it is physiologically possible to contemplate a reduction, or even a complete abortion, of the sterile lamina to such a condition as that shown in O. simplex. And here the anatomical evidence detailed above will come in. It has been above pointed out that the upper part of the appendage in O. simplex shows structure as well as form characteristic of the Ophioglossaceous spike; also that towards the base the structure is comparable with that found at the leaf-base in O. pendirium , while the insertion of the vascular supply upon that of the axis is characteristic of the leaf of that same species : thus the structural evidence falls in with a theory of abortion of the sterile lamina. This, however, presumes that there is a transition from spike to leaf-stalk as the length of this apparently undifferentiated stalk is traversed. Any theoretical difficulty which such a presumption may occasion will be relieved by comparison of the case of the stamina 1 flowers in the genus Euphorbia : there the transition from floral axis to filament (a transition which is not less essential than that from spike to leaf-stalk) is only marked in the slightest way by the well-known articulation. In both cases, on the view above put forward, the simple condition has been arrived at by a process of abortion. Of the two explanations of O. simplex thus put forward, I am disposed to prefer the second, but in all the circumstances of this peculiar case it is not possible to come to any certain conclusion, one way or the other. This can only be done when more material shall be available. Meanwhile the case cannot be held to invalidate the view of Celakovsky, so far as to dictate its rejection. The systematic position of O. simplex will be, according to the charac- ters above described, in Prantl’s § Ophioderma. But while accepting Prantl’s division of the genus (1. c., p. 299), it would I think be well to add to the diagnosis the fact that the insertion of the leaf-trace differs in § Ophioderma from that in § Eu ophioglossum. In all species of the latter 1 Phil. Trans. T903, vol. cxcvi, pp. 228 and 233-4. 215 Bower . — Ophioglossum simplex , Ridley . which have been examined anatomically the leaf-trace unites at the base into a single strand before insertion on the system of the axis — a condition which would appear to be the more primitive, especially if the general view of the family be that they represent an ascending series from a small-leaved, polyphyllous ancestry1. But in § Ophioderma, of which O.pendidum and O. simplex have been examined (it was impossible to investigate the unique specimen of O. intermedium anatomically), the leaf-trace does not unite into a single strand at the base, but the individual strands are separately inserted on the system of the axis : this would appear to be the derivative con- dition, on any phyletic theory of the family as an ascending series of leaf complexity. I should propose, therefore, to add to Prantl’s diagnoses as follows: — for § Euophioglossum, ‘ petioli fasciculi basi tres, deinde in unum conjuncti, in rhizomae fasciculos insertum/ and for § Ophioderma ‘ petioli fasciculi numerosi, separatim in rhizomae fasciculos inserti.’ The case is still open for § Cheiroglossa , in which I am not aware that the stock has yet been examined anatomically. The anatomical difference thus brought forward, though not one which can be readily applied in ordinary syste- matic work, is more distinctive and trustworthy than the number of bundles at the base of the petiole, and for this reason it is to be preferred, if such a character is to figure at all in the diagnosis of the sections of the genus. The provisional conclusion which may be drawn from the study of this new species, together with the two others which are grouped with it, in the § Ophioderma , is this : that they form a natural group, anatomically dis- tinct, which illustrates three phases of proportion of the spike to the subtending leaf-lamina : in O. pendulum the sterile lamina is large, and sometimes irregularly branched ; in O. intermedium it is small and simple, while the spike is still of considerable dimensions ; in O. simplex it is absent, at least in the mature state, while the spike is still large. These three species may illustrate either a descending or an ascending series ; the more probable view seems to be that they illustrate a decrease of the sterile leaf, and the extreme condition of O. simplex is to be attributed to the presence of mycorhiza, which makes nutrition of the large spike still possible in the dense, wet forest in which it grows, notwithstanding that the usual assimilating organ is functionally non-existent. Reduction is, how- ever, not apparent in the spike itself, for, provided nutrition be kept up from whatever source, it would still maintain its character, being essentially a spore-producing, and not a nutritive member. The specimen — at least what remains of it after the anatomical investi- gation above described — together with the original drawing by Professor Groom, will be deposited, according to the wish of Professor Groom, in the botanical department of the Natural History Museum, South Kensington. 1 Compare Studies in Morphology of spore-producing members, Phil. Trans. 1903, B. vol. cxcvi, pp. 233-7. Q 216 Bower . — Ophioglossum simplex, Ridley . DESCRIPTION OF THE FIGURES IN PLATE XV. Illustrating Professor Bower’s paper on Ophioglossum simplex. Fig. I. The only specimen available of 0. simplex , from a drawing by Professor Groom, showing the two spike-like appendages, the short stock below, and small root. A. A. and B. B. indicate the levels at which sections were taken. Natural size. Fig. 2. Base of the specimen on a larger scale after soaking out in water: the course of the vascular bundles in the appendages has been traced, x 4. Figs. 3-5. Transverse sections of leaf-stalks of 0. pendulum : Fig. 3, from the base of a sterile leaf; Fig. 4, from the base of a fertile leaf ; Fig. 5, higher up on a fertile leaf, x 8. Figs. 6-13. Successive transverse sections through the stock of 0. pendulum , illustrating the mode of insertion of the leaf-trace upon the system of the axis, x 4. Whether the leaf be a sterile one (as that on the left side of each of these figures) or a fertile one (as were the other two), the strands do not coalesce to a single strand before insertion, but remain distinct till they join those of the axis. Figs. 14-16. Transverse sections of the leaf of 0. palmatum , showing the origin of the vascular supply to the lowest of its four spikes, x 4. Fig. 17. Transverse section of the stalk of that spike. Figs. 18-23. Successive sections higher up on the same leaf, showing the origin of the vascular supply into the second and third spikes, x 4. Fig. 24. Transverse section of the larger appendage of 0. simplex , at the level A. A. shown in Fig. 1. x 8. Fig. 25. Transverse section at the level B. B. x 8. Fig. 26. Transverse section just above the insertion on the axis, x 8. Figs. 27-29. Successive transverse sections of the stock of 0. simplex showing the insertion of the foliar strands upon the system of the axis. It will be seen that the foliar strands do not unite into one before insertion on those of the axis, x 8. The Extra-floral Nectaries of Hevea brasiliensis, Miill.- Arg. (the Para Rubber Tree), an Example of Bud-Scales 1 serving as Nectaries. BY JOHN PARKIN, M.A., Trinity College, Cambridge. With Plate XVI. WHILE engaged in economic work on india-rubber in Ceylon during 1898-9, the Para Rubber Tree (Hevea brasiliensis ) 2 was constantly under my observation, and peculiar nectaries occupying the position of bud- scales on its young shoots attracted my attention. Less conspicuous nectaries also occur on the foliage leaves proper. Though these latter are incidentally mentioned by systematists, the bud-scales generally, as well as their nectariferous nature, appear to have escaped their notice. This is not surprising, for the adult tree only puts forth fresh foliage annually, and the bud-scales being caducous, are merely evident while the shoots are in the immature condition ; thus, unless the tree be ex- amined during the short period of leaf-renewal, no bud-scales would be seen. In the account of the genus Hevea in Martius’ Flora of Brazil 3, the nectaries of the foliage leaves are mentioned, but no reference is made to the bud-scales. Delpino 4 in his elaborate work on extra-floral nectaries dismisses this genus in a few words, referring apparently only to the nectaries of the foliage leaves. In a recent paper by Huber 5 on the periodicity in growth of Hevea 1 The term £ bud-scale ’ is here used in the sense of a reduced leaf-structure situated on the shoot below the true foliage leaves. The author does not necessarily wish to imply that such structures in Hevea serve or have ever served as protective coverings to the bud. 2 Introduced into Ceylon in 1876. 3 Martius, Flora brasiliensis, vol. xi, pars. II, 1873. On p. 298 the genus (nine species including H. brasiliensis) is described, as having ‘ petioli communes apice supra glanduligeri.’ 4 Delpino, Mem. Accad. Bologna, viii, 1887, p. 635. In giving examples of extra-floral nectaries in the Euphorbiaceae he refers to Hevea as follows : 1 8 specie di Hevea (petioli ima basi patell ari-glanduligeri). ’ 5 Huber, Bot. Centralb. lxxvi, 1898, pp. 259-64. [Annals of Botany, Vol. XVIII. No. LXX. April, 1904.] Q % 218 Parkin . — Extra-floral Nectaries of Hevea brasiliensis brasiliensis the bud-scales are just mentioned, but their nectariferous nature is not pointed out. The description to follow is the result partly of observations made while resident at the Royal Botanic Gardens, Peradeniya, Ceylon, and partly of the examination of some spirit-preserved young shoots from adult trees brought back to England. The following account does not claim to be at all exhaustive. The object of this paper is chiefly to bring to notice a somewhat peculiar type of extra-floral nectary. Morphology of the shoot . The adult trees at Peradeniya shed their leaves early in the year and remain bare for some days before the new foliage appears. On February 16, 1899, the new shoots had almost gained their full length, but their foliage leaves were still very immature. At this stage of growth the bud-scales are fully developed and their nectaries active. On the leaves attaining maturity these structures shrivel and drop off. According to Huber 1 the adult trees in their natural habitat, the Amazon valley, produce likewise only one crop of leaves in the year, but the time they are bare is about June ; hence the trees in Ceylon appear to have changed the time of the annual renewal of their foliage. This may be due to climate. The early months of the year constitute a moderately marked dry season in that part of Ceylon where Peradeniya is situated, and dryness is considered to have a direct bearing on leaf-fall. Yet on this idea the Hevea trees at Peradeniya ought not to burst into fresh leaf till about April, when the rains of the little monsoon commence : as it is they renew their foliage about the driest time of the year, while they cast off the old in January, a wetter and cooler month than either February or March. Though mature trees produce only one set of leaves during the year, young trees — saplings — put forth several, showing a periodicity, which has been described by Huber2. Such saplings may produce fresh shoots about every month. My attention was first called to the nectariferous bud-scales by noticing one day insects busy on the young shoots of some saplings growing in a plot. At a short distance away they looked as if they were devouring the immature foliage, leaving behind the stumps of the petioles, but on closer inspection I saw that they were a hairy kind of ant (?) imbibing the honey secreted by special foliar organs situated on the lower part of the shoot. Owing to the internodes between these structures having lengthened considerably, the general impression conveyed a little distance away was that of short petioles with the foliaceous part nibbled off. The true foliage leaves, however, were quite intact on the upper part of the shoot with their laminas as yet feebly developed. My first observations were made on these saplings, as I had to wait till the 1 Huber, loc. cit. 2 Huber, loc. cit. ( Para Rubber Tree) : Bud-Scales serving as Nectaries. 2 1 9 proper time of the year to see if the young shoots of mature trees likewise possessed these nectariferous scales. Such was found to be the case. The foliage leaves of Hevea brasiliensis are not evenly distributed along the whole length of the shoot, but are crowded together on the upper portion. The stretch of stem, which in the mature shoot appears to be a long internode below the foliage leaves, is really composed of several, the nodes of which were occupied in the young state by the nectariferous scales. That is to say, the internodes between the upper nectariferous scales have increased considerably in length — this is especially well seen in saplings. The leaf is trifoliate with a long petiole. The leaflets are large and lanceolate in shape, and are joined to the apex of the petiole by very short stalks (Plate XVI. Fig. 3). As a rule the length of the petiole and the size of the leaflets of a shoot decrease from the base upwards. For example, the petiole and leaflet of the lowest leaf may have a length of 30 cm. and 23 cm. respectively ; whereas these measurements for the uppermost leaf may be only 3-5 cm. and 4*5 cm. respectively. The direction of the petioles is such that their laminas tend to be in one plane, and thus do not overshadow one another. The number of foliage leaves to a shoot varies, but is commonly twelve. They as well as the scales have a three-eighth arrangement on the stem-axis. The nectary of the foliage leaf is situated on the upper surface just at the point of union of the leaflets with the petiole (Fig. 3 n). It may consist of either three contiguous saucer-shaped glands, one corresponding to each leaflet, or of only two of these, as in the figure, or it may assume the form of an irregular depression due to their fusion. In any case, they are not prominent structures, and do not differ as a rule in colour from the surrounding surface. The bud-scales permit of division into two categories, viz. (1) the basal scales which are very small, non-nectariferous, and usually few in number ; (2) the upper scales which are conspicuous, nectariferous, and numerous. Both kinds of scales as well as the foliage leaves possess each a pair of insignificant stipules (Figs. 1 and 2 st). The basal non-nectariferous bud-scales. In the spirit material brought home for examination two types of young shoots could be distinguished, viz. those which had very few— one to three — non-nectariferous scales, and those possessing a great number, twenty or so. A drawing of each kind of shoot is shown in Figs. 1 and 2 respectively. A few shoots were intermediate in this respect, having several non-nectariferous scales, but not such an imbrication of them as represented in Fig. 2 s. No mention is made in my Ceylon notes of any large number of these basal scales having been noticed, but the shoots are referred to as possessing not 220 Parkin . — The Extra-floral Nectaries of Hevea brasiliensis more than two or three each. Whether or not the possession of a large number of these scales be a common feature of young Hevea shoots cannot well be decided from the few examined, but the supposition is that such a shoot as the one shown in Fig. 2 is exceptional, and that as a rule only two or three non-nectariferous scales occur. The nectariferous bud-scales . These vary considerably in number. The average for twenty-two shoots of adult trees examined was seven, ranging from five to twelve. The shoot from which Fig. I was drawn possessed eight, while that of Fig. 2 was exceptional in having twelve. Naturally, only part can be represented in the drawings. The lower nectariferous scales are small and short. The middle ones are usually the largest and possess the best developed nectaries, while the upper ones, though quite as long, are not so thick, and have the honey-secreting part reduced in extent ; in fact in the uppermost one of all this part may be restricted to the apex, or perhaps even absent. The internodes increase in length as a rule from the base upwards ; thus the lower nectariferous scales are near together, while the upper ones are some distance apart. The inflorescences are borne in the axils of the nectariferous scales as well as in those of the lower foliage leaves (Figs, i and 2 ft). Sapling. The young shoots of saplings resemble in most respects those of the adult trees, but being longer the internodes between the middle and upper nectariferous scales are more marked. From an ex- amination of thirty-eight young sapling shoots the following numbers were obtained : — non-nectariferous basal scales ranged in number from 0-3, aver. 1. nectariferous scales ,, „ 4-7, „ 5. foliage leaves — average 10. Nine of these thirty-eight young shoots possessed each an arrested leaf between the nectariferous scales and the foliage leaves proper. This bore three leaflets well defined but quite small, while the nectariferous scales have mere points to indicate the remains of the leaflets. The nectary appeared to be absent 1. This vestigial leaf did not persist, but withered and fell off with the scales. Seedling. In germination the two cotyledons remain in the testa in the soil, so that what looks like a hypocotyl is really the epicotyl ; it is quite long, 25 cm. or so in length. The first two foliage leaves formed quit the stem about the same level, and are similar in shape to those of older plants. Then comes an internode of about 3 cm., followed by two more foliage leaves situated at nearly the same level on the stem and alternating with the first pair ; sometimes there may be only one 1 Not microscopically examined — might possibly possess a trace of glandular tissue invisible to the naked eye. ( Para Rubber Tree): Bud- Scales serving as Nectaries . 221 leaf, or even three at this point. Occasionally the first pair of leaves may be vestigial, or only one of them fully developed. If the plumule be fatally injured then the bud in the axil of one of the cotyledons develops into a shoot, bearing first three to four reduced leaves apparently without nectaries, before the true foliage leaves appear ; sometimes the buds in both axils so sprout. The shoot arising from the axillary bud of the cotyledon simulates that derived from the plumule, but in the one case the length of stem produced before the foliage leaves are emitted is really composed of several internodes, the nodes being occupied by inconspicuous scale-leaves ; while in the other it consists of one internode only, the epicotyl. Unfortunately my notes do not connect the seedling with the sapling- stage, so as to see when the nectariferous scales first arise. This is probably at the second period of foliation. They apparently do not appear in the seedling, but rather later in the development of the plant. Structure of the individual bud-scales. The structure of the non- nectariferous scales requires little description. A glance at Fig. 4 sf shows their size and shape. They are each accompanied by a pair of lateral bodies — stipules. In the mature or sprouting bud they are brown dead objects. The nectariferous scales are fairly long, often bent structures and somewhat circular in transverse section ; they project from the stem at right angles or with a downward inclination. Each bears at its apex three minute points, the sole remains of the leaflets. Their upper convex surface is covered with yellow honey-secreting tissue, and has often a median longitudinal groove. In the lower and middle scales the whole length of the upper surface is glandular. In the upper scales the glandular portion tends to recede from the proximal part, and in the uppermost one it is confined to the apex (Fig. 6 ne). From a structural point of view the nectar-secreting tissue of plants can be divided into two classes 1, viz. (1) that consisting of small epidermal cells of the usual shape with thin hardly cuticularized outer walls, over- lying a mass of closely packed cells full of contents, and secretory in function, and (2) that in which the epidermis itself assumes the form of a secretory epithelium with greatly thickened cuticle. In the first class the nectar reaches the surface by passing through the thin walls, while in the second class it escapes by bursting the cuticle. The extra-floral nectaries of Hevea brasiliensis present a modification of the second type of structure, in that many of the original epithelial cells become divided in the mature nectary by tangential walls into two or three daughter-cells. That is, in the immature state the epidermis 1 Bonnier, Les nectaires, £tude critique, anatomique et physiologique, Ann. d. Sci. Nat., 6e ser., T. viii, 1879, P* 222 Parkin. — The Extra-floral Nectaries of Hevea brasiliensis is a simple epithelium, but on approaching maturity it becomes in places two or three layered (Fig. 7). Conspicuous nuclei and much cytoplasm without prominent vacuoles are present in the epithelial cells, as well as in the small cortical cells below. The cuticularized part of the outer wall is quite thick, as is shown in the drawing (Fig. 7 at). Examples of extra-floral nectaries with an epithelium divided in places are to be met with in Homalanthus populifera and Clerodendron Bungei 1 ; also a regularly two-layered epithelium exists in Prunus avium 2. The diagram (Fig. 5) shows the position of the nectar-secreting epi- thelium in a transverse section of a typical median bud-scale ; while that of Fig. 6 represents the epithelium as restricted to the apex in the uppermost scale. The minute structure of the nectaries of the foliage leaves is similar to that of the scale ones. General Remarks. This case of Hevea brasiliensis is about the first example cited of bud-scales — cataphyllary leaves — serving as nectaries. The only other instance I have found at all comparable is that mentioned by Reinke 3. He points out that the bud-scales, as well as the foliage leaves of Prunus avium, have glandular teeth which are honey-secreting. But here the transformation is very partial. The scales are not so modified as to be merely nectaries. Their primary function is still that of bud-protection. The Euphorbiaceae are rich in examples of plants with extra-floral nectaries. Baillon 4, in his work on this natural order, enumerates the various types, showing that their situation may be various, such as on the stem, petiole or lamina ; and that different organs may be wholly transformed into them, such as stipules and leaflets. Hevea brasiliensis affords a still further type, viz. that of bud-scales serving as nectaries. Two or three questions suggest themselves as to the origin of these cataphyllary nectaries of Hevea. Are they connected by descent with the petiolar glands, or are they a fresh production of glandular tissue in the evolution of the plant ? What is the relationship between the non-nectariferous and nectariferous scales ? Have they been derived in- dependently at different periods from foliage leaves, or have the former arisen by further retrogression from the latter ? From an identity in structure between the petiolar and scale nectaries and from the situation of the glandular tissue in the uppermost scale it looks as if the two classes of nectaries were directly connected. The petiolar glands have perhaps become much more developed in the scales, so that the function of these latter is now wholly that of secreting honey. 1 Morini, Contribute) all’ anatomia ed alia fisiologia dei Nettarii Estranuziali, Mem. Accad. Bologna, 1886, vii. a Reinke, Secretionsorgane, Prings. Jahrb., 1876, p. 125. 3 Reinke, loc. cit. 4 Baillon, Etude gen^rale des Euphorbiees, p. 230. ( Para Rubber Tree): Bud- Sc ales serving as Nectaries. 223 The view of the evolution of the shoot of Hevea that suggests itself to the author is as follows. Originally the base of the shoot had one, two, or three non-nectariferous bud-scales such as occur now ; the rest of the foliar organs were true foliage leaves arranged equidistantly along the axis. Assuming that their laminas gradually increased in size towards the middle of the shoots and then decreased, the lowest and highest leaves would in consequence be the smallest and the middle ones the largest — a condition often occurring in shoots. That of the Beech ( Fagus sylvatica ) is a case in point. Providing that the Hevea shoot had an upward tendency, as it has at the present day, the large median leaves would tend to overshadow the lower smaller ones, and thus render these latter to a great degree functionless as assimilating organs, and through disuse a gradual reduction in their laminas might follow. The nectaries on the petiolar apices still remaining would be the first to secrete. Viewing their service as one of attracting ants to keep off leaf-destroying insects, it would be an advantage to the plant to retain the nectaries on these retrograde leaf-structures, and further to increase their size and consequently their secretion, in order to protect the expanding foliage leaves, till their nectaries became functional. Thus gradually a condition which now occurs would be brought about. The Beech shoot has a few scales at its base without any lamina, which may be comparable, though not homologous as they are stipules, to the non-nectariferous scales of Hevea ; then come the foliage leaves increasing in size as far as the middle of the axis, and then diminishing towards the apex. There is a tendency in some of its shoots for the small lower leaves to wither and fall early. This may be partly due to their being shaded by the higher leaves, though this overshadowing is largely guarded against by the shoot as a rule having a horizontal direction, and as a consequence the leaf-blades are in one and the same plane. If the shoot, on the other hand, were inclined considerably to the vertical as in Hevea , then the middle leaves would shade the lower ones much more effectually. Such a shoot as that of the Rhododendron demonstrates this. It is obliquely erect and has the foliage leaves crowded together on its upper part, thus resembling the shoot of Hevea. The length of stem below the rosette of foliage leaves is not a single internode, but composed of several, the nodes of which in the young state were occupied by small leaves which have shrivelled and disappeared. These, being perhaps originally smaller than the middle leaves and thus subject to shade, now no longer persist as functional foliage leaves ; they have most likely decreased still further in size, and now apparently serve as protective scales to the bud. Consequently as a rule in horizontal shoots the lowest foliage leaves are the smallest, or at any rate smaller than the middle ones ; while in shoots inclined to the vertical the lowest leaves are the largest, because 224 Parkin . — The Extra-floral Nectaries of Hevea brasiliensis they represent probably the middle leaves of the primitive shoot, the lowest having ceased to act as foliage leaves. The reason why in an ordinary shoot such as that of the Beech the middle leaves should generally be the largest is perhaps owing to the intensity of growth during development, first rising gradually to a maximum, then falling again till growth ceases. This would result in the first and last formed leaf-blades being the smallest. The only other species of Hevea I have been able to examine is H. sprnceana , Mull.-Arg., a very closely allied one. It possesses similar nectariferous scales. Summary. 1. Hevea brasiliensis possesses two kinds of extra-floral nectaries : — (a) Small inconspicuous glands situated on the upper surface of the foliage leaves, where the three leaflets join the petiole (Fig. 3 n). (b) Large conspicuous glands borne on vestigial foliar structures — 4 bud-scales 7 — which are situated on the shoot below the foliage leaves proper (Figs. 1 and 2 ns). 2. The ■ bud-scale 5 nectaries are a prominent feature of the young expanding shoot, and are functional till the foliage leaves are mature, when they wither and drop off. They are present in saplings, as well as in adult trees, but were not observed in seedlings. 3. Besides these nectariferous structures, one or more insignificant bud- scales without nectaries may be present at the base of the shoot (Figs. 1 and 2 .s'). 4. The minute structure of the foliar and ‘ bud-scale ’ nectaries is the same. Each consists of a well-defined secretory epithelium with a thick cuticle. The original cells of this epithelium may be divided here and there by one or two tangential walls to form in places a two- or three- layered epidermis (Fig. 7). The nectar escapes by the bursting of the cuticle. 5. The two kinds of extra-floral nectaries are considered as homo- logous ; that is to say, the ‘ bud-scale ’ one may be regarded as a further development of what was at one time a petiolar nectary. 6. These nectariferous structures, occupying relatively the same position on the shoot as ordinary bud-scales, probably never had a protective function, but have been derived directly from what were once foliage leaves by the disappearance of the lamina and an increase in size of the nectary. 7. According to the usual view taken of the function of extra-floral nectaries, the ‘bud-scale’ glands may be looked upon as attracting ants to keep off insects injurious to the developing foliage. As soon as the foliage ( Para Rubber Tree): Bud- Sc ales serving as Nectaries. 225 leaves mature, their own nectaries become functional, and the scale ones being no longer required wither and drop off. 8. This case of Hevea brasiliensis is the first striking instance recorded, as far as the author is aware, of bud-scales — cataphyllary leaves — serving solely as nectaries. Postscript. Just on the completion of this paper a communication on the extra- floral nectaries of Hevea , read before the Academy of Sciences, Paris, on November 9, 1903, came to my notice. The article1 resulting from it in the corresponding number of the ‘ Comptes Rendus ’ deals wholly with the structure of the petiolar nectaries, and makes no reference whatsoever to the nectariferous bud-scales ; hence the chief subject-matter of my paper is not in the least affected. The authors state that the number of indi- vidual glands composing the petiolar nectary of Hevea brasiliensis may vary between two and five, but is usually three. They point out that the secretory epidermis of the nectary is two-layered in places, and lay stress on the two following structural features: (1) the presence of a ring of lignified parenchyma in the interior of the raised border which surrounds the secretory surface of each gland ; (2) the laticiferous tubes, occurring in fair abundance in the specialized parenchyma of the gland, either end just below the secretory epidermis, or even pass between the epidermal cells to the exterior. J. P. Cambridge, December , 1903. 1 Daguillon et Coupin, Sur les nectaires extra-floraux des Hevea , Comp. Rend, cxxxvii, No. 19, 1903, pp. 767-9. 226 Parkin. — The Extra-floral Nectaries of Hevea brasi liens is . EXPLANATION OF FIGURES IN PLATE XVI. Illustrating Mr. Parkin’s paper on the Extra-floral Nectaries of Hevea brasiliensis. Fig. i. Young shoot from adult tree bearing the immature foliage leaves on its upper part. Natural size, s, single non-nectariferous basal bud-scale ; ns3, ns5, ns6, ns8, represent respectively the 3rd, 5th, 6th, and 8th nectariferous bud-scales ; this shoot possessed eight of these structures ; the other four are not depicted in the drawing ; the shaded areas show the position of the nectar- secreting parts of the scales ; in ns8, the uppermost scale, this part is restricted to the apex ; /, lowest foliage leaf ; pn , position of the petiolar nectary ; st, stipules ; Ji , young inflorescences in the axils of the upper scales and lower foliage leaves ; Is, leaf-scars of previous year’s shoot. Fig. 2. Young shoot from adult tree with many basal bud-scales. Natural size, s, imbrication of non-nectariferous basal scales ; ns, length of stem bearing nine nectariferous bud-scales (the shoot possessed twelve altogether) ; Jl , inflorescences in the axils of the nectariferous scales ; /, lowest foliage leaf ; st, stipules ; Is, leaf-scars of previous year’s shoot. Fig. 3. Part of the upper surface of a foliage leaf. Natural size, n, twin-nectary; p, upper part of petiole, l, lower part of one of the three leaflets ; s, short stalk of leaflet. Fig. 4. Individual bud-scales from shoot represented in Fig. 1. Natural size, s, single non- nectariferous basal scale, a, dorsal view showing pair of stipules, b, side view ; nsx — ns7 (inclusive), side views of seven nectariferous bud-scales in order of succession from base upwards, the position of the honey-secreting tissue is indicated by shading ; ns8, uppermost nectariferous bud-scale, a, side view, b, ventral view showing three points, the vestiges of the three leaflets — the nectar-secreting part is restricted to the apex and is shown by the small shaded area in this region. Fig. 5. Diagram of a median transverse section of a nectariferous bud-scale. x 30. e, ordinary epidermis ; ne, honey-secreting epithelium on the upper surface with groove, g\ fv, ring of fibro-vascular bundles. Fig. 6. Diagram of a longitudinal section of the apical part of the uppermost nectariferous bud-scale ( nss in Figs. 1 and 4) showing the restricted distribution of the glandular tissue, x 30. e, ordinary epidermis ; ne, honey-secreting epithelium in a position corresponding to that occupied by the nectary of the foliage leaf ; p, one of the three apical points — vestige of a leaflet ; h, hairs ; v, vascular strands. Fig. 7. Section of the honey-secreting epithelium of a bud-scale. x 400. e, undivided epithelial cell ; ex, epithelial cell divided into two daughter-cells ; e2 , epithelial cell divided into three daughter-cells ; ct, the thick cuticle ; n, nuclei ; c, small cortical cells full of contents and without intercellular spaces. The empty areas in the figure represent cells in the section from which the contents had accidentally disappeared during preparation. ■Annals of B A J. P a.rldu del. Annals of Botany. Voi.xmirim. PARKIN.- HEVE A B R A S 1 LI E N S I S . The Principles of Phyllotaxis. by ARTHUR H. CHURCH, M.A., D.Sc., Lecturer in Natural Science, Jesus College , Oxford. With seven Figures in the Text. IN a preliminary note published some time ago \ exception was taken to the conventional methods adopted for the description and even interpretation of phyllotaxis phenomena, and a suggestion was made that appeared to be not only more in accord with modern conceptions of the phenomena of energy distribution, but it was further indicated that such a theory when carried to its mathematical limits threw a strong light both on the mechanism of shoot production and the inherent mathematical properties of the lateral appendage usually described as a ‘leaf-member,’ as opposed to any secondary and subsidiary biological adaptations. As publication of the entire paper has been delayed, and the new standpoint has not received any special support from botanists to whom the mathematical setting proved possibly a deterrent, the object of the present note is to place the entire argument of the original paper in as concise a form as possible 2. The preliminary discussion is sufficiently familiar 3. The conventional account of phyllotaxis phenomena involves a system of ‘ fractional expressions ’ which become interpreted into angular diver- gences ; and in practice the appearance of ‘ orthostichies ’ has been taken as a guide to the determination of the proper ‘ fractional expression.’ This method, elaborated by Schimper (1830-5), has more or less held the field to the present time ; and, for want of something better, has received the assent, though often unwilling, of such great investigators as Hofmeister and Sachs, to say nothing of lesser lights. Although elaborated into a system by Schimper and Braun, who added the peculiar mathematical properties of the Fibonacci series to the academical account 1 Note on Phyllotaxis, Annals of Botany, xv, p. 481, 1901. 2 On the Relation of Phyllotaxis to Mechanical Laws. Part I, Construction by Orthogonal Trajectories, 1901. Part II, Asymmetry and Symmetry, 1902. 3 Descriptive Morphology-Phyllotaxis. New Phytologist, i, p. 49. [Annals of Botany, Vol. XVIII. No. LXX. April, 1904.] 228 Church . — The Principles of Phyllotaxis . of the subject, the geometry of the system is based solely on a mathematical conception put forward by Bonnet and Calandrini in 1754; and this mathematical conception applied only to adult shoots and adult members of equal volume arranged in spiral sequence, and thus involved a system of intersecting helices of equal screw-thread, or, reduced to a plane expression, of spirals of Archimedes, also with equal screw-thread. A system of helical mathematics was thus interpolated into botanical science, and these helical systems were correctly tabulated by * orthostichies ’ and ‘ divergence angles * obtained from simple fractional expressions themselves deduced from the observation of orthostichies. But in transferring the study of phyllotaxis to the ontogenetic sequence of successively younger, and therefore gradated , primordia at the apex of a growing plant-shoot which was not cylindrical, these mathematical expressions were retained, although the helices originally postulated have absolutely vanished ; and it is somewhat to the discredit of botanical science that this simple error should have remained so long undetected and unexpressed. As soon as one has to deal with spirals which have not an equal screw-thread, the postulated orthostichies vanish as straight lines ; the fractional expressions therefore no longer present an accurate statement of the facts ; and the divergence angles, calculated to minutes and seconds, are hopelessly out of the question altogether ; while any contribution to the study of phyllotaxis phenomena which continues the use of such expressions must only serve to obscure rather than elucidate the inter- pretation of the phenomena observed. That the required orthostichies were really non-existent at the growing point, a feature well known to Bonnet himself, has thus formed the starting-point for new theories of displacement of hypothetically perfect helical systems, as, for example, in the contact-pressure theory of Schwendener. But once it is grasped that the practice of applying helical mathematics to spiral curves which, whatever they are, cannot be helices, is entirely beside the mark, it is clear that the sooner all these views and expressions are eliminated the better, and the subject requires to be approached without prejudice from an entirely new standpoint. r The first thing to settle therefore is what this new standpoint is to be ; and how can such a remarkable series of phenomena be approached on any general physical or mathematical principles ? Now in a transverse section of a leaf-producing shoot, at the level of the growing point, the lateral appendages termed leaves are observed to arrange themselves in a gradated sequence as the expression of a rhythmic production of similar protuberances, which takes the form of a pattern in which the main construction lines appear as a grouping of intersecting curves winding to the centre of the field, which is occupied by the growing point of the shoot itself. As the mathematical properties JiruiaLs of Botany. VolXWlPUV. Bower del. Univ exsity Press, Oxford. BOWER - OPHIOGLOSSUM SIMPLEX. 229 Church. — The Principles of Phyllotaxis. of such intersecting curve systems are not specially studied in an ordinary school curriculum, a preliminary sketch of some of their interesting features may be excused, since geometrical relationships have clearly no inherent connexion with the protoplasmic growth of the plant-shoot, but are merely properties of lines and numbers. Thus, by taking first, for example, a system in which spiral curves of any nature radiate from a central point in such a manner that 5 are Fiq. 35, Curve-system (5 + 8): Fibonacci series. A. full contact-cycle of eight members is represented by circular primordia. turning in one direction and 8 in the other, giving points of intersection in a uniform sequence, a system of meshes and points of intersection is obtained, and to either of these units a numerical value may be attached. That is to say, if any member along the ‘5’ curves be called 1, the next inmost member along the same series will be 6, since the whole system is made of 5 rows, and this series will be numbered by differences of 5. 230 Church. — The Principles of Phyllotaxis. In the same way differences of 8 along the £ 8 ’ curves will give a numerical value to these members ; and by starting from 1, all the meshes, or points, if these are taken, may be numbered up as has been done in the figure (Fig. 35> (5 + 8)). Observation of the figure now shows what is really a very remarkable property : all the numerals have been used, and 1, 2, 3, 4, &c., taken in order, give also a spiral sequence winding to the centre. This is merely Fig. 36. Curve-system (6 + 8) : Bijugate type. Contact-cycle as in previous figure. a mathematical property of the system (5 + 8), in that these numbers are only divisible by unity as a common factor ; but the single spiral thus obtained becomes in a botanical system the genetic-spiral which has been persistently regarded as the controlling factor in the whole system, since if such a construction be elongated sufficiently far, as on a plant-shoot, this spiral will alone be left visible. The first point to be ascertained in phyllotaxis is the decision as to Church —The Principles of Phyllotaxis. 231 which is to be the prime determining factor ; that is to say, does the possession by the plant of a ‘genetic-spiral’ work out the subsidiary pattern of the parastichies, or are the parastichies the primary feature, and the genetic-spiral a secondary and unimportant consequence of the construction ? Now, other systems may quite as easily be drawn ; thus take next a system of 6 curves crossing 8. On numbering these up by differences of 6 and 8 respectively in either series, it will be found that this time all the numerals are not employed, but that there are two sets of 1, 3, 5, &c., and i\ 3', 5', &c., showing that pairs of members on exactly opposite sides of the system are of equal value. There is thus no single genetic spiral now present, but two equal and opposite systems — a fact which follows mathematically from the presence of a common factor (2) to the numbers 6 and 8. The existence of such factorial systems in plants has created much confusion, and the term bijngate applied to such a construction by the brothers Bravais may be legitimately retained as its designation (Fig. 36, system (6 + 8)). Again, on constructing a system of 7 curves crossing 8, and numbering by respective differences, this time of 7 and 8 ; as in the first case, since these numbers have 1 only as common factor, all the numerals are utilized in numbering the system ; the genetic-spiral may be traced even more readily than in the first example, the adjacent members along it being now in lateral contact, so that the resulting spiral obviously winds round the apex. This effect is common among Cacti, and is the result of a general property of these curve systems which may be summed up as follows : — -Given a set of intersecting curves, the same points of inter- section (with others) will also be plotted by another system of curves representing the diagonals of the first meshes, and the number of these curves, and also of course the difference in numerical value of the units along their path, will be given by the sum and dijference of the numbers which determine the system, for example, 5 and 8 have as complementary system 3 and 13 ; and also other systems may be deduced by following the addition and subtraction series, e. g. : — 5- « 3-13 2 — 21 i-34. Whereas the (7 + 8) system gives only 1 and 15 ; the single so-called ‘genetic-spiral,’ which includes all the points, being reached at the first process. Thus a Cactus built on these principles would show an obvious ‘ genetic-spiral ’ winding on the apex and 15 ridges, which in the adult state become vertical as a true helical construction is secondarily produced as the internodes attain a uniform bulk (Fig. 37 (7 + 8)). R 232 Church . — The Principles of Phyllo taxis. Finally, take the case of 8 curves crossing 8, and number in the same way by differences of 8 along both series. It immediately becomes clear that there are 8 similar series : all other spirals have been eliminated; there is no ‘genetic-spiral’ at all, but only a system of alternating circles of members of absolutely identical value in each circle. We have now, that is to say, systems of true whorls , and also learn in what a true whorl consists — the members must be exactly and Fig. 37. Curve-system (7 +8) : anomalous type. mathematically equal in origin — while the expression a successive whorl is a contradiction in terms. From such simple and purely geometrical considerations it thus follows that the so-called ‘genetic-spiral’ is a property solely of inter- secting curve-systems which only possess 1 as a common factor, and is therefore only existent in one case out of three possible mathematical forms (Figs. 35, 36, 38). While if these four systems were subjected to 233 Church — The Principles of Phy Hot axis. a secondary Zone of Elongation , No. i would pull out as a complex of spirals in which four distinct sets might be traced ; No. 2 as two spiral series leaving paired and opposite members at each f node 5 ; No. 3 as a spiral series with two complementary sets only; while No. 4 would give the familiar case of alternating whorls with 8 members at each ‘node.’ Further these cases are not merely arbitrary: they may all occur in the plant-kingdom, though the first is admittedly Fig. 38. Curve-system (8x8): symmetrical type. the most frequent ; but any theory which interprets one should equally well interpret the others. Similarly all changes of system may be discussed with equal readiness from the standpoint of the addition or loss of certain curves, and only from such a standpoint ; since it is evident that once it is granted that new curves may be added to or lost from the system, the numerical relations of the members may be completely altered by 234 Church . — The Principles of Phyllotaxis. Fig. 39. System (5 + 8) : eccentric construction in the plane of No Chitrch . — The Principles of Phyllotaxis . 235 the addition of one curve only, as in the difference between the systems (7 + 8), (8 + 8), &c. (Figs. 35-38) \ Thus the hypothesis of a genetic-spiral since it entirely fails to account for the arrangement of the members of all phyllotaxis systems in a single spiral, may be conveniently wholly eliminated from future discussions of these systems. It remains as a mere geometrical accident of certain intersecting curve-systems, and the fact that such systems may be very common in plant construction does not affect the main principle at all. On the other hand, it may be urged that in these special cases one cannot get away from the fact that it does actually represent the building- path as seen in the visible ontogeny of the component members, and must therefore ever remain the most important feature of these systems as checked by actual observation apart from theoretical considerations. But even this view is not absolute ; and such a case in which the ontogenetic sequence of development is not the single spiral obtained by numbering the members in theoretical series would naturally confuse the observer of direct ontogeny. For example, in the previous cases figured the proposition of centric growth systems was alone considered, as being the simplest to begin with ; it is obvious that even a small amount of structural eccentricity will produce a very different result. Thus in Fig. 39 the (5 + 8) system is redrawn in an eccentric condition, the so-called ‘ dorsiventrality J of the morphologist ; on numbering the members in the same manner as before it is clear that the series obtained is very different from any empirical ontogenetic value which would be founded on the observation of the relative bulk of the members at any given moment. The occurrence of such systems in plant-shoots — and it may be stated that this figure was originally devised to illustrate certain phenomena of floral construction in the case of Tropaeolum — gives in fact the final proof, if such were any longer needed, of the simple geometrical generalization that such systems of intersecting curves are always readily interpreted in terms of the number of curves radiating in either direction, and not in any other manner. The presence of a circular zone (whorl) or a genetic- spiral is a wholly secondary geometrical consequence of the properties of the numerals concerned in constructing the system. The preference of any individual botanist, either in the past or at present, for any particular method 1 Cf. Relation of Phyllotaxis to Mechanical Laws. Part II, p. 109, Rising and Falling Phyllotaxis. Part IV, Cactaceae. Though the figures (35-38) have, as a matter of fact, been drawn by means of suitable ortho- gonally intersecting logarithmic spirals, because these curves are easily obtained and the schemes are subsequently held to be the representation of the true construction system of the plant-apex, the nature of the spirals does not affect the general laws of intersection so long as this takes place uniformly. 236 Church. — The Principles of Phyllotaxis. of interpreting any of these systems has little bearing on the case: the subject is purely a mathematical one ; and the only view which can be acceptable is that which applies equally well to all cases, in that the question is solely one of the geometrical properties of lines and numbers, and must therefore be settled without reference to the occurrence of such constructions in the plant. If all phyllotaxis systems are thus to be regarded solely as cases of intersecting curves, which are selected in varying numbers in the shoots of different plants, and often in different shoots of the same plant, with a tendency to a specific constancy which is one of the marvellous features of the plant-kingdom, it remains now to discuss the possibility of attaching a more direct significance to these curves, which in phyllotaxis construction follow the lines of what have been termed the contact-par as tichies ; that is to say, to consider I. What is the mathematical nature of the spirals thus traced ? II. What is the nature of the intersection ? and III. Is it possible to find any analogous construction in the domain of purely physical science ? The suggestion of the logarithmic spiral theory is so obvious that it would occur naturally to any physicist: the spirals are primarily of the nature of logarithmic spirals ; the intersections are orthogonal ; and the construction is directly analogous to the representation of lines of equipotential in a simple plane case of electrical conduction. In opposition to this most fruitful suggestion, it must be pointed out however that the curves traced on a section are obviously never logarithmic spirals, and the intersections cannot be measured as orthogonal. But then it is again possible that in the very elaborate growth-phenomena of a plant- shoot secondary factors come into play which tend to obliterate the primary construction ; in fact, in dealing with the great variety of secondary factors, which it only becomes possible to isolate when the primary construction is known, the marvel is rather that certain plants should yield such wonderfully approximately accurate systems. To begin with, logarithmic spiral constructions are infinite , the curves pass out to infinity, and would wind an infinite number of times before reaching the pole. Plant constructions on the other hand are finite , the shoot attains a certain size only, and the pole is relatively large. The fact that similar difficulties lie in the application of strict mathematical construction to a vortex in water, for example, which must always possess an axial tube of flow for a by no means perfect fluid, or to the distribution of potential around a wire of appreciable size, does not affect the essential value of the mathematical conception to physicists. And, though the growth of the plant is finite, and therefore necessarily subject to retarding influences of some kind, there is no reason why a region may not be postulated, Church . — The Principles of Phyllotaxis . 237 however small, at which such a mathematical distribution of ‘growth- potential ’ may be considered as accurate ; and such a region is here termed a ‘ Growth- Centre l Since the interpretation of all complex phe- nomena must be first attacked from the standpoint of simple postulates, it now remains to consider the construction and properties of as simple a centre of growth as possible. Thus in the simplest terms the growth may be taken as uniform Fig. 40. Scheme for Uniform Growth Expansion : a circular meshwork of quasi- squares. Symmetrical construction from which asymmetrical homologues are obtained by the use of logarithmic spirals. and centric', the fact that all plant growth is subject to a retardation effect or may be frequently eccentric , may at present be placed wholly on one side, since the simplest cases evidently underlie these. The case of uniform centric growth is that of a uniformly expanding sphere ; or, 238 Church —The Principles of Phyllotaxis. since it is more convenient to trace a solid in separate planes, it will be illustrated by a diagram in which a system of concentric circles encloses a series of similar figures, which represent a uniform growth increment in equal intervals of time. Such a circular figure, in which the expanding system is subdivided into an indefinite number of small squares repre- senting equal time-units, is shown in Fig. 40, and presents the general theory of mathematical growth, in that in equal times the area represented by one ‘ square 5 grows to the size of the one immediately external to it1. Now it is clear that while these small areas would approach true squares if taken sufficiently small, at present they are in part bounded by circular lines which intersect the radii orthogonally ; they may there- fore be termed quasi-squares : and while a true square would contain a true inscribed circle, the homologous curve similarly inscribed in a quasi- square will be a quasi-circle. It is to this quasi-circle that future interest attaches ; because, just as the section of the whole shoot was conceived as containing a centric growth-centre, so the lateral, i. e. secondary, appendages of such a shoot may be also conceived as being initiated from a point and presenting a centric growth of their own. These lateral growth-centres, however, are component parts of a system which is growing as a whole. The con- ception thus holds that the plane representation of the primary centric shoot-centre is a circular system enclosing quasi-circles as the representatives of the initiated appendages. To this may now be added certain mathematical and botanical facts which are definitely established. I. Any such growth-construction involving similar figures (and quasi- circles would be similar) implies a construction by logarithmic spirals. II. A growth-construction by intersecting logarithmic spirals, and only by curves drawn in the manner utilized in constructing these diagrams (Figs. 35-38), is the only possible mathematical case of continued orthogonal intersection 2. III. The primordia of the lateral appendages of a plant only make contact with adjacent ones in a definite manner , which is so clearly that of the contacts exhibited by quasi-circles in a quasi-square meshwork, that Schwendener assumed both a circular form and the orthogonal arrangement as the basis of his Dachstuhl Theory : these two points being here just the factors for which a rigid proof is required, since given these the logarithmic spiral theory necessarily follows. A construction in terms of quasi-circles would thus satisfy all theo- 1 The same figure may also be used to illustrate a simple geometrical method of drawing any required pair of orthogonally intersecting logarithmic spirals. 2 For the formal proof of this statement I am indebted to Mr. H. Hilton. Church. — The Principles of Phyllotaxis . 239 retical generalizations of the mathematical conception of uniform growth, and would be at the same time in closest agreement with the facts of observation ; while no other mathematical scheme could be drawn which would include primordia arranged in such contact relations and at the same time give an orthogonal construction. If, that is to say, the quasi- circle can be established as the mathematical representative of the primordium of a lateral appendage, the orthogonal construction, which is the one point most desired to be proved, will necessarily follow. Fig. 41. Quasi-circles of the systems (2 + 2), (1 + 1) and (1 4- 2) arranged for illustration in the plane of median symmetry. C' , C" , Cr,f , the centres of construction of the respective curves. (After E. H. Hayes.) It remains therefore now to discuss the nature of the curves denoted by the term quasi-circles ; their equations may be deduced mathematically, and the curves plotted on paper from the equations. These determinations have been made by Mr. E. H. Hayes. Thus a general equation for the quasi-circular curve inscribed in a mesh made by the orthogonal inter- 240 Church — The Principles of Phyllotaxis . section of m spirals crossing «, in the manner required, is given in such a form as, log c + 1-36438 log r — — 5 — .000030864 62, m2 + 0 where the logarithm is the tabular logarithm, and 0 is measured in degrees ; or where the logarithm is the natural logarithm and 6 in circular measure : From these equations the curve required for any phyllotaxis system can be plotted out ; and a series of three such curves is shown in Fig. 41, grouped together for convenience of illustration, i. e. those for the lowest systems (2 + 2), (1 + 2) and (1 + 1). It will be noticed immediately that the peculiar characters of these curves are exaggerated as the containing spiral curves become fewer: thus with a larger number than 3 and 5, the difference between the shape of the curve and that of a circle would not be noticeable to the eye. While in the kidney-shaped (1 + 1) curve the quasi-circle would no longer be recognized as at all comparable in its geometrical properties with a true centric growth-centre. But even these curves, remarkable as they are, are not the shape of the primordia as they first become visible at the apex of a shoot constructing appendages in any one of these systems. The shape of the first formed leaves of a decussate system, for example, is never precisely that of the (2 + 2) curve (Fig. 41), but it is evidently of the same general type ; and it may at once be said that curves as near as possible to those drawn from the plant may be obtained from these quasi-circles of uniform growth by taking into consideration the necessity of allowing for a growth-retardation. Growth in fact has ceased to be uniform even when the first sign of a lateral appendage becomes visible at a growing point ; but, as already stated, this does not affect the correct- ness of the theory in taking this mathematical construction for the starting-point ; and, as has been insisted upon, the conception of the actual existence of a state of uniform growth only applies to the hypothetical ‘ growth-centre.’ On the other hand, the mere resemblance of curves copied from the plant to others plotted geometrically according to a definite plan which is however modified to fit the facts of observation, will afford no strict proof of the validity of the hypothesis, although it may add to its general probability, since there is obviously no criterion possible as to the actual nature of the growth-retardation ; that is to say, whether it may be taken as uniform, or whether, as may be argued from analogy, it may exhibit daily or even hourly variations. Something more than this is necessary before the correctness of the assumption of quasi-circular leaf-homologues can Church . — The Principles of Phyllo taxis. 241 be taken as established; and attention may now be drawn to another feature of the mathematical proposition. It follows from the form of the equation ascribed to the quasi-circle that whatever value be given to m and n, the curve itself is bilaterally symmetrical about a radius of the whole system drawn through its centre of construction. That it should be so when m—n , i. e. in a symmetrical ( whorled ) leaf-arrangement, would excite no surprise ; but that the primor- dium should be bilaterally symmetrical about a radius drawn through its centre of construction, even when the system is wholly asymmetrical and spiral, is little short of marvellous, since it implies that identity of leaf-structure in both spiral and whorled systems, which is not only their distinguishing feature, but one so usually taken for granted that it is not considered to present any difficulty whatever. Thus, in any system of spiral phyllotaxis, the orientation of the rhomboidal leaf-base is obviously oblique , and as the members come into lateral contact they necessarily become not only oblique but asymmetrical, since they must under mutual pressure take the form of the full space available to each primordium, the quasi-square area which appears in a spiral system as an oblique unequal-sided rhomb (Fig. 35). Now the base of a leaf (in a spiral system) is always such an oblique, anisophyllous structure, although the free appen- dage is isophyllous , bilaterally symmetrical, and flattened in a horizontal plane 1. The quasi-circle hypothesis thus not only explains the inherent bilaterality of a lateral appendage, but also that peculiar additional attri- bute which was called by Sachs its * dorsiventrality] or the possession of different upper and lower sides, and what is more remarkable, since it cannot be accounted for by any other mathematical construction, the isophylly of the leaves produced in a spiral phyllotaxis system 2. It has been the custom so frequently to assume that a leaf-primordium takes on these fundamental characters as a consequence of biological adaptation to the action of such external agencies as light and gravity, that it is even now not immaterial to point out that adaptation is not creation , and that these fundamental features of leaf-structure must be present in the original primordium, however much or little the action of environment may 1 These relations are beautifully exhibited in the massive insertions of the huge succulent leaves of large forms of Agave : the modelling of the oblique leaf-bases with tendency to rhomboid section, us opposed to that of the horizontal symmetrical portion of the upper free region of the appendage, may be followed by the hand, yet only differs in bulk from the case of the leaves of Sempervivum or the still smaller case of the bud of Pinus. 2 Anisophylly is equally a mathematical necessity of all eccentric shoot systems. It will also be noted that the adjustment required in the growing bud, as the free portions of such spirally placed primordia tend to orientate their bilaterally symmetrical lamina in a radial and not spiral plane, gives the clue to those peculiar movements in the case of spiral growth systems, which, in that they could be with difficulty accounted for, although as facts of observation perfectly obvious, has resulted in the partial acceptance of Schwendener’s Dachstuhl Theory. This theory was in fact mainly based on the necessity for explaining this ‘ slipping ’ of the members, but in the logarithmic spiral theory it follows as a mathematical property of the construction. 242 Church. — The Principles of Phyllo taxis. result in their becoming obvious to the eye. The fact that the quasi-circle hypothesis satisfies all the demands of centric growth systems, whether symmetrical or asymmetrical, as exhibited in the fundamental character of foliar appendages, and that these characters may be deduced as the mathematical consequences of the simple and straightforward hypothesis of placing centres of lateral growth in a centric system which is also grow- ing, may be taken as a satisfactory proof of the correctness of the original standpoint. And it is difficult to see what further proof of the relation between a leaf-primordium as it is first initiated, and the geometrical properties of a quasi-circle growth system is required ; but it still remains to connect this conception with that of orthogonal construction. This however naturally follows when it is borne in mind, firstly that no other asymmetrical mathematical growth-construction is possible, except the special quasi-square system which will include such quasi-circles ; and secondly, that the contact-relations of the quasi-circles in these figures are identical with those presented by the primordia in the plant, and could only be so in orthogonal constructions. It thus follows that with the proving of the quasi-circle hypothesis, the proof is further obtained that the intersection of the spiral paths must be mutually orthogonal ; and it becomes finally established that in the construction of a centric phyllotaxis system, along logarithmic spiral lines, the segmentation of the growth system at the hypothetical growth-centre does follow the course of paths intersecting at right angles ; and the principle of construction by orthogonal trajectories, originally suggested by Sachs for the lines of cell-structure and details of thickened walls, but never more fully proved, is now definitely estab- lished for another special case of plant-segmentation, which involves the production of lateral appendages without any reference to the segmentation of the body into ‘ cell ’ units. But even this is not all ; the point still remains, — What does such construction imply in physical terms? Nor can it be maintained that the present position of physical science affords any special clue to the still deeper meaning of the phenomena. The fact that the symmetrical con- struction in terms of logarithmic spirals agrees with the diagram for dis- tribution of lines of equipotential and paths of current flow in a special case of electric conduction, while the asymmetrical systems are similarly homo- logous with lines of equal pressure and paths of flow in a vortex in a perfect fluid, the former a static proposition, the latter a kinetic one, may be only an ‘ accident.’ On the other hand it must always strike an unprejudiced observer that there may be underlying all these cases the working of some still more fundamental law which finds expression in a similar mathematical form. In conclusion, it may be noted that if the proof here given of the principle of plant construction by orthogonal trajectories is considered satis- Church. — The Principles of Phyllotaxis. 243 factory, it adds considerably to the completeness of the principles of proto- plasmic segmentation, and may be extended in several directions with further interesting results. It is only necessary to point out that the case of centric-growth is after all only a first step ; and the most elaborate growth forms of the plant-kingdom, as exhibited for instance in the seg- mentation of the leaf-lamina, may be approached along similar lines, and by means of geometrical constructions which are consequent on the more or less perfect substitution of eccentric and ultimately wholly unilateral growth- extension, which again must ever be of a retarded type. The subject thus rapidly gains in complexity ; but that the study of growth-form, which after all is the basis of all morphology, must be primarily founded on such simple conceptions as that of the ‘ growth-centre * which has here been put forward, should I think receive general assent, and in the case of the quasi- circle, there can be little doubt as to the extreme beauty of the results of the mathematical consideration. - ■ ’ - : ' • ' •' ' - - • . -J:: ' .-.fir; - •••'■• • V- ' : ' , < - : v:/ •. •• ; ; - ;i ri -■ .. • •• ; ' ' : ' • -X ' ; \ ■ p: - ■ ' :• ' : ' ; v.; ' M:j vdj ’ ' ■ • ■ • ■ ' ' •• \ . v ! (v, - , : - ; ,ri ' ' ■ ‘ ' : ' ' . • ' - ■ • i ■ - !!■: ; r | ■' : j ' Lrrj? y. f, . ; 1 G^ofra Jy!j'*,vnp^ " ' ; . [■ - ■> • • > / i - The Development of the Spermatozoid in Chara. BY DAVID M. MOTTIER, Professor of Botany in Indiana University. With Plate XVII. WHILE the discovery of the fact that the cilia-bearing part of the spermatozoid in Chara and in certain Archegoniates originates from a cytoplasmic structure resembling a centrosome has aroused a keen interest in the study of spermatogenesis among plants, yet the diversity of opinion and the controversies that have arisen concerning the probable relation existing between the centrosome and the cilia-bearer, or blepharo- plast, have become an equal stimulus to research. Because of the fact that in certain Gymnosperms and Pteridophytes the cilia-bearing structure is derived from a centrosome-like body, the investigator naturally expects to find a similar origin for this structure in the spermatogenesis of other plants possessing male gametes that may be called spermatozoids. In this respect the expectations of the writer were not realized in a study of the development of the spermatozoid of Chara fragilis , and as certain important details observed differed from the accounts given by Belajeff (’94) and others, the publication of the following seems not superfluous. The mature spermatozoid of Chara , as is well known, consists of a thread-like body, making two or more spiral turns, and bearing two long cilia inserted a short distance behind the anterior end. This body is composed of a nucleus occupying the middle portion and a cytoplasmic band or thread, the blepharoplast, which bears the two cilia. The blepharo- plast, therefore, extends some distance posteriorly as well as anteriorly beyond the nuclear portion. Up to the present time Belajeff (’94) has given the most complete account of the development of the spermatozoid in the Characeae, using chiefly Chara foetida as well as species of Nitella . As the earlier part of the development described by Belajeff differs somewhat in detail from the results of my own observations, a brief outline of the process as described by this author will be given first for the sake of greater clearness. [Annals of Botany, Vol. XVIII. No. LXX. April, 1904.] 246 Mottier . — Development of the Spermatozoid in Chara. The first indication of the development of the spermatozoid is the movement of the nucleus from near the centre to one side of the cell and a slight contraction of the entire protoplast. The side towards which the nucleus moves is designated by Belajeff as the ‘dorsal/ and that directly opposite as the ‘ ventral ’ side of the cell. By the side or lateral wall is meant that which is parallel with the longitudinal axis of the spermogenous filament. The contraction of the protoplasm is most pronounced at the sides, while it is withdrawn very little or not at all from the transverse walls. About this plasma-cylinder there now appears a ring-shaped groove (loc. cit., p. 33) which, it is stated, is to play an important role in the further development of the spermatozoid, since in this groove the cilia are to be developed. There now appears a small plasmic protuberance at the boundary between the nucleus and the cytoplasm, i. e. at the dorsal side of the cell, or that which will become the convex side of the developing sperm, from which the cilia arise. ‘ Den Beginn der Entwickelung des Spermatozoidenkorpers kennzeichnet das Auftreten eines kleinen Plasmahockers an der Grenze zwischen Kern und Cytoplasma. Dieser Hocker begiebt sich in die ringformige Rinne und ist der Seitenwand der Zelle zugewandt. Am deutlichsten ist das Hervortreten dieses Hockers von einer der flachen Seiten der spermato- genen Zelle aus zu verfolgen (loc. cit., Fig. 1 5 a). Dieser Hocker entspricht augenscheinlich den “ glanzenden Piinktchen ” des Mettenius und den von Goebel beschriebenen “ Kopfen.” Aus dem Hocker wachsen zwei kurze, elastische Faden hervor. die beide parallel der Seitenwand, aber in ent- gegengesetzter Richtung verlaufen.’ With the further development of the spermatozoid the protuberance or nodule (Hocker) undergoes a change in position, whereby it gradually moves to the side of the cell opposite the nucleus. From the protuberance to the nucleus extends a delicate thread staining intensely red with fuchsin. With the further growth of this thread the protuberance is pushed farther away from the nucleus (loc. cit., Fig. 16). This thread lies in the cytoplasm, from which only the protuberance, bearing the cilia, protrudes into the groove. The thread with its cilia-bearing protuberance forms the anterior end of the spermatozoid. Simultaneously with the development of the anterior end, that of the posterior extremity is to be observed. At the side of the nucleus opposite the point of origin of the plasmic protuberance, there appears in the cytoplasm a homogeneous thread or band, which grows parallel with the side-wall of the cell and in the opposite direction to that of the anterior end. This thread is considerably thicker than that which bears the cilia. Its free or posterior extremity protrudes out of the cytoplasm into the groove, and forms a beak-like outgrowth (loc. cit., Figs. 1 6, 17, and 19). The cilia-bearing filament continues its growth within the cytoplasm and until it has reached a middle point on the Mottier. — Development of the Spermatozoid in Char a. 247 ventral side of the cell. The cilia are fastened to its anterior end. From this time onwards the point of insertion of the cilia is unchanged, so that the further growth of the thread leads to the formation of that part of the sperm anterior to the point of insertion of the cilia. The trans- formation taking place in the nucleus during the development, as described by Belajefif for Char a foetida , differs in no important detail from my own observations on Chara fragilis , a detailed account of which is to follow. From the foregoing it will be seen that, according to Belajeff, the cilia arise from a cytoplasmic protuberance or knob (Hocker), lying near the nucleus at the side which is to become the convex side of the mature spermatozoid. The formation of a cytoplasmic thread now carries the cilia-bearing protuberance some distance from the nucleus, and constitutes the anterior cytoplasmic part of the sperm. At the same time a similar, though somewhat broader, thread or band develops from the opposite side of the nucleus. This constitutes the posterior end of the spermatozoid. Judging from Belajeff’s figures these cytoplasmic extremities of the sperm seem to be direct transformations of the plasma-membrane (Hautschicht), although he does not make a direct statement to that effect. He leaves the impression also that the two ends of the sperm are separate pieces, being fastened to the respective ends of the nuclear portion. Belajeff’s study was made on material stained and observed in toto after mounting in glycerine — a method that cannot be regarded at present as altogether reliable, yet more improved methods show that many of the finer details were brought out with remarkable accuracy. In my own study Chara fragilis was used exclusively. The material was fixed in chrom-osmic-acetic acid, embedded in paraffin, sectioned and stained on the slide with aniline-safranin, gentian-violet and orange G. For the most part the sections were cut from three to five microns in thickness, but in order to bring whole cells into observation thicker sections were prepared. Of all the plants below the Ferns, the Characeae are probably the most suitable objects for the study of spermatogenesis. The mature spermatozoids are relatively large, and the protoplasmic structures of their mother-cells can be fairly well differentiated by present cytological methods. Yet even here, as in the most favourable cases, there are phases in the process which present extreme difficulties, and consequently there are points concerning which there is always more or less doubt. This I have found true in some of the earlier stages in which cytoplasmic differentiation begins. As is well known, the spermogenous cells of Chara are arranged in long filaments coiled up in the globular antheridium. The cells are in the form of flat, cylindrical disks, two or three times as broad as long S 248 Mottier. — Development of the Spermatozoid in Chara. (Figs. 2, 3). Observed from the end they are, therefore, circular in outline, and in longitudinal section somewhat rectangular. Before differentiation begins, the nucleus is spherical or elliptical, and lies near the middle of the cell. The diameter of the nucleus is almost as great as the length of the cell, so that there appears to be room enough only between nuclear membrane and either transverse wall for the plasma-membrane. The chromatin is arranged in the form of a somewhat regular but interrupted spirem, i. e. there are thin places in the nuclear thread in which little or no chromatin is present (PL XVII, Figs. 1-3). There is no definitely recognizable differentiation in the cytoplasm at first. This appears to be a uniformly granular network or alveolar structure. The first indication of the development of the spermatozoid, as has been correctly observed by Belajeff, is the withdrawal of the nucleus towards one side of the cell, and a slight contraction of the entire protoplast. That part of the lateral wall towards which the nucleus moves will be next to the convex side of the developing sperm. This is spoken of as the dorsal side by Belajeff. The first indication of a cytoplasmic differentiation observed by me appears in the form of a very delicate thread or band extending partly around the cell and embracing the nucleus in its arc (Fig. 1). This thread seems to be merely a differentiation of the plasma-membrane. The end of this thread, which will become the anterior end of the sperm, is thinner than the opposite or posterior end. In Fig. 1 the anterior end is below the nucleus, and the posterior above. The row of sharply defined granules included in the concave side of the posterior part of the spermatozoid, and which can be so readily observed in later stages, has just begun to appear. This line of granules marks the line of separation between the posterior end of the sperm and the general cytoplasmic body of the cell. I do not agree with Belajeff that these ends are separate, but that they are the ends of a continuous thread, the blepharoplast, which in the majority of cases is obscured where it lies next to the nucleus. That this interpretation is correct is seen in later stages, where the cell has been shrunken by the reagents (Fig. 5). I was not able to observe any cytoplasmic knob or protuberance (Hocker) from which the cilia arise as described by Belajeff, nor does the band or thread extend around the middle of the cell in a groove. In this respect my observations agree with those of Strasburger (’92), who did not observe any groove. About the same developmental stage shown in Fig. 1, or perhaps a little later, is represented in Fig. 2, as seen from the side or in longitudinal section. The band is seen to lie along one end of the cell. In this figure it will be further noticed that, on the side near the band, the cytoplasm has become denser, while on the opposite side there are a very few granules in the very delicate cytoplasmic network. This portion of the cell, which is comparatively free from granules, might easily be taken Mot tier. — Development of the Spermatozoid in Char a. 249 for a vacuole. Fig. 1 seems to be the same stage as represented by BelajefFs Fig. 12. Fig. 3 is a median longitudinal section of a cell in the same stage as Fig. 1, but passing through the region between the ends of the thread, or on a line in the plane of ab , Fig. 1. The cytoplasm on the side opposite the nucleus, or that which will become the concave side of the mature sperm, is somewhat concave at this time, becoming much more so in a later stage. I observed nothing to correspond with Belajeff’s Fig. 15 a. In the stage of Fig. 1 no trace of cilia was seen. At this stage the chromatin is disposed in a hollow spirem with numerous interruptions, i. e. places free from chromatin but continuous as to linin. The chromatin is in the form of larger or smaller lumps or pieces. A little later in the development, the blepharoplast is more distinctly differentiated (Figs. 4, 5). It has become longer, and extends almost entirely around the cell. The anterior end is thinner and narrower than the posterior. In Fig. 5 it seems that the posterior end broadens rather abruptly. In this instance it seems that the thread, by shrinkage, has become separated from the nucleus at its point of contact. As a rule, the condition of things is as shown in Figs. 4 and 6, in which case the thread lies in such close contact with the nuclear membrane that it is not definitely distinguished there. This gives the impression that the thread consists of two separate pieces growing in opposite directions from a point or points on the nucleus as described by Belajeff. The two ends of the blepharo- plast rarely lie in the same plane at this stage. In the stage of Fig. 6, the cilia are present and of considerable length. In all cases in which their point of origin could be made out with certainty, that was always some distance from the anterior end. In no case were they found attached at the extremity, as figured by Belajeff in his Fig. 11. At this stage the entire protoplast has become more contracted. The diameter of the nucleus is also shorter, but the difference between that in Fig. 1 and in Fig 4 is due in some measure to a difference in the size of the cells. The surface of the cytoplasm at the point between the ends of the blepharoplast now begins to become concave (Figs. 6, 8), although this substance assumes, at a later stage, the form of a rounded vesicle (Figs. 9-13). A marked change is also manifested by the nucleus. It loses its spherical or elliptical shape, becoming denser and flattened on one side (Fig. 8), and assuming the form of a half moon or crescent. In structure it is dense and uniform. Both posterior and anterior ends of the blepharo- plast have become longer and thicker, projecting much beyond the cyto- plasmic mass of the cell. At a later stage (Figs. 11 and 12) the nucleus increases in length, becoming sausage-shaped, and makes about one complete turn in a spiral. In Fig. 13, the body of the sperm makes about three turns, the anterior end describing sharper curves than the posterior end. The cytoplasm of the cell now consists of a finely granular mass 250 Mottier . — Development of the Spermatozoid in Char a. embraced largely by the nuclear portion of the spermatozoid (Figs. 9, 11, 13). The blepharoplast is of a uniformly homogeneous structure, staining well with gentian violet of the triple stain. Along the concave side of the posterior end is the conspicuous row of granules mentioned in a preceding paragraph. These granules stain densely, appearing almost black in preparations otherwise well stained to bring out the remaining parts distinctly. Fig. 10 a and b represent two spermogenous cells as seen from the side (in lateral view) at the stage of development shown in Fig. 9 : a represents the sperm as seen from the surface, while b is an optical section. The dots on the right and left of the body of the sperm represent cross-sections of the cilia. In b , it is seen that the nucleus is cylindrical, being circular in section. The cytoplasmic mass is strongly concave on the side opposite the nucleus. In this concavity are shown two crescents which are sections of the ends of the blepharoplast. The larger crescent, above in the figure, is a section of the posterior end, and the lower that of the anterior end. From optical sections it is clearly seen that the blepharoplast is a band with the outer surface convex and the inner surface somewhat concave. The section of the blepharoplast at the left of the nucleus shows that the cilia-bearing band is closely applied to the nuclear membrane at the place of contact. The cilia are relatively very long, and lie rather loosely coiled in the spermogenous cell. Sometimes they seem to lie in contact with each other for certain distances (Fig. 14). In the final stages of the development of the spermatozoid the general cytoplasmic portion seems to be reduced to a mere vestige. - Unfortunately material was not available at the time of my study to enable me to fix the living spermatozoids upon the slide, and to stain them in that condition. This I hope to be able to accomplish sometime in the near future. In discussing the probable relation of the blepharoplast in Chara and in some of the lower Archegoniates, the observations recently published by Ikeno (’03) upon spermatogenesis in Marchantia polymorpha are of special interest. His excellent paper is the most thorough and complete account thus far given for any Liverwort. Only those who have attempted cyto- logical work in these plants can fully realize the difficulties encountered, because of the small size of the spermogenous cells, and the difficulty with which cytoplasmic structures, especially, are differentiated. Ikeno begins his study with the last two or three mitoses in the sper- mogenous tissue. In these mitoses he finds that centrosomes are present during certain stages, and that this structure functions finally as a cilia- bearer. In 1898 the writer (Mottier, ’98) called attention to the presence of centrosomes in the vegetative cells of the gametophyte of Marchantia polymorpha , and further study was made by Van Hook (1900) with similar Mottier . — Development of the Spermatozoid in Char a. 251 results. Although we found such a body present during certain stages of mitosis, yet we could not trace it from one cell-generation to the next, and our conclusion was that it was only a temporary organ. There is no question but that Ikeno has observed a similar phenomenon in the sper- mogenous cells, for in all known cases wherever centrosomes are present in vegetative cells, they are to be observed also in reproductive cells of the same organism. The behaviour of these structures in the spermogenous tissue, as described by Ikeno, seems to be similar to that which I have mentioned for the vegetative cells; for Ikeno (’03, pp. 70, 71) states explicitly that, in the ‘ aster ’ and c diaster * stages of mitosis, the cen- trosomes are only occasionally to be observed, and that they disappear at the end of each mitosis and reappear again at the beginning of the next successive karyokinesis. In the last division of the spermogenous tissue the cubical cells divide obliquely, and the two resulting cells, which are triangular in shape, are not separated by a cell-wall. This division differs from the preceding division in that the centrosomes are always present (loc. cit., pp. 74, 75), so that each triangular protoplast, which is to become transformed into a spermatozoid, has a centrosome that eventually passes into an angle of the cell (loc. cit., Figs. 22-26). This centrosome now elongates somewhat and places itself in close contact with the inner contour of the cell, so that it appears to have arisen as a local thickening of the plasma-membrane (Hautschicht). Out of this elongated centrosome the two cilia now grow. As soon as the cilia have begun their development, a thread or band-like cytoplasmic differentiation appears, which grows in the direction of the anterior end of the future sperm, and finally connects the nucleus with the centrosome (loc. cit., Fig. 31). According to this statement, therefore, the centrosome-like body gives rise to the cilia only, whereas the specially differentiated cyto- plasmic thread, or band, is of another origin. £ Bald nachdem die Cilien sich zu entwickeln begonnen haben, rundet sich die Spermatide ab (loc. cit., Figs. 29, 30), . . . der cytoplasmatische Fortsatz beginnt sich auszubilden und wachst nach der Richtung des vorderen Endes des jetzt sich bildenden Spermatozoids hin, um schliesslich das Zentrosom zu erreichen, so dass dieser Fortsatz den letzteren mit dem Zellkern verbindet (loc. cit., Fig. 31).’ This statement does not harmonize with Ikeno’s figures. It would seem from his Figs. 30 and 31 that the ‘Fortsatz5 mentioned, connecting nucleus with the elongated, cilia-bearing band is derived from the cilia- bearer itself, as represented in his (loc. cit.) Figs. 29 and 30, and not as a separate formation. If Ikeno’s statement be correct, then the development of the specialized cytoplasmic band, which is known as the blepharoplast, differs in Mar- chantia from that known in all other Archegoniates, for, in the Ferns and zooidogamous Gymnosperms, the centrosome-like body gives rise 252 Mottier. — Development of the Spermatozoid in Char a. directly to this band. Ikeno figures in colours a mature spermatozoid. in which the central, or nuclear portion, is coloured green, while the anterior part with the cilia and the posterior end are coloured red. I do not find in his text any statement concerning the development of the posterior extremity which is undoubtedly of cytoplasmic origin. Ikeno discusses at some length the question of the probable homology of the cilia-bearer, or blepharoplast, and the centrosome, concluding that these structures are homologous in the phylogenetic sense. In the elabora- tion of his argument, it seems to me that he ignores or leaves out of con- sideration the most fundamental principle, namely, that structures to be homologous must have, at least, morphological rank in the cell, i. e. they must be organs in the morphological sense. As a matter of fact the centrosome is not an organ of the cell with morphological significance. In Char a , certain Pteridophytes and spermogenous Gymnosperms, where the development of the blepharoplast is best known, there are no centro- somes with which to homologize these structures. Even in Marchantia where there seem to be both centrosome and blepharoplast, it has not been shown that the former is a permanent structure in the cell, since it cannot be followed from one cell-generation to the next. Ikeno ex- pressly states that, in the aster and diaster stages of mitosis, the centrosome is only occasionally to be seen, and in this he is in accord with the observa- tions of the writer and Van Hook. Ikeno has, of course, brought forth in Marchantia the strongest evidence that has, as yet, been advanced, pointing to a relationship between centrosome and blepharoplast, but this evidence cannot be accepted as final. The fact that two bodies look alike and stain alike in different cell-generations, is not conclusive proof that they are the same. Personally I cannot admit that all the bodies that Ikeno figures in cells of Marchantia are centrosomes, as similar bodies with exactly the same behaviour have been found in cells of plants in which it is known with absolute certainty that centrosomes, as understood among plants, do not exist. If, on the contrary, we attribute a morpho- logical rank and phylogenetic relationship to the substance of which centro- somes and blepharoplasts are made, such as kinoplasm, or whatever we wish to call it, and if we admit that a morphological differentiation exists in the cytoplasm — a view that has much in its favour — then our theory will bring all known facts into line. Otherwise it seems that we shall be obliged to be contented with facts as they are known, and patiently await other facts and unquestionable evidence. Summary. The spermatozoid of Chara is a spirally coiled body consisting of a nucleus and a specially differentiated part of the cytoplasm which exists in the form of a thread or band, the blepharoplast, and bears two long Mottier. — Development of the Spermatozoid in Chara. 253 cilia. The nucleus occupies the middle part of the spermatozoid, while the blepharoplast extends its entire length. The anterior end of the blepharo- plast is thinner than the posterior. The two cilia are borne some distance back of the anterior end. The posterior portion is thicker, and usually ends bluntly. In cross section the blepharoplast is crescentic, being convex on the outside and concave within. It is of a homogeneous structure, except- ing a strip of granular substance along the concave side of the posterior end. The entire spermatozoid, as far as I was able to observe, makes two and one-half or three turns in a spiral. The blepharoplast arises as a delicate thread-like differentiation of the cytoplasm at the surface of the cell, extending some distance along the cell from the nucleus and on opposite sides of the latter. Whether the terminal portions of this thread extending from the nucleus originate as a single piece, or as two pieces, could not be determined with certainty in the earliest stages, but, later, the blepharoplast is clearly seen to be one piece, extending the entire length of the sperm. The blepharoplast seems to be a modification of the plasma-membrane, i. e. it did not seem to be formed within or without this membrane, but as a direct transformation of it that increased in thickness inwardly. No centrosome-like body, or 4 Plasmahocker/ was observed from which the cilia develop, as described by Belajeff, Strasburger and others. The nucleus is transformed from an elliptical or oval body, with a hollow chromatin spirem, to a dense, homo- geneous, sausage-shaped structure making one or more spiral turns. The cilia were always found attached some distance back of the anterior extremity of the blepharoplast. They did not seem to grow at the expense of a protuberance or centrosome-like body. Literature Cited. Belajeff, W., 1894 : Ueber Bau und Entwickelung der Spermatozoiden der Pflanzen. Flora 1-48, 1894. Van Hook, J. M., 1900 : Notes on the Division of the Cell and Nucleus in Liverworts. Bot. Gaz., xxx, 394-399* I9°°- Ikeno, S., 1903 : Beitrage zur Kenntnis der pflanzlichen Spermatogenese : Die Spermatogenese von Marchantia folymorpha. Beihefte zum Bot. Centralblatt, xv, 65-88, 1903. Mottier, D. M., 1898 : The Centrosome in Marchantia. Proc. Indiana Acad. Sci., 1898. Strasburger, E., 1892 : Histologische Beitrage, iv, 107, 1892. 254 Mottier. — Development of the Spermatozoid in Char a. EXPLANATION OF FIGURES IN PLATE XVII. Illustrating Professor Mottier’s paper on the Spermatozoid of Chara. All figures were drawn with the aid of the camera lucida and with Zeiss apochromatic homo- geneous immersion 2 mm., apert. 1-40 with compensating ocular 8. All are magnified about 2,000 diameters. Fig. 1. Sporogenous cell seen from the end at the beginning of the development of the spermatozoid. The blepharoplast appears as a heavier line at the surface of the protoplast above and below the nucleus ; the anterior end being on the side below the nucleus. Fig. 2. Two spermogenous cells in about the same stage of development as Fig. 1, seen from the side, or in longitudinal section. The blepharoplast is seen along the lower transverse wall in each cell. The cytoplasm has accumulated into a denser mass along the side next the blepharoplast. Fig. 3. A median longitudinal section of a cell in the same stage as Fig. 2, passing through the cell in the plane a-b of Fig. 1. The blepharoplast is, therefore, not visible. Figs. 4, 5. Later stages in the development as seen from the end. The blepharoplast is well differentiated and extends almost or quite around the nucleus. The posterior end, which is on the upper side in the figure, is seen to be much broader than the anterior end ; its extremity projects as a beak beyond the surface of the protoplast. The anterior end is a very delicate thread curved or hooked at its extremity. In Fig. 5, the blepharoplast is separated from the nucleus, probably on account of shrinkage, and it is seen to be a continuous thread. The cilia are present at this stage. Figs. 6, 7. Similar to the preceding. Fig. 6 shows the mode of attachment of the cilia ; in Fig. 7 only a part of one cilium is shown. Fig. 8. A later stage. The spermatozoid has begun to assume its characteristic spiral form. The ends of the blepharoplast do not lie in the same plane, and were drawn by changing the focus. The nucleus is now in the form of a half moon or crescent, and the cytoplasm is correspondingly concave. The cilia lie in contact for a part of their length. Fig. 9. A later stage than Fig. 8. The nucleus has become sausage-shaped. The granular cytoplasm is embraced by the middle portion of the sperm. Fig. 10 a and b. Optical longitudinal sections of two cells in about the stage of Fig. 9. a is a surface view, and b an optical section. The dark dots in the cells on the right and left of the sperm represent sections of the cilia. In b the cytoplasm is concave at the right to correspond to the concave side of the sperm. The two crescents seen in this concavity are the transverse sections of the ends of the blepharoplast, that of the posterior end being above. These ^sections show that the blepharoplast is a convexo-concave band. At the left of the nucleus is seen a section of the blepharo- plast closely applied to the nuclear membrane. Figs. 11, 12, 13. Successively older stages. The sperm has now become spirally coiled by the growth or extension in length of both nucleus and blepharoplast. The cytoplasmic vesicle adhering chiefly to the nuclear portion is being gradually diminished. Fig. 14. A nearly ripe spermatozoid. The spermatozoids are usually more closely coiled up than this one. The granular cytoplasm is more reduced. g 'fhmals of Botany. Voimu^pimi. Fig.7. Fig. 1. Fie A. O University Press, Oxford. M OTHER. -SPERMATOZOID OF CHARA. A Mycorhiza from the Lower Coal-Measures. BY F. E. WEISS, D.Sc., F.L.S., Professor of Botany in the Victoria University of Manchester. With Plates XVIII and XIX and a Figure in the Text. ALL who have investigated the microscopic structure of fossil plants are familiar with traces of fungal hyphae and occasional fungal sporangia in and around the plant-remains. An excellent critical account of our knowledge of such fossil Fungi will be found in Seward’s ‘ Manual of Fossil Plants ’ (’98), in which he has not only recorded the Fungi de- scribed by Williamson, Renault, Conwentz, and other observers, but discusses their possible systematic position. In summing up our knowledge of this group of plants, he remarks that ‘we have fairly good and conclusive evidence of the existence in Permo-Carboniferous times of Phycomycetous Fungi.’ Judging from the appearance of the tissues in which these Fungi are found, one is led to the conclusion that they were for the greater part of a saprophytic nature. This would seem more particularly so in the case of the fossil plants from the English Coal-Measures, the internal structure of which is so fully known from the remains found in the nodular concretions, the so-called ‘coal-balls’ of the Bullion Coal. In these coal-balls, which, according to Lomax (’02), were probably not formed in situ, the plant- remains are often of a very fragmentary character, and show traces of having undergone considerable decomposition. The tissues are often penetrated by Stigmarian rootlets, and show signs of having been bored by wood-eating animals. They also show not infrequently internal mycelia, while apparent fungal sporangia are found both within the fossil plants and in the debris lying between them. Indeed, the conditions under which these nodules were formed would seem to have been most favourable for the growth of saprophytic Fungi. Some of the fossil Fungi, however, from the silicified nodules at Grand Croix which have been described by Renault (’83) and Bertrand (’85), and more recently by Oliver (’03), seem to have been of a parasitic nature and to have belonged probably to the group of Chytri- diaceae. One form, indeed, which appears to have been parasitic on the Annals of Botany, Vol. XVIII. No. LXX. April, 1904.] 256 Weiss.— A Mycorhiza from the Lower Coal-Measures . fronds of Alethopteris aquilina , Magnus (’03) considers sufficiently like the recent form Urophlyctis to warrant its inclusion in that genus. From the observations so far made we are able to picture to a certain extent the modes of life of the Fungi in Palaeozoic times, and we come to the conclusion that they differed very little from those of recent Fungi. But, besides leading saprophytic or parasitic existence, Fungi are at the present day also found living together with green plants in a state of sym- biosis, in which they do not destroy the tissues of the green plant, but seem rather to be of some use to it, while at the same time they derive some benefit from the green plant. Living in such mutual relations with algal cells, the Fungi form the group of organisms known as Lichens, while when they inhabit the roots or root-stocks of many higher plants they form the so-called mycorhiza, the significance of which is still under discussion. Remains of Lichens are, according to Schimper and Schenk (’90), known from the Tertiary period, some of them being preserved in Amber ; but none have so far been recorded from Secondary rocks. Mycorhizae, on the other hand, have to my knowledge not been described in a fossil condition. Yet, at the present time this peculiar association of Fungi with the roots of higher plants is a fairly widespread phenomenon. This is perhaps more particularly the case in tropical forests, where, according to Janse’s investiga- tions, sixty-nine plants out of seventy-five chosen from various divisions of the vegetable kingdom had their roots inhabited by apparently symbiotic Fungi. Whatever may ultimately turn out to be the significance of these endo- phytic Fungi, there can be no doubt that this mutual adaptation represents a considerable specialization of the two organisms forming the Mycorhiza. The latter might, therefore, be expected to have arisen at a comparatively recent period in the evolution of plants. But apparently this form of symbiosis is of considerable antiquity, for it seems to have existed as far back as the Lower Coal-Measures. I am conscious that this announce- ment will very naturally meet with some scepticism ; nevertheless, I venture to think that the evidence which will be brought forward in the following pages warrants the conclusion that this highly specialized mutual adaptation of Fungus and cormophyte did actually exist in the Palaeo- zoic age. The root or rhizome in question I found on two slides in the Cash collection of Coal Measure plants in the Manchester Museum, Owens College (slides No. Q527 and Q539), both from the Halifax Hard Bed, which, according to Binney (’62), must be correlated with the Bullion Mine of the Burnley district and the Gannister Mine of Dulesgate, Todmorden. These three seams of the Lower Coal-Measures are characterized by the possession of the nodular concretions (coal-balls) referred to above, and it is from one of these that the preparations were made. On communicating my view of the nature of these specimens to Dr. Scott, he very generously placed at my disposal Weiss . — A Mycorkiza from the Lower Coal-Measures. 257 a similar specimen from the collection of the late Mr. James Spencer, of Halifax, which he had purchased some years ago. It turned out to be undoubtedly of the same nature as the specimens in the Cash Collection, and is in all probability from the same locality, as most of Mr. Spencer’s material came from the Halifax Hard Bed. The Host Plant. The only remains that we have of the host plant are a few delicate root-like organs ranging from about 1 to 2 mm. in thickness. In all the specimens so far discovered the tissues are well preserved in comparison with the surrounding plant-remains, which have for the most part under- gone considerable destructive changes. Among the debris are seen spore- cases, apparently of a fungal nature, and numerous opaque, rounded masses (see PL XVIII, Fig. 1), which are generally taken to be excrements of wood- boring animals. The whole has the aspect of a mass of humus in which Fungi and animal organisms were causing a gradual breaking-down of the vegetable remains. Only a few hard macrospores, some Stigmarian rootlets, and the mycorhiza in question seem unaltered — a fact which is suggestive of their having grown in the humus-like mass. The internal structure of the mycorhiza at once suggests a root of the diarch type, but differs in several respects from other diarch roots found among Coal Measure plants. It should be added that these root-like remains are very constant in character, and therefore easily recognizable. As seen in PI. XVIII. Fig. 1, the root has two groups of wood distinctly separated by well-marked ground-tissue cells. These are present in all the specimens, and as the larger ones are evidently mature we may consider that the xylem-rays did not meet in the adult organ, as is generally the case in the roots of Ferns. In a smaller specimen than that in Fig. 1 these medullary cells are filled with curious granules (PI. XVIII, Fig. 2, and Text-fig. 42), the nature of which is uncertain. At first sight they closely resemble a number of starch-grains, but it is somewhat doubtful whether starch would be pre- served with quite so definite an outer boundary. On the other hand, they do not seem in any way connected with the Fungus, which does not as a rule penetrate to such depths ; nor does the Fungus in other parts of the plant, where it occurs, produce granules quite of this kind. It would, therefore, seem more probable that the granules are the normal cell-contents at a certain stage in the development of the organ or under certain conditions of nutrition. With regard to the xylem-groups, it is not always easy, or indeed possible, to distinguish the protoxylem-elements. In Fig. 2 the smaller elements seem to be on the outside of the wood ; but, as will be seen from Fig. 1, this is not always the case. In one group, indeed, the smaller elements appear to be on the inside. Their position is in fact irregular, and 258 Weiss. — A Mycorhiza from the Lower Coal-Measures. in that respect, as well as in others to which reference will be made later on, the stele resembles somewhat that of the rhizome of Psilotum. It is true that in this latter plant the xylem-rays, whether two or more, are generally connected in the centre ; but in certain parts of the plant, accord- ing to Bertrand’s (’82) figures and description, the xylem- groups may remain separate. Owing to the uncertainty as to the position of the proto- xylem, I prefer, therefore, to leave it an open question whether the organ under consideration was a root or a rhizome. If the latter, then it must have Fig. 42. Enlarged drawing ot stele from specimen represented in Fig. 2, PL XVIII. The protoxylem elements (fix) are on the outside of the groups of wood. The ground tissue is filled with granules (starch ?). The stele is surrounded by a distinct endo-cortex (eri) within which is a phloem-sheath L? .). been a leafless rhizome of the type found in the Psilotaceae or in Corysantkes, Corallorhiza , and some other Saprophytes. In longitudinal section, as seen in Plate XVIII, Fig. 3, which is taken from Dr. Scott’s specimen, it will be seen that the xylem consists mainly of scalariform tracheids. I have in fact not been able to discover any other elements in the longitudinal section ; but it is of course possible that the latter did not pass through the spiral elements if such were present in the plant. There are no very clearly-marked phloem elements, though certain cells on either side of the xylem-groups might be considered as bast cells {ph). In this respect we have another agreement with Psilotum and also with other Lycopodiaceous plants in which the phloem is not made up of well-defined sieve- tubes. Similarly in some cases, as in Plate XVIII, Fig. 2, Weiss.- — A Mycorhiza from the Lower Coal-Measures. 259 there is an indication of a phloem-sheath (fl.s.) or pericycle. The charac- teristics of these tissues are not well pronounced, and except for the xylem-groups the structure of the stele is somewhat unsatisfactory. The cortical tissues are well developed and consist of thin-walled cells. There is no such thickening of the walls as one frequently finds in the roots of Filicineae, nor any such specialized lacunar tissue as is typical of the stem and roots of most of the Lycopodiales, particularly of the Lepidodendraceae. It will be seen from Fig. 1 and Fig. 3 that some of the inner layers of the cortex are slightly more elongated in the radial direction, while the outer layers appear in the transverse section somewhat flattened owing to the tangential extension of the cells. It will be convenient, therefore, to use the terms exo- and medio-cortex for these two portions of the cortex in the same sense in which they have been used by Groom (’95) in his descriptions of the roots of monocotyledonous saprophytes where they show a similar differentiation. The difference between these two layers of the cortex can most readily be seen in the longitudinal section (Fig. 3) in which it will be further seen that the innermost row of cortical cells (endo-cortex, possibly endodermis) is drawn out longitudinally, while the tangentially and radially elongated cells of the exo- and medio-cortex are short, more particularly the latter. The radially elongated cells, as can be seen both in the transverse and in the longitudinal sections (Figs. 1 and 3), are also characterized by their very dark contents. These will be described in detail later on, but it may be mentioned at present that they show indications of fungal hyphae and closely resemble in their appearance and in their position in the cortex the curious contracted masses (clumps) described by various authors in the aerial roots of Orchids and in the absorptive organs (roots or rhizomes) of saprophytic Monocotyledons and of P silo turn. The exo-cortex, though also containing hyphae, possesses none of these ‘ clumps/ and consequently looks at first sight devoid of the Fungus. This specialization of the hyphae in two different regions of the cortex is a very common phenomenon in mycorhizae and is constantly met with in the plants mentioned above. Fungal hyphae are very rarely met with in the endo-cortex. Only in one transverse section were a few hyphae discovered in this layer. This, again, is in conformity with the behaviour of the fungal mycelium in the mycorhizae of living plants, and supports the conclusion I have arrived at, that the Fungus is not of a destructive nature ; or it would probably have penetrated into all the living tissues of the plant. The cells of the epidermis are smaller in size than the cortical cells as can be seen in Plate XVIII, Figs. 1 and 3, and are often drawn out into long absorptive hairs (Plate XIX, Fig. 2). This piliferous layer may possibly be regarded as evidence of the root-nature of the organ, but it 260 Weiss.— A Mycorhiza from the Lower Coal-Measures. must be remembered the rhizome of Psilotum and Tmesipteris , as well as some of the modified absorbing rhizomes of saprophytes, bear such hairs. It is indeed difficult, as Groom (’95, II) has shown, to distinguish between roots and rhizomes in these plants, as they may be very considerably altered both externally and internally. Not only may the leaves be considerably atrophied but, as Groom points out, cthe structure of the stele in absorbing rhizome-axes of hemi- and holo-saprophytes is frequently remarkably like that of a root (Corysanthes, Burmannia , Corallorhiza :), so the root-like structure of the stele of the absorbing organ is no proof of its root-nature/ It would seem, therefore, best to suspend our judgement with regard to the nature of the organ in question, and similarly we shall have to remain in doubt as to the systematic position of the plant until other portions of it have been discovered. The Fungus. The general distribution of the mycelial threads within the host-plant has been mentioned above as well as the different appearance of the Fungus in the exo-cortex and medio-cortex respectively. In the epidermis too hyphae may be noticed, but their occurrence in a cell does not necessarily mean that this has been a centre of infection. Sometimes, as in the cell bearing a root-hair in Plate XIX, Fig. 2, hyphae may be running longitudinally through the cell. Sometimes, however, numerous hyphae are seen running radially across the epidermal cells ( h in Fig. 2) ; but in no case have I been able to trace them beyond the outer wall. Even in such cases, however, we have no proof that it was through these cells that the Fungus had gained admission, as it is known that hyphae may grow out from the mycorhiza into the surrounding medium. Thus Groom (’95, II), in his description of Thismia , mentions (p. 354) that it is possible to observe ‘ that frequently free hyphae are deserting, not entering, the host-plant.’ Whether the particular hyphae shown in Fig. 2 are entering or leaving the tissues must of course remain an open question in the case of a fossil plant. In the outer cortical layers the course of the Fungus is somewhat irregular, both horizontally and vertically running hyphae being met with. On the whole, however, the mycelium seems to grow along the mycorhiza just as Janse (’97) observed in many of the roots which he investigated. Plate XIX, Fig. 1, which is a portion of the transverse section represented on Plate XVIII, shows the majority of the hyphae cut more or less trans- versely, and also exhibits them running, as they seem generally to do, close along the inner face of the cell-walls. This is often found to be the case in the living mycorhizae, where, however, the Fungus often forms more or less definite coils on the inside of the cell-walls. From these thicker coiled hyphae very thin haustorial filaments are sent into the Weiss. — A Mycorhiza from the Lower Coal-Measures. 261 cell-contents in ' Neottia , Psilotum and other plants according to Werner Magnus (’00) and Shibata (’02). But as these are very delicate, and soon disappear, we should not expect to find them in the exo-cortex of the fossil plant. Otherwise the latter presents very much the same appearance as it does in recent plants. In one or two cases the numerous hyphae attached to the inner walls of these cortical cells remind one of the mycelial pegs described by Groom (’95) in the roots of Galeola. For the most part the hyphae seem to be intra-cellular, but there are indications that a few of them run between the cells. In no case, how- ever, is there any sign that the Fungus was in any way destroying the tissues of the host-plant. In some of the cells of the exo- cortex curious pear-shaped bodies are found at the ends, or apparently at the ends, of certain hyphae. These may be fairly numerous, as in Plate XIX, Fig. 3, or there may be only two or three in a cell. They are generally most numerous in the sub-epidermal cells, while in the deeper layers of the exo-cortex they are fewer in number, larger in size, and more rounded. They resemble somewhat the pear-shaped swellings described and figured by Williamson (’81) on the hyphae of Peronosporites antiquarius , a Fungus found in the bark of Lepidodendron. Such pear-shaped bodies are, however, of very common occurrence in the outer cortical layers of recent myco- rhizae. Thus in Psilotum , Janse (’97) figures a dozen of them in a sub- epidermal cell, just as they occur in the fossil mycorhiza. The nature of these vesicles, formed by endophytic Fungi, is a much-disputed point. Some authors look upon them as possibly of the nature of reproductive organs. Thus Bruchmann (’85) considered that they might possibly be oosporangia, while Goebel (’87) thought they might be gonidia (Dauer- gonidien) in the case of the Fungus inhabiting Lycopodium, , which they supposed to be related to Pythium. Groom (’95), on the other hand, who made a very careful examination of these bladders and their mode of formation in the mycorhiza of Thismia , found that they were not terminal but intercalary dilatations, though appearing terminal by hypertrophy. He is consequently inclined to attribute a nutritive function to them, and regards them as of importance in increasing the absorptive area of the Fungus which is supposed to feed upon the host-plant in the outer cor- tical layers. This view of the purely vegetative character of the vesicles had also been entertained by Mollberg, who was apparently the first observer of these intercalary hypertrophies of the Fungus (in Platanthera and Epipactis ). Groom found that the vesicles accumulated a large amount of densely- staining cytoplasm, which afterwards became vacuolated and diminished in amount, ultimately degenerating and depositing ‘ a homogeneous yellow substance in which are rod-like bodies which remind one of the regular rod-like bacteroids in leguminous tubercles.’ In the fossil plant the vesicles, whether large or small, are generally 262 Weiss . — Mycorhiza from the Lower Coal-Measures. devoid of contents, but a few of the larger more rounded ones show homo- geneous contents. In one or two instances vesicles were found among the cells containing the dark clumps and these contained what appear to be spores (Plate XIX, Fig. 6), but as they are really in the medio-cortex it is possible that they were formed in a different manner from the vesicles described above. In some cells lying near those with large vesicles (see Plate XIX, Fig. 5) there are found curious granules distributed very evenly through the cell and apparently attached to the cell-wall. The nature and formation of these I was not able to elucidate from the specimens at my disposal. In the medio-cortex, as described above, we find the characteristic clump formation (see Plate XVIII, Figs. 1, 3, and Plate XIX, Fig. 4) — the clumps consisting no doubt partly of the cell-contents, partly of fungal filaments ; but they are as a rule so dark in colour that no details of their structure can be made out. They are connected to the cell-walls by threads, which are sometimes very delicate and appear as if they were protoplasmic filaments, though they are probably contracted hyphae, as these can in some cases be seen very clearly as shown in Plate XIX, Fig. 4. These hyphae are, however, usually thinner and more delicate than those in the outer layers of the cortex. In this particular the fossil mycorhiza agrees with recent ones in which various observers have noted this difference. The fossil mycorhiza can of course give us no clue as to the formation and significance of these clumps, but their excellent preservation in a fossil condition may be considered to support the view of Werner Magnus (’00), that they consist of the non-digestible and unalterable remains of the Fungus after the host-plant has derived from it all possible nutriment. For the fact that these clumps are so well preserved would indicate that the Fungus had passed into an unalterable condition before fossilization. Should the same degenerative changes have taken place in the medio- cortex of the fossil plant as take place in recent mycorhizae these would readily explain such appearances of degeneration as one meets with in some of the specimens. Thus in Fig. 8 will be seen curious vacuolated masses which, as indicated by the presence of delicate radiating filaments, are probably produced in a similar way to the mycelial clumps. They obviously correspond to the so-called ‘ traubenformige Korper ’ described by Bernatzky (’99) in the rhizome of P silo turn. Ultimately they would seem to break up into separate particles not unlike bacteroids (Fig. 7), but which may also be compared to the curious ‘ Eiweisshyphen ’ described by Magnus (’00) in some of the cells of the medio-cortex (Verdauungszellen) in which the host-plant is digesting the Fungus. The obvious resemblance between these clumps in the fossil plant and those of recent mycorhizae, together with the close agreement in the structure and behaviour of the Fungus in the outer layers of the cortex, Weiss.— A Mycorhiza from the Lower Coal-Measures . 263 with those of the Fungus in recent mycorhizae will, I think, be regarded as sufficient evidence for the conclusion that we are dealing in the case of this fossil plant with a mycorhiza or my corhizome. The Fungus differs materially in its manifestations from other cases of endophytic Fungi so far observed in fossil plants, and in no way suggests that it was living either sapro- phytically or parasitically upon the host-plant. The excellent preservation of both the Fungus and the host-plant and the specialization of the cortex into two layers comparable with the ‘ Pilzwirthzellen ’ and £ Verdauungs- zellen 1 of recent mycorhizae, would suggest that, as in the case of the latter, the host-plant is deriving some benefit from the presence of the Fungus. We cannot of course expect from the investigation of a fossil mycorhiza to elucidate the difficulties that surround the mycorhiza-question, but it is of no little interest to find that already at the Coal-Measure period Fungi and cormophytes exhibited a mutual adaptation of such com- plexity as that involved by the formation of a mycorhiza. Of the systematic position it is difficult to say much on the slender evidence before us. I have not been able, in any of the longitudinal sections, in which one sees occasionally considerable lengths of the Fungus, to detect any transverse walls, and this would incline me to the belief that the Fungus belonged to the group of Phycomycetes. This view would be supported by the fact stated by Seward (’98), that the Phycomycetes certainly existed in Permo-Carboniferous times. Among recent endo- phytic Fungi showing symbiotic adaptation some apparently also belong to this group, according to Bruchmann (’85) and Goebel (’87). The systematic position of the host-plant is almost as difficult to determine as that of the Fungus. Leaving out of consideration the Gymnosperms and Cycadofilices with which it shows no affinities, we may confine ourselves to the consideration of the Vascular Cryptogams. Of the four divisions of these it seems, as mentioned above, to have most affinity to the Lycopodiales, though it differs considerably from most of these. It has not the specialization of the cortical tissues characteristic of the Lepidodendraceae and, if it is a root, does not possess the usual monarch arrangement found in that group. The absence of large inter- cellular passages such as are found in the roots of Lepidodendraceae and Calamarieae would lead us to infer that it was not rooted in marshy ground as were probably these larger fossils, and its association with a P'ungus would rather point to a saprophytic existence in rich leaf- mould or to an epiphytic existence like that of Tmesipteris on the stems of Tree-ferns. In either case it would be likely to become infected with fungal hyphae, and might develop the special adaptation which its mycorhiza exhibits. That a mycorhiza has not been found in other Coal-Measure plants should not astonish us when we remember that T 264 Weiss. — A Mycorhiza from the Lower Coal-Measures . the greater number of Lycopodiales and Equisetales were probably rooted in marshy places in which the conditions would not be very favourable to the formation of mycorhizae. In our present state of uncertainty as to the nature of the host-plant, of which the root or rhizome only is known, it would not be advisable to give it more than a provisional name, merely to facilitate reference to the specimen. For such use it would be best to employ a non-committal designation, like that of Rhizonium — which was invented by Corda (’45) to describe certain roots which he took to be those of an Orchidaceous type, and which was afterwards used by Williamson (’89) for the roots of an unknown plant. In consideration of the peculiar character of the fossil described above, I would suggest that, until we know more about the plant to which it belonged, it should be referred to as Mycorhizonium. List of Works referred to. Bernatzky, ’99 : Beitrage zur Kenntniss der endotropen Mycorhizen. Term^szetrajzi Fiizetek, 1899. Bertrand, C. E., ’82 : Recherches sur les Tmesipteridees. Archives botaniques du Nord de la France, 1882. ’85 : See Renault and Bertrand. Binney, E. W., ’62 : Trans, of the Manchester Geol. Soc. 1862, vol. vi. Bruchmann, H., ’85 : Das Prothallium von Lycopodium. Bot. Centralblatt, Bd. xxi, 1885. Corda, A. J., ’45 : Beitrage zur Flora der Vorwelt, 1845. Goebel, K., ’87 : Ueber Prothallien und Keimpflanzen von Lycopodium inundatum. Bot. Zeit., 1887. Groom, P., ’95, I: Contributions to the knowledge of Monocotyledonous Saprophytes. Journal of the Linn. Soc., 1895. ’95, II : On Thismia Aseroe and its Mycorhiza. Annals of Bot., ix, 1895. Janse, J. M., ’97 : Les endophytes radicaux de quelques plantes javanaises. Ann. du Jard. de Buit. xiv, 1897. Lomax, J., ’02 : On the occurrence of nodular concretions (Coal Balls) in the Lower Coal Measures. Artn. of Bot., xvi, 1902. Magnus, P., '03 : Ein von F. W. Oliver nachgewiesener fossiler parasitischer Pilz. Ber. d. deutsch . Bot. Ges., xxi, 1903. Magnus Werner, ’00: Studien an der endotropen Mycorhiza von Neottia Nidus avis. Jahrb. f. wiss. Bot., 1900. Oliver, F. W., ’03 : Notes on fossil Fungi. New Phytologist, vol. ii, 1903. Renault, B., ’83 : Cours de Bot. Fossile, vol. iii, 1883. Renault, B. and Bertrand, C. E., ’85 : Grilletia sphaerospermi, Chytridiac^e fossile du terrain houiller supdrieur. Comptes rendus, tome c, 1885 . Schimper, W. P„, and Schenk, A., ’90 : Palaeophytologie. In Zittel’s Handbuch der Palaeontologie, 1890. Seward, A. C., ’98 : Fossil Plants I, 1898 : Cambridge Nat. Science Manuals. Shibata, K., ’02: Cytologische Studien iiber endotrophe Mycorhizen. Jahrb. f. wiss. Bot., 1902. Williamson, W. C., ’81 : Organization of the Fossil Plants of the Coal Measures, Part XI, 1881. ’89: „ „ „ „ Part XV, 1889. Weiss. — A Mycorhiza from the Lower Coal-Measures . 265 EXPLANATION OF PLATES XVIII AND XIX. Illustrating Professor Weiss’s paper on a Mycorhiza from the Coal-Measures. (Photographs by Mr. Abraham Flatters, Manchester.) PLATE XVIII. Fig. 1. Transverse section of a mycorhiza from the Cash Collection in the Manchester Museum (Q. 827), about 1 mm. diameter. Above the section are seen some dark oval masses, probably the excrements of a wood-eating animal. In the exo-cortex (ex) traces of the fungal hyphae can be seen projecting from the cell-walls into the cell-spaces. In the medio-cortex (me) are seen dark masses (clumps) consisting largely of fungal remains. Fig. 2. Vascular cylinder of another mycorhiza 2 mm. diameter from the same slide as Fig. 1 (Q. 827), showing a well-marked endo-cortex (en). Between the two groups of wood-elements are a few parenchymatous cells containing numerous granules, ps = phloem-sheath. Fig. 3. Longitudinal section of a mycorhiza from Dr. Scott’s preparation (No. 1527), showing the vascular cylinder, the endo-cortex (en), the radially-elongated medio-cortex ( m.c ) with dark mycelial clumps, the exo-cortex (ex) with scattered hyphae. Fig. 4. Portion of the medio-cortex from a tangential longitudinal section on Dr. Scott’s slide (No. 1527), showing the dark mycelial clumps (cl). An enlarged drawing of a portion of this photograph is shown in Fig. 4, Plate II. PLATE XIX. Fig. 1. A portion of the external tissues of the root figured in Plate I (Cash Collection, No. 527), showing the fungal hyphae (h) in the exo-cortex and one of the cells of the medio-cortex with clump formation (c). Fig. 2. A portion of the epidermis from a transverse section of another root, from the same slide as Fig. 1, showing the radial course of the fungal hyphae (h) in an epidermal cell. r. h— root-hair. Fig. 3. Formation of vesicles by the fungal hyphae in the sub-epidermal layer. Portion of the longitudinal section from Dr. Scott’s collection shown in Plate XVIII, Fig. 3. Fig. 4. Enlarged view of a portion of medio-cortex shown in photograph 4 on Plate XVIII, to show the hyphae connecting the mycelial clumps with the cell-walls. Fig. 5. Portion of the cortex of transverse section on Slide No. 1527 (Dr. Scott’s Collection), showing one of the large vesicles found near the medio-cortex, and also a cell containing curious granulations (see p. 262). Fig. 6. Portion of the tangential section on Dr. Scott’s preparation (No. 1527), showing sporangia (?) (sp) and spores (?) in a cell of the medio-cortex. The other cells contain mycelial clumps (cl). Fig. 7. A cortical cell from a transverse section on Slide 529 of the Cash Collection, in which the cell-contents seem to have undergone degeneration. Fig. 8. Two cortical cells from the same section as Fig. 7, showing earlier stages of degeneration of cell-contents with indication of previous formation of mycelial clumps. ;‘;V: A ■ ' ' : •:=:> v; • ' • ' i'J •' •'• •• V; • i? \.n: : ■ : ■ ■ 1 . : ft ■ . ... p, ■ i a-; • .'•••••. •! . . I.:.-:',.;. .. -'o' .p)* . ‘i- /I-: -V /.. . ' . - V. i. .V,'. • ; . . ii • • J-.ii V;./. /.-II : 5 K-: " ;l I ■; i- . .. \,i f ,£• .r-r'i \- j el) >7 L> .. * . '-.v. . .... r' v,:..- :i ,v j-: .I-? t(;- !i'8 ^ •' •• J’.-Ai . _.:-vr -vl'I :i Oil, vjj 4*ui.-'ijivo .-••.'•■'V:: I? ' . ; . 1 ‘y.\ ' i :.i.i t .agitia ; i-1' «• • -j •• hiv j i.A .{Vi : -qr;:. .. hi; \r fir. , k.o i;--; V c V ,, ; .i ,;nf i : ... y'\) a :'0 . si nU Ui-ii 1 .(•{ : : j u.il lo -j,' . ; ;v. i. : i. ,1 y-^ 'io iiui.io'l .'iJVilU . .r;. id ■; i.v i>;]j ,jl , 10 £U iiiuruo j 'M i. ' ; H / . ' sod;1, riwi j:;^L.roo .iCl iuoil not.hj-J.'i JUVX 1 •' > ; f i ' ?cr '. :•* i’i ■■■ ■. ’' V ^ •••' •i-.u . . : 1 i'. < !y^a;Vuy 'j.r.-i'cf rd c-.ili v/Offa — A Study of the Enzyme- secreting Cells in the and their subsequent discharge. After seventy-two hours the periods of secretion are not well marked. At the end of eleven days there are signs of degeneration in some of the epithelial cells as indicated by an abnormal swelling and vacuolization of the cytoplasm. After twenty-two days the cytoplasm is very scanty. II. Material and Methods. In my work I have attempted to study the morphology of the enzyme- secreting cells in the scutellum of Zea Mats and in the ‘ absorbing organ * of the seedling of Phoenix dactylifera . It has been proved by the work of Brown and Morris (’90), Hansteen (’94), and Griiss (’97) that the diastase produced by the scutellum of the Gramineae is formed and secreted by the columnar epidermal cells of that organ. In Phoenix dactylifera the production and secretion of enzymes occurs in the columnar epidermal cells of the absorbing organ (Griiss, ’94, Puriewitsch, ?98). It is to a consideration of the morphological changes that this paper is devoted ; the physiological changes will be made the subject of a subsequent study. My work was carried out at the Botanical Labora- tory of the University of Michigan during the years 1902 and 1903. It is with great pleasure that I take this opportunity of expressing my thanks to Professor F. C. Newcombe for his invaluable suggestions and criticisms. I used the large variety of Zea Mais known in agriculture as ‘ White Dent,’ and the seeds of Phoenix dactylifera obtained from the dates of commerce. The resting embryos were cut from the dry seeds and killed in strong alcoholic killing fluids. The embryos of different ages were obtained by germinating the seeds in moist sawdust or between layers of moist filter-paper. Light was always excluded in order to prevent any manufacture of food by photosynthesis. The embryos of Phoenix were grown in an incubator at a temperature of 30° C., those of Zea were grown at a temperature of 2d to 250 C. The study of fixed and stained material was supplemented by that of living cells in both plants. Sections of the living material were cut with a razor and mounted in water, or in dilute sugar-solution if they were to be studied for any length of time. Methylene blue in aqueous solution was used for intra vitam staining. In many respects the use of living material was not as satisfactory as one would expect, owing to the difficulties in identifying the different substances in the cell. It was, nevertheless, valuable as a check on the artificial appearances produced by killing fluids and other reagents. It is well known that certain methods of fixing and staining give characteristic appearances to the tissues upon which they are used, especially upon cells containing large amounts of plastic material in a fluid state. Many of the discrepancies between the descriptions of different investigators are doubtless due to different processes of fixing and 271 Seedlings of Zea Mais and Phoenix dactylifera . embedding. An interpretation of the effects produced awaits an extension of our knowledge of the chemical and physical reactions between proto- plasm and the various reagents used in micro-technique. In order to obviate as far as possible particular effects due to chemical or physical action of the killing fluids, a number of different fluids were used and different stains employed after each method of fixing. The technique was rather difficult on account of the delicacy of the material. The following results were observed with the respective killing fluids. Saturated Solution of Picric Acid in 50 per cent. Alcohol. The mate- rial prepared for study by this reagent was unsatisfactory. The proto- plasmic structures were not fixed well enough to show with any definiteness. On the other hand, this is a good reagent for fixing the proteid granules. (Zimmermann, ’93, p. 216.) Aqueous Picro-corrosive Fluid. I used the following modification of Mann’s method as given by Huie (’97). One volume of a saturated aqueous solution of mercuric bichloride was added to three volumes of a saturated aqueous solution of picric acid. The material was allowed to lie in this fluid for twelve to eighteen hours, then washed in water and dehydrated in the usual way with alcohol. This proved to be a very satis- factory killing fluid. The amount of picric acid present was sufficient to fix perfectly the granules, while the mercuric bichloride hardened the protoplasm and precipitated the soluble proteids. In a few cases there was a tendency to contract the cell-contents in the scutellum of Zea. Kleinenherg s Picro-sulphuric Acid. The results obtained by the use of this reagent indicated that it was better than picric acid alone, but not so good as the picro-corrosive fluid. The material seems to be insuffi- ciently hardened. Chromo-Osmo- Acetic Acid. Mottier’s formula, Pring. Jahrb. Bd. 30, p. 170. This fluid is a good fixing agent for the protoplasmic structures of the cell, but not for the granular structures. It produced no shrinking in any of the cases where it was used. Iridium chloride in Acetic Acid. The formula used was — Iridium chloride 1 per cent, aqueous solution . . 25 cc. Glacial acetic acid . . . . . . . 75 cc. It leaves the tissue in better condition for the stain than the following killing fluid, but in other respects the two act similarly. Worcester’s Killing Fluid. This fluid gave uniformly good results when- ever used. The formula according to which it is made is as follows : — Mercuric bichloride, saturated aqueous solution . 9 6 parts. Formalin (40 per cent, formaldehyde) . . 4 „ Acetic acid, 10 per cent. 10 „ Formic acid, to each litre of solution 5 drops. 272 Reed . — A Study of the Enzyme-secreting Cells in the The tissue is allowed to remain in the killing fluid from ten to twenty hours, then transferred for washing to 70 per cent, alcohol, which contains about 1 per cent, of potassium iodide. If the killing fluid is not completely removed, it interferes with the action of the stains, more particularly the basophil stains. Probably the good results obtained with this reagent are due to the large amount of soluble chloride it contains, which precipitates the proteids in the cells. Saturated Solution of Mercuric Bichloride in Absolute Alcohol . This reagent was used to kill resting embryos in the dry seeds, where the presence of air in the tissues hinders the penetration of heavy liquids. It produced good preparations, but did not leave the nuclei of the cells as susceptible of staining as Worcester’s fluid. It will be seen from the foregoing accounts that the killing fluids which gave the best results were those containing a large proportion of soluble chloride and a small proportion of acid. Those fluids which contained a large proportion of acids appeared to have a corrosive action upon the proteid granules in the cells. In small amounts the presence of acids appeared to facilitate the penetration of the reagent without corroding the granules. Most of the material was embedded in paraffin by the ordinary method. In the case of dry seeds I used a method for which I am indebted to Mr. B. J. Howard (’03). By the use of this method I completely infiltrated the pieces of seeds (which were always small) with paraffin, and succeeded in obtaining uniformly good preparations. The results obtained by the use of different stains were as variable as those of the different killing fluids. No one stain could be depended upon in all cases. The use of stains is twofold — to render the objects more opaque for study, and to give an indication of their acidity and alkalinity. While I did not find it possible to apply the terms ‘ cyanophil ’ (baso- phil) and ‘ erythrophil ’ (eosinophil) with precision, yet the absorption of different stains serves in a general way to indicate the nature of cell-contents. The use of the different stains gave the results described below. Picro-Nigrosin. This was found to be a very satisfactory stain for general cytological purposes. It brings out the granules, nuclei, and cyto- plasts very plainly, but does not give them a differential stain. Kleinenberg s Haematoxylin . The results obtained by the use of this stain were much the same as the preceding. It is a very valuable stain for chromatin, and was used chiefly on that account. Haidenhain s Iron Alum Haematoxylin. The method described by Torrey (’02), using Congo Red as a counter-stain, was employed. It is a very valuable stain when used in connexion with others, but cannot be 273 Seedlings of Zea Mais and Phoenix dactylifera . relied upon when used alone. It gives no indication of the acidity or alkalinity of the cell-contents, nor does it differentiate the granules in the cytoplasm from those in the nucleus. Z winter m ami s Fuchsin- Iodine- Green. Repeated attempts with this otherwise valuable stain were not successful in producing a single good preparation. The stain seems to have no affinity for the granules, and not even the protoplasmic structures were stained satisfactorily. A nilin- Gentian- Violet-1 odine-Eosin . Grants Method. The sections were not stained deeply enough when this stain was used to afford any satisfactory study. Mantis E osin- T oluidin Blue. This proved to be the most valuable stain used, and was employed more extensively than any other in my work. The method described by Huie (’97) was employed in staining the sections. It not only differentiates cell-walls and cell-contents, but differentiates one kind of granules from another in the same cell. The cell-walls were stained blue ; starch-grains were stained bluish-green ; cytoplasm, blue (red in cells where secretion had progressed for some time) ; zymogen granules, blue (or purple) ; chromatin, reddish-purple. The best results were obtained in material which had been fixed by Mann’s aqueous picro-corrosive killing fluid. Eosin and Anilin Blue. This stain was used to good advantage with sections of Zea> where a differentiation of starch- and proteid-grains was desired. Eosin and Gentian Violet. The same methods were employed with this stain as for the two preceding, but it did not give as good results. The cell-contents were not plainly differentiated. Flemming s Triple Stain. Good preparations were obtained by the use of this stain, but it did not differentiate the different granules in the cell sufficiently to make it a valuable stain. It works best after the use of Chromo-osmo-acetic acid, but does not give good results after the use of fluids containing mercuric bichloride. III. Observations on the Scutellum of Zea Mais. A cross-section of the scutellum, when examined under the microscope, is seen to consist of large, nearly isodiametric cells, which are bounded on the side next the endosperm by a single layer of columnar epidermal cells. There is not only a noticeable difference in the size and shape of the two kinds of cells, but also in their contents. The granules found in the epi- dermal cells are always small, and are composed of proteid ; the granules in the large cells of the scutellum may be either starch or proteid ; and the proteid may be in the form of large or small granules. 274 Reed . — A Study of the Enzyme- secreting Cells in the A. Studies on Living Material. Twenty-four hours old. The protoplasmic body of the epidermal cells of the scutellum still shows the characteristics of the resting condition ; it does not entirely fill the cell-cavity. The cells contain a large amount of granular substance which is coloured yellowish-brown by iodine. The under- lying cells of the scutellum contain two kinds of granules — proteid and starch. Two days old. The cells of the epidermal layer are still closely packed with fine granules, but the larger proteid granules have disappeared from the underlying layers of cells. The nuclei of the epidermal cells, where observable, present an uneven contour. The scutellar cells in the vicinity of the plumule contain a small amount of starch. Three days old. This lot of seeds had made rapid growth ; when the embryos were removed for study the radicles were 1*5 to 2*5 cm. long, but the condition of the secreting-cells was only slightly more advanced than that described for the second day. The epidermal cells were full of small proteid granules, but the large proteid granules had disappeared from the first three or four layers of hypodermal cells. There was again an evident accumulation of starch about the plumule. Fotir days old. The radicles of this lot of seeds only average i cm. in length, but the cytological changes in the scutellum are farther advanced than in the last set. The epidermal cells contain proteid granules only in the distal 1 half or third of the cell. A few cells are entirely free from granular material. Seven days old. Cotyledons i cm. long, radicles 2 cm. long. The epidermal cells are nearly free from granules. In those cells where granules are found they are small, and occur only in the distal end of the cell. The scutellar cells have lost all of the large proteid granules, but still contain large numbers of the small granules. In favourable cases the vacuolate protoplasm may be seen. When the small granules of the epidermal cells disappear, the small granules of the scutellar cells, which in their turn have probably resulted from the disintegration of the large granules, also begin to disappear. The cells of the scutellum show an increasing amount of starch in the region of the plumule. After the application of iodine to the sections the consequent darkening can be seen plainly by the naked eye. The presence of this starch has two possible explanations — it may arise indirectly from the breaking down of proteid matter in the scutellum (cf. Timberlake> ’01), or it may be formed by the anhydration of some soluble carbohydrate derived from the endosperm. 1 In the subsequent descriptions of epidermal cells, the term * distal ’ refers to the end of the cell in contact with the endosperm, 1 proximal ’ refers to the opposite end. 275 Seedlings of Zea Metis and Phoenix dactylifera . Nine days old. Nearly all the epidermal cells are devoid of granular material ; the remainder contain granules evenly distributed throughout the cytoplasm. The starch in the vicinity of the plumule is beginning to disappear. The substance of the endosperm is almost completely dissolved and absorbed at this time. Apparently the epidermal cells of the scutellum have ceased to be actively secreting cells and are now principally vegetative in function. A little later the contents of these cells disappear ; possibly they are absorbed by the growing plant. B. Studies of Microtome Sections. Cells in the Resting Condition. The protoplasm of the epidermal cells is contracted away from the lateral walls and often from the distal wall of the cell. With a low -magnification the protoplasm appears as a fine granular substance, homogeneous throughout. In a number of the cells there are irregular vacuoles in the proximal end of the cell. With higher magnification one can distinguish the granules arranged on the protoplasmic network. There is no regularity as to the position of the nucleus in the cell at this time. It has a slightly irregular, elliptical outline. The nucleus contains fine granular material which renders it darker than the cytoplasm. There is only a small amount of chromatin present, and it occurs in the form of small spheres at the surface of the nucleus. These can be distin- guished easily from the other granular material, because the latter appears to be evenly distributed throughout the interior of the nucleus, while the chromatin is at the surface. The faint outline of a nucleolus may be distinguished in favourable sections. (Plate XX, Fig. i.) The other cells of the scutellum are so densely filled with granules of proteid matter that none of the protoplasmic structures are visible except the nucleus. Cells after imbibition with water for three hours. The contents of many of the epidermal cells do not yet completely fill the cell-cavity (Fig. d). The cytoplasm is full of small, flocculent granules which stain blue with Mann’s Eosin-Toluidin Blue. The nucleus contains two kinds of nucleo-proteid matter in the form of granules. Undoubtedly the larger granules are chromatin and the smaller ones may be a reserve product stored in the nucleus ; because as the cell-metabolism proceeds these fine granules rapidly disappear from the nucleus. The nucleolus stains red, the chromatin and fine granules, dark purple. The large isodiametric cells of the scutellum are completely filled with two kinds of granules at this stage, a number of large and small proteid grains which nearly fill the cell and stain red with Mann’s Eosin- Toluidin Blue, and larger spherical starch grains which stain blue. 276 Reed . — Z? Study of the Enzyme- secreting Cells in the Condition of the scutellum after thirty hours of activity (Fig. 3). The cells of the epidermal layer are noticeably larger than those in the stage last studied, and are completely filled with granular protoplasm, which is slightly denser at the distal end of the cell and stains purple with Mann’s Eosin-Toluidin Blue. The granules show some conformity in their arrangement to the strands of the cytoplasmic network. The nuclei are of the sort one finds in cells where active metabolism is going on. Nearly every nucleus is surrounded by a vacuole and lies in the proximal end of the cell. The karyolymph, or nuclear plasm, contains, as in the last stage, two kinds of material. However, there is much less of the fine granular material than in any preceding stage, and the larger bodies of nucleo-proteid, chromatin, have increased in size. It can also be seen that the chromatin is arranged on the linin network. The nucleolus, although distinct, is not prominent. The staining properties of the granules in the cytoplasm indicate that they are of a different character from those in the nucleus ; the former stain blue with Mann’s Eosin-Toluidin Blue, the latter stain red. Part of the proteid granules have disappeared from the more deeply lying cells of the scutellum. There is a slight increase in the amount of starch in the scutellum, the greatest increase being near the plumule. Morphology of the cells at the end of two days of activity. Radicles 1 cm. long. There is a further decrease in the amount of the granular substance in the cytoplasm and an increase in the amount of chromatin in the nucleus (Fig. 4). The nucleolus is hardly to be distinguished because it has become smaller and lost most of its ability to absorb stain. The nucleus is usually located in the proximal half of the cell. The con- tents of the subepidermal cells of the scutellum are beginning to diminish. They are probably being used for the nourishment of the plant or to form enzymes. Morphology of the cells at the end of the third day of activity. Tem- perature 350. On the third day the scutellum appears to be secreting diastase very actively. The epidermal cells contain a fine granular substance which stains bluish-purple and contains embedded in it larger granules, which are of almost the same size as the chromatin granules of the nucleus, but differ from them in staining properties. This granular matter is quite evenly distributed throughout the cell, but is nowhere so dense as in the preceding stages. The nuclei are at the middle of the cell or in the proximal half. The finely divided granular substance which was present in the nucleus has all disappeared at this stage, but the chromatin of the nucleus continues to increase. It exists in the shape of spherical and rod-like masses at the surface of the nucleus (Fig. 5). I have not seen any cases where I thought that solid particles were passing from the nucleus into the 277 Seedlings of Zea Mais and Phoenix dactylifera . cytoplasm. On the contrary I think there is evidence that the nucleo- proteids leave the nucleus in a fluid condition ; because the nuclei at this stage have a swollen and distorted appearance (Fig. 6). In a few cases a vacuole surrounds the nucleus. A nucleolus is present in nearly every case. Apparently some of the contents of the subepidermal cells are used in the formation of diastase, because those cells in the vicinity of the epidermal layer are showing more signs of depletion than any others ; in fact those in the vicinity of the young plant show little sign of depletion. The content of starch has increased over any previous stage. Morphology of the cells at the end of the fifth day of activity. At this time there is a still greater depletion of the granular substance in the cytoplasm of the epidermal cells, which causes it to appear lighter coloured in all instances. The greatest scarcity of granules is at the proximal end of the cells. The nuclei are large and contain a large amount of chromatin in the form of spherical masses at the surface of the nucleus. They show a difference from the last stage in being found no longer in the proximal half of the cell, but in the distal half, a short distance above the centre. The nucleolus has diminished in size until it is no longer visible. Apparently there is some relation existing between nucleolus and chromatin, because as one increases the other decreases. Some later experiments give more light on this subject. The same condition was found in Drosera by Rosenberg (’99), although there were numerous exceptions. The morphology of the cells on the ninth day of activity (Fig. 7). The condition of the cells in the epidermis is quite similar to that described in the last stage, except it is very evident that increasingly greater quan- tities of proteid grains are disappearing. The nuclei are quite similar to those last described, both in appearance and position. The proteid granules which remain are most abundant in the distal end of the cell. The morphology of the cells after thirteen days of activity (Fig. 8). The endosperm of the seeds from which this material was taken was nearly exhausted. The cytoplasm of the scutellar epidermal cells is compact and fills nearly the entire cell-cavity. It contains a small number of flocculent granules of a different sort from those appearing in the cells when enzymes are being actively secreted. The nuclei are situated near the centre of the cell. The nucleo-plasm stains but slightly different from the cytoplasm, making it difficult to determine the exact boundary between the two. Large nucleoli and a small amount of chromatin are the only substances which can be distinguished in the nuclei. This condition of affairs suggests that the cells have ceased their active metabolism and are at this time in a passive state, and that perhaps even at this time a few have been partly absorbed by the growing plant. From this time on there is probably very little enzyme produced. Taking a general survey of the morphological changes, we see that 2j8 Read. — A Study of the Enzyme- secreting Cells in the at first the cytoplasm of the secreting-cells contains fine, granular proteid material, which shows an affinity for basic stains. The nuclei of these cells contain finer granules, which stain differently from those in the cytoplasm and disappear much sooner. With the high power they can easily be distinguished from the chromatin. The amount of chromatin is small in the earliest stages, but the nucleoli are quite prominent. At the beginning of the second day of activity the cytoplasm is densely filled with granular material, which is distributed on the cytoplasmic reticu- lations. A comparison of Figs. 2 and 3 will show the relative increase in size of the cell during the first day’s activity. From this time forward there is a constant depletion of the granules of the cytoplasm, accompanied by a slow elongation of the epidermal cells. The nuclei are almost invariably nearer the proximal end of the cell. After two days of activity, one can notice that part of the starch of the endosperm has disappeared. In the nuclei of the epidermal cells there are two noticeable changes — a continued increase in the amount of chromatin and a decrease in the size of the nucleolus, which from this time forward is very inconspicuous. The processes described above continue without much variation until the endosperm is exhausted. On the third or fourth day the nuclei move from their position in the proximal end of the cell toward the centre, and on the fifth or sixth day are found in the distal end of the cell (Fig. 7). The cytoplasm is nearly free from granules on the tenth day, and now shows an affinity for acid stains. Just before the final dissolution occurs, the cyto- plasm of the epidermal cells becomes abundant and stains similarly to the nucleus. C. The Effect of Inhibited Growth upon the Secreting-Cells. Seedlings which had grown at a temperature of 23° C. for fifty hours, and whose roots had attained a length of 3 to 4 cm., were transferred to a temperature of 8-io° C., after having removed and fixed a number of scutella. When they had remained twenty-four hours at the low tempera- ture, the amount of growth shown by the roots was very small. At this time more material was removed and fixed in different killing fluids ; the remaining seedlings were then placed at a temperature of 20° C. for forty hours longer. At the expiration of this time the roots showed that growth was progressing normally again. The different lots of material were then sectioned and stained for microscopical study. The sections showed that the cells of the epidermal layer were in an active condition when transferred to the cold room. They were full of fine granules, which stained blue with Mann’s Eosin-Toluidin Blue. -The densest accumulation of granular material was in the distal end of the cells. The nuclei, which were in the proximal end of the cell, contained chromatin in Seedlings of Zea Mais and Phoenix dactylifera . 279 the form of fine granules. The nucleoli were spherical, well-defined bodies which stained dark purple. After the growth had been checked by low temperature for twenty-four hours there was a marked change in the appearance of the cells. The cytoplasm contained large lumps which stain dark purple. The staining property of these aggregations was more intense than that occurring in ordinary secretion. The nuclei had lost most of their chromatic substance and appeared quite clear. They contained at least one large nucleolus which had lost neither in size nor staining qualities. When normal conditions were restored, the cells appeared to regain activity. The fixed and stained material showed two things very clearly : (1) the aggregations had almost entirely disappeared from the cytoplasm, leaving it homogeneous purple ; and (2) the nuclei contained numerous large granules of chromatin, but the nucleoli had disappeared. It hardly seems possible that any of the nucleo-proteids are absorbed by the nucleus when activity is resumed. The fact that the nucleus contains granules only during resting periods, or periods of arrested growth, and is large, hyaline, and intimately connected with the cytoplasm during the time of most active secretion, indicates to my mind that the nucleus is not a storehouse of diastase in any form, but is the source of energy by which the diastase is produced. The view here taken is that the diastase is manufactured from other proteids and turned into secretion by the activity of the epidermal cells, in which processes the nuclei probably perform the largest part of the work. My own work fails to confirm many of the observations made by Torrey (’02). I did not find, at the beginning of secretion, the nuclei filled with dark staining granules. Nor did I find the granules being extruded in a solid state into the cytoplasm through breaks in the nuclear membrane. There were, it is true, in the resting stage, fine granules in the nucleus ; but they disappeared during the first two days and did not reappear. By far the greater amount of proteid granules was found in the cytoplasm of the epidermal and subepidermal cells before secretion began. Moreover, the granules in the cytoplasm differed both in size and staining qualities from those in the nuclei, and gradually disappeared as secretory activity pro- gressed. I likewise failed to find that the process of secretion in Zea is an inter- mittent one, in which periods of activity alternate with periods of rest. If the seedlings are grown at a constant temperature, the process of secretion is continuous while enzymes are being produced. In repeating Torrey’s methods I found that the use of Iron Haematoxylin as a stain was the cause of many differences in our observations, because it is not reliable in differentiating the various cell-constituents. As his own paper states, it stains everything alike in a very deceptive manner, and is the cause of u 280 Reed. — A Study of the Enzyme-secreting Cells in the many artefacta in the sections. When part of the material was stained with Mann’s Eosin-Toluidin Blue and part with Iron Haematoxylin, the resulting sections gave very different appearances. The preparations made with the former stain were relied upon because of their similarity to the sections of living cells. IV. Observations on the Absorbing Organ of Phoenix dactylifera. The absorbing organ is a button-shaped structure which is located in the date-seed on the side opposite the furrow. By means of the enzyme it produces, it dissolves the hard, ivory-like endosperm of the seed, and absorbs the soluble material for the use of the young plant. At first the absorbing organ is about the size of the head of a pin, but as germination progresses it enlarges and finally fills all the space formerly occupied by the endosperm. If we make a longitudinal section of this organ after the radicle has begun to protrude, it will be found to have a mushroom shape ; under the microscope it is seen to be composed of thin-walled parenchyma-cells with large inter- cellular spaces. In the radicle the cells are elongated, but they approach a spherical shape in the head of the absorbing organ. Near the margin of such a longitudinal section the cells are smaller and have contents different from the other cells. The entire surface of the head of the organ is covered by a layer of short, columnar cells. A. Studies of Living Material. Cells in the resting seed. The cells of the epidermal layer contain, in addition to the large spherical nuclei, fine hyaline granules in the cytoplasm. The other cells of the absorbing organ contain similar small granules and, in addition, numerous larger granules, all of which give the test for proteid with re-agents. The epidermal cells do not contain as much granular material as the other cells of the absorbing organ. Observations upon seedlings five days old. At the end of this time the absorbing organ has increased to nearly twice its original size. The epidermal cells contain approximately the same amount of proteid matter as before, but in the form of larger granules. The nuclei of these cells are large and distinct, and each is at the centre of the cell. Observations upon seedlings twelve days old. The radicles have not yet appeared outside of the seed. Except for continued enlargement, the sec- tions of the absorbing organs are much the same as in the previous stage. The epidermal cells contain more proteid in the form of fine granules, but the nuclei are unchanged. Observations upon seedlings twenty -two days old. The radicles of the seed- lings average 2-5 cm. in length. The epidermal cells of the absorbing organ 28i Seedlings of Zea Mais and Phoenix dactylifera. have lost most of the granules which they contained in the previous stage. Those which remain are small and quite evenly distributed in the parietal layer of cytoplasm. The nucleus is smaller than in previous cases. It is situated near the centre of the cell, and connected to the lateral walls of cytoplasm by radiating strands. The deeper-lying layers of cells are nearly empty of granular material at this time. The elements of a fibro-vascular system are beginning to appear among the hypodermal cells. Observations upon seedlings twenty-nine days old. The parts of the embryos outside the seed average 4 cm. in length. The cells and cell- contents are much the same as in the last stage, except that they are more depleted of granular material. The granules which yet remain are nearly all confined to the epidermal and first hypodermal layers. Observations upon seedlings thirty-three days old. The scanty granular material which yet remains in the cells of the absorbing organ exists in the form of large granules. Each epidermal cell contains a large vacuole surrounded by a thin layer of cytoplasm. The nuclei of these cells are smaller than in previous stages. Observations upon seedlings fifty days old. The cells of the absorbing organ, both epidermal and hypodermal, are empty of granules, so far as can be ascertained by staining with iodine or methylene blue. The cytoplasmic body and nucleus retain the same size and relative position as in the previous stage. B. Studies of Microtome Sections. The morphology of cells in seedlings six days old (Fig. 9). There are very few changes from the condition which has been described for the resting stage in the living material. The epidermal cells of the absorbing organ show no elongation as yet. The cytoplasm contains a large amount of fine granular material which may represent the zymogen, because it disappears as enzymes are formed. The spherical nuclei, whose average diameter is about two-thirds the width of the cell, are usually situated near the centre of the cell. When stained with Kleinenberg’s Haematoxylin, the chromatin is demonstrated as very small grains on the nodes of the linin network at the surface of the nucleus. The karyolymph does not contain granular matter as in the case of Zea. A small nucleolus is present in each nucleus. The morphology of cells in seedlings nine days old. The cells show certain well-marked changes from the conditions described in the preceding stage. The zymogen granules in the cytoplasm have not only increased in size, but in their affinity for stain. All the cells of the embryo are filled with proteid granules which stain more cyanophil than at any subsequent time. The densest accumulation of granules is in the hypo- dermal layers of cells. There is also a larger amount of chromatin present in the nucleus, the difference being due to an increase in the size of the u a 282 Reed. — A Study of the Enzyme-secreting Celts in the granules already present rather than to an increase in their number. The nucleoli show a slight increase in size. The morphology of cells in embryos fourteen days old (Fig. 10). With the exception of a few layers of marginal cells, the granular contents of the absorbing organ have disappeared, and from this time forward there is no indication of metabolic activity in any except the surface-layers of cells. The epidermal cells are not only increasing in size, but numerous examples of karyokinetic division indicate that they are increasing in number. The staining reactions indicate that they are not so cyanophil as in the stage last described. The cytoplasm is beginning to show a diminution in the amount of granular material present. The finer granules seem to be the first to disappear. After their number has been diminished, it can be seen that the granules lie on the cytoplasmic network, not in its meshes. The nuclear chromatin does not increase as fast with increased activity as in the case of Zea, yet there is an increase. The nucleolus, instead of diminishing, has up to this time retained its original size, and shows a strong affinity for stain. The morphology of cells in seedlings eighteen days old (Fig. 11). The cells are much the same, except for continued elongation, as in the pre- ceding stage. Their staining qualities indicate that they are becoming more erythrophil. The morphology of cells in seedlings twenty -six days old (Fig. 12). At this stage the cells of the absorbing organ are practically free from granular material. There are a few erythrophil granules in the marginal cells. The deeper-lying cells which originally contained proteid material have enlarged to several times their former size, and the protoplasm remaining forms a thin parietal layer. The elements of a vascular system have begun to make their appearance among the marginal cells of the absorbing organ. There are three morphological differences between the nuclei of this and preceding stages — (1) they are smaller in volume; (2) the nucleoli are also smaller and often surrounded by a vacuole ; (3) there is an increased amount of chromatin in the nucleus, which occurs in the form of small grains on the linin network at the surface of the nucleus. The morphology of cells in seedlings thirty-three days old. At this age the absorbing organs are white elliptical disks about 8 or 10 mm. diameter and 3 mm. thick. The epidermal cells have lost much of the cytoplasm which they previously contained, many of them containing only a parietal layer. The most noticeable change is in the number of granules present in the cytoplasm (Fig. 13). It may be that part of the protoplasm has broken down to form an enzyme. The nuclei are not much smaller, but the nucleoli are diminished in size. The amount of chromatin is less also. The morphology of cells in seedlings four months old. The seedlings from which the material for this study were obtained were grown in earth 283 Seedlings of Zea Mais and Phoenix dactylifera . in the plant-house. At the end of four months I found that the reserve cellulose of the seed was entirely consumed, and that the absorbing organ filled all the space occupied by the cellulose. The amount of cytoplasm in the cells is very small ; in none is there more than a thin parietal layer, and in many the nuclei have broken down and been absorbed. In the epidermal cells the amount of cytoplasm is very small. Strands radiate from the nucleus to different parts of the cell. The layer of cytoplasm on the distal wall is usually thicker than on the other walls. In some cells the cytoplasm has begun to break down into a disorganized mass of granules. The nuclei, situated in various parts of the cell, have a smooth hyaline appearance and are devoid of chromatin. They still have a small, distinct nucleolus, which is about the same size as when activity began. The whole condition of affairs suggests that the cells no longer possess the function of actively secreting enzymes, but are now breaking down and being absorbed by the growing plant. The cells in the centre of the absorbing organ were the first to disappear, but the process of dissolution goes on until finally the epidermal cells are reached. Taking a general view of the changes in the secret ing-cells in Phoenix , we first find them short and thick, containing large spherical nuclei and densely granular cytoplasm which is distinctly basophil. During the first five or six days the cells increase in size, due to the absorption of water, but the contents show scarcely any change in composition. When secretion begins, the nuclei contain small granules of chromatin and small nucleoli. As secretion progresses they increase in size slowly until near the end of the third week ; then the nucleoli begin to diminish, followed a little later by the chromatin granules. The cells and their contents are strongly erythrophil at this stage, and the proteid granules have nearly disappeared from the cyto- plasm. The cytoplasm itself begins to disappear at the end of the fourth week, and finally the cells contain only a disorganized mass of substance. It is quite evident that if my observations have been correct there are some differences between the secreting cells of Zea and those of Phoenix . The nuclei of the epidermal cells in Zea contain, at the beginning of activity, varying amounts of granular substance which soon disappear, leaving only the chromatin and the nucleoli. The nuclei of similar cells in Phoenix show no such substance. The position of the nucleus in the cells of the epidermal layer is different in the two cases. In Phoenix it is almost always found at the centre of the cell ; in Zea it moves to the distal end of the cell as the activity of secretion progresses. It may be that the behaviour of the nucleus in the latter instance is in accord with the views of Haberlandt (’87), Townsend ('97), Harper (’99), and others, that the nucleus is usually situated in that part of the cell where the most active metabolism occurs. But nothing definite can be stated 284 Reed . — A Study of the Enzyme- secreting Cells in the until more plants have been examined. The behaviour of the nuclei in Phoenix appears to contradict any assumption which could be made concerning Zea. The nuclei in the two plants increase in size for a time as germination pro- gresses. In Phoenix this increase ceases when the embryo is but a few centi- metres long, and the nucleus then appears to become smaller, though it is probable that enzyme-formation is greater after the nucleus begins to diminish. The nuclei may increase in size merely because the cells increase in size. There are also differences in the behaviour of chromatin and nucleoli in the nuclei of the two plants. In Zea the nucleoli are large when the activity of secretion begins, and diminish in size as it progresses until they are no longer visible. The quantity of chromatin, on the other hand, increases as the nucleoli diminish. In Phoenix the nucleoli are present in all stages of secretory activity, attaining their maximum size about the time the cyto- plasmic granules begin to disappear. The chromatin, scanty at all times in comparison with Zea> shows slight changes in quantity analogous to those occurring in Zea. In the secreting cells of Zea the greatest apparent activity was reached when the nucleoli had disappeared and the chromatin had greatly increased. When the activity ceased, there was a formation of large nucleoli simul- taneous with a disappearance of chromatin (Fig. 8). When the activity of secreting cells was checked by means of low temperature, it was found that the chromatin disappeared and large nucleoli were formed. If these facts be interpreted to mean that there is some kind of re- lationship existing between the chromatin and nucleolus, such as has been postulated by Rosen ('95), Dixon (’99), van Wisselingh (’00), Gardner (’01), and others, then the difference between chromatin and nucleolus appears to be one of degree rather than of kind. The observations would seem to indicate that the material in the form of nucleoli was in a less active state than when in the form of chromatin, and as the cell-metabolism increased the latent substances became active. Our ideas of the relations of the nucleolus to the other constituents of the cell are, as yet, entirely hypothetical. There are opportunities for studying the nucleolus of secreting-cells which seem capable of yielding better results than those hitherto obtained during the process of mitosis, because the former are performing their normal metabolism instead of being interrupted by the process of division. The fact that these cells are both absorbing and secreting organs gives added importance to a point raised by Rosenberg (’99) concerning the causes for an increase of chromatin during secretory activity. He thought it might follow either as the result of the formation of an enzyme or as the result of absorbing abundant nutrition. The ease with which the scutellum of Zea may be separated from the endosperm suggests an experiment (which, Seedlings of Zea Mais and Phoenix dactylifera . 285 unfortunately, has not been performed) for determining which of the two hypotheses (if either) is correct. Concerning the nature of the various granules in the cells little can be said with any degree of certainty. They appear to arise as products of cellular activity, yet Muller (’96) believes that they are the elementary organs of the cell and are capable of growth and division. It is hoped that a wider application of the methods of physiological chemistry will supple- ment the results thus far obtained by the methods of histology. The extrusion of solid substance from the nucleus has been reported by many observers, but I have not been able to find any indications of it in my study. In every case it appeared as though the exchange of substances was accomplished when they were in a liquid state. It must be borne in mind that the soluble proteids would be precipitated by the reagents used in killing and dehydrating, and therefore granules in the prepared sections are not necessarily indicative of their presence in the living cell. Throughout the progress of the work I have been impressed by the similarities between enzymes and protoplasm in their manner of origin, action, and ultimate dissolution. Many of the similarities have been pointed out by Bokorny (’00). The two substances react similarly to most stimuli, except that enzymes show a greater resistance to destructive agents such as temperature, light, chemicals, &c. The decomposition of certain substances, e. g. sugar, by enzymes and protoplasm is very similar. The passage of two currents through the epidermal cells, one toward the endosperm and one toward the seedling, is undoubtedly brought about by osmosis modified by the selective action of the protoplasmic membrane. The existence of starch in the scutellum of Zea Mais shows that very little diastase passes toward the young plant. V. Summary. 1. The results obtained in fixed and stained material are dependent to some extent upon the technique. 2. In the resting condition the secreting-cells of both Zea and Phoenix are crowded with relatively small proteid granules. As secretion begins these granules gradually disappear. In Zea this disappearance coincides closely with the consumption of the endosperm ; in Phoenix , however, the granules disappear long before the endosperm is dissolved. 3. The chromatin of the nuclei is small in amount at the beginning of secretion and increases as germination progresses. The nucleolus diminishes in size as germination progresses. These changes are more noticeable in the case of Zea than in Phoenix. 4. There is no evidence that solid matter is extruded from the nucleus. 5. At the close of secretory activity the protoplasm of the secreting- cells breaks down and the products of disintegration disappear from sight. 286 Reed. — A Study of the Enzyme- secreting Cells in the Bibliography. An asterisk denotes that I have not examined the original paper. Bokorny (’00) : Empfindlichkeit der Fermente : Bemerkungen iiber die Beziehung derselben zu dem Protoplasma. Chemiker-Zeitung, 1900. Brown and Escombe (’88) : On the depletion of the endosperm of Hordeum vulgare during germina- tion. Proc. Roy. Soc. Ixiii, 3, 1888. Brown and Morris (’90) : On the germination of some of the Gramineae. Jour. Chem. Soc. Trans, lvii, 458, 1890. Correns (’96) : Zur Physiologie von Drosera rotundifolia. Bot. Zeit., 1896. Darwin, C. (’75): Insectivorous Plants. London, 1875. Darwin, F. (’76) : The process of aggregation in the tentacles of Drosera rotundifolia . Quart. Jour. Micr. Sci., N. S., xvi, 309, 1876. (’77) : On the protrusion of protoplasmic filaments from glandular hairs on the leaves of the common teasel. Quart. Jour. Micr. Sci., N. S., xvii, 1877. (’78): Experiments on the nutrition of Drosera rotundifolia. Jour. Linn. Soc. Bot., xvii, 17, 1878. De Fries (’86) : Ueber die Aggregation im Protoplasma von Drosera rotundifolia. Leipzig, 1886. Dixon (’99) : The possible function of the nucleolus in heredity. Ann. Bot., xiii, 269, 1899. *Fischer, R. H. (’03) : Ueber Enzymwirkung und Garung. Sep.-Abdr. a. d. Sitzungsber. d. Nieder- rheinischen Gesells. fur Natur- und Heilkunde. Bonn, 1903. From ann (’84) : Untersuchungen fiber Struktur, Lebenserscheinungen und Reactionen thierischer und pflanzlicher Zellen. Jenaische Zeit., xvii, 1, 1884. Gardiner (’83) : On the changes in the gland cells of Dionaea muscipula during secretion. Proc. Roy. Soc., xxxvi, 180, 1883. (’85) : On the phenomena accompanying stimulation in the gland cells of Drosera dichotoma. Proc. Roy. Soc., xxxix, 229, 1885. Gardner (’01) : Studies on growth and cell division in Vicia Faba. Contributions from the Bot. Lab. of the Univ. of Pennsylvania, ii, No. 2, 1901. Garnier (’00) : De la structure et du fonctionnement des cellules glandulaires sereuses. Jour, de 1’Anat., xxxvi, 22, 1900. Gruss (’93) : Ueber den Eintritt von Diastase in das Endosperm. Ber. d. d. Bot. Ges., xi, 286, 1893. (’94): Ueber die Einwirkung der Diastase auf Reservecellulose. Ibid., xii, 60, 1894. (’97) : Ueber die Secretion des Schildchens. Jahrb. wiss. Bot., xxx, 645, 1897. (’99) : Beitrage zur Enzymologie. Festschrift ffir Sehwendener, 1899. (’92) : Ueber den Umsatz der Kohlenhydrate bei der Keimung der Dattel. Ber. d. d. Bot. Ges., xx, 36, 1902. Haberlandt (’87) : Ueber die Beziehung zwischen Function und Lage des Zellkernes bei den Pflanzen. Jena, 1887. Hansteen (’94) : Ueber die Ursachen der Entleerung der Reservestoffe aus Samen. Flora, Ixxix, 419, 1894. Harper (’99) : Cell division in sporangia and asci. Ann. Bot., xiii, 467, 1899. Howard (’03) : Journal of Applied Microscopy and Laboratory Methods, vi, p. 2498, 1903. Huie (’97) : Changes in the cell organs of Drosera rotundifolia produced by feeding with egg- albumen. Quart. Jour. Micr. Sci., N.S., xxxix, 387, 1897. (’99) : Further study of cytological changes produced in Drosera. Ibid., xiii, 203, 1899. Klebs (’88) : Beitrage zur Physiologie der Pflanzenzelle. Unters. a. d. Inst, zu Tfibingen, ii, 489, 1888. Korscheldt (’89) : Beitrage zur Morphologie und Physiologie des Zellkernes. Zool. Jahrb., Abth. fur Anat. iv, 1, 1891. Macfarlane, (’01) : Current problems in plant cytology. Contributions from the Bot. Lab. of the Univ. of Pennsylvania, ii, No. 2, 183, 1901. Mathews (’99) : The changes in the structure of the pancreas cell. Jour. Morph., xv, 1899, Supplement. Montgomery (’99) : Comparative cytological studies with especial regard to the morphology of the nucleolus. Jour. Morph., xv, 265, 1899. MtlLLER (’96) : Drfisenstudien, I. Arch. f. Anat. u. Physiol. Anat., 1896. (’98) : Drfisenstudien, II. Zeitsch. f. wiss. Zool., lxiv, 1898. Seedlings of Zea Mais and Phoenix dactylifera . 287 Puriewitsch (’98) : Physiologische Untersuchungen iiber die Entleerung der Reservestoffbehalter. Jahrb. wiss. Bot., xxxi, i, 1898. Rosen (’95): Beitrage zur Kenntniss der Pflanzenzellen. Cohn’s Beitrage, vii, 225, 1895. Rosenberg (’99) : Physiologisch-cytologische Untersuchungen iiber Drosera rotundifolia . Upsala, 1899. Schimper (’82) : Notizen iiber insectenfressende Pflanzen. Bot. Zeit., 1882. *Schniewind-Thies (’97) : Beitrage znr Kenntniss der Septalnectarien. Jena, 1897. Timberlake (’01) : Starch formation in Hydrodictyon utriculatum. Ann. Bot., xv, 619, 1901. Torrey (’02) ; Cytological changes accompanying the secretion of diastase. Bull. Torrey Bot. Club, xxix, 421, 1902. Townsend (’97) : Der Einfluss des Zellkernes auf die Bildung der Zellhaut. Jahrb. wiss. Bot.,xxx, 484, 1897. Van Gehuchten (’90) : Le mecanisme de la secretion. Anat. Anzeig., 1, 1891. Van Wisselingh (’00) : Ueber Kemtheilung bei Spirogyra. Flora, lxxxvii, 355, 1900. Zimmermann (’93) : Botanical Microtechnique. (Trans, by Humphrey, New York, 1893.) (’96) : Die Morphologie und Physiologic des pflanzlichen Zellkernes. Jena, 1896. explanation of figures in PLATE XX. Illustrating Mr. Reed’s paper on Enzyme-secreting Cells. All figures were drawn with Abba’s camera lucida, Zeiss 1/16 oil immersion objective, and compensating oculars 8, 12, and 18. Figs. 1-8 were drawn from the epidermal cells of the scutellum of Zea Mais. Figs. 9-13 were drawn from the epidermal cells of the absorbing organ of Phoenix dactylifera. Fig. 1. Nucleus in the resting condition. Saturated solution of mercuric bichloride. Kleinen- berg’s Haematoxylin. x 1550. Fig. 2. Cell after absorbing water for three hours. Mann’s Piero -corrosive killing fluid. Mann’s Eosin-Toluidin Blue, x 1550. Fig. 3. Cell after thirty hours’ activity. Worcester’s killing fluid. Mann’s Eosin-Toluidin Blue, x 1550. Fig. 4. Nucleus after forty-eight hours’ activity. Worcester’s killing fluid. Mann’s Eosin- Toluidin Blue, x 1550. Fig. 5. An unusually swollen nucleus after three days of activity. Worcester’s killing fluid. Mann’s Eosin-Toluidin Blue, x 1550. Fig. 6. Normal nucleus after three days of activity. Worcester’s killing fluid. Mann’s Eosin- Toluidin Blue, x 1550. Fig. 7. Cell after eight and one-half days of activity. Worcester’s killing fluid. Mann's Eosin-Toluidin Blue, x 1600. Fig. 8. Cell after thirteen days of activity. Worcester’s killing fluid. Mann’s Eosin-Toluidin Blue, x 1600. Fig. 9. Cell after six days of activity. Worcester’s killing fluid. Kleinenberg’s Haematoxylin. x 1550. Fig. 10. Cell after fourteen days of activity. Worcester’s killing fluid. Kleinenberg’s Haematoxylin. x 1550. Fig. 11. Cell after eighteen days of activity. Worcester’s killing fluid. Mann’s Eosin-Toluidin Blue, x 1550. Fig. 12. Cell after twenty-six days of activity. Mann’s Picro-corrosive fluid. Kleinenberg’s Haematoxylin. x 1550. Fig. 13. Cell after thirty-three days of activity. Worcester’s killing fluid. Flemming’s triple stain, x 1550. .fJ •• H# ... ’’ ’ • • ' 'VO- : : . • • M : ■ . v .•••••, ■ • • ii; r: .oYY J o i ,7 . :■ : '■ •.[) I . jy ) i to : j A'J j: or Y: Y; ft.-.-Rj * • i ! A ‘tij .! :. iloY ■ i \ ’< tWiv . >V , . ,Y ..O' t _ Y • •, ..." ' jr. / ’kT ,7 .dftrfa't ..J ■ ;.Y ■" - - ' .71 • ■ . j . loc i . : : f f 7* r . -7 v,<7 . 7 _ i 7. 5 Y-Y. ; .•*•. ;Y \<\ j ! y;7 Yl I r: ; : k .<■ ■' tr. .. • YY 1 fY :::• J. I : Yl .>!'! ... ! ■■Y Y r TO f : .- i? iXr‘ ■ . : >7V : . yyyh vr; .'•T.V 7 . ; Y '■ : ./.i ...' ...- ■. Y^oY i'.'l ■ ■ ; ■ . : ? ' }{$n r; . : i s ■ ' - Y ioHiY/U' r -.7 Y7 Y:-;;) fain" ' moil 1: .ViViY /it'tr :' -i .v;7'J . Y . ’V:n xr. :• }■; • • _ • :• 7.- ,\YY 1. .77 ■ ■ . • Y! 7... . .I'/,.’.' • ,. YO : ' v;":i • ViY;. .177 •r 7 7 7 , ' Y: ■ A 7"7'..': •’ ' 7 .77 r ’ • • 7 ,Y . ' 7 7o •* ;7v77 .... ."Y ,c‘ . t >: ■ Y !]■:.; / ' ;• Y- . ' 1 ■ . ( ' Y . Y . ; ; -,Y .! ' .O'l'j :< . -= jfoT . Ht < . 1 ■ ■ 1 i . .. : ; ' . od i x U i Y‘ • ■ V- :Y ■ Y . . .Hi- ;V ’ 'i. YY. YY ’-: .' ... ■ ,o:;Y7 ' I \7 :i\z 1 1) .)*! ! fi / ' .'I . . : hi " ' ' ■ . ' ■; ' ' ' ' . * . : Y . .vii.-ji 1o ;■ I ..Y-v y.v} Y . ! .vi'i • .0 . ;< ■ • .V?!v!.Y;: <> .. '! j 7 J'iiY f Ti 1> ] ! • !• •; ! .V,!’ ftiubals of Botany. Vol.XViqPUX. REED - ENZYME-SECRETING CELLS. The Proteases of Plants, BY S. H. VINES, Sherardian Professor of Botany in the University of Oxford. INCE the publication in these pages (June, 1903) of my last paper O (1) on this subject, I have been continuing my investigations, and I have also come across several important papers by other observers, so that a considerable amount of further information has accumulated of which some account may now well be given. Before discussing the new facts, I propose to indicate very briefly the desirability of somewhat modifying the current method of describing the phenomena of proteid-digestion, and to suggest a terminology more in harmony with the present state of knowledge. Hitherto the proteases of both plants and animals have been classified as ‘ peptic * or as ‘ tryptic/ in accordance with their general resemblance to either the pepsin or the trypsin of the animal body ; and a digestion has been described as ‘ peptic ’ when it went no further than the conversion of the higher proteids into albumoses and peptones, and as ‘ tryptic ’ when the peptones formed were decomposed into non-proteid bodies such as leucin, tyrosin, &c. But with the discovery of erepsin by Cohnheim, this simple classification of the proteases has become inadequate, for erepsin is neither ‘ peptic * nor ‘ tryptic.’ Of the two, it is more nearly allied with trypsin than with pepsin, inasmuch as it actively decomposes peptones : but it differs widely from trypsin in that it cannot peptonize the higher proteids, such as albumin and fibrin. It is, in fact, a representative of a new, third, class of proteases, which may be described as ‘ ereptic.’ Now as to the terms employed in describing the digestive process. The word ‘ proteolysis ’ is in common use, but not always in the same sense : it is sometimes applied to peptonization by pepsin : at other times, and more accurately, it is applied to the disruption of the proteid-molecule into non-proteid substances, and it is in this sense that I have used it in my more recent papers. But the most appropriate use of the word is its application to the sum-total of the processes involved in proteid-digestion, to all the changes determining the conversion of the higher proteids into such substances as leucin, tyrosin, &c. Accepting this connotation of [Annals of Botany, Vol. XVIII. No. LXX. April, 1904.] 290 Vines. — The Proteases of Plants. i proteolysis/ the successive stages of the process may, I would suggest, be conveniently distinguished as — (a) peptonization , the conversion of the higher proteids into albumoses and peptones ; and ( b ) peptolysis , the decomposition of peptones into nitrogenous but non-proteid substances. This terminology offers the prospective advantage of simplifying the classification of the proteases. But before attempting this, it is necessary to draw attention to a recent paper by Vernon (2) in which he announces the important discovery that the peptolytic activity hitherto attributed to trypsin is largely due to an ereptic enzyme associated with it. This enzyme, which may be distinguished as pancreato-erepsin, is not identical with the entero-erepsin found by Cohnheim in the small intestine, though it belongs to the same group, other members of which will no doubt be dis- covered in due time. The effect of this discovery is somewhat to alter the position of trypsin proper — that is, trypsin free from pancreato-erepsin — in a classification of the proteases, bringing its peptonizing activity into relatively greater prominence. Taking this into account, and neglecting the somewhat conflicting views as to the possible peptolytic activity of pepsin — which may, after all, be due to an associated erepsin hitherto undis- covered— the proteases of the animal body may be classified as follows : — A. Actively peptonizing, but not at all peptolytic : pepsin. B. Actively peptonizing and peptolytic : trypsin. C. Feebly peptonizing, actively peptolytic : erepsins. There is a question bearing upon the relation between trypsin and erepsin that requires special consideration. Trypsin, it is well known, forms tryptophane as one of the products of its peptolytic activity : but does erepsin produce this substance ? It is not inconceivable that a pepto- lytic enzyme might produce leucin and tyrosin without, however, forming tryptophane ; and if this were found to be true of any form of erepsin, it would afford a clear distinction between tryptic and ereptic proteases. It is unfortunate that, so far as I have been able to ascertain, the available information on this important point is not altogether conclusive. Cohn- heim’s account of the products of digestion by entero-erepsin conveys the impression that tryptophane was not among them : but it does not appear that the presence or absence of this substance was made the subject of special investigation. On the other hand, Dr. Vernon informs me by letter that he has detected tryptophane in a digestion of peptone by entero- erepsin. For the present, at any rate, I accept the positive rather than the negative evidence, adopting the view that tryptophane is a product of peptolysis by erepsin as well as by trypsin. I have not included the proteases of plants in this survey, as I propose to discuss their nature in the concluding section of this paper. In dealing with the papers on proteolysis in plants, to which I have alluded, I will take first those relating to cases that I have not myself 291 Vines, — The Proteases of Plants, examined. There is, to begin with, an elaborate investigation by Butkewitsch (3) into the digestive action of certain of the lower Fungi [Aspergillus niger , Penicillium glaucum , and species of Mucor , M, stolonifer, M. racemosus , M. Mucedd) upon proteids. The Fungi in question were cultivated, in previously sterilized vessels, on a substratum consisting of proteid matter (Witte-peptone or fibrin) either with or without cane-sugar, together with a small proportion of suitable mineral ingredients acidified with phosphoric acid. The duration of an experiment varied from five days to over a month. The results show that these Fungi can peptolyse Witte-peptone, with formation of leucin and ty rosin, and can proteolyse fibrin, thus confirming the observations of earlier observers such as Malfitano (4) and others. A remarkable feature of the proteolysis effected by Aspergillus was the formation of a large proportion of ammonia (NH3), though it was much smaller in the presence than in the absence of cane-sugar in the culture. There is, further, a laborious research by Weis (5) into the nature of the proteolytic enzymes of malt. The author recognizes that in the germination of barley both peptonization and peptolysis take place — or as he puts it, there is a ‘ phase pepsique ’ and a * phase trypsique * — whence he infers the presence of two distinct proteases which he respectively terms peptase and tryptase. The peptic action is apparently rapid, and soon comes to an end, whilst the tryptic action is slower and continues until the complete decomposition of the products of the peptic stage. The tryptic action was found to be only slight, at most, in a neutral liquid ; rapid in the presence of a small quantity of added acid (e. g. lactic acid 0-2 °/o ; HC1 0-04 °/o), and much retarded by the addition of alkali. The author is of opinion that the effect of acid and alkali upon the activity of proteolysis is to be explained, in accordance with the views of Fernbach and Hubert (Comptes rendus, t. 131, 1900, p. 393), who regard the primary (acid) and secondary (basic) phosphates present in the malt-extract as determining the course of proteolysis, the former promoting, the latter retarding it. The author found both the peptic and the tryptic activity of malt to be interfered with by certain antiseptics, such as thymol, chloroform, formol, whilst toluol had but a slight effect. In the paper already referred to (1) I also have drawn attention to the influence of antiseptics on proteolysis in the special case of papain. It was further ascertained that the proteases of malt-extract could digest various vegetable proteids other than the glutin of wheat ; such as its own proteids, proteids of ungerminated barley, of rye, and of oats, legumin, vegetable casein ; as also, among animal proteids, the fibrin of ox-blood, whilst the action on egg-albumin was slight. 292 Vines . — The Proteases of Plants . Yeast (Saccharomyces Cerevisiae). In a previous paper (6) I expressed the opinion, as the result of a few experiments, that yeast contains a proteolytic enzyme which is active in neutral and in acid liquids but not in alkaline. At that time I had not seen the paper in which Hahn and Geret (7) have given a full account of their very thorough investigation of this subject : their results are of such interest that a brief re'suml of them will not be out of place. They worked with the expressed juice of fresh yeast, a liquid that contains a considerable amount of proteid coagulable on boiling, and is also spon- taneously coagulable on being kept in the incubator at 37°C. for two hours. They ascertained that this liquid digested fibrin within twenty-four hours ; but their investigation was directed more especially to the self-digestion (autolysis) of the liquid. Their method of estimating digestive activity was the comparison of the weights of the coagulum obtained from a given quantity of juice before and after digestion. For instance, in one case the weight of coagulum obtained before digestion was taken as 100, the weight after autolysis for twenty hours was 9*1. By this means they ascertained (a) that the natural acid juice digests actively ; ( b ) that its activity is diminished, though not to any great extent, by such antiseptics as chloroform, thymol, toluol, salicylic acid, and hydrocyanic acid (HCN) ; (c) that it is increased by the presence of neutral salts, such as NaCl 3 °/o, KN03 i°/0, KN03 io°/o; (d) that it is increased by the addition of HC1 from 0*05 °/o up to 0-3 °/o, o-2°/o HC1 being the optimum, and that it is diminished in the presence of o*5°/o HC1, and almost destroyed by HC1 i°/o ; (e) that the activity is diminished by neutralization, and still more so by alkalinity of o-2-o*5 °/o NaHO. The inferences that they draw as to the nature of the proteolytic enzyme will be discussed in the concluding section of this paper. Since the publication of the paper by Hahn and Geret, the only other contribution to the study of yeast-proteolysis is, so far as I am aware, that of Bokorny (8). He investigated the action of dried yeast, used in the solid form, upon various proteids either of animal or of vegetable origin : his experiments were made exclusively with liquids containing from 0-2-2 °/o of added acid, chiefly phosphoric, without, apparently, any antiseptic, their duration varying from three hours to three days. The measure of digestive activity was the amount of the precipitate obtained on treating the concen- trated digestion-liquid with excess of alcohol : the nature and relative quantity of the products was determined by dissolving the alcohol-precipi- tate in water, when any albumose present could be precipitated by saturation with ammonia sulphate or zinc sulphate, and any peptone by precipitation with phosphotungstic acid from the filtrate obtained after the separation of the albumoses. 293 Vines. — The Proteases of Plants. The main conclusion at which Bokorny arrives is that the acid reaction is essential to the digestive activity of yeast, and that the degree of acidity has an important influence upon the character of the digestive process as indicated by the products: thus, when the acidity is less than 0-5 °/o, only a little albumose is formed, but a relatively large quantity of substances that are not precipitated by zinc sulphate or ammonia sulphate. It can hardly be said that the paper adds material facts to existing knowledge of digestion by yeast, nor can the conclusion as to the relation between acidity and proteolysis be regarded as convincing. In the first place, the objection may be raised that no antiseptic was used ; and though it may be urged that in many experiments the amount of acid present (0*5-1 °/J was sufficient to prevent bacterial action, yet in those cases where the acidity was less strong, and the digestion prolonged (24-48 hours), the possibility of such action is obvious : in one case, indeed (Expt. 1), an offensive odour was noted and the development of mould. In the second place, no account is taken, in estimating the digestive products, of the proteids contained in the yeast itself. I have found that a watery extract of dried yeast, after boiling, filtering, and concen- trating, yielded a mainly proteid precipitate with alcohol amounting to something like 20 °/o of the original weight (see p. 298) : hence, when it is borne in mind that in Bokorny’s experiments the weight of yeast employed amounted to to, 20, or even 4o°/o of the proteid supplied for digestion, it is clear that the omission to take the yeast itself into account is a serious one. Finally, it is doubtful if any material amount of proteo- lysis was effected when the proportion of added acid was 0*5 °/o or more : for, as Hahn and Geret have pointed out, and as I have myself found, the digestive activity of yeast rapidly diminishes with increasing proportions of added acid (see p. 302). I give now a selection of my experiments to illustrate the digestive activity of yeast under various conditions. I have employed fresh brewers’ yeast, also yeast that I myself preserved in the dry state, but chiefly the dried yeast that is now obtainable as an article of commerce (prepared by the Granular Yeast Company Limited, London, E.C.), which is con- venient to use, with the great advantage that it is possible to make a number of experiments with uniform material. The experiments include observations on self-digestion (autolysis), on the peptolysis of Witte- peptone, and on the proteolysis of fibrin. The test applied in the autolytic and peptolytic experiments was that for tryptophane, the presence of this substance being taken as evidence of peptolysis. When the experiments were comparative, the test had to be applied with certain precautions. Thus, in each set of observations, it was necessary to ascertain in some one case what quantity of chlorine-water had to be added to a given quantity of the digested yeast-liquid in order to produce a tryptophane-reaction 294 Vines. — The Proteases of Plants. of maximum intensity; thus a standard of comparison for the other experi- ments in the same set was obtained. The quantity of chlorine-water required varies considerably, in relation, apparently, with the amount of tryptophane present : since an excess of chlorine destroys the reaction, it may be concluded that the more chlorine-water required, the greater the amount of tryptophane present. For instance, in an autolysis of a watery liquid containing 5°/0 dry yeast, I found that the addition of an equal volume of chlorine-water (say 5 cc. of each) gave the maximum trypto- phane reaction, when the digestion had been short (say 4-6 hours) : but when the digestion had been more prolonged (say 24 hours), it required twice the volume of chlorine-water to obtain a reaction as intense as that given as the result of the shorter digestion. It is also necessary to allow the tested liquid to stand for several minutes before estimating the intensity of the reaction, for it takes an appreciable time to develop. The liquid to be tested must, of course, have an acid reaction. In the experiments with fibrin, the main object was the determination of peptonizing activity : accordingly the crucial test was the total dis- appearance of the fibrin, which was consequently supplied in small quantity (usually about 0-5 grm. to 100 cc. of liquid). The fibrin had been preserved in 5° % glycerin. The dried yeast, previously to an experiment, was ground to a fine powder in a hand-mill, and was then thoroughly triturated in a mortar with the water necessary to prepare the required digestive liquid. The resulting mixture was then either used as it was, or it was filtered at a low tempera- ture so as to prevent autolysis during the somewhat lengthy process, and the clear filtrate was employed. Toluol, to about 1 °/o, was found to be the most unobjectionable antiseptic, though I obtained good results with others, such as chloroform, sodium fluoride, and hydrocyanic acid. It is important to state definitely that in no case did the freshly-prepared yeast-liquid, whether mixture or extract, give any tryptophane-reaction, thus proving that the yeast used contained no tryptophane to begin with. Autolysis. The fact that, under certain circumstances, the yeast can digest its own proteids is a familiar one. My object in experimenting upon it was to ascertain something more definite as to the conditions determining the activity of autolysis, and as to the nature of the enzyme by which it is effected. The following experiments will give an idea of the method adopted and of the results attained. Experiment 1. 1 grm. dried yeast placed in each of 3 bottles with 40 cc. distilled water: to No. 1 nothing further was added; to No. 2 was added 0*5 grm. precipitated chalk, to neutralize any free acid present; to No. 3, o-i grm. citric acid (=o-25%)’ Vines. — The Proteases of Plants . 295 After about 20 hours in the incubator (temp. 38-40° C.) the tryptophane- reactions were: No. 1, marked; No. 2, strong; No. 3, faint. A repetition of the experiment, using chloroform-water as the liquid, gave the same results. A somewhat similar experiment was made with the object of ascertain- ing if so prolonged a period of digestion were necessary for autolysis. Experiment 2. 40 cc. of an intimate mixture of 20 grms. ground yeast with 200 cc. of distilled water were placed in each of 5 bottles, to each of which toluol (1%) was added; the contents of the bottles were then varied as follows: — No. 1, nothing further added; No. 2, HC1 to 0-05 % ; No. 3, HC1 to o*i %; No. 4, HC1 to 0-2 % ; No. 5, Na2C03 to 0-5 %. After 2\ hours in the incubator the tryptophane-reactions, on treating 5 cc. of the liquids with equal vol. of chlorine-water, were No. 1, very strong ; No. 2, dis- tinct; No. 3, marked; No. 4, distinct; No. 5, which was distinctly alkaline, strong: further addition of chlorine-water did not intensify the reaction in any case. The following morning, after 20 hours more in the incubator, the tryptophane- reaction of Nos. 2 and 4 had become strong. From this it appears that autolysis is a rapid process, a conclusion confirmed by another experiment in which one of two bottles, each contain- ing 2 grms. dried yeast and 40 cc. distilled water, was kept for one hour in the incubator at 38°C., whilst the other bottle remained on the laboratory table at about n°C. At the end of this time the contents of the former gave a distinct tryptophane-reaction, whilst those of the latter gave no reaction. It appears, further, that autolysis can proceed within a wide range of alkalinity and acidity. The limits of this range were more nearly approached in the following experiment, which was of short duration : — Experiment 3. 40 cc. of a mixture of 20 grms. dried yeast with 400 cc. distilled water, and toluol to 1 %, were placed in each of 9 bottles, the contents of which were varied as follows: — No. 1, nothing further added; No. 2, added 2 grms. precipitated chalk ; No. 3, added Na2C03 to 0-5 % ; No. 4, added Na2C03 to 1 % ; No. 5, added Na2C03 to 2 %; No. 6, added HC1 to 0-05 % ; No. 7, added HC1 to o-i % ; No. 8, added HC1 to 0-2 % ; No. 9, added HC1 to 0-5 %. After 4 hours in the incubator at 38° C., 5 cc. of each of these various liquids, treated with an equal volume of chlorine-water, after acidification with acetic acid where necessary, gave the following tryptophane-reactions : — Nos. 1 and 2, very strong; No. 3, which was neutral before acidification, gave a strong reaction, as did also No. 4, which was alkaline ; No. 5, which also was alkaline before acidification, gave only a distinct reaction ; Nos. 6 and 7 gave a strong reaction ; No. 8 a distinct reaction ; No. 9 no reaction. The limit of acidity is here definitely indicated, the absence of the tryptophane-reaction proving that proteolysis did not take place in the presence of HC1 added to 0-5 °/o. The limit of alkalinity was not actually reached, though the retarding effect of 2 °/o Na2C03 was sufficiently x 296 Vines . — The Proteases of Plants. marked to justify the inference that a small further addition of alkali would arrest autolysis altogether. As, however, the time of digestion in this experiment was short, it was necessary to ascertain whether similar results were obtainable with more prolonged digestion. In the case of HC1, a repetition of the foregoing experiment showed that no tryptophane-reaction was developed in a 0-5 °/o HC1 liquid after digestion for 20 hours, and only a slight reaction after 72 hours. In yet another experiment, with 5 °/o yeast-mixtures containing respectively 0-2 °/0, 05%, o-8°/o, and i°/oHCl, the tryptophane-reactions were — at the end of 24 hours’ digestion — strong in the first, faint in the second, and none in the third or fourth ; and at the end of 72 hours, strong in the first, distinct in the second, and still none in either the third or the fourth. Hence it appears that autolysis was much retarded by o*5°/oHC1, and altogether inhibited by o-8°/o*or 1 °/o. In fact the limit of proteolytic activity in the presence of HC1 lies between 0-5 °/o and o-8 °/o, probably about o*6°/0, for a mixture containing 5 °/o yeast. With regard to alkali, I found that a similar yeast-mixture, to which 2 °/Q Na2C03 had been added, gave no tryptophane-reaction after digestion for 72 hours. In this case I endeavoured to ascertain as nearly as possible the minimum time of exposure to the action of alkali required to arrest proteolytic action, by the following method : — Experiment 4. 50 cc. of a 5 % yeast-mixture, with toluol, were placed in each of 3 bottles, Nos. 1, 2, and 3, to which Na2C03 was added to the extent of 1 %, 2 %, 3 % respectively. The bottles were then placed in the incubator for a certain time. At the end of this time the contents of each bottle were divided into two equal portions, one of which was left alkaline, whilst to the other half HC1 was added to slight acidity, and the 6 bottles were then returned to the incubator for any required number of hours, after which the tryptophane-reactions were compared. In the decisive experiment of this kind, the 3 alkaline bottles were digested for 2 hours, when the tryptophane-reactions were, in No. 1, distinct ; in No. 2, none; in No. 3, none. Half of the contents of each bottle having been acidified, the 6 bottles were further digested for 2 2 hours, when the reactions were : — No. 1, still alkaline, faint acidified, strong: „ 2, „ none „ distinct : „ 3, „ none „ none. This experiment showed that digestion for two hours with 3 °/o Na2C03 entirely destroyed proteolytic activity. A similar experiment, in which the period of exposure to this degree of alkalinity was only one hour, showed that this time did not suffice to destroy proteolytic activity, though it was much diminished. The foregoing experiments were made with mixtures containing usually 5% dried yeast: in none was the proportion less. It seemed 297 Vines. — The Proteases of Plants . important to determine whether or not the results given by such a mixture apply equally to others containing less yeast, and it was found that they do not apply. The following experiments were made with mixtures containing 2°/o dried yeast : — Experiment 5. Acid. 50 cc. of the mixture were placed in each of 3 bottles, Nos. 1, 2, and 3, acidified respectively with o-i %, 0-2%, and 0-5% HC1. After 3 hours in the incubator none gave any tryptophane-reaction; after 24 hours the reaction was faint in No. 1, none in either of the others. Alkaline. In a similar experiment, in which the contents of the 3 bottles had been rendered alkaline by the addition of 1 %, 2 %, and 3 % Na2C03 respectively, no tryptophane-reaction was obtained after digestion for 4 hours, or for 24 hours. Here proteolytic activity was destroyed by o-2°/oHC1, as also by 1 °/o Na2C03, degrees of acidity and alkalinity which produced no such effect in mixtures containing 5 °/o yeast. It may be concluded that there is a definite ratio between the amount of yeast present in a mixture and the amount of acid or alkali necessary to prevent autolysis. On the evidence of the tryptophane-reaction, it results from the fore- going observations that autolysis is very active at the natural acidity of the yeast-mixture : anything more than a slight addition of either acid or alkali tends to diminish it. The acidity of yeast is partly due to the presence of organic acids ; but not chiefly, for I have observed that it is impossible to neutralize a mixture or extract of yeast by adding excess of chalk. In view of the large proportion of phosphoric acid (about 50 °/o) and of potash (about 35%) in the ash of yeast, it may be concluded that the acidity is mainly due to the presence of acid phosphate of potash. Naegeli (9) has in fact suggested that the cell-sap contains KH2P04 and K2HP04. Repeated digestions of mixtures to which excess of chalk had been added (see Experiment 3) have shown me that autolysis is even more active when the free organic acid present has thus been neutralized than at natural acidity. The conclusion to be drawn is that the most favourable degree of acidity is that afforded by the acid phosphates, a conclusion agreeing with that of Weis (see p. 291) in the case of malt. The influence of added acid on autolysis would appear to be, in accordance with the views of Fernbach and Hubert with regard to malt, that it is favourable so long as it merely suffices to convert the dibasic (K2HP04) into monobasic (KH2P04) phos- phates, but unfavourable when free acid begins to accumulate. Similarly, the action of added alkali is favourable so long as it merely neutralizes any free acid present, but unfavourable when it begins to convert the monobasic into dibasic phosphates. The study of autolysis necessarily involves the consideration of the proteids contained in the yeast-cell. I am not aware of any more recent 298 Vines. — The Proteases of Plants. investigation in this direction than that of Naegeli (10), who stated the proteid content of yeast containing 8 °/o of nitrogen as follows : — Ordinary albumin ..... 36 °/o. Gluten-casein, soluble in alcohol . . . 9 „ Peptone, precipitated by lead acetate . . 2 „ This statement is not altogether clear. The first item probably means that 36 °/o of the dry weight of the yeast consisted of coagulable proteid. The significance of the second item is doubtful : it is not im- possible that it may really be albumose, or perhaps a mixture of albumoses and peptones. For some albumoses are relatively soluble in alcohol, precipitation only beginning with an alcoholic strength as high as 80 °/o ; moreover, some of them possess the property, specially mentioned as characteristic of Naegeli’s ‘ gluten-casein,’ of readily giving off sulphuretted hydrogen when treated with caustic soda or potash. Again, peptone is to some extent soluble in alcohol when at all dilute ; in fact one form (amphopeptone B) of it is soluble in 9 6°/o alcohol. Finally, the substance described as 4 peptone, precipitated by lead acetate,’ is possibly not 4 peptone ’ at all, since peptone proper (amphopeptone) is only partially precipitated by lead acetate : it is more probably one of the albumoses which are readily precipitated by this reagent. In view of the rapidity with which autolysis took place, as indicated by the tryptophane-reaction, it seemed probable that the dried yeast used in my experiments contained albumoses or peptones, or a mixture of these, to begin with ; and I have only so far investigated the proteids as to determine this point. A filtered watery extract was slightly acidulated and then boiled, when a precipitate of the coagulable proteids (albumin, &c.) was obtained. The filtrate was concentrated by evaporation and then treated with excess of alcohol, when a considerable precipitate was given. The precipitate was filtered off, dried, and dissolved in water on a filter ; the solution was then saturated with ammonic sulphate, after the method of Kiihne, in both alkaline and acid reaction, giving a considerable precipitate which consisted of albumoses. The filtrate, after appropriate treatment, still gave the biuret-reaction, indicating the presence of amphopeptone. Hence it is clear that the dried yeast contained both albumoses and peptones, the former in larger quantity than the latter. What still remains to be done is to determine the nature of the proteids that are coagulated on boiling. Since there is evidence that the dried yeast contained albumoses and peptones, and since the test of proteolysis was the presence of tryptophane, my experiments do not throw light upon the peptonization of the higher proteids of the yeast in the course of autolysis. The conclusion to be drawn from them is that there is present in yeast a peptolytic enzyme which is most active at or near the natural acidity of a watery mixture or extract, which is due to the presence of acid phosphates. Vines . — The Proteases of Plants. 299 Peptolysis. Inasmuch as the foregoing experiments on autolysis were gauged by the tryptophane-test, they were essentially experiments in peptolysis. Nevertheless, I thought it necessary to institute experiments as to the peptolytic action of yeast upon added albumoses and peptones, as con- tained in the substance known as Witte-peptone : the results, as might perhaps be expected, were similar to those of the autolysis-experiments. In the first place it was ascertained that a filtered watery extract of yeast was always peptolytically active, however short the period of extraction might be, even when the quantity of yeast used was small ; and further, that peptolysis was rapidly effected. The following experiment, in which the period of extraction was limited to the time necessary for filtration, illustrates both these points : — Experiment 1. 2 grms. of the dried yeast were extracted on a filter with 100 cc. of distilled water containing 1 % toluol : within an hour 50 cc. of liquid were obtained, to which 0-5 grm. Witte-peptone was added, and were then placed in the incubator. The liquid gave no tryptophane-reaction. After digestion for one hour the liquid gave a marked tryptophane-reaction. Here, notwithstanding the short duration of both extraction and diges- tion, a dilute extract gave unmistakable evidence of peptolytic activity. The next experiment was made with the object of demonstrating the effect upon peptolysis of various strengths of acid and alkali. Experiment 2. 20 grms. of dried yeast were extracted with 400 cc. of toluol- water (1 %), and left to filter for several hours in a cold room. The filtered liquid, which gave a faint tryptophane-reaction, was distributed as follows : — 40 cc. were put into a bottle (No. 1) without further addition; in the remainder of the liquid 10 grms. of Witte-peptone were dissolved, and 40 cc. of the solution were put into each of 8 bottles (Nos. 2-9): to No. 2, nothing more was added; to No. 3, added 2 grms. of precipitated chalk; to No. 4, Na2C03 to 1 % ; to No. 5, Na2C03 to 2%; to No. 6, Na2C03 to 3 % ; to No. 7, HC1 to o-i %; to No. 8, HC1 to 0*2 %; to No. 9, HC1 to 0-5 %. After 4 hours in the incubator the tryptophane-reactions were : — No. 1, distinct; No. 2, marked; No. 3, strong; all three being acid: No. 4, distinct; No. 5, faint; No. 6, none ; all three being alkaline : No. 7, strong ; No. 8, marked ; No. 9, distinct. After 25 hours in the incubator the reactions were : — No. 1, distinct; Nos. 2 and 3, strong; No. 4, marked; No. 5, distinct; No. 6, none; Nos. 7 and 8, strong; No. 9, marked. These results are in general agreement with those of the corresponding autolysis experiment (p. 295). It was further ascertained that here also the retarding or inhibiting effect of added acid or alkali was the more marked the more dilute the yeast-extract. 300 Vines.— The Proteases of Plants. Experiment 3. Acid. 4 grms. dried yeast were extracted for several hours in the cold with 200 cc. toluol- water (1 %) : the filtered liquid gave a faint tryptophane-reaction : 2 grms. of Witte-peptone were dissolved in the liquid, 50 cc. of which were then put into each of 3 bottles acidified as follows: — No. 1, HC1 to o-i %; No. 2, HC1 to o-2 %; No. 3, HC1 to 0-5 %. After 3 hours in the incubator the tryptophane-reactions were : — No. 1, distinct ; No. 2, faint; No. 3, none: after 29 hours they were — No. 1, marked; No. 2, distinct ; No. 3, none. Alkali. 50 cc. of an exactly similar yeast-extract, containing the same per- centage of Witte-peptone, were placed in each of 4 bottles: to No. 1, Na2C03 to 1 % was added ; to No. 2, Na2C03 to 2 % ; to No. 3, Na2C03 to 3 % ; to No. 4, nothing. After 2 hours’ digestion the tryptophane-reactions were : — in Nos. 1, 2, 3, faint; in No. 4, marked : the reactions were the same after the bottles had remained in the incubator for 25 hours. On comparing the results of Expt. 3, where the strength of the yeast-extract may be taken as 2°/o, with those of Expt. 2, where the strength of the extract may be taken as 5 °/o, it appears that the retarding action of added acid and alkali was more marked in the former than in the latter : for instance, in the case of the 2 °/Q extract, peptolysis was inhibited by the addition of HC1 to 0*5 °/o, and by all strengths of added alkali ; whereas the 5°/o extract peptolysed actively with HC1 0-5 °/o and with 1 °/o Na2C03. This is very much the same relation as that indicated by the corresponding autolysis-experiments : such differences as exist are due to the different chemical composition of the liquids in the two sets of experiments. The fact that the retarding action of added acid and alkali is less marked in peptolysis than in autolysis is clearly brought out by a com- parison of the results obtained by a peptolytic experiment on the same lines as the autolytic experiment (Expt. 4, p. 296), which had as its object to determine the effect of exposure to the action of alkali of different strengths for a short time, and showed that autolysis was inhibited by treatment for two hours with 3 °/o Na2COs liquid at 38°C. : complete inhibition was not produced in the peptolytic experiment under similar conditions. Experiment 4. 10 grms. of dried yeast were extracted for several hours with 200 cc. toluol-water (1%) and filtered: the filtered liquid gave faint tryptophane- reaction. 50 cc. of the liquid were piaced in each of 3 bottles, to which Na2C03 was added to 1 %, 2 %, 3 % respectively, and then the bottles were kept in the incubator for 2 hours. The contents of each bottle were then divided into 2 equal parts, in separate bottles, and one half was slightly acidified with HC1, whilst the other half remained alkaline : 0-2 grm. of Witte-peptone was added to each bottle, and a little more toluol. 3° 1 Vines. — The Proteases of Plants. After 24 hours’ digestion the tryptophane-reactions were : — Na2C03 ; I % 2 % 3% Alkaline . faint faint none Acid .... strong strong marked 24 hours later they were : — Alkaline , marked faint none Acid .... very strong very strong strong. These experiments with Witte-peptone confirm the conclusion arrived at from the autolysis-experiments (see p. 298) — that yeast contains an actively peptolytic enzyme, most active at or near the natural acidity of the extract, becoming less active, to total arrest, on the addition of either acid or alkali. Further, they show that this protease can be very readily extracted with water, and that the peptolytic action of a watery extract is marked as well as rapid, even when (as in Expt. 1) the extract is dilute (2%)- Peptonization. Having ascertained that yeast is actively peptolytic, I proceeded to investigate its peptonizing capacity, the test being the complete dis- integration of a small quantity of fibrin. The experiments were made with (a) solid yeast substance, (b) aqueous extracts, (e) extracts made with 2 °/o NaCl solution. (a) Experiments with solid yeast substance. The first of these experi- ments was of a general character, with the object of ascertaining definitely if digestion of fibrin were effected by yeast, and how digestion would be influenced by added alkali and acid. Experiment 1. In each of 6 bottles were placed 40 cc. distilled water and 5 grms. of partly dried brewers’ yeast, with toluol as the antiseptic; 0-5 grm. of fibrin was added to each, and the bottles were severally treated as follows : — to No. 1, nothing further was added ; to No. 2, 1 grm. chalk (reaction remained acid); to No. 3, Na2COs to 0-5 % ; to No. 4, HC1 to 0-04 % ; to No. 5, HC1 to o-i % ; to No. 6, HC1 to o-2 %. After 20 hours in the incubator the fibrin had not disappeared in any bottle, though in some it had diminished; 24 hours later it had disappeared in Nos. 1 and 2 ; 24 hours later it had disappeared in No. 3, whilst most of it remained in the others. In a repetition of this experiment (omitting bottle 6), with 10 grms. of yeast (25 %) in each bottle, the fibrin disappeared in all the bottles within 48 hours. These experiments show that yeast can digest fibrin ; and that the activity of any given mixture, as also its resistance to the retarding action of added acid or alkali, depends upon the amount of yeast that it contains. These two points were then further investigated. The material used in the following experiments was the dried ‘ granular * yeast already mentioned. 302 Vines. — The Proteases of Plants. With regard to the relation between the digestive activity of a mixture and the quantity of yeast contained in it, I found to begin with that 50 cc. of a mixture of chloroform-water with 1-25% yeast did not digest 0*2 grm. fibrin in 70 hours. This relation, as well as the action of acid and alkali, is further determined in the following comprehensive experiment : — Experiment 2. Mixtures were prepared of toluol-water (1 %) with 2-5 %, 5%, 10 %, and 20% dried yeast respectively. 40 cc. of each of these mixtures were put into each of 3 bottles, to one of which nothing was added, to the second Na2C03 to 2%, to the third HC1 to 0-5%, and 0-3 grm. fibrin to each bottle. After 22 hours’ digestion the results were : — Added nothing . 20 % bottles; fibrin gone 10 „ „ „ gone 5 „ „ „ not gone in any : 2-5 ,, „ „ not gone in any : after further digestion for 25 hours — 20% bottles; fibrin — ,, ,, — 5 » » » gone 2*5 „ ,i „ gone NafX)y gone gone not gone not gone HCl not gone not gone not quite gone not gone not gone not gone after further digestion for 28 hours— 20% bottles; fibrin — — 10 ,, ,, ,, 5 » ,» „ — not gone 2.5 „ „ „ — not gone 24 hours later, when the experiment closed, the results were the gone not gone not gone not gone same. These results suffice to indicate the relation between the digestive action on fibrin of yeast-mixtures of different strengths, and the degree to which digestive activity is affected by fairly strong acid and alkali in each case. Those afforded by the bottles containing 0-5 °/o HCl are of special interest in relation to Bokorny’s experiments, in which, as I have already pointed out (see p. 293), the conditions seem to have been such as to prevent any digestion at all of the added proteid. This criticism applies more particularly to those of his experiments (Nos. 1-8) in which the amount of yeast present was 5 °/o, the strength of acid o. 5-1 °/o H3P04 , and the proteid to be digested (ten times the weight of the yeast employed) the meat-residue from the preparation of Liebig’s extract. The facts upon which I base this criticism are supported by other results, subsequently described, obtained in experiments with yeast-extracts. It is more difficult to criticize Bokorny’s further experiments (Nos. 9-15), in which proteids of vegetable origin (prepared from Pea-flour, Soja-bean-meal, and Rape-cake) were the material to be digested, since the quantitative relations are not clearly stated : but, in view of the small amount of digestive products obtained, 303 Vines . — The Proteases of Plants . and the possibility that a considerable proportion of these may be attributed to the relatively large quantities of dried yeast added, amounting to 30 °/o or more of the weight of the proteid to be digested, they appear to be open to the same objection as the others. In all the foregoing experiments unboiled fibrin was used, so that a possible source of error was present. In order to eliminate this, some experiments were made in which the fibrin had been boiled for a few minutes. I found that 50 cc. of both a 10 °/o and a 20 °/o mixture of yeast with toluol-water digested 0-3 grm. fibrin in four days. (b) Experiments with aqueous extracts. The foregoing experiments with solid yeast afforded no information as to the solubility of the digestive protease ; so I had recourse to filtered watery extracts in the first instance. It was soon ascertained that active extracts can be obtained under suitable conditions. The following experiment proves this, and gives some indica- tion of the effect of acidity and of alkalinity : — Experiment 1. i2-| grms. of dried yeast were extracted (24 hours) with 250 cc. distilled water (yeast = 5 %). 50 cc. of the filtered liquid were put into each of 4 bottles, with 0-2 grm. fibrin, and treated thus : — to No. 1, nothing further added ; to No. 2, Na2C03 to 0-5 % ; to No. 3, HC1 to 0-05 % ; to No. 4, HC1 to 0-2 %. After 21 hours’ digestion in the incubator the fibrin in Nos. 1 and 3 showed diminution, in Nos. 2 and 4 it was not affected : 23 hours later it had disappeared in Nos. 1 and 3, but remained unaffected in Nos. 2 and 4. In the next place I tested the digestive activity of stronger yeast- extracts, whether of natural acidity, or alkaline, or with added acid. Experiment 2. Extracts of 10 grms. and of 20 grms. dried yeast with 100 cc. of toluol-water (1 %) were prepared and filtered. 30 cc. of each extract were put into each of 3 bottles with 0-3 grm. fibrin, one bottle of each having nothing added, another bottle having Na2COs added to 2 %, and the third bottle HC1 to 0*5%. After 2 7 hours’ digestion in the incubator the condition of the fibrin was — Nat. acid. Na2COs. HCL bottles 10 % yeast gone unaltered unaltered „ 20 % „ 21 hours later it was — gone attacked gone bottles 10 % yeast — unaltered attacked „ 20 — . attacked — 48 hours later the fibrin was still unaltered in the 10 % extract with Na2C03, and had not disappeared in either the 10 % extract with HC1 or the 20 % extract with Na2C03. Hence it appears that both io°/o and 20°/o watery yeast-extracts actively digest fibrin, and that, as might be expected, the latter are less affected by added acid and alkali than the former. Comparing the results of this experiment with those of the corresponding experiment (see p. 302) 304 Vines . — The Proteases of Plants . with solid yeast, it is clear that the resistance of solid yeast to acid and alkali is greater than that of the filtered extracts. The following experiment shows that a dilute and rapidly prepared watery extract has little or no action on fibrin : — Experiment 3, 2 grms. of dried yeast were extracted on a filter with 100 cc. toluol-water (1 %), the whole process being completed in 2 hours. 50 cc. of the filtrate were put into a bottle with 0*2 grm. fibrin. The fibrin had not disappeared after digestion in the incubator for 5 days. Having observed that the digestion of boiled fibrin took place in the presence of solid yeast, I made an experiment of this kind with yeast- extract, and with the same result: 60 cc. of a 2o°/o yeast-extract digested 0-2, grm. fibrin in 4 days. (c) Experiments with 2, °/o NaCl extracts. It occurred to me that it should be possible to prepare yeast-extracts that would digest fibrin more actively than did the aqueous extracts, by the use of some solvent other than distilled water. I found such a solvent in a 2°/o solution of common salt (NaCl): extracts made with this liquid, even when rapidly prepared, can be depended upon to digest fibrin, and are therefore specially suitable for the investigation of the digestive action of yeast upon this proteid. The following experiment, with a io°/o yeast-extract, gives a general indication of the effect of various antiseptics and of HC1 upon the digestion of fibrin : — Experiment 1. 20 grms. of dried yeast were extracted for several hours on a filter with 200 cc. of 2 % NaCl solution. 25 cc. of the filtrate were placed in each of 6 bottles, treated respectively as follows: — No. 1, nothing further added; No. 2, added HCN to 0-2 % ; No. 3, added NaF to 1 % ; No. 4, added chloroform to 0-5 %; No. 5, added toluol to 0*5 % ; No. 6, added HC1 to 0*2 % : to each added 0*2 grm. fibrin. After 26 hours in the incubator the fibrin had disappeared in all the bottles except No. 6, where it seemed to be quite unaltered. The following experiment demonstrates the inhibiting action of added acid and alkali : — Experiment 2. 40 grms. of dried yeast were extracted for several hours with 400 cc. of 2 % NaCl solution containing 1 % toluol. 40 cc. of the filtered liquid, with the addition of a little more toluol, were put into each of 9 bottles with 0-3 grm. fibrin : the further additions were — to No. 1, nothing ; to No. 2, 2 grms. precipitated chalk; to Nos. 3, 4, 5, HClto o* 1 %, 0.2 %, 0-5 % respectively ; to No. 6, H3P04 to 0-5 % ; to Nos. 7, 8, 9, Na2C03 to 1 %, 2 %, 3 % respectively. After 5 hours in the incubator the fibrin in Nos. 1, 2, 3 showed signs of solution. After 24 hours it had disappeared in Nos. 1, 2, 3, and 7 : it had not perceptibly diminished in any of the others, nor had it done so 72 hours later. 305 Vines . — The Proteases of Plants . These results show that addition of Na2C03 to 2°/0, or of H3P04 to °‘5°/o5 or to o -2% inhibits the digestive action of a io°/o yeast- extract made with 2% NaCl solution. On comparing them with those obtained in the corresponding experiments with io°/o solid yeast (p. 302), and with io°/o aqueous extract (p. 303), there is complete agreement as regards the effect of added acid, a matter of importance in relation to Bokorny’s experiments ; and as regards the effect of added alkali, the only preparation that withstood the action of 2°/o Na2C03 was that containing solid yeast. The inhibiting effect of added alkali was further investigated by the method employed in the corresponding experiments in autolysis (p. 296) and peptolysis (p. 300). Experiment 3. 10 grms. of dried yeast were extracted on a filter for 3-4 hours with 200 cc. 2 % NaCl solution containing 1 % toluol. 50 cc. of the filtrate were placed in each of 3 bottles, to which Na2C03 was added to 1 %, 2 %, 3 % respectively; the bottles were then placed in the incubator for 2 hours. On being removed, the con- tents of each bottle were divided into 2 equal parts, one of which remained alkaline, the other being made slightly acid with HC1. There were then 6 bottles, 3 acid and 3 alkaline, each containing 25 cc. of liquid : to each was added 0-2 grm. fibrin and a little more toluol. The 6 bottles were placed in the incubator, together with another bottle, No. 7, containing 25 cc. of the original extract, which had not been treated with Na2C03, and o° 2 grm. fibrin. After 20 hours’ digestion the fibrin had disappeared in bottle No. 7 (natural acidity) : but it had not undergone any apparent change in Nos. 1-6, nor had it done so after 48 hours’ digestion. Thus treatment for two hours with even 1 °/o of Na2C03 sufficed to destroy the digestive activity of a 5°/o yeast-extract made with salt- solution, a result that more closely defines the action of alkali than those of the preceding experiments, in which it was ascertained that neither a 5% yeast-mixture (p. 302) nor a io°/o watery extract (p. 303) digested fibrin in the presence of 2°/o Na2C03. But it has not yet been made clear what is the advantage of a NaCl extract over a watery extract of yeast. When the extracts are strong, say io°/o, the advantage is not very marked: the digestive action of both is vigorous and rapid, more especially when the period of extraction and filtration has been prolonged for many hours. When, however, dilute and rapidly prepared extracts are employed, the greater activity of the NaCl extract is most apparent, as the following experiment shows. Experiment 4. 1 grm. of dried yeast was treated with 50 cc. toluol-water (1 %); 1 grm. of yeast was treated with 50 cc. of 2 % NaCl solution : the mixtures were at once filtered, so that the preparation of the filtered liquids did not last more than 2 hours. To each bottle 0-2 grm. fibrin was added, and then they were both put into 306 Vines. — The Proteases of Plants. the incubator. After 20 hours’ digestion the fibrin in the watery extract was unaltered, whilst that in the NaCl extract was much diminished, and 4 hours later had dis- appeared. The fibrin in the watery extract had not disappeared after digestion for 4 days longer. In another experiment in which extraction was prolonged for several hours, the activity of stronger aqueous and 2°/0 NaCl extracts of yeast was compared. It was found that 30 cc. of a 10 °/o yeast NaCl extract digested o-2 grm. fibrin within 18 hours, and the same quantity of 5 °/o yeast NaCl extract digested the fibrin in 24 hours ; whereas the times of digestion by the corresponding aqueous extracts were 46 and 66 hours respectively. The object of the next experiment was to ascertain in what way NaCl affects digestion. Does it directly promote it, or does it do so indirectly by dissolving out of the yeast something that distilled water fails to extract or extracts less completely ? Experiment 5. 3 grms. of dried yeast were extracted with 60 cc. of 2 % NaCl solution (toluol 1%); 6 grms. were also extracted with 120 cc. of toluol- water : extraction and filtration occupied about an hour and a half. 40 cc. of the NaCl extract were put into a bottle (No. 1) with 0-2 grm. fibrin : of the aqueous extract, 40 cc. were put into a second bottle (No. 2), and other 40 cc. into a third bottle (No. 3), to which NaCl to 2 % was added, as also 0-2 grm. fibrin to both 2 and 3. After 18 hours in the incubator the fibrin in No. 1 had disappeared: this was not the case in either No. 2 or No. 3, nor had it disappeared after digestion for 24 hours. This result, in the first instance, confirms the conclusions as to the superior activity of NaCl extracts as compared with aqueous extracts ; and, in the second place, it gives an explicit answer to the question pro- pounded above. It clearly shows that the presence of NaCl is of importance in the process of extraction rather than in the process of diges- tion ; and it may be inferred that the NaCl solution dissolved out of the yeast something, no doubt a protease, that water alone failed to extract. An experiment was made to test the action of a NaCl extract on boiled fibrin. It was found that 60 cc. of a 20°/o yeast-extract with NaCl digested 0-2 grm. fibrin in 3 days: the action was slow, but it was more rapid than that of a watery extract of the same strength (see P. 304). Summary of Experiments on Yeast. Evidence has been adduced to prove that yeast can effect both peptolysis and peptonization. The fact that these processes can be carried on by filtered extracts makes it clear that they are not due to the yeast- plant as a living organism, but to one or, perhaps, more substances that can be dissolved out of it ; and there can be no doubt that the active substance is, in any case, a protease. An important issue is thus raised Vines. — The Proteases of Plants. : 307 that may be expressed in the two questions — (1) is there, as is now generally held, a single protease in yeast, or is there more than one, and in the latter case, how many ? (2) What is the nature of the protease or proteases ? The results of my experiments will be briefly considered with a view to a reply. Peptolysis. The most important fact that has been brought to light is the rapidity with which this process is effected : thus in an autolysis- experiment (p. 395) it was found to have proceeded actively in 2I hours ; and in a Witte-peptone experiment (p. 299) in 1 hour. Moreover, in the latter experiment the watery- extract was very dilute (2 °/o) and the time of extraction very short (1 hour): therefore the protease concerned is readily soluble in water. It has been shown, further, that peptolysis is most active at or near the natural acidity of the liquid, at a degree of acidity determined by the presence of acid phosphates. It is retarded, and eventually arrested, by any deviation from this degree of acidity, in the direction either of alkalinity or of increased acidity : the effect of added alkali or acid varies with the amount of solid yeast present, or with the strength of the extract, as also with the length of the exposure to its action. Thus, in the case of a 5°/0 yeast-mixture, peptolysis was found to be inhibited in the presence of either about o *6°/o HC1, or 2% Na2C03, for 24 hours, as also by ex- posure to 3°/o Na2C03 for 2 hours (p. 296). Similar results were obtained with 5 °/o watery extracts acting on Witte-peptone (p. 299). Peptonization. Under this heading I include the experiments upon the digestion of fibrin. It has been made clear, in the course of these experiments, that peptonization takes place much less rapidly than peptolysis. Even with relatively strong yeast-extracts several hours were required for the digestion of a small quantity of fibrin : thus 40 cc. of a 5 °/o yeast-mixture did not digest 0-3 grm. fibrin at all in 22 hours, though the fibrin disappeared within 24 hours more (p. 302). The next point of importance is the relation between watery extracts of yeast and NaCl extracts. When the extracts were strong (5% and upwards) and the time of extraction long, the difference in the activity of the two kinds of extracts was not found to be important ; but when the extracts were dilute and the time of extraction short, the difference was striking. A rapidly prepared 2 °/Q watery extract did not digest fibrin at all (p. 304), whilst a similarly prepared NaCl extract (p. 305) digested the fibrin in about 24 hours. The inference to be drawn is that the peptonizing enzyme is not readily soluble in distilled water, but is readily soluble in 2 °/o NaCl solution. Peptonization was found, like peptolysis, to proceed most actively at or near the natural acidity of the liquid, and to be arrested or retarded by the addition of either acid or alkali. But a comparison of the results 308 Vines. — The Proteases of Plants. shows that the two processes do not exactly agree in the latter respect : it appears that the range of reaction is rather more limited for peptoniza- tion than for peptolysis. Thus, with regard to the action of added acid, the only case in which digestion of fibrin took place in the presence of as much as 0*5 °/o HC1 was one in which the mixture or extract was very strong (20% see p. 302) ; on the other hand, 0-5 °/o HC1 did not inhibit peptolysis in a 5 °/o yeast-mixture (p. 29 6) or in 5 °/o yeast-extract (p. 299). Similarly, with regard to the action of added alkali, whilst it is true that in the experiments in which solid yeast was used, peptonization and peptolysis were equally affected by the addition of Na2C03 to 2 °/o (compare Expt. 2, p. 302, with Expt. 4, p. 296), there was a marked difference in favour of peptolysis when extracts were employed (compare Expt. 2, p. 303, and Expt. 2, p. 304, with Expts. 2, p. 299, and 4, p. 300). In the two former digestion of fibrin by io°/o or 20 °/o extracts was inhibited, whilst in the two latter peptolysis was effected by 5 °/o extracts treated with the same amount of alkali. In Expt. 3, p. 305, digestion of fibrin was inhibited by 1 °/o Na2C03. Conclusions . The chief results of the investigation are these : — • (1) dilute yeast-mixtures or aqueous extracts rapidly effect peptolysis, as indicated by the tryptophane reaction, but do not digest fibrin ; (2) dilute NaCl extracts of yeast readily digest fibrin ; (3) peptolysis and peptonization are influenced in the same manner, but not in the same degree, by the addition of acid or alkali. I infer that these two digestive processes are not effected by one and the same protease. On the contrary, the facts indicate the presence of two proteases : the one exclusively peptolytic, readily soluble in water ; the other peptonizing, less soluble in water, but readily soluble in 2°/o NaCl solution. The Mushroom. Agaricus (Psalliotd) campestris. The discovery of proteases in Basidiomycetous Fungi seems to have been made by Hjort (11), who found that watery extracts of them digested fibrin. In the case of Agaricus (. Pleurotus ) ostreatus , digestion was most active when the liquid was neutral ; less active when acidified with 0-5 °/o oxalic acid, and was altogether inhibited by alkalinity. The fibrin entirely disappeared in 40 hours ; the digested liquid then giving no biuret, but strong tryptophane-reaction, and containing leucin and tyrosin. In the case of Polyporus sulfur eus, the naturally acid extract readily digested fibrin, as did also extracts acidified with HC1 to 0*2 °/0 or with oxalic acid to 0-25 °/Q ; but neutralized or alkaline extracts did not digest at all. In a 12-hours’ digestion the liquid contained albumoses and peptones, but no amido-acids or hexon-bases. 309 Vines. — The Proteases of Plants. Shortly afterwards the matter was investigated by Bourquelot and Hdrissey (12). They found that a filtered watery extract of Agaricus {Amanita) muscarius digested five-sixths of the caseinogen of skim-milk within four days, and they detected tyrosin in the digested liquid. Similar results were obtained with Polyporus sulfureus. In the course of a few experiments with the mushroom, I obtained evidence (13) that the tissue can both peptonize fibrin and peptolyse the lower proteids, thus confirming in a general way the conclusions of my predecessors. Somewhat more recently the matter has been taken up by Delezenne and Mouton (14), and with widely different results. They prepared extracts, using normal saline solution (o*8°/o NaCl) with chloroform or toluol, of the dried pilei of various species (the Mushroom, Amanita muscaria, Amanita citrina, Hypholoma fasciculare) , which readily peptolysed peptone, and digested gelatine and casein, but had no action on fibrin. This last result seems so contradictory to previous observations that I have thought it necessary to make some further experiments to test its accuracy. Peptonization. The test applied was the disappearance of a relatively small quantity of fibrin. The laminae were in all cases removed from the pileus. In the first instance the actual tissue, reduced to a pulp, was made use of : provided that the material was mature, the pileus being fully expanded, the result was that digestion of fibrin took place. Experiment 1. io grms. fresh mushroom-pulp were digested with ioocc. dis- tilled water containing chloroform (0-25 %) and 2 grms. fibrin for 26 hours ; at the end of this time the fibrin was completely disintegrated. Experiment 2. 15 grms. fresh mushroom-pulp were digested with 1 00 cc. chloroform- water (0-5 %) and 2 grms. fibrin : after 24 hours’ digestion the fibrin was dissolved. Experiment 3. In an experiment similar to and contemporaneous with the preceding, where the antiseptic was HCN (0.2 %), the fibrin was disintegrated and mainly dissolved. When a watery extract of the fresh ripe pileus was used, from which the solid matter had been removed either by straining through muslin or by filtering through paper, the result was less certain : in most cases the fibrin seemed to be somewhat diminished in quantity, but rapid and com- plete solution was not constant. The following are some of the more successful experiments : — Experiment 4. 1 grm. of fibrin was digested with 50 cc. of a watery extract of 40 grms. mushroom-pulp with 200 cc. distilled water, with HCN to 0-2 % : within 24 hours the fibrin was completely disintegrated. 3io Vines . — The Proteases of Plants . Experiment 5. In this experiment 0-2 grm. of fibrin was completely digested by 50 cc. mushroom-extract, to which NaF to 1 % had been added, in 18 hours. These results clearly indicated that watery extracts of the Mushroom digest fibrin, in agreement with those of Hjort. However, I was not altogether satisfied, as digestion of fibrin did not occur in every experiment. With the object of obtaining active extracts with greater certainty, I had recourse to the method of extraction with 2 °/0 NaCl solution that had proved serviceable in the case of yeast, and with the same success. Experiment 6. 120 grms. fresh mushroom-pulp were extracted with 300 cc. 2 % Na Cl solution for 2 1 hours ; the liquid was then strained off through muslin, and placed in 8 bottles each holding 40 cc. with 0-2 grm. fibrin. The treatment of the bottles was — No. 1, nothing added; No. 2, liquid boiled; No. 3, added HCN to o-2 %; No. 4, added NaF to 1 % ; No. 5, added chloroform to 0-5 % ; No. 6, added toluol to 0-5 % ; No. 7, added HC1 to o*i % ; No. 8, added HC1 to 0-2 %. After 20 hours’ digestion in the incubator the fibrin had completely disappeared in Nos. 1, 3, 4, 5, 6 ; it had not undergone any apparent diminution in Nos. 2, 7, 8. I found that it is also possible to obtain active extract from dried Mushroom. Experiment 7. 8 grms. of dried pileus (had been kept for over 6 months) extracted for 4 hours with 100 cc. 2 % NaCl solution; filtered, and added toluol to 1 %. 30 cc. of this liquid had completely digested 0-2 grm. fibrin within 19 hours. I incidentally observed that 20 cc. of the expressed juice of fresh Mushroom digested 0-2 grm. fibrin within 24 hours in the presence of 1 °/o toluol. In another experiment a comparison was instituted between the relative activities of aqueous and 2°/o NaCl extracts which were (a) of natural acidity, or ( b ) acidified to o-i °/o HC1, or (c) made alkaline by adding NazC03 to 0-5 °/o, in the presence of toluol. Experiment 8. 60 grms. fresh mushroom-pulp were extracted for about 24 hours with 250 cc. distilled water ; a similar quantity of pulp was extracted for the same time with 250 cc. 2% NaCl solution. 40 cc. of the filtered aqueous extract were put, with some toluol and 0-2 grm. fibrin, into each of 3 bottles: to No. 1, nothing was added ; to No. 2, HC1 to o*i % ; to No. 3, Na2C03 to 0-5 %. 40 cc. of the NaCl extract were also put into each of 3 bottles, and similarly treated. After 24 hours in the incubator (38-40^.) the result was that the fibrin had not been digested, or apparently diminished, in any one of the bottles containing the aqueous extract, whilst it had disappeared in the bottle containing the NaCl extract alone, but not in either the Na2C03 or the HC1 bottles. The superior activity of the NaCl extract is more marked, as was the case with yeast, the more dilute the liquids. Experiment 9. 2-5 grms. of partly dried Mushroom were extracted (4 hours) with 50 cc. 2 % NaCl solution, and an equal quantity with 50 cc. distilled water. Vines. — The Proteases of Plants. 31 1 25 cc. of the NaCl extract were put into each of 2 bottles with the addition of some toluol and 0-2 grm. fibrin, the fibrin having been previously boiled in one case : 25 cc. of the watery extract were put, with toluol and unboiled fibrin, into each of 2 bottles, and to one of them 1 grm. NaCl was added. After 19 hours in the incubator, the unboiled fibrin in the bottle containing NaCl extract had disappeared ; the fibrin had not disappeared in any of the other three. This experiment demonstrates not only the superior activity of the NaCl extract, but also the solvent action of the NaCl (compare Yeast, P. 306). The next experiment relates to the action of acid and alkali upon a rapidly prepared NaCl extract. Experiment 10. 100 grms. fresh mushroom-pulp were extracted with 300 cc. of 2 % NaCl solution containing toluol, and filtered, the whole process of preparation occupying about an hour. 40 cc. of the extract were put into each of 5 bottles, with 0-2 grm. of fibrin, treated thus — to No. 1, nothing further added; to No. 2, Na2C03 to 1 % ; to No. 3, Na2C03 to 2 % ; to No. 4, HC1 to o-i % ; to No. 5, HC1 to 0-2 %. After 24 hours in the incubator, the fibrin had disappeared in No. 1, was unaltered in Nos. 2 and 3, and seemed to be attacked in Nos. 4 and 5 : 24 hours later the fibrin had not disappeared in any one of these four bottles. These results, as also those of Expts. 6 and 8, show that the pepto- nizing activity of a mushroom-extract, of the strengths employed, is destroyed by the addition to the naturally acid liquid of o-i °/o HC1, or of 0.5-1 % Na2COs. So far it has been assumed that the disappearance of the fibrin in the experiments implied peptonization. In order that there might be certainty on this point, the following experiment was made : — Experiment 11. 60 cc. of NaCl extract (toluol 1 %) were put to digest 2 grms. of fibrin: in 20 hours the fibrin had disappeared, and the liquid, after boiling and filtering, gave a well-marked biuret-reaction. At the commencement of the experi- ment, a sample of the extract gave no biuret-reaction. The results of these experiments on fibrin are such as to lead inevitably tothe conclusion that the mushroom contains a peptonizing enzyme capable of digesting fibrin : it is therefore remarkable that Delezenne and Mouton (see p. 309) should have expressed the contrary opinion. The reason for this contradiction is that these observers used boiled fibrin in their experiments. This precaution, it is true, obviates a possible source of error by eliminating any self-digestion of the fibrin : but it is doubtful if this advantage compensates for the disadvantage involved, the disadvantage of missing altogether the presence of the peptonizing enzyme. Proteids coagulated by heat offer, as is well known, considerable resistance to the digestive action even of animal proteases ; so that it is not surprising Y 312 Vines . — Proteases of Plants. that the protease of the mushroom should have failed to act upon them. This precaution is not, however, absolutely indispensable : for it is a simple matter to check possible self-digestion of the fibrin by control-experiments. Thus, in the foregoing Expt. 6, digestion of fibrin took place in Nos. i, 3, 4, 5, 6, but not in No. 2 where the liquid (but not the fibrin) had been boiled : had the results given by Nos. 1, 3, 4, 5, 6 been due to self-digestion, then the probability is that the same result would have been given by No. 2, which was not the case. All the fibrin used in these experiments was prepared and preserved at one time and in the same manner : hence the fibrin may be regarded as a constant factor, all the variations being due to the mushroom-liquids, as affected by the various substances added to them. I may add that I have not succeeded in observing digestion of boiled fibrin, though strong mixtures and extracts were used, and the experiment was continued for a week. Peptolysis. There is already a certain amount of evidence that the mushroom contains an actively peptolytic enzyme, which is to be found in all the papers that I have previously cited (Nos. 11, 12, 13, 14). My main object in making further experiments has been not so much to establish this fact, as to determine the conditions that affect the activity of the protease and so to arrive at some conclusion as to its nature. I may say, however, that I have never failed to obtain from mushrooms, whether ripe or immature, and with great facility, a watery extract — in some cases a glycerine- extract — that readily peptolysed Witte-peptone, as indicated by the trypto- phane-reaction. It should be explained that, on account of the deep colour of the liquids, it was not possible to apply the tryptophane-reaction directly : the liquids had first to be boiled with animal charcoal and then filtered. The following experiment gives a general idea of the method employed and of the results obtained : — Experiment 1. 90 grms. of fresh mushroom-pulp were extracted for 18 hours with 200 cc. chloroform-water: on straining through muslin, a red, opalescent, acid liquid was obtained. 40 cc. of the liquid were placed in each of 6 bottles, with 0-5 grm. Witte-peptone and a little toluol, the bottles being then treated as follows: added to 1, nothing further; to 2, 1 grm. chalk to neutralize any free acid ; to 3, Na2C03to 1*25 % ; to 4, HC1 to 0-04 % ; to 5, HC1 to o-i % ; to 6, HC1 to 0-2 %. After 23 hours in the incubator the tryptophane-reactions were — 1, marked; 2, marked; 3, strong; 4, marked ; 5, strong; 6, faint. There was thus evidence of active peptolysis having taken place within a distinctly alkaline and a distinctly acid range of reaction, and of its arrest in the presence of stronger acid. Vines. — The Proteases of Plants. 313 The next experiment gives some idea of the rapidity with which peptolysis was found to be effected by a dilute mushroom-extract, and shows further that peptolysis is a much more rapid process than is peptonization. Experiment 2. 5 grins, fresh mushroom-pulp were extracted on a filter for about an hour with 100 cc. 1 % toluol-water. The filtered liquid gave no trypto- phane-reaction. 50 cc. of it were put into a bottle with 0*3 grm. fibrin, and 50 cc. into another bottle with 0*5 grm. Witte-peptone. In I hour the contents of the Witte-peptone bottle gave a distinct tryptophane- reaction : 24 hours later the reaction was strong. In the same time the fibrin in the other bottle had not disappeared ; but it disappeared within the next 24 hours. The following experiment brings out clearly the relative rapidity of peptolysis and of peptonization, and of the effect of added acid and alkali on these processes respectively: — Experiment 3. 5 grms. of dried powdered mushroom were extracted with 200 cc. toluol-water (1 %) : 5 grms. were also extracted with 200 cc. of 2 % NaCl solution containing 1 % toluol. 40 cc. of the NaCl extract were put into each of 3 bottles, with 0-2 grm. of fibrin; also 40 cc. into each of 3 bottles with 0-5 grm. Witte-peptone : to 1 fibrin bottle and 1 Witte-peptone bottle (Nos. 1), nothing was added ; to another pair of bottles (Nos. 2) Na2C03 to 1 % was added ; to a third pair (Nos. 3), HC1 to on % was added. An exactly similar series of bottles containing the aqueous extract was prepared. After 24 hours’ digestion the results were — Nos. 1 fibrin gone tryptophane strong NaCl ext. 2 3 not gone not gone strong marked Aq. ext . 1 2 not gone not gone very strong strong 3 not gone strong. Hence it is apparent that peptolysis and peptonization are inde- pendently affected by the addition of acid and alkali. Conclusions. tr . (a, \ :■ ; • • , . , i}.' .... . . l .. ....... i Although my investigation of the mushroom has not been so minute as in the case of yeast, the results obtained suffice to draw similar con- clusions. In the first place, the conclusion is justified that the mushroom contains a peptolysing enzyme which is readily extracted by water and acts with rapidity. Secondly, it is equally clear that the mushroom contains a peptonizing enzyme capable of digesting fibrin. As in the case of yeast, so here, the question arises as to whether both processes are effected by one and the same protease, or whether there are not two proteases in the mushroom, the one especially peptolytic, the other especially peptonizing. 3H Vines.— -The Proteases of Plants. The observed facts are, on the whole, favourable to the latter conclusion. To begin with, the results of the peptolysis-experiment No. 2, suggest that the rapid peptolysis and the slow peptonization should be interpreted as being due to the presence of two proteases, the one readily soluble in water, the other less soluble. Again, the superior peptonizing activity of NaCl extracts compared with watery extracts (see p. 310), suggests that the protease concerned is more readily soluble in 2 °/o NaCl solution than in distilled water. The importance of NaCl as a solvent is demonstrated in Expt. 9 (p. 310). The inference drawn from these observations on solubility is supported by the observations upon the effect of added acid and alkali on peptolysis and peptonization respectively. Taking the limits of peptonization as HC1 on °/o and Na2C03 1 °/o (p.311), those of peptolysis are less restricted, extending beyond these limits in both directions. The Nature of the Proteases. The generally accepted opinion with regard to the two Fungi in question is that they contain a single protease. In the case of yeast, Hahn and Geret, Bokorny, and others, regard this protease as a trypsin : and Hjort has made the same suggestion in the case of the Basidiomycetous Fungi investigated by him. My observations, as already explained, lead me to the conclusion that two proteases exist in these plants, the one peptolytic, the other peptonizing. It remains now to consider what the nature of these proteases may be. At a meeting of the Linnean Society of London, on November 20, 1902 (Proceedings, 1902-3, p. 42), I announced the discovery in many plants and different parts of plants of a peptolytic enzyme analogous to the recently discovered entero-erepsin of the animal body : a more complete account of my researches was soon afterwards (January, 1903) published in this periodical (13), the mushroom having been one of the plants investigated. The further observations, of which an account has now been given, confirm me in the conclusion that the mushroom contains an erepsin, that is, a peptolytic enzyme which is unable to peptonize the higher proteids such as fibrin and albumin ; and they justify the extension of the conclusion to yeast. This vegetable erepsin is not, however, identical in properties with either the entero-erepsin discovered by Cohnheim or the pancreato-erepsin discovered by Vernon (see p. 290). The action of both these animal erepsins is limited to neutral or feebly alkaline liquids, whilst I have found that vegetable erepsin can act through a fairly wide range of acid and alkaline reaction, its greatest activity being manifested when the reaction of the liquid is at or near natural acidity. Hence vegetable erepsin affords a new type of ereptic action. Vines. — The Proteases of Plants. 315 Now as to the nature of the peptonizing enzyme. It may be either a pepsin or a trypsin, but the question is, which ? This question is more easy to put than to answer, because there is at present no method by which the enzyme can be obtained free from the associated erepsin ; and until that is done, no direct answer can be forthcoming. But it is possible to form an opinion upon indirect evidence. It should be borne in mind that, as I have elsewhere stated, there is at present no well-established instance of the occurrence of a merely peptonizing enzyme in the Vegetable Kingdom. A more important point is, however, that of the reaction of the liquid in which the protease will work. The activity of animal pepsin is limited to acid liquids ; whilst animal trypsin, though most active in a distinctly alkaline liquid, can nevertheless work in a neutral or even in a slightly acid liquid. In its range of reaction, the vegetable peptonizing enzyme resembles animal trypsin rather than pepsin ; but with this difference, that whereas animal trypsin is most active in a distinctly alkaline liquid, the vegetable protease is most active in a distinctly acid liquid. It seems therefore probable that the protease in question may be a trypsin of a new type, characterized by its activity in an acid, rather than an alkaline, liquid. On these grounds it is suggested that the yeast and the mushroom contain two associated proteases, vegetable erepsin and vegetable trypsin, an association that finds its analogue in the pancreatic secretion of animals which, as Vernon has recently shown (2), contains both erepsin (pancreato- erepsin) and trypsin proper. The term f vegetable trypsin 1 is already in common use, but in a wider sense than that in which I have just employed it. Hitherto it has been applied to the vegetable proteases without taking the presence of erepsin into account, whereas I limit the term to the peptonizing enzyme apart from the erepsin. Vernon’s results have introduced the same distinction into animal physiology ; formerly the term ‘ trypsin ’ was applied to the protease of the pancreas, but this, as he has shown, is really a mixture of pancreato-erepsin with true trypsin. A few lines may be devoted, in conclusion, to the consideration of the question as to how far these views are applicable to plants in general. It can hardly be doubted that, at some period in their existence, all plants and all parts of plants contain a peptolytic enzyme concerned in promoting the distribution of proteids in the temporary form of amido-acids, &c. But it is not clear that a peptonizing enzyme is of such general occurrence : on the contrary, as I have already pointed out (13, p. 262), many parts of plants failed to digest fibrin in my experiments. It is possible that in those experiments the precise conditions most favourable to peptonization were not provided : it may be that, for instance, the use of NaCl extracts that have given such good results with the Yeast and the Mushroom, will give similar results in other cases. I have not yet had time to make 3 1 6 Vines. — The Proteases of Plants . extended investigations in this direction ; but what already I have done is, I think, of sufficient interest to be mentioned here. In a previous paper (13, p. 254), I gave an account of some peptonizing experiments with the bulbs of the hyacinth, the tulip, and the onion, the bruised bulb-tissue being employed. The results were not conclusive, but indicated that whilst the onion did not digest fibrin, the hyacinth and the tulip did so to some extent in a slightly alkaline liquid. I have since resumed these experiments, using watery or NaCl extracts of the bulbs, and taking the disappearance of a small quantity of fibrin as the test of digestion, following the method adopted in the investigation of the yeast and the mushroom, with interesting results. Experiment 1. A hyacinth bulb, weighing about 75 grms., was reduced to pulp and extracted for two hours with 100 cc. of 2 % NaCl solution : the liquid was strained through muslin. 30 cc. were put into each of 3 bottles with o 2 grm. fibrin and some toluol : to No. 1, nothing further was added ; to No. 2, HC1 to on % ; to No. 3, Na2C03 to 0.5 %. After 23 hours in the incubator, the fibrin had disappeared in Nos. 1 and 3, and was attacked in No. 2 : the tryptophane-reactions were, strong in No. 1, distinct in No. 2, marked in No. 3 : 28 hours later the fibrin had disappeared also in No. 2. Experiment 2. Similar extracts were prepared of the tulip and onion bulbs : 40 cc. of the extract were in each case put, with 0-2 grm. fibrin, into each of 4 bottles, with the following additions: to No. 1, nothing; to No. 2, HC1 to 0-05 % ; to No. 3, HC1 to 0*2 % ; to No. 4, Na2C03 to 1 %. Within 48 hours the results were : tulip, fibrin gone in No. 1 ; nearly gone in Nos. 2 and 4 ; unaltered in No. 3 : onion, fibrin unaltered in all. All the bottles gave more or less strong tryptophane-reaction. The positive results given by the hyacinth and the tulip point to the probability that a peptonizing enzyme is more generally present in plants than is at present recognized. But the negative result given by the onion is even more suggestive ; clearly peptolysis (autolysis) had occurred in this experiment without peptonization of the added fibrin. It appears, there- fore, that erepsin is present in the onion without any other protease. If this be so, it is important evidence in favour of the existence of an ereptic protease in plants, and strengthens the conclusion, already expressed, that in those plants that can digest fibrin there is also present a distinct peptonizing enzyme. Vines. — The Proteases of Plants. 3i7 List of Papers Referred to. 1. Vines : Proteolytic Enzymes in Plants (II) ; Annals of Botany, vol. xvii, 1903, p. 59 7 (June). 2. Vernon: The Peptone-splitting Ferments of the Pancreas and Intestine; Journ. Physiol., vol. xxx, 1903, p. 330. 3. Butkewitsch : Umwandlung der Eiweissstoffe durch die niederen Pilze, etc. ; Jahrb. f. wiss. Bot. xxxviii, 1902, p. 147. 4. Malfitano : Sur la protdase de 1’ Aspergillus niger; Ann. Inst. Pasteur, t. xiv, 1900, p. 420. 5. Weis : Etudes sur les enzymes protdolytiques de l’orge en germination ; Compte-rendu des travaux du Laboratoire de Carlsberg, v, 1903, p. 133. 6. Vines: Tryptophane in Proteolysis; Annals of Botany, vol. xvi, 1902, p. 13. 7. Hahn und Geret: Ueber das Hefe-Endotrypsin ; Zeitschrift fiir Biol., Bd. xl, 1900, p. 117. 8. Bokorny : Die proteolytischen Enzyme der Hefe ; Beiheftezum Bot. Centralblatt, Bd. xiii, 1902, P- 235- 9. Naegeli : Ernahrungschemismus der niederen Pilze; Bot. Mittheilungen, Bd. iii, 1881, p. 464 (Sitzungsber. der K. Bay. Akad. d. Wiss. in Miinchen, 5. Juli 1879). 10. : Ueber die chemische Zusammensetzung der Hefe ; ibid. p. 270. 11. Hjort : Neue eiweissverdauende Enzyme; Centralblatt fiir Physiol., x, 1897, p. 192. 12. Bourquelot et Herissey : Recherche et presence d’un ferment soluble proteo-hydrolytique dans les Champignons ; Comptes Rendus de la Soc. de Biol., ser. 10, t. v, 1898, p. 972. 13. Vines : Proteolytic Enzymes in Plants (I) ; Ann. Bot., vol. xvii, 1903, p. 254 (January). 14. Delezenne et Mouton : Sur la presence d’une kinase dans les Champignons Basidiomycetes ; Comptes Rendus, t. cxxxvi (Jan. 19), 1903, p. 167 : also, Sur la presence d’une erepsine dans les Champignons Basidiomycetes, ibid. p. 633 (Mar. 9, 1903). . ■ . . ■ NOTES ON THE ORIGIN OF PARASITISM IN FUNGI1.— Up to the present no definite explanation has been offered as to why a given parasitic Fungus is often only capable of infecting one particular species of plant. This, however, is well known to be the case, for although the spores of Fungus-parasites germinate freely on the surface of any plant when moist, infection only takes place when the spores germinate on the particular species of plant on which the Fungus is known to be parasitic. This apparently selective power on the part of the Fungus I consider to be due to chemotaxis. An extensive series of experiments was conducted with various species of Fungi, including saprophytes, facultative parasites, and obligate parasites, and the results are given in tabulated form in the full paper. The chemotactic properties of substances occurring normally in cell-sap were alone tested ; among such may be enumerated saccharose, glucose, asparagin, malic acid, oxalic acid, and pectase. In those instances where the specific substance, or combination of substances, in the cell-sap assumed to be chemotactic could not be procured, the expressed juice of the plant was used. These experiments proved that saprophytes and facultative parasites are positively chemotactic to saccharose, and this substance alone is sufficient in most instances to enable the germ-tubes of facultative parasites to penetrate the tissues of a plant, unless prevented by the presence of a more potent negatively chemotactic or repellent sub- stance in the cell-sap. As an illustration, Boirytis cinerea , which attacks a greater number of different plants than any other known parasite, cannot infect apples, although saccharose is present, on account of the presence of malic acid, which is negatively chemotactic to the germ-tubes of Botrytis. In the case of obligate parasites the cell-sap of the host-plant proved to be the most marked positive chemotactic agent. Malic acid is the specific substance that attracts the germ-tubes of Monilia fructigena into the tissues of young apples; whereas the enzyme pectase performs the same function for the germ-tubes of Cercospora cucumis , an obligate parasite on the cucumber. Immune specimens of plants belonging to species that are attacked by some obligate parasite owe their immunity to the absence of the substance chemotactic to the parasite. Purely saprophytic Fungi can be educated to become parasitic, by sowing the spores on living leaves that have been injected with a substance positively chemotactic to the germ-tubes of the Fungus experimented with. By a similar method of pro- cedure, a parasitic Fungus can be induced to attack a different species of host-plant. 1 Abstract, reprinted from the Proceedings of the Royal Society. [Annals of Botany, Vol. XVIII. No. LXX. April, 1904.] Z 320 Notes. These experiments prove what has previously only been assumed, namely, that parasitism in Fungi is an acquired habit. A series of experiments prove that infection of plants by Fungi occurs more especially during the night, or in dull, damp weather. This is due to the greater turgidity of the cells, and also to the presence of a larger amount of sugar and other chemotactic substances present in the cell-sap under those conditions. GEORGE MASSES, Kew. CULTURAL EXPERIMENTS WITH ‘ BIOLOGIC FORMS’ OF THE ERYSIPHACEAE k — In the introductory remarks the author points out that through specialization of parasitism 4 biologic forms ’ have been evolved in the Erysiphaceae which, both in their conidial (asexual) stage and ascigerous (sexual) stage, show specialized and restricted powers of infection. The powers of infection, characteristic of each 4 biologic form,’ are under normal conditions sharply defined and fixed, and hitherto the result of the experiments of numerous investigators — both in regard to the present group of Fungi and to the Uredineae, where the same specialization of parasitism occurs — has been the accumulation of evidence tending to emphasize the immutability of £ biologic forms/ The second part of the paper gives the result of cultural experiments with ‘ biologic forms ’ of Erysiphe Graminis DC., carried out during the past summer in the Cambridge University Botanical Laboratory. It has been found that under certain methods of culture, in which the vitality of the host-leaf is interfered with, the restricted powers of infection, characteristic of ‘ biologic forms,' break down. In the first method of culture adopted, the leaf, which was either attached to a growing plant, or removed and placed in a damp chamber, was injured by the removal of a minute piece of leaf-tissue. In this operation the epidermal cells on one surface, and all or most of the mesophyll tissue, were removed at the cut place, but the epidermal cells on the other surface (opposite the cut) were left uninjured. Conidia were sown on the cuticular surface of the uninjured epidermal cells over the cut. In a few experiments the conidia were sown on the internal tissues of the leaf exposed by the cut, and these gave the same results. Using this method of culture, over fifty successful experiments, of which details are given, were made. In these the conidia of certain ‘ biologic forms ’ were induced to infect ‘ cut ’ leaves of host-species which are normally immune against their attacks. The experiments proved that the range of infection of a ‘ biologic form ' becomes increased when the vitality of a leaf is affected by injury, and also that species of plants ‘ immune ' in nature can be artificially rendered susceptible. Further experiments showed that the conidia of the Fungus produced on a ‘ cut ' leaf are able at once to infect fully wiinjured leaves of the same host-species. In other experiments, a method suggested by Professor H. Marshall Ward with the object of avoiding lesion of the leaf, was adopted. Leaves were injured by touching the upper epidermis for a few seconds with a red-hot knife, and conidia were 1 Abstract, reprinted from the Proceedings of the Royal Society. [Annals of Botany, Vol. XVIII. No. LXX. April, 1904.] Notes . 321 sown on the injured place. It was found that the cells immediately surrounding the place of injury were rendered susceptible to the attacks of a * biologic form 1 which is unable to attack uninjured leaves of the plant in question. In the third part of the paper, dealing with general considerations, the following hypothesis is advanced as to the actual manner in which the injury to a leaf causes it to become susceptible to a ‘ biologic form ’ otherwise unable to infect it. It is supposed that the leaf-cells of each species of host-plant contain a substance or sub- stances— possibly an enzyme — peculiar to each species which, when the leaf is uninjured and the cells are vigorous, are able to prevent the successful attack of any mildew except the one ‘ biologic form 5 which has become specialized to overcome the resistance. When the vitality of the leaf, however, becomes affected by injury, this substance is destroyed, or becomes weakened, in the leaf-cells in the neighbourhood of the injury, so that the conidia of other 1 biologic forms ’ are now able to infect them. The author suggests that injuries to leaves, caused in nature by hail, storms of wind, attacks of animals, &c., may produce the same effect as the artificial injuries described above in rendering the injured leaf susceptible to a Fungus otherwise unable to infect it. Conidia produced on these injured places would be able to infect uninjured leaves, and would spread indefinitely. Such may be the explanation of a common phenomenon — the sudden appearance of disease caused by parasitic Fungi on plants hitherto immune. A case is described which, it is believed, gives evidence that the injuries produced by Aphides caused leaves previously ‘ immune ’ to become susceptible. In the concluding remarks, reference is made to the antagonistic forces concerned in the evolution of a ‘ biologic form/ viz. ‘ specializing factors ’ and ‘ generalizing factors.’ Attention is also drawn to the close parallel between (1) the behaviour of the Fungus in the experiments in which the conidia were sown on the tissues of the leaf exposed by the cut ; and (2) the biological facts obtaining in the class of parasitic Fungi known as ‘wound parasites’ (. Nectria , Peziza willkommii , &c.), which are able to infect their hosts only through a wound. ERNEST S. SALMON, Cambridge. ON THE STRUCTURE OF THE PALAEOZOIC SEED LAGENO STOMA LOMAXI, WITH A STATEMENT OF THE EVIDENCE UPON WHICH IT IS REFERRED TO LYGINODENDRON. — The present communication deals with the structure of Lagenosioma Lomaxi> a fossil seed from the lower coal-measures, and with the evidence upon which the authors refer it to the well-known carboniferous plant, Lyginoden dr on . It is found that this species of Lagenosioma , especially in its young form, was enclosed in a husk or cupule, borne on a short pedicel. The seed, which is of Cycadean character, is fully described, and its relation to other fossil and recent seeds discussed. 1 Abstract, reprinted from the Proceedings of the Royal Society. [Annals of Botany, Vol. XVIII. No. LXX April, 1304.] 322 Notes . The cupule enclosing the seed was borne terminally on a pedicel ; it formed a continuous, ribbed cup below, and divided above into a number of lobes or segments. Externally, both pedicel and cupule were studded with numerous prominent multi- cellular glands of capitate form. The anatomy indicates that the whole organ was of a foliar nature. A comparison with the vegetative organs of Lyginodendron O/d/iamium, with which the seeds are intimately associated, demonstrates a complete agreement in the structure of the glands and in the anatomy of the vascular system. Where vegetative and reproductive organs, presenting identical structural features, not known to occur in other plants, are thus found in close and constant association, the inference that the one belonged to the other appears irresistible. As regards the position of the seed on the plant, two possibilities are discussed ; the cupule, with its pedicel, may either represent an entire sporophyll, or a modified pinnule of a compound leaf. Either view is tenable, but various comparative con- siderations lend a somewhat greater probability to the second alternative. In the concluding section of the paper, the systematic position of Lyginodendron is discussed. On the whole of the evidence, the position of the genus as a member of a group of plants transitional between Filicales and Gymnosperms appears to be definitely established. While many Filicinean characters are retained, the plant, in the organization of its seed, had fully attained the level of a Palaeozoic Gymnosperm. There are many indications that other genera, now grouped under Cycadofilices, had likewise become seed-bearing plants. It is proposed to found a distinct class, under the name Pteridospermae , to embrace those Palaeozoic plants with the habit and much of the internal organization of Ferns, which were reproduced by means of seeds. At present, the families Lyginodendreae and Medulloseae may be placed, with little risk of error, in the new class, Pteridospermae. F. W. OLIVER and D. H. SCOTT. c ANNALS OF BOTANY, Vol. XVIII. No. LXIX. January, 1904. Contains the following Papers and Notes: — Lawson, A. A. — The Gametophytes, Archegonia, Fertilization, and Embryo of Sequoia semper- vivens. With Plates I-IV. Wager, H. — The Nucleolus and Nuclear Division in the Root-apex of Phaseolus. With Plate V. WoRSDELL, W, C. — The Structure and Morphology of the 4 Ovule.’ An Historical Sketch. With twenty-seven Figures in the Text. Cavers, F.— On the Structure and Biology of Fegatella conica. With Plates VI and VII and five Figures in the Text. Potter, M. C. — On the Occurrence of Cellulose in the Xylem of Woody Stems. With Plate VIII. Williams, J. Lloyd. — Studies in the Dictyotaceae. I. The Cytology of the Tetrasporangium and the Germinating Tetraspore. With Plates IX and X. Benson, Miss M. — Telangium Scotti, a new Species of Telangium (Calymmatotheca) showing Structure. With Plate XI and a Figure in the Text. NOTES. Hemsley, W. Botting.— On the Genus Corynocarpus, Forst. Supplementary Note. Weiss, F. E. — The Vascular Supply of Stigmarian Rootlets. With a Figure in the Text. Ewart, A. J.— Root-pressure in Trees. THE JOURNAL OF BOTANY, BRITISH AND FOREIGN. A MONTHLY JOURNAL, EDITED BY JAMES BRITTEN, F.L.S. Usually a Plate with each Number . Price Is. 8d • Contains Original Articles by Leading Botanists , Notices of Books, and Botanical News genet ally. Contents of April Number Notes on Mycetozoa from Japan, by Arthur Lister, F.R.S. ; Mons. A, Robert’s Matto Grosso Plants, by Spencer le M. Moore; Epilobium colli mini , Gmel., by C. E. Salmon ; Radnorshire Lichens, by W. H. Wilkinson ; Botanical Exchange Club Report, 1902 ; Short Notes, Notices of Books, Book Notes, News, & c. SuPPLEMENT^Notes on the drawings for ‘ English Botany,’ by F. N. A. Garry (continued). LONDON : WEST, NEWMAN & CO., 54 HATTON GARDEN. OXFORD UNIVERSITY PRESS. » >»<♦ The Physiology of Plants. A Treatise upon the Metabolism and Sources of Energy in Plants. By Dr. W. Pfeffer, Professor of Botany in the University of Leipzig. Second fully Revised Edition, translated and edited by Alfred J. Ewart, D.Sc., Ph.D., F.L.S. With many Illustrations. Royal Svo, half-morocco, Vol. II, i6j. Previously published, Vol. I, 28j. Schimper’s PlanUGeography. A Summary of the Existing State of Knowledge about the Geographical Distribution of Plants so far as dependent on Conditions now present; with Special Consideration of the Principles determining the Habits and Structure of Plants, the Evolution of their Adaptive Characters, and the Congregation of Plants to form Communities. Authorized English Translation by W. R. Fisher, M.A. Revised and Edited by P. Groom, M A., D.Sc., F.L.S., and I. Bayley Balfour, M.A., M.D., F.R.S. With a Portrait of the late Dr. A. F. W. Schimper, Four Maps, Four Cojlotypes, and 497 other Illustrations. Royal 8vo. In Four Parts, at gs. net per Part. Complete Volume, half-morocco, 42 s. net. Scotsman .-—‘This work on Plant-Geography is undoubtedly Professor Schimper’s magnum opus. It is already recognized on the Continent as the standard work on oecologieal botany, and its translation into English will give an immense stimulus to a fascinating branch of botanical study in this country.’ On the Physics and Physiology of the Proto= plasmic Streaming in Plants. By Alfred J. Ewart, D.Sc., Ph.D., F.L.S., Lecturer on Botany in the Birmingham Technical Institute. Communicated to the Royal Society by Francis Gotch, D.Sc. (Oxon.), F.R.S. Royal Svo, half-morocco, with 17 Illustrations. 8j". 6d. net. Plant World (Washington): — ‘The present volume not only contains a summary of what has already been learned about protoplasmic streaming, but also the results of many experiments and observations made by the author. The work is comprehensive, and will be indispensable to all physiological botanists.’ REPRINTED FROM THE ‘ANNALS OF BOTANY.’ A Sketch of the Life and Labours of Sir William Jackson Hooker, K.H., D.C.L. Oxon., F.R.S., F.L.S., &c., late Director of the Royal Gardens of Kew. By his son, Joseph Dalton Hooker. Royal 8vo, paper covers, 3 s. 6d. net. LONDON: HENRY FROWDE, OXFORD UNIVERSITY PRESS WAREHOUSE, AMEN CORNER, E.C. Vol. XVIII. No. LXXI. July, 1904. Price 14s. Annals of Botany EDITED BY ISAAC BAYLEY BALFOUR, M.A., M.D., F.R.S. KING S BOTANIST IN SCOTLAND, PROFESSOR OF BOTANY IN' THE UNIVERSITY AND KEEPER OF THE ROYAL BOTANIC GARDEN, EDINBURGH D. H. SCOTT, M.A., Ph.D., F.R.S. HONORARY KEEPER OF THE JODRELL LABORATORY, ROYAL BOTANIC GARDENS, KEW WILLIAM GILSON FARLOW, M.D. PROFESSOR OF CRYPTOGAMTC BOTANY IN HARVARD UNIVERSITY, CAMBRIDGE, MASS., U.S.A. ASSISTED BY OTHER BOTANISTS // Ns. A/aifona\ Bonbon HENRY FROWDE, AMEN CORNER, E.C. CLARENDON PRESS DEPOSITORY, 116 HIGH STREET I 904 Printed by Horace Hart, at the Clarendon Press, Oxford. PAGE CONTENTS. Blackman, V. H. — On the Fertilization, Alternation of Generations, and general Cytology of the Uredineae. With Plates XXI- XXIV 323 Darbishire, O. V.— Observations on Mamillaria elongata. With Plates XXV and XXVI . . . 375 Lawson, A. A. — The Gametophytes, Fertilization, and Embryo of Cryptomeria japonica. With Plates XXVI I-XXX . . . 417 Gregory, R. P. — Spore-Formation in Leptosporangiate F erns. With Plate XXXI and a Figure in the Text 445 Massee, G. — A Monograph of the genus Inocybe, Karsten. With Plate XXXII 459 Boodle, L. A. — On the Occurrence of Secondary Xylem in Psilotum. With Plate XXXIII and seven Figures in the Text . . . 505 NOTE. Scott, D. H. — On the Occurrence of Sigillariopsis in the Lower Coal- Measures of Britain 519 NOTICE TO SUBSCRIBERS. The subscription-price of each volume is thirty shillings, payable in advance : the Parts, four in number, are supplied as they appear, post free to subscribers in the United Kingdom, and with a charge of is. 6d. per annum for postage to subscribers residing abroad. The price of individual Parts is fixed at a higher rate. Intending subscribers should send their names, with subscription, to Henry Frowde, Oxford University Press Ware- house, Amen Corner, London, E.C. As the earlier volumes of the Annals of Botany are becoming scarce, Vol. I will only be sold as part of a complete set; and Parts will not as a rule be sold separately, after the publication of the volume to which they belong. A few extra copies of particular Parts at present remain on hand, for which special application must be made to the Editors, Clarendon Press, Oxford. NOTICE TO CONTRIBUTORS. Contributors in America should address their communications to Professor Farlow, Harvard University; and all other contributors, to the Editors, at the Clarendon Press, Oxford. Papers sent in with a view to publication must be type-written ; and the illustrative figures should be planned so as to properly fill a 4to or an 8vo plate. Attention to this will conduce to the rapid publication of the papers if accepted. Each contributor to the Annals of Botany is entitled to receiye gratis twenty-five separate copies of his paper, and may purchase additional copies if he informs the Editors of his wishes in this respect when he returns corrected proofs. The price of these additional copies will depend upon the amount of text and the number of plates in the paper. On the Fertilization, Alternation of Generations, and General Cytology of the Uredineae1. BY VERNON H. BLACKMAN, M.A., F.L.S., Fellow of St. John's College , Cambridge ; Assistant , Department of Botany, British Museum. With Plates XXI-XXIV. HE question of the sexuality of the Uredineae has been a vexed one JL ever since the suggestion put forward by Meyen (34), more than sixty years ago, that the spermogonia and aecidia represented the male and female organs. This view seemed to receive support from the later observations of Tulasne (53), and De Bary (3, 4), who showed that not only were spermogonia and aecidia closely associated in a large number of forms, but also that the spermatia produced in the spermogonia were apparently wanting in any power of germination. It appeared, then, possible that the aecidium was the result of the fertilization by a sper- matium of a definite female reproductive organ. All workers, however, failed either to trace this process or even to produce any evidence of the existence of sexual organs at any stage in the development of the aecidium. These results, together with the observation that aecidia were sometimes produced in the entire absence of spermogonia (see De Bary (4) ; also Klebahn (27), p. 194), seemed to negative the view of any actual fertilization by the spermatium. Owing in part to these observations and to the discovery by Cornu (12), in 1875, that the spermatia were capable of germinating to a slight degree in nutritive solutions, and also to the later observations of Moller (36) that the spermatia of some lichens could develop a mycelium under similar con- ditions, Brefeld (10) was led about 1889 to put forward the view that the Uredineae (and all the higher Fungi) were without any trace of sexuality and that the group should be considered as a family of the Basidiomycetes. In his view the spermogonia and spermatia, in this group and in the lichens, 1 A short preliminary account of the chief results of this work appeared in the New Phytologist, vol. iii, 1904, pp. 24-27. [Annals of Botany, Vol. XVIII. No. LXXI. July, 1904.] A a 324 Blackman. — On the Fertilization , Alternation of are nothing more than accessory asexual reproductive organs to which the terms pycnidia and conidia should more aptly be applied. These ideas were followed by Van Tieghem (54) and by Vuillemin, and have of late years gained considerable acceptance 1. The great objection, however, to the view that the spermatia are of conidial nature is that they seem quite incapable of causing infection ; the fact that they appear about the same time as the aecidiospores is also an argument against this view, as will be shown later. Although it is clear that a careful cytological study would most likely throw great light on the vexed question of sexuality in this group, yet for a very long time our knowledge of the cell structure of the Uredineae was confined to a few scattered observations by Schmitz (49) and Rosen (45). About ten years ago, however, the researches of Poirault and Radiborski (43) and of Sapin-Trouffy (48), a pupil of Dangeard’s, threw very considerable light on the cytology of this group. The former observers showed that in many stages of the Uredineae the nuclei were to be found closely associated and dividing in pairs (the conjugate nuclei of these authors), each pair in a separate cell. Sapin-Trouffy carried the matter further, and showed that the following very interesting cycle of nuclear development was to be observed in the forms possessing an aecidium (Eu- and -opsis forms). The mature teleutospore is always uninucleate and gives origin to four uninucleate sporidia from which a mycelium arises in which the nuclei are arranged singly, usually in separate cells. In the aecidium, borne by this mycelium, the nuclei, however, become paired ; the aecidiospores thus contain two nuclei, and the paired condition of the nuclei is retained throughout ensuing mycelia and uredospores (if present) up to the teleuto- spores, which in the young state are binucleate, but at maturity become uninucleate by the fusion of their nuclei. In all cases of division of the paired nuclei the two are very closely associated, and a half of each nucleus goes to the new cell, so that the two nuclei in the cells produced by division are never daughter-nuclei. These very striking observations throw no light on the nature of the spermatia, for though the latter were shown to be uninucleate, like the cells of the mycelium on which they are borne, Sapin-Trouffy assumed their conidial nature and paid little attention to them. For him the most important stage in the cycle just described was the fusion of nuclei in the teleutospore, a fusion which he considered to be of the nature of a true sexual process2, and to be comparable with the fusion in the ascus and in the basidium. On the facts observed, however, there does not seem sufficient evidence for such a view ; for a process of fusion which takes place in a cell 1 A full account of the older observations on the nature of the spermatia and the sexuality of the group is given by Klebahn (27, pp* 194-202). 2 Dangeard and Sapin-Trouffy (18) had earlier considered it as a ‘ pseudofecondation.’ Generations , and General Cytology of the Uredineae . 325 which shows no signs of specialization as a female reproductive cell, and one in which the ancestors of the fusing nuclei have been associated together in the same cell for many thousands of generations, lacks all of the characters of a sexual process except that of mere nuclear fusion. It would, in fact, seem obvious that the critical point for investigation in relation to the question of sexuality is the early development of the aecidium, for it is there, in the full life-cycle, that the two nuclei first become associated, and the transition from the single to the paired con- dition takes place. Sapin-Trouffy, however, imbued with the idea of the importance of the fusion in the teleutospore, paid little attention to this point, but merely states that from the mycelium with single nuclei, binucleate hyphae grow up which cut off a series of binucleate aecidiospore-mother- cells, from which by division the binucleate aecidiospores are derived. The other results obtained by this worker, such as those on the details of nuclear division (in which he describes the regular presence of two chromosomes and the absence of a spindle) and on chromosome-reduction, left much to be desired, owing to the insufficiency of the methods employed. The figures which he gives are also of a very diagrammatic nature. In spite, however, of the obvious need of further work, the only later contributions of importance are those of Maire (29, 30), published in part after this work was in progress. Maire accepts the results of Sapin-Trouffy, but states that the cells of the aecidium which give origin to the aecidiospore-mother- cells are, in Endophyllum Sempervivi , De Bary, and in Puccinia Bunii , DC., at first uninucleate, but later become binucleate by a process which he believes to be one of simple division. The result of cytological observations, as far as they go, is thus to suggest that the Uredineae are totally wanting in any ordinary form of sexuality or of sexual organs, and that the aecidiospores are produced by a structure of the nature of a conidiophore, in which, however, a peculiar association of nuclei takes place. The spermatia, however, still remain as very puzzling structures on which these investigations throw no light, for they are apparently completely wanting in function. It is interesting to note that during last year Arthur (1), while admit- ting our ignorance of the question of sexuality in the group, has put forward the view that the aecidium is ‘ a device to restore vigour to the fungus/ as after this stage the parasite usually spreads and develops very rapidly ; the sexual nature of the aecidium is thus suggested on physiological grounds. An investigation for the purpose of studying the vexed question of sexuality and for settling some of the points left doubtful by Sapin- Trouffy had been projected ever since the appearance of Sapin-T rouffy ’s work, and was carried out during the last two years. It was clear that what was required was not so much a rapid survey of the whole group, as had been done by Sapin-Trouffy, but the careful investigation of a few A a 2, 326 Blackman. — On the Fertilization, A Iter nation of forms with special attention to aecidium-development and the character of the spermatia. With this object there were chosen as likely to be favourable objects, Phragmidium violaceum , Wint. 1, a common autoecious ^-form found on the various forms of Ruhns fruticosus , and Gymno- sporanginm clavariaeforme , Rees, a heteroecious form with the spermogonial and aecidial stage on various species of Crataegus , and the teleuto- spore stage (uredospores are wanting) with a perennial mycelium on Juniperus communis , L. Material of the last-named form was obtained in abundance from the neighbourhood of Crockham Hill, Kent, where the infected Hawthorn and J uniper are often to be found in close proximity. A careful study of these forms, especially of Phrag. violaceum , shows very clearly that a definite fertilization is to be observed in the aecidium, that the Uredineae show a well-marked alternation of sexual and asexual generations, and that the fusion in the teleutospore, the nature of which has been so much disputed, is relatively of unimportance, being a mere preliminary to reduction. Methods. The material used was fixed chiefly with either Flemming’s fluid (usually the weak formula) or with acetic alcohol which was allowed to act only for a short time. All material, except that for the study of teleutospore- germination, was fixed in the field, an air-pump being used in the case of the non-alcoholic fluids to remove air from the surface of the pieces and so facilitate the penetration of the fluid. The material after fixing, washing, and dehydrating was preserved in equal parts of alcohol, glycerin, and water. For staining, Flemming’s triple stain, Benda’s iron-haematoxylin, and brazilin were all found very useful ; for the latter stain it was found convenient to use, instead of the ordinary alcoholic iron-alum solution, which goes bad very quickly, a solution made by diluting Benda’s 5 Liquor ferri ’ with about nine parts of 70 °/o alcohol. This mixture remains unpre- cipitated for many months. For the study of teleutospore-germination G. clavariaeforme was found very favourable. The masses of teleutospores will retain their power of germination in a cool place for over a month. If pieces of the bark with the masses attached are placed for a few seconds in water under an air-pump so as to get them thoroughly wet, and then in Petri -dishes over damp filter-paper at a temperature of 20° C., germination takes place very readily and sporidia are formed in less than three hours. Very good results in the study of nuclear divisions in the germ-tube were obtained by the use 1 The characters which distinguish this species from that of Phrag. Rubi, Wint., are very un- satisfactory and the two forms should probably be united. The form here investigated is obviously the same as that studied by Sapin-Trouffy, which he calls P. Rubi , but as there are usually four cells to the teleutospore and only a slight projection at the apex the other name seems preferable. Generations , and General Cytology of the Uredineae . 327 of Flemming’s weaker fluid diluted with an equal quantity of water. After fixing and washing, the material was brought up to the clearing fluid by the method of Overton (41). It was first placed in 10 °/o glycerin, which was allowed to evaporate either in a warm place or in a desiccator over calcium chloride ; the glycerin was then removed with absolute alcohol and the material placed in a io°/o solution of cedar oil in alcohol. After the alcohol had evaporated in a desiccator the material was obtained in cedar-wood oil without the least distortion of the delicate germ-tubes, and with far less trouble and risk than by the passage through various strengths of alcohol, and mixtures of alcohol and clearing fluid. The material can then either be treated by the method of fixing to the slide (see Blackman, 8), or by imbedding small teased pieces in hard paraffin very successful sections can then be made through the teleutospores and germ- tubes. By the latter method alone can the details of nuclear division be successfully studied. For staining the cell-walls of the hyphae a 1 °/o watery solution of Congo -red, either neutral or slightly alkaline, was found of great use, as it stains the walls of the hyphae without affecting the host-cells. I have to thank Miss H. C. I. Fraser for very considerable help in preparing material during the latter part of this work. TELEUTOSPORE AND PROMYCELIUM. Phragmidium violaceum. The peculiar nuclear cycle with its sudden transition in the aecidium, as described by Sapin-Troufify, was fully confirmed in the case of the two forms investigated. It will thus be convenient to start with the first stage of the series with single nuclei — the mature teleutospore — and then work by way of the spermogonia and aecidia to the series with paired nuclei, and after dealing with the uredospores end with the development of the teleuto- spores and the fusion of the two nuclei into one. Mature Teleutospore . The teleutospore of this form appears towards autumn, and is to be found in groups on the leaves which remain attached throughout the winter. It is a large structure consisting usually of four thick-walled cells, and a stalk which is enlarged below, this part of the stalk being capable of swelling up in water to a great extent (Fig. 1). The wall of the uppermost cell is provided with a rounded peg-like projec- tion of cell-wall substance which is sometimes perforated by a narrow canal connected with the cavity of the cell below. The whole spore is covered with a thin cuticular layer which is raised into numerous rounded bosses (Figs. 1 and 2). Each cell possesses a delicate endosporium, a continuous, thick mesosporium, and an outer exosporium, which is to be found only as 328 Blackman. — On the Fertilization , Alternation of a sheath round the two middle cells, but exists also round the upper part of the top cell and the base of the lowest cell (Fig. 2). The stalk has a narrow lumen which widens out below (Figs. 5 and 86 e) ; its wall shows only- slight differentiation of layers. Each cell has usually four pores which are closed on the outside by the cuticle only, the protoplasm and endosporium projecting slightly into the cavity in the mesosporium and exosporium (Fig. 2). Each cell contains a single central nucleus with a well-marked nucleolus, and dense cytoplasm containing a quantity of yellow oily material which gives to it an orange colour. In the dry state very little can be made of the structure of the nucleus, and except for the nucleolus it appears almost homogeneous. Germination . Spores taken in March from the ‘ wintered ’ leaves germinate readily in moist air or water, but will not do so earlier. The first beginning of the germ-tube (promycelium) is the welling out through one of the pores of a spherical mass of yellow cytoplasm surrounded, of course, with a thin wall. This stage (Fig. 1) is very striking, and is also to be observed in G. clavariaeforme. From this spherical mass the cylindrical germ-tube grows out away from the spore (Fig. 5). The nuclei undergo an interesting series of changes when the spores have been lying in water some time. After twelve hours in this condition the nucleus increases in size, and there is to be observed a somewhat flattened nucleolus pressed close against the nuclear membrane, which often projects at this point. The faintly staining, and hitherto almost homo- geneous, mass of the nucleus begins to stain more darkly, and faint indications of a chromatin thread are to be seen (Fig. 3). Soon the chromatin part becomes clear as a single much-twisted thread, at first closely coiled ; but as increase in size of the nucleus still goes on the coil becomes a more open and thinner spireme thread. This stage, how- ever, is not a preparation for division, for it appears to last only a short time, and very soon the nuclei return to the ordinary condition and show a nucleus with a chromatin network, and, instead of a single large nucleolus, a few smaller ones (Fig. 4). The nucleus has a large cavity and is only partly filled with the network, so that in the living teleutospore examined in water before germination it appears as a distinct, clear { vacuole 5 (Fig. 5), as it was termed by the older observers. In some few cases two smaller nuclei are to be found in the cell of the teleutospore (Fig. 4 a) instead of one larger one. These are, no doubt, the original paired nuclei of the teleutospore, which for some reason have delayed their fusion. Their fate is unknown, as two nuclei were never observed to pass into the germ-tube ; probably they fuse later. There was no evidence that they represented the results of a precocious division. The nucleus passes into the germ-tube, being constricted in its passage through the pore, which is only of small diameter (Fig. 6). If the germ-tube Generations , and General Cytology of the Uredineae . 329 does not reach the air, it usually grows on continually (Blackman, 7), the protoplasm with the nucleus being found at the end of the tube (Fig. 7). Division of nucleus. This form is not a favourable object for the study of this process on account of the irregularity of the germination and the difficulty of cutting the necessary sections. The divisions in the promy- celium were studied fully in the form to be next described, but here it was merely observed that not two (as Sapin-Trouffy states), but numerous chromosomes were formed on division, and that there was a definite spindle with centrosomes. Sporidia- formation. The four cells of the promycelium contain only small nuclei with a very small amount of chromatin, The sterigmata produced may either be short, as in the typical condition seen in Fig. 8 a ; or elongated, with more of the appearance of germ-tubes (Fig. 8). The difference appears to depend upon the degree of moisture present (vide infra under Gymnosporangium ). The whole of the protoplasm does not pass into the sporidia, but a small quantity remains in the promycelial cell (Figs. 8 a and 9). The sporidium is a globular or pear-shaped structure with a wall of some slight thickness and irregular in outline (Fig. 8 a), a nucleus which sometimes shows no nucleolus, but a well-marked chromatin network (Fig. 10). The sporidia germinate very readily, sometimes while still attached to the sterigma (Fig. 9), and often form a secondary sporidium which seems always to be binucleate (Fig. 11), and in one case four nuclei were apparently to be observed (Fig. 13). This binucleate condition is also sometimes to be observed in the primary sporidium even while it is still attached to the sterigma (Fig. 12). The condition with two nuclei appears to be fairly common in the primary or secondary sporidia of the group. It has been described in Puccinia Malvacearum by Sapin-Trouffy (48), in Endophyllum Sempervivi by Maire (29), and for Coleosporium Euphrasiae by Poirault and Raciborski (43), and it is to be sometimes found in G. clavariaeforme. It appears to be entirely without significance and to be merely a precocious division of the nucleus in which the usual wall- formation is delayed. The nuclei are certainly not paired (conjugate), for in Phragmidium and the first two species mentioned above the sporidia have been observed to give rise to a mycelium with single nuclei as in all other known cases ; in fact in E. Sempervivi Maire states that the two nuclei can be observed to pass into the infecting germ-tube and there become separated by a septum. Gymnosporangium clavariaeforme. M attire Teleutospore. The two-celled teleutospores of this form, with their very long gelatinous stalks by means of which the spores are held together in compact yellow masses, are well known (Fig. 14). As noticed by 330 Blackman . — the Fertilization , Alternation of earlier observers they are of two kinds, thick-walled and thin-walled, the latter being found more usually in the interior of the mass ; it is possible, also, that there may be some connexion between the thickness of the wall and the time of year at which the spores are formed, for the first- formed spores seem to be almost all thick-walled 1. Each cell of the teleutospore contains dense vacuolate cytoplasm with a quantity of yellow oily material which gives the colour to the spore. There is a single nucleus of considerable size with a chromatin network arranged superficially round the cavity and one or more small nucleoli (Figs. 14 and 15). The wall of the spores shows usually no clear distinction into layers. Near the septum, the wall of each cell (of the thick- walled spores) is thinner at one spot, and into the cavity so produced the protoplasm projects slightly (Fig. 14). It would seem that the wall is never thickened at these spots, for such thin places in the wall are found in comparatively young spores ; they are thus to be considered of the nature of pores. These pits are not to be found in the thin-walled spores (where they are obviously unnecessary), but the naturally thin wall may project slightly at the corresponding points. Germination . Under suitable conditions of moisture and temperature germination takes place very readily, in a moist atmosphere at 20° C. the promycelium being formed and the sporidia developed in less than three hours. The cytoplasm forces its way through the pore described above, and, after swelling out into a spherical mass, as in Phragmidium , attains a considerable length before the nucleus migrates into it (Fig. 16). The nucleus, very much constricted and condensed in its passage through the pore, takes up a position in the middle of the tube and is there to be seen as a narrow elongated body with a few small nucleoli, and granular chromatin which forms usually only an indistinct network. Under suitable conditions of air-supply the nucleus then proceeds to divide. Nuclear division. The first sign of division is the gathering of the granular chromatin material towards one end or the middle of the nucleus, while at the same time a small, deeply staining, spherical body, the centrosome, is to be observed in the cytoplasm and connected with the nucleus by a portion of slightly differentiated cytoplasm or kinoplasm (Fig. 1 7). The chromatin then forms a closely coiled spireme thread in the middle of the nuclear area and the outline of the nucleus becomes very faint (Fig. 18); the centrosome was not traced at this stage. The thread then becomes apparently broken up into a number of narrow, elongated 1 The suggestion put forward by Kienitz-Gerloff (Bot. Zeit., xlvi, 1888, p. 388) and Dietel (Hedwigia, xxviii, 1889, p. 99), that the thin- walled spores represent in this genus the endospores, cannot be accepted. Both kinds of spores are uninucleate, and both are able to form normal sporidia ; the long, undivided tubes put out by the thin-walled spores were no doubt the result of special conditions of growth (see Blackman, 7). Generations , and General Cytology of the Uredineae . 331 chromosomes (Fig. 19 a), but they lie so close together that their exact number and length are difficult to make out ; there are certainly at least ten. At this stage the spindle was first observed as a very short structure with well-marked centrosomes at each pole, lying among, but without any distinct relation to, the chromosomes (Fig. 19 b). From the fact that the centrosome with a mass of kinoplasm was first observed outside the nucleus, and from analogy with the second division, there can be no doubt that the spindle is formed in the cytoplasm, between two centrosomes which arise by division of the one observed earlier, and that it later comes into association with the chromosomes. The spindle gradually increases in length, and the closely packed, ill- defined chromosomes come into close relation with it (Fig. 20). As the spindle lengthens the chromosomes become confined to the more central part of the spindle (Figs. 21, 22), which, if originally oblique (Fig. 22 ), takes up a position parallel to the long axis of the germ-tube. The spindle still continues to elongate and the chromosomes to spread out, so that when the spindle has attained its full length, they occupy usually two-thirds of its length (Figs. 23-5). This stage, at which the spindle has reached its full length, seems to correspond with the metaphase, being the one most frequently met with ; but no distinct equatorial plate is ever formed. At this stage the form of the chromosomes is usually almost completely lost, and they appear to have partly fused together, a few free ends (Fig. 23) or the presence of irregular lumps of chromatin (Figs. 24, 25) alone suggesting their existence. The chromatin material then begins to move towards the poles, but rarely, as in Fig. 26, are two distinct polar groups to be seen. Usually this stage cannot be sharply distinguished from the last, and the chromatin lumps merely become strung out from pole to pole and often showing a tendency to form two elongated masses (Fig. 29). Sometimes two irregular twisted threads forming two clumps can be seen moving towards each pole (Fig. 28), and occasionally the masses moving towards the poles can be seen to have formed two chromatin networks (Fig. 27). After this stage, the suggestion of separate chromosomes becomes usually completely lost and the chromatin merely stretches out from pole to pole as two elongated dumb-bell-shaped masses (Fig. 30). The chromatin then collects at the poles, forming there one irregular mass (Fig. 31), or the distinction into separate masses may be clearly maintained at each pole (Fig' 33)- During these stages the centrosomes show clear but faintly staining polar radiations, which are often of very considerable length (Fig. 24), and in the case shown in Fig. 31 extended to the free end of the germ-tube. When the chromatin has collected at the poles the two (or four) masses move apart 332 Blackman . — On the Fertilization , Alternation of and the spindle becomes stretched out and very narrow ; the centrosomes and radiations are however still clearly visible (Fig. 31). The daughter-nuclei are then formed ; each is a somewhat pear-shaped body and consists of a mass of perfectly homogeneous staining material. At the pointed end of each, the centrosome can be clearly observed ; it is not attached directly to the nucleus, but is connected with it by means of a small portion of kinoplasmic material which seems to correspond with the end of the former spindle (Fig. 33). The centrosome shows distinct radiations (F>g- 33)- Under the conditions under which the teleutospores were germinated the two second divisions follow immediately on the first. The centrosome becomes divided into two (Fig. 34), and between them the spindle appears (Fig. 34 a ). At the same stage the nucleus becomes larger and irregular in shape, and a few granules appear in it ; in Fig. 34 a they are to be seen just where the spindle is in contact with the nucleus, over the edge of which it lies. The spindle then becomes arranged in the long axis of the germ-tube, and shows distinct polar radiations, while the nucleus, which has shown no definite wall from its first formation, is seen as a mass of chromatin lying in an irregular way over the spindle (Fig. 35). The chromatin then becomes more symmetrically arranged round the spindle, and is seen to be granular throughout (Figs. 36, 37). The chromatin then takes the form of a distinct network , which at first covers only part of the spindle (Fig. 38 a)t but soon spreads over the whole of it, leaving only the centrosomes visible (Fig. 38 b). The network of chromatin then becomes drawn apart as two portions towards the poles (Fig. 39). In a later stage the threads which connected the two main masses may remain visible for a time at the poles, and resemble chromosomes (Fig. 40). The chromatin then collects at each pole usually as a single mass (Fig. 41), but not infrequently it forms two distinct masses (Fig. 41 a). The four daughter- nuclei when they are first formed show, as in Fig. 42, a somewhat pear- shaped mass of lightly staining homogeneous material, which bears at the pointed end a mass of kinoplasm and a distinct centrosome, as in the corresponding stage of the first division. The promycelium then becomes divided into four cells, and the nuclei gradually take on a normal appear- ance, a chromatin network becoming visible and the centrosome apparently disappearing (Fig. 43). It is clear that though the division in the promycelium of Gymno- sporangium is much more typical than Sapin-Trouffy supposed, yet there is certainly no chromosome-formation in the second division, and in the first division, though distinct chromosomes appear to be present, they seem soon to lose their individuality, so that of their splitting or regular separation there is very considerable doubt. The two chromatin-groups, which are often to be observed both in Generations , and General Cytology of the Uredineae. 333 the first and second division, probably represent the chromatin derived respectively from the two nuclei which fused in the teleutospore. It was these two masses which Sapin-Trouffy mistook for two chromo- somes, as a comparison of his figures with Figs. 30, 32, and 41 a clearly show. The method of spindle-formation, in which the spindle is formed free in the cytoplasm between two centrosomes and later comes into relation with the dividing nucleus, is of great interest. It is clearly of the type described, in animals, by Hermann and others for the Centralspindel. The only case in plants with which it is at all comparable is the peculiar method of spindle-development described by Lauterborn (28) for Diatoms 1. Since the work of Sapin-Trouffy, who failed to discover either a spindle or chromosomes in the divisions in the promycelium, it has been shown by Juel (25), in Coleosporium Euphrasiaey that a spindle and polar radiations were present, and that the division was of a much more typical nature than Sapin-Trouffy had described. He was not, however, able to determine the behaviour of the chromatin part of the nucleus. When this work was complete a paper appeared by Holden and Harper (22), in which it was shown that in the divisions of the developing teleutospore of a form of Coleosporium on Callistephus, &c., to which they give the name C. Sonchi - arvensis , a spindle with centrosomes, polar radiations, and showing numerous chromosomes was to be observed, as in Gymnosporangium. They did not, however, trace the origin of the spindle, and they give but a few figures of the behaviour of the chromatin, for Coleosporium does not seem to be nearly so favourable an object of study as the form here investigated. They are of the opinion that both divisions are typical in nature. Sporidia formation. There is nothing worthy of special comment in the normal formation of sporidia. A primary sporidium is shown, in section, at Fig. 25. In some cases the original four cells of the promycelium may round themselves off, separate, and put out germ-tubes, thus behaving like sporidia. Long germ-tubes may also be put out by the four cells while they are still attached, as shown in Fig. 44. Although the exact conditions were not investigated, this method of shortened development seems to be dependent on conditions of moisture, and is probably a response to growth under water or at least in a very moist atmosphere. It has been shown in an earlier paper (7) that promycelia actually growing under water develop to a great length, but rarely divide, and never form sporidia. A single case was there figured for Phragmidium, where the promycelium had divided 1 P. Denke (Beihefte z. Bot. Centralbl., xiii, 1902, p. 182) has described the development of an extra-nuclear spindle, though, of course, without centrosomes, in the division of the microspore and megaspore mother-cell of Selaginella. 334 Blackman . — On the Fertilization , Alternation of under water, and one of the four cells was putting out a germ-tube 1. Inside the gelatinous mass of teleutospores access to free air must be difficult, and it is probably in such a situation that the promycelial cells tend to germinate directly. Such cases as these of Gymno sporangium and Phragmidium show that the real germmating unit is the promycelial cell , and that under certain conditions it can become a separate spore, put out a germ-tube and behave as a sporidium, and no doubt cause infection. When growing normally, however, in air the development of the germ-tube is arrested, and it becomes a mere sterigma, on the end of which the primary sporidium is developed. The primary sporidium thus appears to be merely the arrested and swollen germ-tube put out by the promycelial cell, just as the secondary sporidium is merely the arrested and swollen germ-tube of the primary sporidium. The sporidia are probably nothing more than special adaptations to the develop- ment of the promycelium in air ; in the absence of air such a development would be unnecessary, for there would be no chance of the wind-dispersal of the sporidia. The primary sporidium is thus really secondary in nature, and bears the same relation to the promycelial cell (the true primary spore) as does the so-called secondary sporidium to the first-formed one. It is evident that the formation of the so-called promycelial cells is really nothing more than the division of the contents of the teleutospore into four spores, which may separate as such, but usually remain united and form four secondary spores — the sporidia. As far as is known, the teleutospore is quite incapable of forming a mycelium ; it seems then time that the term promycelium should be dropped, and the process of develop- ment considered merely as one of spore-formation. That the later development of the teleutospore is really a process of spore-formation has been obscured not only by the formation of sporidia (really the secondary spores), but also by the fact that, owing to the thick- ness of the spore-wall, a process of germination is a necessary preliminary to division. In Coleosporium , however, where the teleutospores are thin- walled and develop in situ , there is no process of germination, but the spores become directly divided into four cells, which normally put out each a sterigma and form sporidia. Holden and Harper (23), however, in this genus have lately observed cases in which the divisions of the teleutospore rounded themselves off and became directly spores. In such cases, the teleutospore, its real nature no longer obscured either by the process of germination or by the formation of sporidia, is clearly seen to be a spore- mother-cell which undergoes a tetrad division to form four spores . 1 A somewhat different interpretation was placed upon this case at the time, but a comparison with Gymnosporangium gives the clue to its real nature. Sapin-Trouffy (48) has described a case in P. Malvacearum in which submerged promycelial cells separate and put out germ-tubes, which become sterigmata (and bear sporidia) only if they reach the air. Generations , and General Cytology of the Uredineae. 335 MYCELIUM AND SPERMOGONIA. Phragmidium violaceum. The mycelium which appears in the leaf soon after germination of the teleutospores is derived by infection from the sporidium, and soon bears the spermogonia ; it consists of numerous hyphae with single nuclei, which appear to be always enclosed in separate cells. Most of the hyphae are intercellular ; some penetrate the cells, but they appear to have no definite connexion with the nuclei of the host-cells, as in the haustoria described by Sapin-Trouffy for some forms. The spermogonia are found usually on the upper side only of the leaf. Like the aecidia they are of indefinite extent, and are to be found as simple layers of parallel spermatial hyphae developed beneath the cuticle. Each irregular layer or spermogonium can be distinguished into a number of slightly projecting ‘hummocks,’ above which the cuticle is perforated to form the ostiole. One of these ‘hummocks’ is shown in section in Fig. 46. The spermogonia have usually no paraphyses, but occasionally one or two spermatial hyphae may grow out as sterile threads and project through the aperture in the cuticle. The spermogonia are formed by hyphae which grow up between the epidermal cells, and form beneath the cuticle a mycelial bed, or plec- tenchyme, of uninucleate cells. From this there arises a compact mass of parallel spermatial hyphae. The spermatial hyphae of any given group all point their free ends towards the aperture in the cuticle in the centre of the hummock (Fig. 46). Each group of spermatial hyphae with its ostiole should probably be looked upon as a single spermogonium, several spermogonia being collected together to form a composite structure. The cuticle, which is of course pushed up by the growth of the hyphae, becomes very much thickened at these spots, and stains very deeply with the safranin of Flemming’s triple stain and with iron-haematoxylin (Fig. 46). Each spermatial hypha contains a single, usually elongated, nucleus with a chromatin network, one or two small nucleoli, and an indistinct membrane. The spermatia are budded off from the tips of the hyphae, a projection being first pushed out from the apex of the hypha, and when it has reached its full size the nucleus passes into it. The end of the hypha has a ring of special cell-wall, staining deeply with Congo-red, as in Gymnosporangium , which offers a much more favourable object for studying the development of the spermatia. The mature spermatia are minute uninucleate cells more or less oval in shape. Each contains a comparatively very large nucleus, which shows a dense chromatin network but no nucleolus; surrounding this is a thin 336 Blackman . — On the Fertilization , Alternation of layer of very finely granular cytoplasm with apparently no reserve-material, and the whole is enclosed in a very thin cell-wall. The spermatia make their escape through the ostiole, and are found spread out on the leaf in the immediate neighbourhood. They took no part in aecidium-formation or in any other process, and numbers of them when observed on the leaf-surface appear to be in a disorganized state, in which the distinction between nucleus and cytoplasm is almost completely lost. Gymnosporangium clavariaeforme. The mycelium in the leaves of the Hawthorn which bears the spermogonia shows clearly the single nuclei apparently always enclosed in separate cells. They are of small size, and usually show no distinct nucleolus, and their nuclear membrane is often indistinct. The spermogonia appear very early after infection, in about seven to ten days. The spermogonia are flask-shaped structures of the usual type, and are developed beneath the epidermis. They arise from a layer of small- celled ‘tissue’ (plectenchyme), which gives origin to a number of upwardly directed and parallel hyphae, the spermatial hyphae. Certain of these hyphae, chiefly the peripheral ones, grow out to form the paraphyses which project through the ruptured epidermis. The others bud off each a series of spermatia, which collect in the cavity of the flask, and later become extruded (Fig. 48). This form is a favourable object for studying the exact development of the spermatia. The spermatial hyphae are narrow, elongated cells with a central elongated nucleus, which shows a granular chromatin network and one or more small nucleoli ; the nuclear membrane, however, is indistinct (Fig. 49). The hypha contains only finely granular cytoplasm. It is furnished at the free end with a curious ring of thickening, which is easily rendered visible by the fact that it takes the Congo-red with more avidity than the rest of the cell- wall (Figs. 49-54). The first beginning of the development of a spermatium is the pushing out of a finger-like projection from the free end of the hypha, thus displacing the last-formed spermatium (Fig. 49). This projection contains exceedingly finely granular protoplasm. When it has attained its full size, or a little earlier, the nucleus of the spermatial hypha begins to undergo a process of condensation (Fig. 50), so that it is transformed into a homogeneous and deeply staining chromatin mass (Figs. 51 and 55), the nucleolus or nucleoli being apparently squeezed out into the general cytoplasm (Fig. 55 a). The condensation of the nucleus may begin at one end and progress downward, as shown in Fig. 50. The mass of chromatin then becomes drawn apart into two masses (Figs. 52, 56), which remain Generations , and General Cytology of the Uredineae. 337 connected by a thread of kinoplasmic substance. The two become finally separated, and the upper one, passing through the thickened girdle, moves into the spermatium (Fig. 53). The lower chromatin mass passes back into the resting state (Fig. 54), and the process is again repeated. The spermatium becomes cut off from the hypha by a wall formed just above the thickening ring, which may be connected with the disjunction of the spermatium. As in the divisions to be described later, the nuclei sometimes show not a single chromatin mass, but two masses, which are drawn out separately towards the respective poles (Fig. 57). The only trace of spindle-formation was the connecting strand between the two separated chromatin masses. It may be that a rudimentary spindle is present as in the case of the conjugate divisions, but is obscured by the chromatin mass. The whole method of division is obviously of an ex- ceedingly simple type. When the spermatium separates from the spermatial hypha its nucleus is an almost homogeneous and deeply staining mass (Fig. 58), but in the mature state becomes larger and shows a dense network without a nucleolus. As in Phrag . violaceum , the mature spermatia (Fig, 59) are small cells with a large, dense nucleus, very little cytoplasm, a thin cell-wall and apparently no reserve-material. The spermatia, when extruded from the spermogonium, collected on the leaf in apparently sticky masses. Many of them seemed, as in Phragmidium , soon to undergo a process of degeneration, in which the staining distinction between nucleus and cytoplasm became lost, and in one case this process of disorganization was observed in some of the spermatia while they were still enclosed in the spermogonium. The spermatia were never observed germinating or taking any part in aecidium-formation. The formation of spermatia has been investigated by Sapin-Trouffy (48) in Uromyces Erythroni , DC., and other forms, and in Puccinia Liliacearum, Duby, by Poirault and Ra6iborski (48), and also by Maire (29, 30) ; the observations here given confirm and supplement their results. Both Sapin-Trouffy and Maire, however, describe two chromatin masses as always present on nuclear division, and consider them to be of the nature of chromosomes. These two structures are not always present in G. clavariaeforme , and they certainly cannot be considered of the nature of true chromosomes, as a comparison with the divisions in the promycelium shows, for these chromosomes are formed and are much more numerous. This matter will be discussed later in dealing with conjugate division. 338 Blackman . — On the Fertilization , Alternation of AECIDIUM-DEVELOPMENT. Phragmidium violaceum. The so-called aecidium in this genus, like that of Caeoma ( = Melempsora in many cases), is characterized by a very simple structure, for unlike that of the other genera of the group it is neither definite in shape nor bounded by a thick- walled pseudoperidium. In the form investigated the aecidium is nothing more than a group, indefinite in extent, of aecidio- spore-bearing cells, bounded at the periphery by a number of thin-walled paraphyses, which are, however, sometimes wanting. It is clear that this aecidium is no more a definite organ than is the sorus of uredospores or teleutospores, which has an exactly similar arrangement — a group of spore-bearing cells surrounded by paraphyses. The aecidiospores are developed immediately beneath the epidermis, and not deeper down in the tissues as in the typical aecidium. The first beginning of the aecidium is the massing of hyphae beneath the epidermal cells of the leaf, usually on the lower side. The hyphae form there a layer of uninucleate cells, two or three cells thick. The cells immediately beneath the epidermis increase somewhat in size, and soon become divided by a transverse wall, parallel with the surface of the leaf, into an upper and lower cell, each with a single nucleus. The upper cell remains more or less cubical, and shows a vacuolate protoplasm and a small nucleus, which has no nucleolus and sometimes remains as a dense structure without returning completely to the resting state (Fig. (56). The lower cell elongates considerably, and shows abundant granular protoplasm and a large nucleus with a well-marked nucleolus and a clear chromatin network (Figs. 61, 66, 70). The upper cell is a sterile cell ; its nucleus becomes disorganized, and it is soon destroyed by the upward growth of the cell below. It is from the lower cell, which may be called the fertile cell that the aecidiospores arise ; for after a pause in its development it becomes binucleate , and proceeds to elongate (Fig. 61) and cut off a series of binucleate aecidiospore-mother- cells (Fig. 60), the pair of nuclei dividing together by the process of conjugate division. The discovery of a sterile cell is alone a very important fact, for it shows that the aecidiospore-producing cell or hypha, at least in Phrag. violaceum , is a very specialized reproductive cell and not simply of the nature of an ordinary conidiophore. The process by which the fertile cell became binucleate was naturally closely investigated, but all attempts to observe the single nucleus of the fertile cell in any stage of division were quite unavailing, in spite of the fact that other divisions were met with fairly frequently, and that in passing from the periphery to the centre of an aecidium of an appropriate Generations , and General Cytology of the Uredineae . 339 age one can observe the transition from the uninucleate to the binucleate condition. This failure led to a close examination of the fertile cells at about the point of transition, and it was observed that in a number of cases the fertile cell was occupied by two nuclei of different size and structure (Figs. 62-65). One was usually a larger nucleus with the characters given above for the original nucleus of the cell, the other, usually a smaller, denser nucleus containing no nucleolus or only a small one, and having, in fact, more the characters of a nucleus of an ordinary cell of the mycelium. The differences in size and structure of these two nuclei was hardly compatible with the view that they were sister-nuclei, and when this fact was further considered in conjunction with the absence of all stages of the necessary division, there seemed no escape from the view that the smaller, denser nucleus must have had some other origin. This view was amply confirmed by the discovery of a number of cases in which a nucleus zvas actually found passing into the fertile cell from one of the smaller cells of the mycelium at its base (Figs. 66-70) 1. The migrating nucleus is reduced to a narrow thread during the process, the actual aperture through which it passes being very small ; neither before nor after its passage could a pit in the wall be observed 2. In many cases there is to be found lying below the fertile cell a cell which obviously belongs to the same cell-row ; it may be called a basal cell though it is in no way specially differentiated. When this basal cell is present it is often from it that the migrating nucleus comes ; and it is interesting to note that the nucleus seems to pass more often into one of the neighbouring fertile cells (Figs. 66, 67, and 71), with which it is in contact at the sides owing to the irregularity in length of these cells, rather than into the fertile cell immediately above it (Figs. 68 and 69). In the former case the relationship between the two nuclei which meet in the fertile cell may be considerably distant, while in the latter case the two nuclei are separated in origin only by one division, that which cuts off the sterile cell. In some cases in which the fertile cell contained two dissimilar nuclei, the smaller, denser one lay in the upper part of the cell, and at the same time the sterile cell was without a nucleus (Fig. 65). It seemed then possible that the nucleus of the sterile cell might have migrated into the fertile cell below ; but no direct evidence could be obtained for this view, and it is rendered improbable by the fact that a similar state of affairs was 1 More than twenty-four cases were observed in which the fertile cell contained two dissimilar nuclei and more than fifteen cases in which a nucleus was in a stage of actual migration. 2 The migration of nuclei in the tissues of Phanerogams which have been placed under abnormal conditions has been described by several observers (cf. Koernicke, Ber. d. Deutsch. Bot. Ges. xxi, 1904, p. 100). 340 Blackman . — On the Fertilization , Alternation of to be observed in the fertile cell while the nucleus of the sterile cell above was still in position (Figs. 63 and 64). The disappearance of the nucleus of the sterile cell is doubtless to be accounted for by early disorganization, and the peculiar position in the fertile cell of the small nucleus by a change in position after migration, or by the fact of its having entered at the side. It would appear that the two nuclei of the fertile cell soon become similar in size and shape, for it is only occasionally that a dissimilarity can be observed. Very soon after this condition has been attained the fertile cell proceeds to elongate, pushing its way through the sterile cell (Fig. 61) and soon completely destroying it. The paired nuclei divide by the process which has been termed conjugate division, and the fertile cell cuts off a series of cells (aecidiospore-mother-cells), which do not develop directly into aecidiospores, but each cuts off a small cell below, known as the inter- calary cell (Figs. 71-74). It has been suggested that the function of this cell, which becomes disorganized and disappears soon after its formation, is to act as a disjunctive apparatus, and so bring about the complete separation of the aecidiospores. A mature aecidiospore is shown in Fig. 75. Most of the fertile cells become binucleate and develop further, for in older stages of the aecidium only a few of the fertile cells are found, here and there, still uninucleate. It is no doubt these cells, which never achieve the binucleate condition, that are found crushed and distorted in still later stages h Although two is the usual number of nuclei in the fertile cell, the number three is nearly always to be found in one or two cells of each aecidium, and in one case it was found no less than four times in a single aecidium ; but in comparison with the normal number it is of course very uncommon. The three nuclei divide together by a process of conjugate division (Fig. 77), and trinucleate aecidiospore-mother-cells (Fig. 76) and aecidiospores are produced, the fate of which is unknown. In one case also a fertile cell with four nuclei was to be observed. How this multinucleate condition is brought about is not quite clear ; it may be due to the migration of more than one nucleus into the cell, or the division of one or both of the nuclei without cell-wall formation. In a fertile cell which was yet undivided one of the nuclei was observed to be apparently undergoing division while the other was in the resting state. When three or four nuclei were observed in an undivided fertile cell it was interesting to note that their size was usually considerably less than that of the normal paired nuclei. 1 The fertile cells are often in contact with several of the cells of the tissue below, owing to the irregularity in length of the former, and the usually narrow form of the latter when they do not torm distinct basal cells ; a more than sufficient number of nuclei is thus at hand to supply all the fertile cells. Generations , and General Cytology of the Uredineae . 341 The development of the aecidium is centrifugal, and when it has reached a certain size the peripheral cells, instead of forming ordinary sterile and fertile cells, grow out into uninucleate paraphyses : these may, however, in some cases may be absent. The process of nuclear migration, by means of which the fertile cell is stimulated to rapid growth and continued division, is clearly a process of fertilization , the exact relation of which will be discussed later1 *. Gymnosporangium clavariaeforme. The first beginning of the aecidium takes place deep down in the hypertrophied tissue of the leaf, & c., after the majority of the spermogonia have faded. The aecidium in this genus is of the typical form, with a definite pseudoperidium. Unfortunately the very early stages, which are much more difficult to obtain than in Phragmidium , were not observed, so that the presence or absence of sterile cells and the exact behaviour of the nuclei could not be investigated. It was clear, however, that the transition from the condition of single to that of paired nuclei took place, as described by Sapin-Troufify, in connexion with the aecidiospore-bearing cells (fertile cells) ; for while in the whole mycelium, and even in the dense layer of fungal cells surrounding the actual aecidium, the nuclei were single, yet the fertile cells and their products showed clearly two nuclei in the paired (conjugate) condition. It is evident that a comparative study of aecidium-development in the Uredineae generally, with special relation to the existence of sterile cells and the origin of the binucleate condition, is much to be desired. The only other forms of which we have any information are Endophyllum Sempervivi , De Bary, and Puccinia Bunii , DC., in which Maire (29, 30) is of opinion that the two nuclei of the fertile cell are daughter-nuclei arising by division of the original single nucleus. The migration of nuclei, however, is very easily overlooked unless special attention is drawn to such a point ; it is interesting to note that Maire does not appear to have observed in Puccinia Bunii the actual division, for he merely states, ‘ A little later, it can be seen that the terminal cells [fertile cells] contain each two nuclei 5 (30, p. 39). It is of course possible that in the aecidium of Phragmidium with its simple structure we have a more primitive condition, and that in some other forms the association of nuclei of two different cells may have been replaced by association of daughter-nuclei of the same cell ; but judgement must be suspended in the matter. 1 Whether any protoplasm passes over with the migrating nucleus could not, of course, be determined. B b % 342 Blackman. — On the Fertilization , Alternation of MYCELIUM AND DEVELOPMENT OF UREDOSPORES. Phragmidium violaceum. The uredospores appear on the leaves about June and replace the aecidia, which are only short-lived. After passing the aecidial stage the parasite appears to have gained greatly in energy, for the degree of infec- tion is now much greater 1. The mycelium in the leaf which gives origin to the uredospores shows, naturally, paired nuclei like the aecidiospores from which it was derived. The details of uredospore-formation were not investigated in the leaf, but from some patches on older stem-portions where the mycelium was perennial and gave rise to uredospores early in the year, even before the aecidia were developed on the leaves. These patches were found to be much more favourable objects of study than those on the leaves, the general form of which has been figured by Sapin-Trouffy. The perennial mycelium was found to form a layer of considerable thickness, in which even thin sections show a considerable number of the nuclei arranged in pairs (Fig. 78), but as the two nuclei of these cells often separate for some distance, the paired condition cannot always be observed. A number of the hyphae penetrate the cells of the host, thus acting as haustoria (Fig. 99), but they did not appear to have that special relation to the nucleus of the host-cell described for several cases by Sapin-Trouffy. The stalked uredospores are borne on somewhat rectangular cells which may be termed basal cells. They grow up from the free surface of mycelium and form a regular layer (Fig. 79). These cells are usually vacuolate, and show two nuclei which are larger than the nuclei of the mycelial cells and exhibit each a well-marked nucleolus and granular chromatin. The uredospore arises from the basal cell as a binucleate outgrowth (Figs. 80, 81), which soon becomes of a somewhat oval shape. The two nuclei then undergo the process of conjugate division (Fig. 82) and the outgrowth becomes divided into two cells (Fig. 83), the upper increasing much in size and forming the uredospore, the lower remaining narrow, but elongating later and forming the stalk. The wall of the uredospore becomes much thickened, but a very definite pit through which there is distinct protoplasmic continuity con- nects, for some time, its cavity with that of the slightly thickened stalk (Fig. 83 a). In the mature state, however, before the uredospore separates from the stalk, the pit becomes obliterated. 1 The discovery of a process of fertilization is, of course, a confirmation of the interesting view, put forward by Arthur (1, 2), on other grounds, ‘that the aecidium is a device to restore vigour to the fungus.’ Generations , and General Cytology of the Uredmeae. 343 The nuclei of the free uredospore have sometimes a peculiar, irregular contour as seen in Fig. 84 ; this appears to be only a temporary phase of development. DEVELOPMENT OF TELEUTOSPORES. Phragmidium violaceum. The teleutospores appear on the leaves towards autumn, and are at first mixed with the uredospores, but later arises in sori which consist of teleutospores only, surrounded by a layer of paraphyses. The mycelium on the leaf from which they arise shows a septate mycelium with a pair of nuclei in each cell. In Fig. 85 the paired nuclei can be distinctly seen in the haustoria in the cells marked * and in several of the other cells, but owing to the thinness of the section the nuclei appear single in a number of cases. From this mycelium there grow up, when the teleutospore sorus is to be developed, a number of rectangular, binucleate basal cells closely packed beneath the epidermis. It is from these special cells (which Sapin-Trouffy has figured in a diagrammatic way) that the teleutospores are developed. The young teleutospore appears first as an elongated, binucleate, cylindrical outgrowth from the basal cell. It increases in diameter, and soon cuts off from the apex downwards a series of three or four superposed cells (Figs. 86, 86 a ), which form the body of the teleutospore, the lowest cell becoming the stalk. The upper cells take on the characteristic shape, and the walls become thickened. The upper cell grows out into a short, narrow projection (Fig. 86 b ), the cavity of which soon becomes completely or almost com- pletely obliterated by the great increase in thickness of its wall ; there is thus produced the knob-like projection characteristic of the mature spore. The lowest cell, the stalk, increases very greatly in length, and its cavity becomes almost completely obliterated by the thickening of its walls, except in the lowest part where it is inserted on the basal cell. Here there is a cavity of considerable size in which the two disorganized nuclei are to be found (Fig. 86 c). Some evidence was obtained of a pit connecting for some time the end of the stalk with the basal cell (as in the case of the uredospore and its stalk) ; the absence of thickening at this point, and the peculiar widening of the cavity shown in Fig. 86 c, strongly suggest such a connexion, though its existence was not established with certainty. During the process of thickening of the wall of the spore the nuclei in the cells show each a single well-marked nucleolus and granular chromatin, and are usually to be found close together (Figs. 86, 86#). When the wall is fully thickened the process of nuclear fusion begins. In the first stage the two nuclei are found in close contact (Fig. 87 a) ; in the second stage instead of two small nuclei one larger one is found, but the two 344 Blackman. — On the Fertilization, Alternation of nucleoli remain for a time without fusion (Fig. 87 b ), as described by Sapin- Trouffy. In the next stage the nucleus has increased somewhat in size, the two nucleoli have been replaced by one large one, and the chromatin, which has hitherto not been well marked, begins to stain more deeply (Fig. 87 c ). The nucleus still increases in size, and the chromatin becomes resolved into a very well-defined thread which is apparently a single spireme (Fig. 88 a). This, however, is not the precursor of division, for the nucleus still increases in size and the chromatin returns to the condition of a fine network with the nucleolus as a somewhat elongated body flattened against the nuclear membrane (Fig. 88 b). The nucleus then goes into the resting state, in which it passes the winter, appearing as an almost homo- geneous body except for the well-marked nucleolus (Fig. 2). The process of increase in size and of formation of a spireme thread seems to correspond with the changes included under the term synapsis in the higher plants and animals, but it is not followed by nuclear division but by a period of rest in which the spireme again disappears. A somewhat similar case has been described by Williams (60) for the tetraspore-mother-cell of Dictyota , where after synapsis a similar disappearance of the spireme thread and a period of rest was found to occur. Gymnosporangium clavariaeforme. The perennial mycelium which inhabits the stems of Juniperus and bears in spring the yellow masses of teleutospores shows very clearly the paired nuclei in the cells of hyphae, which are thick-walled as described by Sapin-Troufify and others. As in the case of the teleutospores and uredospores of Phrag. violaceum , the teleutospores of this form are not borne directly on the mycelium but arise from comparatively large rectangular cells, which form a close-set layer on the surface of the mycelium, at the points where the teleutospores are developed. These each contain two nuclei which are somewhat larger than those of the ordinary cells of the mycelium, and show each a well-marked nucleolus (Fig. 89). They are similar to the teleutospore-bearing cells described by Sapin- Trouffy for G. Sabinae , but their side and lower walls are of considerable thickness. Each of these cells gives origin to a number (not more than three or four) of narrow outgrowths which develop into the stalked, two- celled teleutospores (Fig. 89). The outgrowths have paired nuclei (Fig. 90) and soon undergo two divisions, the first cutting off the stalk-cell, while the second divides the upper cell into the two cells of the teleutospore. The teleutospore then increases in size, the wall becomes thickened, and the cytoplasm vacuolar and filled with yellow, oily reserve-material. The two nuclei in the teleutospore fuse, earlier than in Phragmidium, before the spore has obtained its full size or thickness of wall. The wall Generations , and General Cytology of the Uredineae. 345 at the point of contact of the two nuclei disappears and the two nuclei fuse, the two nucleoli being distinguishable for some little time after fusion (Fig. 91 a,b, and c). The fusion-nucleus then increases in size and goes through changes exactly similar to those of Phragmidium . A single large nucleolus appears, and the chromatin becomes resolved into a thick, continuous, spireme thread (Fig. 92). Here, also, this condition is merely temporary, for the thread after a time becomes replaced by a fine network, with usually only small inconspicuous nucleoli (Figs. 14 and 1 5), as described under the mature teleutospore. As in the case of Phrag. violaceum , these nuclear changes must be looked upon as a process of synapsis. METHOD OF ‘CONJUGATE5 DIVISION. The method of division of the paired nuclei is of the same simple type as that described for the single nuclei of the spermatial hyphae, but the two nuclei are always to be found in the same stage of division and in very close association during the process. This division can best be observed in the development of the aecidiospores, uredospores, and teleutospores. The first step in the process is the disappearance of the nuclear membrane and the condensation of the chromatin of each nucleus into a homogeneous, irregular mass, with the natural result that the single nucleolus (which is always present at least in the nuclei of the cells connected with spore- formation) comes to lie free in the cytoplasm (Figs. 93 95 a). Sometimes the chromatin can be observed in a state where it is not yet completely homogeneous and still retains in part the form of the original nucleus (Fig. 98 a). In the next stage the chromatin mass becomes more regular (Figs. 93 95 &)t and in favourable cases can be seen to be connected with a thread-like structure (Figs. 72, 93 b)} which has the staining reactions of kinoplasm, and is no doubt to be considered as of the nature of a simple spindle. The chromatin mass then becomes elongated to form a thick rod which is apparently stretched out on the spindle, for the latter is completely obscured (Figs. 93 c, 95 c). Each chromatin rod then becomes drawn out into two more or less pear-shaped masses (Figs. 93 d , 95 d), which at first remain connected by a fine thread which has the staining reactions of kinoplasm (it takes the gentian in the triple stain), and is no doubt the drawn-out spindle (cf. divisions in promycelium). As the two masses move still further apart the connecting thread becomes very faint and ceases to be continuous (Fig. 93 d). At this stage, or somewhat earlier, the rejected nucleoli, which have hitherto lain almost unaltered in the cytoplasm, begin to decrease in size and finally disappear, though they may remain visible up to the stage in which the new nucleoli have begun to appear in the daughter-nuclei (Fig. 93 e\ 346 Blackman. — On the Fertilization , Alternation of In some cases, as in the case of the divisions of the single nuclei, it can clearly be observed that the chromatin does not form a single rod but a double rod (Fig. 96), which becomes drawn out into two pairs of chromatin masses, so that at this stage of the anaphase the nuclei show two opposite groups, consisting of four chromatin masses (Figs. 97 and 98). How far the presence of these two masses in the nucleus is of common occurrence it is difficult to say, for if the plane of separation of the two lies at right angles to the direction of view they would easily be overlooked. There would seem to be no doubt, however, that they are only occasionally present in the divisions in Phragmidium ; in the teleutospore in Gymnosporangium , however, the double chromatin mass seems to occur more frequently. It is the appearance of these two chromatin masses which led Sapin- Trouffy (48) and Maire (29, 30) to the belief that the nuclei in the Uredineae possess in all cases two chromosomes, while Poirault and RaSiborski (43), observing the single mass, considered it to represent a single chromosome. It is obvious, however, that neither the origin nor behaviour of these structures is that of chromosomes, and the observation of numerous chromosomes, described earlier for the promycelial divisions, is sufficient to negative such a view. The real nature of these chromatin masses is, however, apparent on comparison of the divisions in the promycelium with the simpler divisions, either single or paired (conjugate), found in other stages of development. In both the promycelial divisions the chromatin (vide supra) shows at times a distinct segregation into two masses (Figs. 30, 32, 41 a). These masses are certainly not themselves chromosomes, and in the first division are clearly produced by the aggrega- tion of numerous chromosomes ; they probably represent, as suggested earlier, the chromatin derived respectively from the two nuclei which fuse in the teleutospore. It is not surprising then that nuclei which are the direct descendants of those in the promycelium should also show a similar tendency to a differentiation of their chromatin into two masses, though they have apparently ceased to form chromosomes. A comparison of Fig. 30 with Figs. 94 and 9 6 shows clearly the strong resemblance between the form taken by the chromatin in the two types of division. NATURE OF SPERMATIA. As the study of Phrag. violaceum clearly shows, the spermatia take no part in aecidium-development, and fertilization is regularly brought about in another way. This result is only in agreement with all earlier work, which has always failed to establish any direct relation between the spermatia and the aecidia. A fertilization of the fertile cells would obviously be difficult in Phragmidium, but it would be almost impossible in the typical aecidium, where the fertile cells are developed deeper down in the tissues. Generations , and General Cytology of the Uredineae. 347 As the spermatia have no power to act as male cells and appear quite incapable of causing infection, there would seem to be no escape from the view that they are now functionless 1. It becomes evident then that a study of their structure is the only means by which one can hope to obtain evidence as to their primitive nature, for originally they must have acted either as male cells or as conidia. Strangely enough, their histological characters have never been considered from this point of view, though they appear to give very definite evidence for a decision of the disputed question. As the observations detailed earlier show, the spermatia, in both the forms investigated, are small, uninucleate cells, with a thin wall, apparently no reserve-material, a very dense nucleus without a distinct nucleolus, and a very small amount of cytoplasm (Figs. 47, 59) 2. That these characters are general throughout the whole group is shown by a glance at the figures of Sapin-Trouffy (48), diagrammatic as they are ; though this observer, assuming that the spermatia were conidia, passed over these characters without particular comment. It is obvious, however, that these characters are not those of conidia nor of any regular asexual reproductive cells ; on the other hand they are to a very striking degree the characters of male cells, in which, as is well known, there is usually little or no reserve-material, the nucleus is often very dense, and the cytoplasm is nearly always much reduced in amount. The reduction in cytoplasm is very well marked in the small sper- matia of Phragmidium and Gy mno sporangium, but it is just as clearly shown in the comparatively large spherical spermatia of Coleosporium Senecionis , Fr., where, according to the figure of Sapin-Trouffy, the diameter of the nucleus is more than two-thirds of that of the whole cell. Such a struc- ture is certainly without parallel in any known asexual reproductive cell, and sufficiently explains the inability of the spermatia to develop in water, for the volume of cytoplasm is insufficient to form a germ-tube of any length. When one considers the peculiar structure of the spermatia of the Uredineae, their total incapacity, as far as is at present known, to bring about infection, their feeble power of development even in nutritive solu- tions, their usual close association with the aecidium, there seems no escape from the view that the spermatia are male cells which have now become functionless. The peculiar process of fertilization in the fertile cells of the aecidium of Phrag. violaceum , which one can hardly but consider as a reduced process, seems only explicable on the view that the spermatia formerly acted as fertilizing agents in connexion with the aecidia. 1 It is satisfactory to note that Klebahn (27) in his recently published work, though unable to throw any light upon their nature, also arrives at the conclusion that they act neither as conidia nor male cells ; he thinks they may in some way be useful to the plant as an excretory product. 2 I have also observed the same structure in the spermatia of Puccinia Poarum , Niels, P. Phalaridis, Plowr., and Uromyces Poae, Rabh. 348 Blackman.— On the Fertilization 3 Alternation of It must be pointed out in relation to the observations of Cornu (12), Brefeld (10), Plowright (42), Sapin-Troufify (48), and others on the ger- mination of the spermatia in nutritive solutions, that a complete absence of power of vegetative development is not a necessary character of male cells, as is shown by the well-known cases of the potential gametes of some algae which can develop either sexually or asexually. It must also be remembered that nutritive solutions are a highly artificial condition for the reproductive cells of such obligate parasites as the Uredineae ; it is difficult to understand how the spermatia could act in nature as infecting organs (conidia) unless they are able to germinate in water like the other spore- forms of the group. Their appearance at about the same time as the aecidiospores is another argument against their conidial nature (as well as an argument in favour of their sexual nature), for, in the presence of such very effective infecting organs as the aecidiospores are known to be, there would seem to be absolutely no need for the production of conidia so structurally ill-equipped for directly carrying on the life-cycle as are the spermatia. The fact that most of the spermatia in the forms investi- gated are found to be disorganized soon after they are shed, and that rarely in G. clavariaeforme some of them degenerate while still in the spermogonium, is evidence against the conidial view and in favour of the view that they are abortive male cells. There is of course a view which might be suggested, that the spermatia are degenerate conidia which have ceased to have any function. Such a view, however, may certainly be dismissed, for it is difficult to imagine the conditions in which the conidia would cease to be of value, and still more difficult to imagine a process of degeneration which brought about a reduction of the cytoplasmic portion of the cell, but left the nucleus untouched. That the spermatia, though perfectly functionless, should still be pro- duced by the majority of forms and in such great abundance is certainly a very striking phenomenon 1. A somewhat similar process, apparently equally wasteful, is to be observed in certain animals, as in the gasteropod, Paludina , where two sorts of spermatozoa, normal and giant ones, are constantly produced, although the smaller ones only are functional. The similarity of the spermogonia of the Uredineae to those of Collema , in which the earlier observations by Stahl on the fertilizing action of the spermatia have of late years been confirmed by Baur (5), is additional confirmation of the view that the spermatia in the Rusts are male cells. It is true that Moller has shown that in certain cases the spermatia Even when the aecidinm is dropped out of the life-history the spermogonium sometimes still remains as the first structure to be developed on the mycelium which arises from the sporidium, being later followed by uredospores or teleutospores ; e. g. in the so-called brachy- forms. Generations , and General Cytology of the Uredineae. 349 of lichens can develop as conidia, but, as pointed out by E. Fischer (Bot. Zeit., 1888, p. 158) in reviewing Moller’s work, and also later by Harper (20), this in no way proves that the spermatia are not potential male cells. That they do not usually develop except in nutritive solutions, that it is often months before they germinate, and that their growth is ex- cessively slow, strongly suggest that they are not simple conidia ; but these facts are quite in keeping with the view that they are male repro- ductive elements. The suggestion lately put forward by Metzger (33) — that the spermatia of lichens are male cells which have retained a certain power of vegetative development, and, now that in many cases the ascus fruit develops without their aid, they sometimes act as conidia, and may have become modified in that direction — seems to be the most satis- factory one1. There seem to be no data for a comparative cytological study of the spermatia of lichens, though it would be of great interest. The figure of the spermatium of Buellia punctiformis given by Istvanffi (24) shows a cell with a much smaller nucleus in proportion to the size of the cell than in the Rusts ; but it is possible that only the nucleolus was observed. FERTILIZATION IN THE AECIDIUM AND THE NUCLEAR FUSION IN THE TELEUTOSPORE. It is very clear that the process which takes place in the young cells of the aecidium of P. violaceum is one which has most of the characters of an ordinary sexual process. A cell is to be observed, which after cutting off a sterile cell above, increases in size and becomes a specialized reproductive cell, exhibiting abundant protoplasm and a large nucleus with a well-marked nucleolus. A pause then occurs in its development, but it is later stimulated to further growth and rapid division by the entrance of a nucleus from without. Such a series of phenomena leave no escape from the view that the fertile cell is a female reproductive cell which undergoes a process of fertilization. There are, however, two points in which the process differs from that of ordinary fertilization. One is, that the two nuclei which become associated in the fertile cell do not fuse, but retain their morphological individuality, though the closeness of their relationship is shown by their always dividing together in close juxtaposition ; a fusion of nuclei does ultimately take place in some of their descendants, but it is confined to those pairs of nuclei which are to be found in the teleutospores. The second point of difference is, that the entering nucleus is derived from an un- 1 Hedlund (22) has described two cases in which the spermatia of lichens germinated and developed a thallus in a state of nature. 350 Blackman . — On the Fertilization , Alternation of differentiated vegetative cell and not from a specialized male cell or organ ; this will be discussed in the next section. A study of the cytological features of fertilization in various animals and plants shows clearly that actual fusion of nuclei is not a necessary part of fertilization as ordinarily understood. Although in Angiosperms the sexual nuclei fuse together before the egg develops further, yet in the Abietineae (Blackman, 6 ; Ferguson, 16), and in most animal eggs, development of the egg begins without the fusion of the nuclei, for the two begin to divide in- dependently and only meet together on the first division-spindle. Again, in the egg of Cyclops , we have cases in which the two nuclei can be distinguished for several cell-generations1. Not only was Rtickert (46) able to observe that in the first segmentation of the egg of Cyclops the two nuclei divide side by side in an association very little closer than that to be observed in the conjugate nuclei in the Uredineae ; but Hacker (17, 18) even observed double nuclei in the cells of the embryo up to the sixteen-celled stage (at what stage actual fusion took place was not determined). Again, in many cases of fertilization the fusion between the two nuclei is more apparent than real, for the chromatin and chromosomes derived from the male and female nuclei respectively have in a number of instances been traced as separate structures through several generations (eggs of Ascaris, Cyclops , &c., and the chromosomes in Pinus). Hacker (17) also believes that in the egg of Cyclops he has been able to trace the maternal and paternal chromatin as separate structures from the time of fertilization up to the formation of the next germ-cells. In fact all the cytological work of recent years tends to show that the chromosomes have a distinct ‘ individuality/ and that the nuclei of the cells of the higher animals and the nuclei of the sporophytic cells of the higher plants are really dual in nature, there being no real mixing of the chromatin from the two sources until the time of chromosome-reduction 2. A study of the varieties of fertilization in the higher animals and plants shows clearly that three morphological stages are to be observed in connexion with the complete sexual cycle — nuclear association within the 1 Conklin (11) also in the egg of Crepidula was able to observe the double character of the nuclei at the telophase in every cleavage up to the twenty-four-cell stage, and in several cleavages up to the sixty-cell stage. 2 See for example the work of Van Beneden, Boveri, Herla and Zoja, and of later years that of Sutton (52), and the observations on chromosome reduction of Montgomery (37, 38), Sutton (51), and Farmer and Moore (15). Sutton describes a very interesting case in which the chromosomes (twenty-three in number) in the spermatogonium of an insect ( Brachystola ) are not only of eleven different sizes, those of the same size being arranged in pairs, but the majority of them in the resting state are each contained in a separate diverticulum of the sacculated nucleus, the ‘ accessory chromo- some ’ lying in a separate closed vesicle and thus forming, virtually, a separate nucleus. The results obtained in certain hybrids in which the sexual cells are apparently ‘ pure ’ in relation to certain maternal characters also suggest that the nuclear material from the two sources remains distinct from the time of fertilization up to the formation of the germ-cells. Generations , and General Cytology of the Uredineae. 351 same mass of cytoplasm and during division , nuclear reduction [fusion), and chromosome-reduction. These processes may either take place together or be separated by a considerable number of divisions. In the case of the higher plants the first two processes usually take place together, i. e. the sexual nuclei fuse in the resting state directly they meet1. In many animal eggs, and a few plants, the second process is somewhat delayed, and there is no nuclear fusion until the end of the first division. In the case of the egg of Cyclops we have a further stage, for nuclear reduction is put off for many cell-divisions, each of the cells showing two nuclei which are, however, closely associated together during division. In the Uredineae there is a still further stage of separation, for in this case nuclear reduction (or fusion) is put off until the stage corresponding with chromosome-reduction (vide infra), so, like that process, it is confined to the reproductive cells (teleutospores = spore-mother- cells) which are to carry on the life-history. The two nuclei, also, instead of dividing together on the same spindle divide on separate (rudimentary) spindles side by side. Fertilization is, of course, a somewhat ill-defined term, and from a strict morphological point of view it is perhaps arbitrary to confine it to any one of the three processes just described, for sooner or later all of them must occur in the complete sexual cycle. A cell is, however, usually considered to be ‘ fertilized ’ directly the stimulus has been given to development and the number of chromosomes doubled by the entry of the male cell or nucleus, and for such a use of the term fertilization it is obvious that the essential point is the first stage — nuclear association in the same mass of cytoplasm and during division. The time at which nuclear reduction (fusion) takes place is seen from the stages given above to be quite unimportant, but, like chromosome- reduction, it must occur sooner or later in the sexual cycle. The view which would consider fertilization as incomplete until nuclear fusion had taken place could hardly be accepted, as chromosome-reduction might equally be considered as the end of the process. It would also tend to divorce the morphological and physiological aspects of fertilization, for in Cyclops and Phragmidium , for example, we should have to consider the female cell as developing under the stimulus of fertilization, but with the process still incomplete. Besides, all recent work at present tends to show that nuclear fusion is mainly a nominal process (hence the term nuclear reduction is preferable), a mere change from association within the same cytoplasm to association within the same nuclear membrane. The view of Dangeard and Sapin-Trouffy, which would consider the fusion in the teleutospore as itself a sexual process, was based on an 1 It is generally believed that in many of the Thallophyta where there is no alternation of generations the three processes all take place together, chromosome-reduction occurring before the fusion-nucleus divides. 352 Blackman. — On the Fertilization , Alternation of ignorance of the origin of the two nuclei and the exaltation of nuclear fusion as a test of fertilization. The fusion is the result of a sexual process (at least in Phrag. violaceum ), but is clearly not in itself a fertilization. Raciborski (44), when he put forward the hypothesis that the conjugate divisions in the Uredineae represented a vegetative phase intercalated in fertilization between the stages of cytoplasmic fusion and nuclear fusion, was much nearer the truth, though, as has been pointed out above, it is not at all necessary to consider nuclear fusion as a stage in fertilization as usually understood. Raciborski, equally with Dangeard and Sapin-Trouffy, was quite ignorant as to the method by which the two nuclei became ‘ conjugate.’ Maire (29) had earlier recognized a similarity between the conjugate nuclei of the developing egg of Cyclops and the paired nuclei of the Uredineae, and applied the term ‘synkaryon5 to the two nuclei. He recognized that the fusion in the teleutospore was not in itself a process of fertilization, but considered it a mere process of ‘ mixie,’ in which he includes both nuclear and chromosome reduction in the sense used above. He believed the ‘ synkaryon ’ to be brought about by the association of two daughter-nuclei of a cell, and at first considered it actually comparable to a sexual process, but in a later paper (30) he comes to the conclusion that between the fertilization in the higher plants and the formation of the ‘ synkaryon ’ in Uredineae (and Basidiomycetes generally) there are only * des relations de cousinage.’ He elaborated his views in some considerable detail, but the discovery of the origin of the two nuclei in the aecidium of Phragmidium , and the evidence here brought forward that the spermogonia and aecidia are in all cases to be considered as sexual organs, throws a completely new light upon the matter ; his views need not therefore be discussed further. ORIGIN OF THE PROCESS OF FERTILIZATION. The process of fertilization as observed in the fertile cells of Phrag. violaceum might be looked upon as merely one of great simplicity in which the male cell remained undifferentiated. The two nuclei which take part in the process are not sister-nuclei though they may sometimes be closely related, and the process as a whole (except for the absence of obvious cytoplasmic fusion) is not much simpler that that of Basidiobolus , in which two neighbouring cells, after forming each a sterile cell, fuse together to form a zygospore. Taking into account, however, on the one hand, the completely un- differentiated nature of the cell from which the acting male nucleus comes, and the peculiar method by which the nucleus passes from one cell to another, and on the other, the presence of the functionless spermatia with their special cytological characters, the only view that seems satisfactory to explain the facts is that the primitive normal process of fertilization by Generations , and General Cytology of the Uredineae . 353 means of the spermatia has been replaced by a fertilisation of the female cell by the nucleus of an ordinary vegetative cell. Such a view receives strong support from the observation of Farmer, Moore, and Digby (14) on apogamy in Ferns, published while this work was in progress. There seems no doubt that the process in the aecidium is to be looked upon as intermediate between that of normal fertilization, where both the cells are specially differentiated, and that observed in apogamous Ferns by these observers, where both acting male and female cells are represented by ordinary vegetative cells. If the view be accepted that the process observed is a reduced form of ordinary fertilization, it seems very probable that the sterile cell also is reduced. Its position above the fertile cell would suggest that it formerly acted as a receptive cell pushing up between the epidermal cells as a trichogyne to which the sticky spermatia could be brought, for example, by insects. Some support is lent to such a view by the fact that occa- sionally cases are to be found in which the sterile cells do push up between the epidermal cells and swell out above, being merely covered by the cuticle (Fig. 77 a). If development were pushed one stage further and the cuticle pierced, a very effective receptive organ would be the result. MORPHOLOGY OF THE AECIDIUM. Although this subject cannot be treated fully in the absence of a comparative study of aecidium development in various members of the group, yet, as has been stated earlier, the aecidium of Phragmidium (and apparently that of Caeoma) is very different from the typical aecidium found in the other genera. It is merely a group of special reproductive organs, indefinite in extent, and merely bounded by paraphyses which are some- times absent ; and thus no more a definite structure than the sorus of uredospores or teleutospores. In the light of our present knowledge the aecidium of Phrag. violaceum is thus to be considered as a group or sorus of female reproductive organs , each consisting of a female (fertile) cell below and a sterile cell above. We probably have in this genus the most primitive form of the aecidium to be found in the group, but of the form of the aecidium in the still more primitive state when fertilization by the spermatia yet occurred, we can say nothing1. With the change to the reduced form of fertilization, with its obvious advantage of certainty, the number of female organs in the group could be indefinitely extended without the necessity of breaking through to 1 It is of course possible that the spermatia may sometimes really act as fertilizing agents in Phragmidium , and perhaps in Caeoma , for they are occasionally found scattered over the surface of the leaf below which the aecidium is developing ; and if a sterile cell should push through to the surface and come in contact with a spermatium fertilization might result : such a process, however, was never observed, and it can only be of very rare occurrence. 354 Blackman. — On the Fertilization , Alternation of the surface. The absence of this necessity might also lead later to a develop- ment deeper down in the tissues in the host, as is found in the typical aecidium, where conditions of nutrition would perhaps be more satisfactory. With this deeper point of origin may also perhaps be associated the develop- ment of the pseudoperidium, which would protect the young structure while pushing its way to the surface. There can be little doubt that the aecidium, when present, is always the seat of the process of transition from the condition of single to that of paired nuclei. Sapin-Trouffy (48) found this to be the case in all the forms (about fourteen) which he investigated. His observations have been confirmed by Maire (29 ; 30) for Endophyllum Sempervivi and for Puccinia Bunii , DC., and by the writer in the two forms here described, and also in Puccinia Poarum , Niels, P. car ids , Rebent, and in Uromyces Poae> Rabh. The conclusions as to the nature of the aecidium in Phragmidium must then apply throughout the group, and the aecidium in all cases must be con- sidered as a group of female reproductive organs 1. Whether in all cases the transition (fertilization) is brought about by a nuclear migration, or whether in some cases the two nuclei are formed in the cell by division (as Maire believes), must be left to future work to decide. Such a condition would, however, be merely a further stage of reduction in fertilization, comparable to the well-known case of fertilization by means of a sister- nucleus (the second polar body) observed in the * parthenogenetic ’ eggs of Artemia. The hypothesis put forward by Poirault and Raciborski, that the sporidium may become binucleate and so start a mycelium with paired (conjugate) nuclei, is quite untenable. The work of Sapin-Trouffy, of Maire, and the observations here detailed have shown that in all cases investigated the mycelium which arises from the sporidium has single nuclei. It is true that in certain cases, such as in Coleosporium Euphrasiae , investigated by 1 In Coleosporium what are usually known as the uredospores are produced in chains and have intercalary cells between them , and are thus liable to be taken for the true aecidiospores though no pseudoperidium is present. In forms such as C. Senecionis , Fr., in which the full life-history is known (about fifteen species of this genus are now known to be heteroecious), there can be no confusion between the two, as the true aecidium is found early in the year on the one host, is associated with spermogonia, and has a definite pseudoperidium ; while the aecidiospore-like uredo- spores are produced later, on the other host, and are intercalated, like ordinary uredospores, in the life-cycle between the true aecidiospores and the teleutospores. Sapin-Trouffy has shown, as was to be expected, that in C. Senecionis the transition from the single to the paired nuclei takes place in connexion with the true aecidium, while the aecidiospore-like uredospores are borne on a mycelium with paired nuclei derived by infection from the aecidiospores. Owing to their position in the life- cycle, the histological nature of the mycelium which bears them, and the absence of a pseudo- peridium, it is clear that the spores in question partake much more of the nature of uredospores than of aecidiospores, so the former name should be retained for them in spite of their development in chains. Bearing in mind the heteroecious nature of nearly all the forms of Coleosporium hitherto investigated the view put forward by Holden and Harper (23) can hardly be accepted, that the uredospores are really aecidiospores in the form to which they give the name C. Sonchi-arvensis , Lev., and of which they observed only the stage with uredospores and teleutospores. Generations, and General Cytology of the Uredineae . 355 Poirault and Raciborski, the sporidium does become binucleate, but this appears to be a mere precocious division of the nucleus without wall- formation, for, as mentioned earlier (see under section dealing with sporidia- formation), the sporidia of Phrag . violaceum and a few other forms have been found to give origin, in spite of their binucleate nature, to the normal mycelium with single nuclei. Poirault and Raciborski, however, observing that the mycelium which gave origin to the uredospores and teleutospores had paired nuclei and that the sporidium also was binucleate, concluded that the former arose from the latter. They completely, however, overlooked the fact that C. Euphrasiae had been shown by Klebahn (26) to be a heteroecious form and that the sporidium really gave rise to the aecidial stage, the so-called Peridermium Pint , which in the case of C. Senecionis Sapin-Troufify had shown to have the typical structure, a mycelium with single nuclei. Holden and Harper (23) in a recent paper, observing that the sporidium becomes binucleate in a form to which they give the name of Coleosporium Sonchi-arvensis , Lev., have again put forward the view that the sporidium starts the stage with paired nuclei. Such a view, for the reasons just given, can hardly be accepted without direct evidence. The authors assume that the binucleated sporidium gives rise directly to the mycelium with paired nuclei which they investigated and found bearing the uredo- spores and teleutospores ; but it has been shown that the species in question, like that of C. Euphrasiae , is heteroecious (see Klebahn, 27), with its aecidial stage on Pinus , like nearly all the other species of the genus yet investigated. The aecidial stage in this form has therefore no doubt a mycelium with single nuclei like C. Senecionis and all the aecidial stages of the Uredineae hitherto known. The mere fact of the sporidium being binucleate cannot be considered as being of any special significance. It is true that the name C. Sonchi-arvensis has been applied loosely to a number of what are now known to be different forms, the original form being really confined to Sonchus in its second stage (see also Klebahn, 27). The Coleosporium on Aster and Callistephus will probably thus prove to be a new form, but there can be no reason to doubt its heteroecism, since nearly all the forms of the genus hitherto investigated appear to have two hosts. There is no evidence that the sporidium of any form ever gives rise to anything but a mycelium with single nuclei. As stated earlier, all the work of Sapin-Trouffy, of Maire, and that here detailed goes to show that the condition with paired nuclei always starts in the aecidium when that structure is present. Even in the absence of the aecidium the paired nuclei only arise later in the mycelium in connexion with the development of the uredospores or teleutospores 1. 1 Arthur (61), in reviewing the work on this subject, concludes that the sporidium regularly starts the condition of paired nuclei in the group, and is naturally at a loss to explain the fact that the spermatia are uninucleate. His views seem based on a misunderstanding of the work of Sapin- 356 Blackman. — On the Fertilization , Alternation of NUCLEAR DIVISION. From the observations of Poirault and Raciborski, Sapin-Trouffy, and Maire- — although they did not observe all the details of the process and their interpretation is clearly at fault — there can be no doubt that the type of nuclear division here described, whether for single or conjugate nuclei, is general for all the Uredineae. It is clear then that in all of the divisions in the group, except those in the promycelium, we have to deal with a process of such a simple nature that chromosome-formation is in complete abeyance ; so that the division actually partakes of the nature of direct division (amitosis). A comparison, in such a form as G. clavariaeforme , of the first and second divisions in the promycelium with the other simple divisions (whether single or conjugate), leaves no escape, however, from the view that the simple method of division is really a reduced form of indirect division. The first promycelial division is a fairly typical form of mitosis with a formation of chromosomes, though the absence of any regular equatorial plate, the fact that no splitting can be observed, and the early fusion of the chromosomes suggest that perhaps even here the process may be reduced from a halving of definite chromatin elements to the more or less direct separation of chromatin material as a whole. In the second division, however, though the spindle structure is perfectly typical, the chromatin instead of forming distinct chromosomes merely forms a network which becomes spread out on the spindle and later draws apart into two portions1. If this second method of division be still further reduced, so that the chromatin instead of forming a network forms a solid mass and the spindle is represented only by a fine, thread-like structure, we have the ordinary method of division characteristic of the nuclei, whether single2 or paired, of the cells other than those of the promycelium. It seems impossible to doubt that these three methods of division represent progressive stages in a process of reduction, for it is hardly Trouffy, who showed clearly (and his work has been confirmed by Maire and in this paper) that the spermogonium was always borne on a mycelium with single nuclei. Arthur also states that it has been shown that both aecidiospores and uredospores arise from a binucleated mycelium in the usual vegetative manner. This is directly contrary to the facts observed by Sapin-Trouffy eight years ago, who in all cases found that the aecidiospores were borne on a mycelium with single nuclei, and that the nuclei only became paired at the time of aecidium-formation. 1 Whether the two promycelial divisions are regularly different in other forms as in G. clavariae- forme remains for future work to show. Holden and Harper (23), in their recent paper, are of the opinion that in the species of Coleosporium which they examined the two divisions are of like nature, though they were not fortunate enough to have such a complete series of stages as were obtained for the form here investigated. 3 In the divisions of the single nuclei, as observed in the spermatial hyphae, the spindle structure is not so clear. A precocious division of the nucleus of one of the four 1 promycelial ’ cells has been observed, however (while this paper was in the press), in G. clavariaeforme , which shows a distinct spindle, though without centrosomes or polar radiations (Fig. ioo). This shows that the achromatic part of the nucleus may not always be as reduced as in most of the divisions observed. Generations , and General Cytology of the Uredineae . 357 conceivable that in such a highly developed group as the Uredineae we should witness the evolution of mitotic division. The second division with its very complex achromatic mechanism but simple method of chromatin segregation points clearly to the process being a reduced one. A comparative study of nuclear division in the Uredineae with that to be observed in the Basidiomycetes would be of great interest, and, con- sidering the obvious relationship of the two groups, might throw some light on the question as to whether the reduction in nuclear mechanism in the former group is to be connected with the marked parasitic habit. The two chromosomes described by Maire (30) as universal for the Basidiomycetes are doubtless merely chromatin masses, comparable to those described above. Wager (55, 56), Ruhland (47), and Harper (21) have all shown that in the forms they investigated the chromosomes in the basidium were much more numerous. Maire seems to have observed such structures in certain cases, but he applies to them the term ‘ protochromosomes,’ and states that they fuse later to form the two structures which he believes to be the real chromosomes. This is exactly what happens in some cases in the promycelium in G. clavariaeforme (Fig. 30), but there can be no doubt, as the figures show, that the first-formed structures are the true chromosomes. With reference to the question of the function of the nucleus and the relation of direct to indirect division, the existence of a group of Fungi, which in the whole life-cycle show but one, or perhaps two, divisions in which there is a distinct formation of chromosomes, is certainly a very interesting fact. The view that direct nuclear division is confined to cells which are in a senile state has of late years been largely modified. It has been shown that a number of Protozoa and certain animal cells divide normally in an amitotic manner, and Nathansohn (39) has shown for Spirogyra , and Shibata (50) for Podocarpus , that cells which have divided amitotically can later divide in the ordinary mitotic manner. Similarly Hacker (19), Wasielewski (58), and Nemec (40) have shown that nuclei which have divided in an abnormal 1 way under the influence of narcotics can again divide in the normal indirect manner when placed under natural conditions. These observations are generally considered to show that the nucleus has remained entirely unaffected by the intercalation of one or more direct divisions in the normal series of mitoses, but they really prove no more than that the effect (if any) has not been sufficiently profound to destroy their power of normal division or to prevent them carrying on their normal development within the limits of the experiment. Nemec has pointed out 1 There seems considerable doubt as to how far these divisions are abnormal. Wasielewski considers them to be of the nature of amitoses, while Nemec brings forward very considerable evidence to show that they are merely modified indirect divisions in which, however, there is still an exact halving of the chromatin. C C 7, 358 Blackman . — On the Fertilization , Alternation of that in the case of the lower organisms, the effect of a division, presumably unequal, of the nature of amitosis might be long delayed ; and in the case of the higher plants might not be visible immediately. It seems quite possible, however, that even in the case of the higher plants a change might be produced which would not be observable at all in the ordinary restricted life-history of cells other than the germ-cells. We have no knowledge as to how unequal an indirect division may be in relation to the chromatin ; no doubt the degree of inequality must be very variable (it may probably be sometimes as equal in effect as indirect division), but it is quite conceivable that the change produced by a direct division in the c hereditary properties 5 of, say, a root-cell may be confined to those con- cerned with flower or leaf development, and so never be visible in the normal course of development of a root-cell. The beautiful experimental researches of Boveri (9) on double fertilized eggs of the Sea-urchin (in which he practically proves the existence of physiological individuality in the chromosomes) lend strong support to such a view, for certain blastomeres of the three- or four-celled stage, though capable of normal division and apparently similar to the other blastomeres, were found later to be deficient in the power of producing certain organs (such as the skeleton, pigmentation, &c.), owing to the unequal sorting of the chromosomes in the first division. It is clear however that in the case of the intercalation of an amitotic division in the series of divisions leading directly up to the germ-cells, such as that described by Meeves (32) and Macgregor (31) for the spermatozoa of Amphibians, the effect of any inequality of division would produce an effect, sooner or later, in the offspring. If such a division be confirmed it seems hardly possible to consider it as other than a true equational one of equal value with mitosis ; at least if the ordinary view be accepted as to the meaning of chromosomes and their splitting. There seems to be some doubt, however, as to whether this form of division is regularly present. Sutton (52) has also suggested that, as the amitotic division in these cases does not apparently lead to cell-division, it may be that the two nuclei meet again on the spindle of the next mitosis, and so the direct division is without effect. The view may perhaps be hazarded that the sufficiency for the cell needs of the Uredineae of such a simple method of division may be connected with the simple organization of such forms as compared with the higher plants ; the ‘ idioplasm 5 must be far less complex. Alternation of Generations. If the view be accepted that the spermogonia and aecidia represent male and female organs respectively (and the observations in Phragmidium appear to leave no escape from such a view) it is clear that the Uredineae Generations , and General Cytology of the Uredineae. 359 with the aecidium in their life-cycle exhibit an alternation of generations as sharply marked as that of the higher plants. For not only are the two generations to be distinguished as sexual and asexual — bearing in the one case the sexual organs, spermogonia and aecidia, and in the other case only asexual spores, aecidiospores, uredospores, and teleutospores — but they are also to be cytologically differentiated, the sexual generation being characterized by the presence of single nuclei, the asexual by the presence of paired (conjugate) nuclei ; the special nuclear condition of the sporophyte being the result, as in the case of the 2 n chromosomes of the higher plants, of a process of fertilization (at least in Phrag . violaceum). Owing to the extreme reduction in the process of nuclear division in the Uredineae, so that the process of chromosome-formation is in abeyance in nearly all the divisions, one cannot distinguish the two generations by their number of chromosomes ; yet one can distinguish in the nuclei of the oophyte two chromatin masses, while the two paired nuclei, which together correspond with the single nucleus of the ordinary sporophyte, show four chromatin masses. There is thus a near approach to the distinction of chromosome number which is to be found in the higher plants, and also in the interesting alternation of generations which Williams (59, 60) has lately observed in the Dictyotaceae. The teleutospore clearly corresponds to the spore-mother-cell, for it is there that the return to the nuclear condition characteristic of the oophyte is brought about. In the teleutospore, although no actual reduc- tion of chromosomes can be observed, there is a reduction from the four chromatin masses of the two nuclei to the two chromatin masses of the single nucleus, each mass, as shown above, appearing to correspond to a group of chromosomes. This reduction in number of chromatin masses is also associated with peculiar changes in the nucleus concerned which correspond to the process of synapsis. As however the two nuclei remain individualized from the time of fertilization up to the time of reduction in number of chromatin masses we have associated with this process a nuclear reduction , i. e. a reduction in the number of the nuclei from two to one. Since this process of nuclear reduction takes place in most organisms at the time of fertilization, or soon after, it has been confused with that process itself, and so has led to the belief that the fusion of nuclei in the teleutospore was itself a sexual process. This view, which was hardly credible even on the facts hitherto known, has been rendered quite untenable by the discovery of an association of nuclei in the aecidium, which must itself be considered as the fertilization process. After the process of synapsis and chromatin-reduction, the cell of the teleutospore undergoes a process of tetrad division to form the four cells of the ‘ promycelium ’( = spores). The teleutospore (or its cell) is thus exactly 360 Blackman . — On the Fertilization , Alternation of comparable to the spore-mother-cell of the higher plants and to the tetra- spore-mother-cell in the Dictyotaceae. The gametophyte generation starts again in the teleutospore and is continued up to the ‘ fertile cell ’ (the cell bearing the mother-cells of the aecidiospores) in the aecidium ; while the sporophyte continues throughout the rest of the life-history, through the aecidiospores, the uredospores (if present), and up to the teleutospores again. Credit must be given to Maire (29 and 30) for having pointed out the resemblance between the nuclear history of the higher plants and that of the Uredineae. He believed, however, that the ‘ synkaryon’ was a special condition always brought about by the association of two daughter-nuclei. He considered the spermogonia and aecidia to be non-sexual in nature, and was ignorant of any process of fertilization such as has just been described for the aecidium of Phragmidium . The existence of actual histological differences between gametophyte and sporophyte is in full agreement with the usual rigid distinction which is to be observed between the two generations, both in heteroecious and autoecious forms. As is well known, the aecidiospore by infection gives rise only to a mycelium bearing uredospores or teleutospores, the uredospore gives rise to a mycelium bearing only uredospores again, or teleutospores. Again, the sporidium arising from the teleutospore gives rise by infection to a mycelium bearing only aecidiospores, if those are present in the life-history (i. e. in the eu- and -opsis forms) 1. Apogamy. In the forms in which the aecidium is absent (< brachy -, lepto -, micro -, and hemi - forms) the sporidium produces a mycelium which gives rise directly to uredospores or teleutospores with paired nuclei. In such a case we have a transition from the sexual to the asexual generation without the intervention of the female reproductive organ, and thus it is clearly comparable to apogamy among the higher plants. Sapin-Trouffy has shown that the change from single to paired nuclei takes place in these cases in connexion with the formation of the uredospores, or of the teleuto- spores, if the former are absent. The exact cytological details have yet to be worked out. It may be that the binucleate condition is brought about by the interaction of two vegetative cells , as in the case of apogamy in Ferns investigated by Farmer, Moore, and Digby (14), where they found the nuclei of neighbouring prothallial cells fusing to form sporophytic tissue. A form like Puccmia Liliacearum , DC., seems, like some of the Ferns, to be only imperfectly apogamous, for though the aecidia have been described they do not seem to be usually present. Apospory . The peculiar cycle of development observed regularly in 1 The very few observed cases of aberrations from the normal life-history, such as aecidiospores giving rise to a mycelium bearing aecidiospores again (A. P. Dietel, Zeit. fiir Pflanzenkrankheiten, iii, 1893, p. 25; Flora, lxxxi, 1895, p. 394), require further investigation; such cases may possibly be due to a separation of the nuclei of the pair, as in Endophyllum. Generations , and General Cytology of the Uredineae . 361 Endophyllum , in which binucleate aecidiospores, produced in a normal way, behave on germination like teleutospores, the two nuclei separating and uninucleate sporidia being produced (see Maire, 29), is just the opposite to the case discussed above. It is the transition from the asexual to the sexual generation without the intervention of the teleutospore (spore- mother-cell). It can therefore be compared with cases of apospory among the higher plants. Maire has described a case in Endophyllum in which the aecidiospore germinated in a normal way, but the life-history of the form is not known. There can be little doubt that the eu- and -opsis forms which possess the aecidium must be considered as more primitive and the other (apogamous) forms as reduced. The view of Dietel, that the lepto- and micro - forms which possess only teleutospores are the most primitive, and that the other forms are derived from them, can hardly be accepted in the light of our present knowledge. Heteroecism . As the heteroecious forms are confined to those possessing the aecidium, i.e. to the more primitive, it seems probable that heteroecism may not be, as generally conceived (see Klebahn, 27), a later adaptation, but may actually be the primitive condition in the group. Although we are ignorant of the origin of the group it is possible to conceive that the sporophyte was first developed in connexion with life on another host, just as the sporophyte in the higher plants seems to have been developed in connexion with a new terrestrial existence. The autoecious eu- forms would then be the first step in reduction — a purely environmental one ; later a morphological reduction of the number of spore- forms would appear to have taken place. Fusion in the Basidium. The fusion of nuclei in the basidium of the Basidiomycetes is exactly comparable to that in the teleutospore, for it has been shown by Maire (30) in the case of a large number of forms, and also by Harper (21) in two cases, that in this group, also, the two nuclei which fuse are conjugate nuclei, the ancestors of which have been associated together for a number of generations. It is clear then that the view of Dangeard and others, which considers this fusion as a process of fertilization, is quite untenable. The stage at which the nuclei first become conjugate is unknown, but it is there that one must look for a process of fertilization, and not to the fusion in the basidium, which, like that in the teleutospore, is a process of mere nuclear reduction, and (like chromosome-reduction) the necessary sequence of fertilization if the sexual cycle is to be completed. As the Basidiomycetes are apparently without anything which could be considered even as a reduced form of sexual organ, it is suggested that the conjugate nuclear condition, the transition from gametophyte to 362 Blackman. — On the Fertilization , Alternation of sporophyte, is brought about in a purely apogamous way (as in the case of the Uredineae without an aecidium), perhaps by the interaction of two vegetative cells, as in the union of vegetative cells of the prothallium mentioned above. Maire (30) has stated that the basidiospore gives origin to a mycelium with single nuclei (thus agreeing with the sporidium, which has always been found to behave in this way), and that later in the mycelium the nuclei become paired, but in what manner is not known. The condition of paired nuclei seems in some cases to arise very early, for Harper (21) was unable to observe in Hypochnus any but binucleate cells ; the germination of the basidiospore, however, was not followed. The Basidiomycetes would thus seem to resemble very closely the lepto - forms among the Uredineae, where the mycelium has at first single nuclei but later develops paired nuclei in connexion with the formation of teleutospores, and like them would appear to have an alternation of genera- tions obscured , however , by the apogamous transition from one to the other. The fusion of nuclei in the basidium, like that in the teleutospore, is followed by a tetrad division, and there seems little doubt that it is also followed by a process of chromosome-reduction, or one corresponding to it ; the basidium, like the teleutospore, seems also of the nature of a spore- mother-cell, and is to be compared to that of the higher plants. The reduction in the basidium from four to two chromosomes described by Maire (30) as general for the Basidiomycetes can hardly be accepted, since these structures appear to represent groups of chromosomes, but the observations show clearly the analogy with Gymnosporangium. Whether the nuclear divisions in the Basidiomycetes will be found sufficiently typical to allow of the observation of an actual numerical reduction remains for future inquiry. Maire has also described a process of synapsis in the fusion-nucleus of the basidium, comparable to that here described for the teleutospore. Relationships of the Uredineae. It is obvious that the Uredineae and Basidiomycetes are very closely related, for both possess single nuclei during the early part of their life- history (starting from the teleutospore and basidium respectively), which later become paired. In both cases, also, the paired nuclei fuse in special reproductive cells, the fusion being followed by a process of synapsis, and then by a tetrad division 1. The Basidiomycetes seem wanting in all trace of definite sexual organs, so that the Uredineae must be considered as by far the more primitive and cannot be treated as a mere class of the Basidiomycetes, as in the well-known classification of Brefeld. As pointed out in the previous section the latter group appear rather to be reduced, apogamous forms of the Uredineae. 1 As is well known the transversely divided teleutospore of Coleosporium is practically indis- tinguishable from a transversely divided basidium. Generations , and General Cytology of the Uredineae . 363 The existence of immotile male cells (spermatia) and of a sterile cell (in Phvagmidium ) which may possibly have formerly had the function of a trichogyne certainly points to a relationship of the Uredineae with the Florideae, a relationship which has been suggested by Meyer (35), chiefly on anatomical grounds, for all the higher Fungi. It is possible that an alternation of generations may yet be found in this algal group, for tetraspore-formation suggests a process of chromosome-reduction; the fact that in the majority of forms there is a distinction of sexual and asexual plants is also very striking in this connexion. Such a discovery — and it must be remembered that Williams (59 and 60) has already observed this alternation in the Dictyotaceae with chromosome reduction in the tetra- spore-mother-cell — -would certainly lend strong support to a view of a relationship between the Florideae, Uredineae, and Basidiomycetes. Summary. The peculiar nuclear cycle described by Sapin-Troufify for the Uredineae was confirmed in the case of Phragmidium violaceum , Wint., and of Gy mno sporangium clavariaeforme , Rees. In this cycle, the mature teleutospore is uninucleate and gives rise to four uninucleate sporidia, from which a mycelium is developed with the nuclei arranged singly, usually in separate cells. The spermatia produced on this mycelium are uninucleate, but in the young aecidium the nuclei become paired (forming binucleate cells) and divide together in very close association. This paired condition is then persistent throughout the rest of the life-cycle (aecidiospores, uredospores, and mycelia produced from them) up to the formation of the teleutospores, which in the young state are binucleate, but when mature become uninucleate by the fusion of the two paired nuclei. This cycle of development seems common to all the Uredineae (except Endophyllum ) which have an aecidial stage in their life-history. A study of the structure of the spermatia of the Uredineae shows that they have the characters not of conidia but of male cells , for they exhibit a large dense nucleus, very little cytoplasm, no reserve material, and a very thin cell-wall. These characters, together with their usual association with the aecidia, their absence of function, and the peculiar, apparently reduced, form of fertilization to be observed in the aecidium of Phragmidium violaceumy point clearly to the view that the spermatia are male cells which formerly took part in a process of fertilization in connexion with the aecidium, but have now become fiinctionless. The aecidium of Phrag. violaceum is developed immediately beneath the epidermis of the leaf and has a very simple structure. It consists of a layer of special cells, each of which cuts off a sterile cell above, which soon disorganizes, while the lower, the fertile cell, increases in size and shows abundant protoplasm, and after a pause in its development is fertilized by 364 Blackman . — On the Fertilization , Alternation of the migration into it of the nucleus of one of the undifferentiated mycelial cells at its base. It then undergoes a series of rapid divisions, cutting off a series of aecidiospore-mother-cells. The ‘ fertile cell 5 has thus the characters of a female cell. The aecidium of Phrag. violaceum is to be considered as a sorus of female reproductive organs , each of which consists of a sterile cell above and a female cell below, the nucleus of an ordinary vegetative cell bringing about fertilization and performing the part which was apparently formerly taken by the nucleus of the spermatium. The female cell thus develops by means of a reduced form of fertilization. It is suggested that the sterile cell is reduced, and that it formerly pushed its way to the surface (as it can sometimes now be observed to do) and acted as a trichogyne to bring the spermatium into relation with the female cell below. In the process of fertilization the two nuclei do not fuse, but merely remain very closely associated (as in the egg of Cyclops) ; there is thus started in the female cell the condition of paired nuclei which continues up to the teleutospore. The aecidium throughout the group must be considered as a sorus of reduced female reproductive organs , for it appears to be always the fertile cell which becomes binucleate. Whether in all cases there is a nuclear migration, or whether in some there is a still further reduction and the process consists of an association of two daughter-nuclei, has yet to be deter- mined, as has also the question of the presence or absence of a sterile cell. As the spermogonia and aecidia are to be considered as male and female reproductive organs, respectively, it is evident that the Uredineae which possess an aecidial stage in their life-history ( eu - and -opsis forms) exhibit a well-marked alternation of generations which are not only to be distinguished as sexual and asexual, but are also to be sharply differentiated cytologically, the sexual generation being characterized by single nuclei (with two chromatin masses on division), the asexual by paired nuclei (with four chromatin masses on division). The transition from the gametophyte to the sporophy te takes place in the aecidium and the transition from the sporophyte to the gametophyte in the teleutospore. The alternation of generations is thus as clearly marked as that of the higher plants or of the Dictyotaceae. The fusion in the teleutospore of the two nuclei— the direct descendants of those which first became associated in the fertile ( female ) cell of the aecidium — is clearly not in itself a process of fertilization (nor the teleuto- spore an egg-cell), as Dangeard and Sapin-Trouffy supposed, but a mere secondary process, the result of fertilization and the preliminary to reduction. It is pointed out that three nuclear stages are to be observed in the sexual cycle of plants and animals — nuclear association , nuclear reduction (so-called fusion)) and chromosome-reduction. Of these three stages, only the first } Generations , and General Cytology of the Uredineae . 365 is the essential part of fertilization (as various plants and animal eggs and that of Cyclops show) ; the second may take place at the same time as the first, or it may be delayed for a time, or, as in the Uredineae, it may be delayed until the stage corresponding to chromosome-reduction. The fusion-nucleus in the teleutospore undergoes changes which correspond to synapsis , and when it divides there is seen to be a reduction from the four chromatin masses of the paired nuclei to two chromatin masses ; owing to the absence of chromosome-formation in most of the nuclear divisions an actual reduction in number of chromosomes cannot be observed. The process of fusion and reduction in the teleutospore is followed by a definite tetrad division in the c promycelium/ so that the teleutospore corresponds exactly with the spore-mother-cell of the higher plants, the { promycelial ’ cells being really of the nature of spores. As the fusion of the two nuclei is delayed throughout the whole of the sporophyte there is associated with the reduction from four to two chromatin masses a process of nuclear reduction from two nuclei to one. Nuclear division in most of the cells of the Uredineae is of an exceedingly simple type. The nuclei (whether single or paired) lose their membrane, the nucleolus becomes extruded, and the chromatin condensed into one, or sometimes two, masses. A rudimentary spindle can sometimes be observed, and upon this the chromatin becomes spread and is drawn apart into two or four pear-shaped masses which separate and form the daughter-nuclei. In the paired (conjugate) state the nuclei divide side by side in close juxtaposition, passing pari passu through the various stages of division, as described by earlier observers. In the promycelium the two divisions are much more typical. In G. clavariaeforme a well-marked spindle with centrosomes and polar radia- tions was observed. The first division showed a formation of numerous chromosomes, though their behaviour on the spindle is not typical and there is doubt as to whether a definite splitting actually occurs. In the second division the chromatin forms no chromosomes, but merely a network which covers the spindle, and is later drawn apart into two portions. The spindle in the case of the second division in the promycelium of G. clavariaeforme , and apparently also in the case of the first, is formed free in the cytoplasm between the two portions of a divided centrosome (like the ‘ Centralspindel * of Hermann), and afterwards comes into close relation with the nucleus. The simple form of division found in the Uredineae is to be considered as reduced from the typical method of karyokinesis. The two structures observed constantly during nuclear division by Sapin-Trouffy and Maire, on which they base their views of chromosome- reduction, cannot be considered of the nature of chromosomes, but are merely chromatin masses. They probably represent the chromatin derived 366 Blackman . — On Ike Fertilization , Alternation of respectively from the two nuclei of the teleutospore, but are not always present. In the first division of the promycelium of G. clavariaeforme they can be seen to be formed by the aggregation of a number of chromo- somes ; in the other divisions they are formed directly, by the condensation of the chromatin of the nucleus. In those Uredineae which have a reduced life-cycle without an aecidium (, brachy hemi -, micro-, and lepto- forms) the transition from single to paired nuclei, and from gametophyte to sporophyte, apparently takes place in connexion with the uredospores, or if these are absent, in connexion with the teleutospores. Such a shortening of the life-cycle is comparable with the case of cipogamy among the higher plants ; the cytological details are not yet known, but it is possible that it is brought about by the association of the nuclei of two vegetative cells (cf. apogamy in ferns). The peculiar shortening of the life-cycle to be observed in Endophyllum , where the binucleate aecidiospore germinates like a teleutospore and the uninucleate condition is brought about by the simple separation of the two nuclei, is clearly a transition from the sporophyte to the gametophyte without the intervention of the teleutospore (spore-mother-cell), and should accordingly be considered as comparable to apospory in the higher plants. The fusion of paired nuclei in the basidium of the Basidiomycetes is exactly comparable with that in the teleutospore, and should also be con- sidered, not as a process of fertilization, but as a purely secondary process of nuclear reduction preliminary to chromosome reduction. How the condition of paired nuclei is brought about in the Basidio- mycetes is not yet clear. The life-history of these forms seems comparable with that of the lepto- forms among the Uredineae, and it is suggested that, in the absence of sexual organs, the transition from single to paired nuclei, i.e. the transition from gametophyte to sporophyte, is brought about apogamously in a similar way. The Uredineae and Basidiomycetes show an obvious relationship, for both have paired nuclei at some stage of their life-history, and in both groups the individuals of certain pairs fuse together in specialized reproduc- tive cells, a fusion to be followed by synapsis and a tetrad division. The Uredineae with their sexual organs would certainly seem to be more primitive and can hardly be classed as a mere subdivision of the Basidiomycetes. The Uredineae appear to show a relationship with the Florideae among the Algae. The aecidium of Phragmidium with superficial position and very simple structure without a pseudoperidium is really not a definite organ, but a mere ill-defined sorus of reduced, female reproductive organs ; it is no doubt primitive , while the typical aecidium with its pseudoperidium and deeper point of origin is a definite organ, and is probably a later development. The binucleate condition of the sporidium exhibited by Phrag . Generations , and General Cytology of the Uredineae . 367 violaceum and some other forms is without particular significance, and is apparently a mere precocious division in which wall-formation is delayed. The view which has been put forward, that the sporidium starts in the life-history the condition with paired nuclei, is without foundation, for in all cases observed the mycelium arising from the sporidium has single nuclei, even in such cases as Phrag. violaceum , where the sporidium is known to be usually binucleate. The teleutospore is really of the nature of a spore-mother-cell, the contents of which, usually after germination (but directly in Coleosporium ), break up by a tetrad division into a series of four primary spores (the so-called promycelial cells). These can separate and put out germ-tubes, and no doubt cause infection ; normally, however, when growing in free air, they remain connected and their germ-tubes become arrested, so that four secondary spores — the sporidia — are formed. The term promycelium is thus a misnomer. NOTE. It is obvious that the method of development of the fertile (female) cells in the aecidium of Phragmidium may be considered as a new type of so-called partheno- genesis; for, just as in Artemia there is a fertilization of c parthenogenetic ’ eggs by the second polar body (nucleus) instead of by a spermatozoon, so in this case the nucleus of a vegetative cell has apparently replaced in function that of the spermatium. The term parthenogenesis is, however, an unsatisfactory one, as it is not very aptly applied to these cases in which there is a process of nuclear fusion ; it would be much more satisfactory to confine the term to cases in which the sexual cells actually develop with the reduced number of chromosomes, without any form of fertilization. There is the same difficulty with apogamy now that we know that in apogamous Ferns the nuclei of prothallial cells may fuse together, for the application of this term suggests a denial of the existence of any process of fusion of cells or nuclei which could be considered of the nature of gametic fusion. It becomes clear that cytological investigations of recent times have practically broken down the distinction between fertilization, parthenogenesis (in the wide, but not in the narrow, sense suggested above), and apogamy. Between the fusion of an egg-cell with a differentiated male cell, the fusion (association) of an egg-cell with a vegetative cell (or nucleus), the fusion of an egg-cell and its polar body, and the fusion of two vegetative cells (or nuclei) no sharp line can be drawn, and they must all be considered as terms in a series of fertilizations. Just as on one side of normal oogamous fertilization with its differentiated male and female cells there are such primitive types of fertilization as isogamy, and that, the most primitive, in which there is no distinction of sexual and vegetative cells ; so, on the other side, there is a series of reduced processes of gradually increasing simplicity, the most simple being practically a return to the most primitive, where the sexual and vegetative cells are alike. Apart from the presence of the functionless spermata and the peculiar migration of the nucleus through a cell-wall, which show that it is reduced in evolu- tion, there is nothing to clearly distinguish the process observed in Phragmidium from such accepted fertilizations as the conjugation of adjacent cells in a Spirogyra filament, the fusion of gametes (which are of the relationship of cousins) in Adinosphaerium , or from such a process as is to be observed in Basidiobolus, or even other Fungi, in which the sexual organs arise on neighbouring cells of a hypha. Blackman . — On the Fertilization, Alternation of 368 List of Papers. 1. Arthur, J. C. : Problems in the Study of Plant Rusts. Bull. Torrey Bot. Club, xxx, 1903, p. r. 2. — : The Aecidium as a Device to restore Vigour to the Fungus. Soc. prom. Agric. Sci. 23rd Ann. Meeting, 1903 (not seen). 3. de Bary, A. : Untersuchungen liber die Brandpilze. Berlin, 1853. 4. : Vergleichende Morphologie u. Biologie der Pilze u.s.w. Leipzig, 1884. 5. Baur, E. : Zur Frage nach der Sexualitat der Collemaceae. Ber. d. deutsch. Bot. Ges., xvi, 1898, P- 363- 6. Blackman, V. H. : On the Cytological Features of Fertilization and Related Phenomena in Pinus silvestris , L. Phil. Trans. Roy. Soc., cxc, 1898, p. 395. 7. : On the Conditions of Teleutospore-germination and Sporidia-formation in the Uredineae. New Phytologist, ii, 1903, p. 10. 8. : On a new Method for facilitating the Staining of microscopically small Objects. Ibid., ii, 1903, p. 103. 9. Boveri, Th. : Mehrpolige Mitosen als Mittel zur Analyse des Zellkems. Verh. Phys.-Med. Ges. Wurzburg, xxxv, 1902, p. 67. 10. Brefeld, O. : Untersuch. aus dem Gesammtgebiete der Mykologie, vii, 1888, p. 60 ; viii, 1889, p. 230; ix, 1891, p. 26. 11. Conklin, E. G. : Karyokinesis and Cytokinesis in the maturation, &c., of Crepidula. Journ. Acad. Philad., xii, Pt. I, 1902. 12. Cornu, M. : fitude de la fecondation dans la classe des Champignons. Compt. Rend., Juin, 1875, p. 14 66. 13. Dangeard, P. A., and Sapin-Trouffy, P„ : Une pseudofecondation chez les Ur<£dinees. Compt. Rend., cxvi, 1893, p. 267. 14. Farmer, J. B., Moore, J. E., and Dxgby, L. : On the Cytology of Apogamy and Apospory. I. : Prelim, note on Apogamy. Proc. Roy. Soc., lxxi, 1903, p. 453. 15. and Moore, J. E. S. : New Investigation into the Reduction Phenomena of Animals and Plants. Proc. Roy. Soc., lxxii, 1903, p. 104. 16. Ferguson, M. C. : The Development of the Egg and Fertilization in Pinus Strobus. Annals of Botany, xv, 1901, p. 435. 17. Hacker, V. : Ueber die Selbstandigkeit der vaterlichen und miitterlichen Kernbestandtheile u.s.w. Archiv fiir mikr. Anat., xlvi, 1895, p. 579. 18. : Die Keimbahn von Cyclops. Archiv fiir mikr. Anat., xlix, 1897, p. 35. 19. : Mitosen im Gefolge amitosenahnlicher Vorgange. Anat. Anzeig., xvii, 1900, p. 9. 20. Harper, R. : Sexual reproduction in Pyronema confluens , & c. Annals of Botany, xiv, 1900, p- 321. 21. : Binucleate Cells in certain Hymenomycetes. Bot. Gaz., xxxiii, 1902, p. 1. 22. Hedlund, T. : Ueber Thallusbildung durch Pyknokonidien bei Catillaria denigrata , Fr. u. C. prasina , Fr., Bot. Centralbl., lxiii, 1895, p. 9. 23. Holden, R. J. and Harper, R. A. : Nuclear Division and Nuclear Fusion in Coleosporiuvi Sonchi-arvensis , Lev. Trans. Wisconsin Acad. Sci., &c., xiv, 1903, p. 63. 24. Istvanffi, G. : Ueber die Rolle der Zellkerne bei der Entwickelung der Pilze. Ber. d. deutsch. Bot. Ges., xiii, 1895, p. 459. 25. Juel, H. O. : Die Kerntheilungen in den Basidien und die Phylogenie der Basidiomyceten. Jahrb. fiir wiss. Bot., xxxii, 1898, p. 361. 26. Klebahn, H. : Kulturversuche mit heterocischen Uredineen. Zeitschr. fiir Pflanzenkrank. ii. 1892, p. 264 ; ibid, v, 1895, p. 69. 27. : Die wirtswechselnden Rostpilze. Berlin, 1903 (dated 1904). 28. Lauterborn, R. : Untersuch. iiber Bau, Kernteilung u. Bewegung der Diatomeen. Leipzig, 1890. 29. Maire, R. : L’evolution nucliaire chez les Endophyllum. Journ. de Bot., xiv, 1900, p. 80. 30. : Recherches cytologiques et taxonomiques sur les Basidiomycetes. Bull. Soc. Mycol. de France, xviii, 1903, pp. 1-209. Generations , and General Cytology of the Uredineae. 369 31. McGregor, J. H. : The Spermatogenesis of Amphiuma. Journal of Morphology, xv, 1899 (Supplement), p. 57. 32. Meeves, Fr. : Uber amitotische Kernteilungen in den Spermatogonien des Salamanders u. s. w. Anat. Anzeig., vi, 1891, p. 626. 33. Mezger, O. : Untersuch. fiber die Entwickelung der Flechtenfrucht. Beitr. zur wiss. Bot., v, 1903, p. 108. 34. Meyen, F. J. F. : Pflanzen-Pathologie. Berlin, 1841, p. 143. 35. Meyer, A. : Die Plasmaverbindungen und die Fusionen der Pilze der Florideenreihe. Bot. Zeit., lx, 1902, p. 139. 36. Moller, A. : Uber die Cultur flechtenbildender Ascomyceten ohne Algen. Untersuch. aus dem Forst-Bot. Inst. Munster, 1887. 37. Montgomery, T. H. : A Study of the Chromosomes of the Germ-cells of the Metazoa. Trans. Amer. Phil. Soc., xx, 1901, p. 154. 38. * : The Heterotypic Maturation Mitosis in Amphibia and its general significance. Biol. Bull., iv, 1903, p. 258. 39. Nathansohn, A.: Physiologische Untersuchungen fiber amitotische Kerntheilung. Jahrb. f. wiss. Bot., xxxv, 1900, p. 48. 40. Nemec, B. : Ueber die Einwirkung des Chloralhydrats auf die Kern- und Zellteilung. Jahrb. f. wiss. Bot., xxxvii, 1902, p. 8. 41. Overton: Mikrotechnische Mitteilungen. Zeitschr. f. wiss. Mikrosk., vii, 1890, p. 9. 42. Plowright, C. B. : British Uredineae and Ustilagineae. London, 1889. 43. Poirault, G., and Raciborski, M. : Sur les noyaux des Uredin^es. Journ. de Bot., ix, 1895, v P- 3*8. 44. Raciborski, M. : Ueber den Einfluss ausserer Bedingungen auf die Wachsthumsweise des Basidio bolus ranarum. Flora, lxxxii, 1896, p. 107. 45. Rosen, F. : Beitrage zur Kenntniss der Pflanzenzellen. Beitr. zur Biol, der Pflanzen, vi, 1892, p. 255. 46. RBckert, J. : Ueber das Selbstandigbleiben der vaterlichen und mfitterlichen Kernsubstanz u.s.w. Archiv ffir mikr. Anat., xlv, 1895, p. 3. 47. Ruhland, W. : Ueber intracell ulare Karyogamie bei den Basidiomyceten. Bot. Zeit, lix, 1901, p. 187. 48. Sapin-Trouffy, P. : Recherches histologiques sur la famille des Uredinees. Le Botaniste, 5e s